International Soil Mechanics and Geotechnic Conference Proceedings

Technical Committee 102
Ground Property Characterisation

from In-Situ Tests
Comit technique 102
Caractrisation des proprits des terrains
par essais in situ
471
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
1
General Report for TC102
In-Situ Testing
Rapport gnral du TC102
Essais in-situ
Giacheti H.L.
So Paulo State University, Dept. of Civil and Environmental Engineering, Bauru-SP, Brazil (wwwp.feb.unesp.br/giacheti)
Cunha R.P.
University of Braslia, Dept. of Civil and Environmental Engineering, Braslia-DF, Brazil (www.geotecnia.unb.br/gpfees)

ABSTRACT: An overview of all the technical papers accepted for the in-situ testing session of the 18
th
ICSMGE is presented. Forty
two papers submitted to this conference were considered as part of this session. The papers were grouped into four major categories:
site characterization, technological advances, geotechnical analyses and behavior, and soil and rocks properties. The objective of this
report is to present an overview of the theme topics and briefly discuss the major contributions achieved by these papers.

RSUM : Un aperu de tous les articles accepts la confrence dans la session Essais in-situ du 18
e
CIMSG est prsent ici.
Quarante-deux crits soumis cette confrence ont t considrs comme faisant partie de cette session. Les documents ont t
regroups en quatre grandes catgories: la caractrisation du site, les progrs technologiques, les analyses gotechniques, les
comportements et enfin les proprits des sols et des roches. L'objectif de ce rapport est de prsenter une vue d'ensemble de tous les
sujets et de discuter brivement des contributions majeures apportes par ces documents.
KEYWORDS: In-situ tests, site characterization, technological advances, geotechnical analysis and behavior, soil and rock properties.

1 INTRODUCTION
Site characterization is the first step on all geotechnical projects
and the objectives generally relate to the definition of the
stratigraphic profile and groundwater level, estimation of the
geotechnical properties from each soil unit, identification of
critical layers, definition of geotechnical design parameters and
indication of required, if necessary, additional laboratory tests.
The traditional methods for site characterization rely
basically on drilling, sampling and laboratory tests. These are
usually time consuming and, in some cases, over budget. The
modern approach, on the other hand, focuses on the rational
use of in-situ penetration tools coupled in some cases with
geophysical techniques. Of course, the success of an efficient
site characterization program depends on clearly defining the
scope or objectives of the enterprise and, in some cases,
combined site investigation techniques are adopted as will be
demonstrated through the papers of this session.
Hence, TC102 sessions of the conference contain papers
with distinct investigative approaches and scopes. Some have
presented new testing devices; others new characterization or
interpretation methods. Some have described real case studies
where the site characterization was a major issue, whereas
others discussed the interpreted soil and rock properties to be
used as input for routine geotechnical analyses.
Most of the contributions deal purely with in-situ
investigation tools, but many have mixed it with laboratory or
numerical investigation techniques. As presented in Figure 1,
the majority of the papers are European in essence, which is
expected for the 18
th
ICSMGE held in this continent.
Asia
Africa
Europe
South
America
North
America
Australia

Figure 1. Paper distribution by continents for this conference session.
Figure 2.a shows the distribution of all papers in this session
that used any particular in-situ testing technique as a major site
investigation tool. Notice that most of the papers used more
than one technique. From this, it is possible to realize that CPT
and SPT were the more widely-employed tools amongst the
published papers. Figure 2.b depicts, from the universe of
papers that solely adopted CPT or SPT (or both) as site tool,
which interpretation techniques were adopted. It is clearly
evident that empirical approaches still form the dominant
interpretation group, although in many papers it has been used
together with other complementary methods.

a) b)
Figure 2. a) Percentage of all papers in the session that used the listed
in-situ technique among others site investigation tools and b) Percentage
of (only) CPT and/or SPT papers in the session that adopted the listed
approach to interpret the data, among other techniques.
Figure 3 shows the types of geotechnical formations that
served as the major soil stratum for the employed investigative
techniques. It is clear that the great majority of the presented
papers are concerned with sedimentary deposits, whereas few of
them focused on less classical materials such as residual
(tropical) soils or man-placed tailings and compacted earth fills.
Sedimentary
Residual/
Tropical
EarthFill/
Tailing
Other

Figure 3. Percentage of geomaterial types addressed in this section.
472
th
th
2
This report is organized into several major topics and subject
areas, as follows: site characterization (4 papers), technological
advances (9 papers), geotechnical analyses and behavior (14
papers), and soil and rocks properties (15 papers), leading to a
total of 42 papers. The main objective is to present an overview
and the advances on the main addressed topics of this
Conference Session, hence summarizing and emphasizing the
important contributions from the papers.
Table 1 presents a summary with the main topics and
subtopics addressed for each of the listed papers, together with
the adopted investigation tools (in-situ or laboratory), the main
soil type, the investigation approach, and a short 1-line
summary of the papers prime objective & contribution. Given
such cited divisions, the papers will be reported under each of
the defined topics and subtopics, as it follows next.
2. SITE CHARACTERIZATION
In this Conference Session, four papers dealt with site
characterization; two of them are related to soil classification
and the other two are more focused on geotechnical modeling.
2.1. Soil classification
The paper written by Serratrice proposes a classification
method for natural soils based on piezocone test data. Two
aspects are considered to classify the soils with liquefaction
potential, the drained and undrained strength via triaxial tests
and the soils density. The method is presented and applied in
two examples where CPTU data are available in homogeneous
clayey deposits.
The paper from Baud & Gambin presents a contribution to
enhance the Pressiorama
diagram with the extra rheological

factor , which was originally introduced by Mnard on his
design method. The authors used very good quality self-bored
PMT tests (STAF technique) in several soil types, from soft
clays to rock to obtain E
M
/p
*LM
values. They proposed a re-
evaluation of the rheological factor and the findings are given
by an equation graphically expressed in the Pressiorama
.
2.2. Geotechnical modeling
The paper from Ivi et al. discusses the applicability of the
RNK-method for spatial engineering & geological and/or
geotechnical modeling. This method was tested on many
landslides in Croatia and it allows the differentiation of the
minimum shear strength zone, or regions of different hydraulic
conductivities and varied soil densities. The proposed model
was verified by measurements of lateral movements in the
landslide area and by results of stability analyses. They
concluded that the RNK-method can be used in the study of
landslides and slope stability by searching the zone of minimum
shear strength.
The paper from Steenfelt et al. presents the use of in-situ
and laboratory tests for site characterization on an important on-
going infrastructure project in China. A very extensive site
investigation campaign was carried out comprising geotechnical
boreholes, CPTUs and seismic testing with associated advanced
laboratory testing. The paper described the results and the
interpretation technique used to provide ground stratification
and stiffness variations to be used in design. They concluded
that the CPTU was a important tool for a clear geological unit
delineation, which also allowed a robust and safe design.
3. TECHNOLOGICAL ADVANCES
In this Conference Section, nine papers were selected to be part
of the technological advances (main) topic; three of them
presented new uses of in-situ testing technologies and six dealt
with new types of in-situ testing tools (or apparatuses).
3.1. New uses of in-situ technology
Kims et al. paper has an environmental appeal since it deals
with the geological CO
2
sequestration as an effective mean of
reducing the emissions of carbon dioxide. The problem pointed
out in the paper is that forward strategies and technologies of
CO
2
sequestration in Korea need to be specified depending on
the geological conditions of potential sites (in Korea). The
authors reviewed the geological characteristics of CO
2
storage
projects around the World and also discuss the suitability for
CO
2
sequestration. A systematic and quantitative evaluation
method to assess the storage and economic efficiencies of
sedimentary basins in Korea using dimensionless values
between 0 and 1 was applied (Figure 4). The paper also
discusses the potential of using geophysical tests to assess the
suitability of field strata for CO
2
-storing, and to monitor CO
2

movement and possible leakages.

Figure 4. Score for suitability for Korean sedimentary basin (Kim et al).
The paper from Fenton & Hicks discusses the uncertainty
associated with site characterization and it focused specifically
on the effect of number of samples on residual uncertainty. The
results can be used to quantitatively select the required number
of samples needed to achieve a target maximum residual
uncertainty level. A statistical approach was used to study this
problem and it was concluded that the accuracy improves as the
number of samples and the correlation length increases.
Somasundaram et al. present the characterization and
settlement modeling of deep inert debris fills. Inert fills can be
considered as a non-text book type geomaterial since they are
difficult to characterize and model by current geotechnical
methods, due to their inherent heterogeneity, very large particle
sizes, and nested and voided structure. The authors presented an
approach to characterize a 54 m deep inert debris fill, to model
its settlement behavior under seismic loading and groundwater
level rise, and to develop remedial measures to render it suitable
for development.
3.2. New in-situ tools
Jacquards et al. paper presents a new probe to overcome the
limitation of Menard type pressuremeter tests, i.e., the difficulty
of reaching large expansion volumes and pressures. This new
device allows for the volume of the hole to be doubled, even
under high pressures. The authors described the technological
innovations that increased the capabilities (and reliability) of the
pressuremeter probe as well as presented comparative tests on
different sites to demonstrate the advocated technical advance in
this enhanced PMT device.
In Rito & Emura paper a new type of sampling method
called Koken wire line system is developed (Figure 5) to
retrieve high depth undisturbed samples in deep Pleistocene
clay and sand layers at the Kansai International Airport area.
The authors also developed two different pore pressure
measuring devices, and concluded that both the sample quality
and the measured values were respectively of high quality and
with reasonably good accuracy to be used in the settlement
design of the subsoil of this airport, in Japan.
473
Technical Committee 102 / Comit technique 102
th
3
T
a
b
l
e

1
.

S
u
m
m
a
r
y

o
f

t
h
e

g
e
n
e
r

a
l

c
h
a
r
a
c
t
e
r
i
s
t
i
c
s

f
r
o
m

e
a
c
h

o
f

t
h
e

p
a
p
e
r
s

f
o
r

T
e
c
h
n
i
c
a
l

S
e
c
t
i
o
n

T
C

1
0
2

(
I
n
-
S
i
t
u

T
e
s
t
i
n
g
)
.

M
a
i
n

I
n
v
e
s
t
i
g
a
t
i
o
n

T
o
o
l

M
a
i
n

T
o
p
i
c

I
n

S
i
t
u

L
a
b
o
r
a
t
o
r
y

S
t
u
d
i
e
d

S
o
i
l

T
y
p
e

I
n
v
e
s
t
i
g
a
t
i
o
n

A
p
p
r
o
a
c
h

Subtopics
A
u
t
h
o
r

/

P
a
p
e
r

SPT / BPT / SWS
CPT
DMT
PMT
Panda/DCP
Geophysics
Borehole
Sampling
Permeability
New Device
Other Type
Index
Oedometer
Uniaxial / Triaxial
Simple / Direct
Ring Shear
Proctor / CBR
Other Type
Sand / Gravel
Silt / Clay
Residual / Tropical
Earth Fill / Tailing
Rock / Saprolite
Other Type
Empirical
Analytical
Numerical
Instrumentation
Calibration
Chamber
Field Data
Probabilistic
Statistical
Experimental
O
b
j
e
c
t
i
v
e

o
r

M
a
j
o
r

C
o
n
t
r
i
b
u
t
i
o
n

S
e
r
r
a
t
r
i
c
e

X

X

X

X

X

C
l
a
s
s
i
f
i
c
a
t
i
o
n

o
f

s
o
i
l

s
e
n
s
i
b
i
l
i
t
y

v
i
a

C
P
T

t
e
s
t
s

S

C

B
a
u
d

&

G
a
m
b
i
n

X

X

X

X

X

X

X

E
n
h
a
n
c
e
m
e
n
t

o
f

s
o
i
l

t
y
p
e

i
n
t
e
r
p
r
e
t
a
t
i
o
n

v
i
a

P
M
T

t
e
s
t
s

I
v
s
i
c

e
t

a
l
.

X

X

X

X

X

X

X

X

M
o
d
e
l

f
o
r

l
a
n
d
s
l
i
d
e

a
n
a
l
y
s
i
s

v
i
a

s
i
t
e

c
o
r
r
e
l
a
t
i
o
n
s

Site Characterization
G M

S
t
e
e
n
f
e
l
t

e
t

a
l
.

X

X

X

X

X

X

X

X

X

X

X

X

X

X

S
i
t
e

c
h
a
r
a
c
t
e
r
i
z
a
t
i
o
n

f
o
r

t
u
n
n
e
l

d
e
s
i
g
n

K
i
m

e
t

a
l
.

X

X

E

X

X

S
t
o
r
a
g
e

o
f

C
O
2

e
m
i
s
s
i
o
n
s

i
n

s
e
d
i
m
e
n
t
a
r
y

b
a
s
i
n
s

F
e
n
t
o
n

&

H
i
c
k
s

X

X

T
e
c
h
n
i
q
u
e

f
o
r

o
p
t
i
m
u
m

s
o
i
l

s
a
m
p
l
i
n
g

N

U

S
o
m
a
s
u
n
d
a
r
a
m

e
t

a
l
.

X

X

L
S

X

X

X

X

X

S
i
t
e

t
e
c
h
n
i
q
u
e
s

t
o

c
h
a
r
a
c
t
e
r
i
z
e

a
n
d

a
n
a
l
y
z
e

d
e
b
r
i
s

f
i
l
l
s

J
a
c
q
u
a
r
d

e
t

a
l
.

X

X

X

X

X

X

X

X

D
e
t
a
i
l
s

o
f

a

n
e
w

h
i
g
h

v
o
l
u
m
e

P
M
T

p
r
o
b
e

R
i
t
o

&

E
m
u
r
a

X

X

X

X

X

X

X

D
e
t
a
i
l
s

o
f

a

n
e
w

h
i
g
h

d
e
p
t
h

s
a
m
p
l
e
r

a
n
d

p
i
e
z
o
m
e
t
e
r

K
a
y
s
e
r

e
t

a
l
.

X

X

N
S
X

X

X

X

X

S
c
o
u
r

e
v
a
l
u
a
t
i
o
n

f
o
r

p
i
e
r
s

v
i
a

n
e
w

i
n
-
s
i
t
u

p
r
o
b
e

G
i
a
c
h
e
t
i

e
t

a
l
.

X

X

X

X

X

X

X

D
e
v
e
l
o
p
m
e
n
t

o
f

s
e
i
s
m
i
c

S
P
T

f
o
r

r
e
s
i
d
u
a
l

s
o
i
l
s

F
r
o
s
t

&

M
a
r
t
i
n
e
z

X

X

X

X

X

X

X

C
P
T
u

u
p
g
r
a
d
e

w
i
t
h

a

n
e
w

m
u
l
t
i

s
e
n
s
o
r

d
e
v
i
c
e

Technological
Advances
N

I

M
o
n
n
e
t

X

X

P

X

X

X

X

X

X

D
e
v
e
l
o
p
m
e
n
t

o
f

a

n
e
w

e
n
h
a
n
c
e
d

P
M
T

p
r
o
b
e

Y
a
s
u
f
u
k
u

e
t

a
l
.

X

X

X

X

X

X

X

X

X

R
a
t
i
o
n
a
l

u
s
e

o
f

i
n
-
s
i
t
u

a
n
d

l
a
b
.

t
e
s
t
s

f
o
r

f
o
u
n
d
a
t
i
o
n
s

C
a
o

e
t

a
l
.

X

S
S

X

N
S
X

X

X

X

X

X

F
i
e
l
d

i
n
s
t
r
u
m
e
n
t
a
t
i
o
n

a
n
d

r
e
s
u
l
t
s

o
f

a

p
i
l
e

c
u
r
t
a
i
n

w
a
l
l

H
o
k
m
a
b
a
d
i

e
t

a
l
.

S
T

X

X

X

D
y
n
a
m
i
c

s
o
i
l
-
s
t
r
u
c
t
u
r
e

a
n
a
l
y
s
e
s

f
o
r

p
i
l
e
s

D

I

A
m
o
r
o
s
o

e
t

a
l
.

X

X

X

X

X

X

X

X

X

X

X

P
o
t
e
n
t
i
a
l

u
s
e

o
f

S
D
M
T

i
n

a

r
e
a
l

c
a
s
e

i
n
v
e
s
t
i
g
a
t
i
o
n

H
a
z
a
-
R
o
z
i
e
r

e
t

a
l
.

S
S

X

X

I
m
p
r
o
v
e
m
e
n
t

o
f

t
h
e

b
e
h
a
v
i
o
r

o
f

a

s
o
i
l

f
o
u
n
d
a
t
i
o
n

S
v
i
n
k
i
n

S
S

X

X

X

I
s
s
u
e
s

o
n

g
r
o
u
n
d

v
i
b
r
a
t
i
o
n

b
y

p
i
l
e

d
r
i
v
i
n
g

M
a
t
e
s
i
c

e
t

a
l
.

X

X

X

L
T

N
S
X

X

X

X

X

F
i
e
l
d

t
e
s
t
s

t
o

m
o
n
i
t
o
r

t
h
e

f
o
u
n
d
a
t
i
o
n

o
f

o
i
l

t
a
n
k
s

J
e
o
n

&

M
i
m
u
r
a

X

X

X

X

X

S
o
i
l

f
o
u
n
d
a
t
i
o
n

d
e
f
o
r
m
a
t
i
o
n

o
f

a
n

o
f
f
s
h
o
r
e

a
i
r
p
o
r
t

C
h
o
u

e
t

a
l
.

X

X

F
i
e
l
d

s
u
r
v
e
y

o
f

a
f
f
e
c
t
e
d

r
e
g
i
o
n

a
f
t
e
r

M
o
r
a
k
o
t

t
y
p
h
o
o
n

L
i
n

e
t

a
l
.

X

X

X

X

X

S
o
i
l

e
r
o
s
i
o
n

v
i
a

m
u
l
t
i
s
c
a
l
e

s
e
d
i
m
e
n
t

m
o
n
i
t
o
r
i
n
g

t
o
o
l

A
l
-
S
a
o
u
d
i

e
t

a
l
.

X

N
S

G
S

X

G
e
o
t
e
c
h
n
i
c
a
l

p
r
o
p
e
r
t
i
e
s

o
f

g
y
p
s
e
o
u
s

s
o
i
l
s

v
i
a

l
a
b
.

t
e
s
t

S
h
u
l
y
a
t
i
e
v

e
t

a
l
.

L
T
X

N
S

X

X

X

P
a
r
a
m
e
t
e
r
s

f
r
o
m

f
i
e
l
d

l
o
a
d

t
e
s
t
s

o
n

b
a
r
r
e
t
t
e

t
y
p
e

p
i
l
e
s

C
h
e
n

e
t

a
l
.

X

X

X

S
t
u
d
y

o
f

c
u
t
t
a
b
i
l
i
t
y

i
n
d
e
x

f
o
r

t
u
n
n
e
l

e
x
c
a
v
a
t
i
o
n

Geotechnical Analysis and
Behavior
F

C

B
e
l
l
a
t
o

e
t

a
l
.

X

X

X

X

X

A
s
s
e
s
s
m
e
n
t

o
f

c
u
t
t
e
r

s
o
i
l

m
i
x
i
n
g

s
a
m
p
l
e
s

i
n

t
h
e

l
a
b
.

B
a
u
d

e
t

a
l
.

X

X

X

X

S
h
e
a
r

m
o
d
u
l
u
s

d
e
g
r
a
d
a
t
i
o
n

a
s
s
e
s
s
m
e
n
t

v
i
a

P
M
T

t
e
s
t
s
B
e
n
z

e
t

a
l
.

X

X

X

X

X

X

X

X

N
e
w

i
n
t
e
r
p
r
e
t
a
t
i
o
n

a
p
p
r
o
a
c
h

f
o
r

P
a
n
d
a

p
e
n
e
t
r
o
m
e
t
e
r

N
i
s
h
i
m
u
r
a

e
t

a
l
.

X

N
S

X

X

X

E
a
r
t
h

f
i
l
l

i
n
v
e
s
t
i
g
a
t
i
o
n

u
s
i
n
g

p
r
o
b
a
b
i
l
i
t
y

a
n
a
l
y
s
e
s

P
o
u
l
s
e
n

e
t

a
l
.

X

X

X

X

X

I
n
f
l
u
e
n
c
e

o
f

C
P
T

p
e
n
e
t
r
a
t
i
o
n

r
a
t
e

i
n

s
i
l
t
y

s
o
i
l
s

G
a
l
a
a

e
t

a
l
.

X

X

X

X

X

X

X

X

X

H
y
d
.

c
o
n
d
u
c
t
i
v
i
t
y

d
e
t
e
r
m
i
n
a
t
i
o
n

o
f

g
l
a
c
i
a
l

d
e
p
o
s
i
t
s

P
h
o
o
n

&

C
h
i
n
g

X

X

F
V

X

X

X

X

X

X

I
n
-
s
i
t
u

p
a
r
a
m
e
t
e
r
s

v
i
a

r
e
l
i
a
b
i
l
i
t
y
-
b
a
s
e
d

a
p
p
r
o
a
c
h

N

T

M
o
t
a
g
h
e
d
i

e
t

a
l
.

X

X

X

X

X

X

X

X

C
P
T
u

s
t
r
e
n
g
t
h

v
a
l
u
e
s

v
i
a

c
a
p
a
c
i
t
y
-
b
a
s
e
d

e
q
u
a
t
i
o
n

T
u
m
a
y

e
t

a
l
.

X

X

X

X

X

X

X

O
r
g
a
n
i
c

c
o
n
t
e
n
t

a
s
s
e
s
s
m
e
n
t

f
o
r

s
e
d
i
m
e
n
t
a
r
y

s
o
i
l
s

M
u
l
a
b
d
i
c

X

X

X

X

X

X

X

X

X

X

C
h
a
r
a
c
t
e
r
i
z
a
t
i
o
n

o
f

a

c
o
m
p
a
c
t
e
d

d
a
m

v
i
a

i
n
-
s
i
t
u

t
e
s
t
s

Z
a
b
i
e
l
s
k
a
-
A
d
a
m
s
k
a

X

X

F
A

X

X

A
s
s
e
s
s
m
e
n
t

o
f

a

c
o
m
p
a
c
t
e
d

s
o
i
l

v
i
a

C
B
R

t
e
s
t
s

C
h
a
p
u
i
s

X

X

X

X

X

X

X

X

S
c
a
l
e

e
f
f
e
c
t
s

i
n

t
h
e

p
e
r
m
e
a
b
i
l
i
t
y

o
f

s
a
n
d
y

a
q
u
i
f
e
r
s

M
l
y
n
a
r
e
k

e
t

a
l
.

X

X

X

X

X

X

X

X

D
e
f
o
r
m
a
t
i
o
n

m
o
d
u
l
i

f
r
o
m

j
o
i
n
t
e
d

C
P
T

&

D
M
T

t
e
s
t
s

L
i
u

e
t

a
l
.

X

X

X

X

X

X

X

P
r
a
c
t
i
c
e

a
n
d

c
o
r
r
e
l
a
t
i
o
n
s

o
f

C
P
T
u

t
e
s
t
s

i
n

C
h
i
n
a

E
s
p
i
n
a
c
e

e
t

a
l
.

X

X

N
S

X

X

X

X

C
o
n
t
r
o
l

o
f

t
a
i
l
i
n
g

d
a
m
s

w
i
t
h

t
h
e

P
a
n
d
a

p
e
n
e
t
r
o
m
e
t
e
r

Soil and Rock Properties
P

E

H
a
n
z
a

&

S
h
a
h
i
e
n

X

X

X

X

X

X

X

C
o
r
r
e
l
a
t
i
o
n
s

o
n

d
r
a
i
n
e
d

c
o
m
p
r
e
s
s
i
b
i
l
i
t
y

p
a
r
a
m
e
t
e
r
s

S
C
=
S
o
i
l

C
l
a
s
s
i
f
i
c
a
t
i
o
n
,

G
M
=
G
e
o
t
e

S
P
T
=
S
t
a
n
d
a
r
d

P
e
n
.

T
e
s
t
,

B
P
T
=
B
e
c
L
T
=
L
o
a
d

T
e
s
t
c
h
n
i
c
a
l

M
o
d
e
l
i
n
g
,

N
U
=
N
e
w

U
s
e
s

o
f

I
n
-
S
i
t
u

T
e
c
h
n
o
l
o
g
y
,

N
I
=
N
e
w

I
n
-
S
i
t
u

T
o
o
l
s
,
,

D
I
=
D
e
s
i
g
n

I
m
p
r
o
v
e
m
e
n
t
,

F
C
=
F
i
e
l
d

C
o
n
d
i
t
i
o
n
s
/
S
i
t
e

P
e
r
f
o
r
m
a
n
c
e
,

N
T
=
N
e
w

T
h
e
o
r
e
t
i
c
a
l

A
d
v
a
n
c
e
s
,

P
E
=
P
a
r
a
m
e
t
e
r

E
v
a
l
u
a
t
i
o
n
.

k
e
r

P
e
n
.

T
e
s
t
,
,

S
W
S
=
S
w
e
d
i
s
h

W
e
i
g
h
t

S
o
u
n
d
i
n
g
,

C
P
T
=
C
o
n
e

P
e
n
.

T
e
s
t
,

D
M
T
=
D
i
l
a
t
o
m
e
t
e
r

P
e
n
.

T
e
s
t
,

D
C
P
=
D
y
n
a
m
i
c

C
o
n
e

P
r
o
b
i
n
g
,

L
S
=
L
a
r
g
e

S
c
a
l
e

D
e
n
s
i
t
y
,

F
V
=
F
i
e
l
d

V
a
n
e

T
e
s
t
,

S
S
=
S
t
r
e
s
s

S
t
r
a
i
n

S
e
n
s
o
r
s
,

,

C
B
R
=
C
a
l
i
f
o
r
n
i
a

B
e
a
r
i
n
g

R
a
t
i
o
,

P
=
P
e
r
m
e
a
b
i
l
i
t
y
,

S
T
=
S
h
a
k
i
n
g

T
a
b
l
e
,

E
=
E
l
e
t
r
o
r
e
s
i
s
t
i
v
i
t
y
,

G
=
G
e
o
p
h
y
s
i
c
s
,

N
S
=
L
a
b

T
e
s
t

a
d
o
p
t
e
d

b
u
t

N
o
n

S
p
e
c
i
f
i
e
d
,

G
S
=
G
y
p
s
e
o
u
s
S
o
i
l
,

F
A
=
F
l
y

A
s
h
.
474
th
th
4

Figure 5. The Koken wire line sampling system (Rito & Emura).
The paper from Kayser et al describes an approach to assess
soil scour potential through the use of the In-Situ Erosion
Evaluation Probe (ISEEP), which is advanced by water jetting.
Soil erosion parameters were assessed for silty sand in terms of
a critical stream power (critical shear stress and detachment rate
coefficient). Scour depths around a circular bridge pier were
also computed using ISEEP data, and compared with an
empirical approach available in literature.
Giacheti et al briefly describes a test which associates the
up-hole technique to the SPT, the seismic SPT (Figure 6).
This hybrid test allows the determination of the maximum shear
modulus (G
0
) together with the N value in a unique test. The
paper also presents and discusses cross-hole, down-hole, SCPT
and SPT test data for a Brazilian tropical sandy soil to
emphasize the advantage of using the interrelationship between
the small strain stiffness (Go) and the ultimate strength (N
value) to identify and characterize different soil behaviors.
Case with
geophones
H1
H2
H3
Hi
DAQ
System
Manual SPT Equipment
Trigger &
Anvil
L1
L2
L3
Li
1
2
3
i

Figure 6. S-SPT test and a seismic refracted path (Giacheti et al).
Frost & Martinez enhances the well-established cone
penetration test with an extra multi-sleeve penetration
attachment (Figure 7). The new CPT probe incorporates a series
of friction sleeves with varying surface textures and a torsional
load sensing capabilities along with a series of pore pressure
sensors, in addition to the standard smooth friction sleeve and
pore pressure sensor located behind the tip. They advocate that
the multiple measurements made with this device allow it to
provide a new insight into the characterization of soil types,
besides of establishing relations between stratigraphic variations
and in-situ shear strength with the texture height of the sleeves.
The authors really consider that the multi-sleeve technology
CPT offers significant benefits over other devices to measure
the mechanical response of soils.
f
s
fa1
u
2
f
a2
f
a3
f
a4
Attachment
Digital
Housing
Digital Housing
u
a1
ua0
u
a4
u
a3
u
a2
q
c
Friction Sleeve
Pore Pressure
Tip Load
Dual Axis
Inclinometer
Digital
Board
Attachment
Digital
Boards
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Mandrel
Mandrel
Mandrel
Mandrel
Mandrel
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Attachment
Sleeve
Mandrel
Replaceable
Attachment
Friction
Sleeve
Attachment
Individual
Load Cell
Attachment
Individual
Piezo Sensor
(0.67)
(0.81)
(0.88)
(1.07)
(1.14)
(1.33)
(1.40)
(1.59)
(1.66)
f
s
fa1
u
2
f
a2
f
a3
f
a4
Attachment
Digital
Housing
Digital Housing
u
a1
ua0
u
a4
u
a3
u
a2
q
c
Friction Sleeve
Pore Pressure
Tip Load
Dual Axis
Inclinometer
Digital
Board
Attachment
Digital
Boards
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Mandrel Mandrel
Mandrel Mandrel
Mandrel Mandrel
Mandrel Mandrel
Mandrel Mandrel
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Attachment
Sleeve
Mandrel
Attachment
Sleeve
Mandrel
Replaceable
Attachment
Friction
Sleeve
Replaceable
Attachment
Friction
Sleeve
Attachment
Individual
Load Cell
Attachment
Individual
Load Cell
Attachment
Individual
Piezo Sensor
Attachment
Individual
Piezo Sensor
(0.67)
(0.81)
(0.88)
(1.07)
(1.14)
(1.33)
(1.40)
(1.59)
(1.66)

Figure 7. The multi-piezo-sleeve friction penetrometer along with a
standard CPT probe (Frost & Martinez).
Monet presents a new in-situ testing device called the
Geomechameter, i.e. an evolution of the pressuremeter. This
new device uses the forces generated by water flow around the
probe. The hydraulic flow allows the control of the level of
vertical stress at the test depth. The influence of this stress is
hence taken into account in the test interpretation. The new
probe can also evaluate the soil permeability and sensibility to
erosion. It was validated by direct comparison with mechanical
properties from triaxial tests and permeability values from
Lefranc type injection tests.
4. GEOTECHNICAL ANALYSIS AND BEHAVIOR
Fourteen papers in this Conference Session were grouped in the
topic of geotechnical analysis and behavior; four of them dealt
with design improvement and the other ten addressed field
conditions and/or site performance.
4.1. Design improvement
The paper from Yasufuku et al. emphasizes the importance of
integrating the geotechnical investigations with pile foundation
design. Figure 8 shows the policy and concept of geotechnical
investigation & design for the studied case, i.e. the construction
of the connecting bridge for New-Kitakyushu airport. A rational
method for evaluating the pile bearing capacity was presented
which reflected the soil characteristic values and the geological
environmental history. They concluded that field and laboratory
investigations with a reasonable geotechnical consideration
sharply decreased the total cost of the bridge in the studied case.
The paper from Cao et al. studied the performance of a deep
excavation in downtown Toronto. They presented field
measurements of soldier pile walls installed into clayey soils
and shaly rock. The authors assessed the method of deducing
wall bending moments from inclinometer measurements, among
other aspects. The paper provides recommendations for such
walls when designed in similar geotechnical conditions.
The paper from Hokmabadi et al. studies the seismic
response of superstructures on soft soils. Shaking table tests and
three dimensional numerical simulations using FLAC3D were
carried out to investigate the influence of the soil-pile-structure
interaction on the seismic response of a 15-storey moment
resisting building, supported by end-bearing pile foundations.
The authors observed a good agreement between the numerical
predictions and the experimental data confirming the reliability
of the numerical approach.
475
th
5
Design of foundation
Assessments based on
Geotechnical considerations
Investigation design
Select of possible
models
Decision of sort and
number of field & lab.
tests
Modeling of ground
Careful selection of soil parameters
Determination of adequate model
Reconsideration of
safety factors et al.
Implementation of rational and
Economical design in total
Verification by
site investigations
Full scale load tests
Field observations
(Collaboration)
Implementation of
site investigation
(Feedback) (Reassessment)

Figure 8. Collaboration of geotechnical investigations with design
(Yasufuku et al).
Amorosos et al. paper presents a case history emphasizing
the use of the seismic dilatometer (SDMT) as a powerful site
investigation tool on the restoration design of an historical
building which was damaged by the 2009 LAquila earthquake.
The investigation of the foundation also included boreholes and
laboratory cyclic simple shear tests. The paper presented the
interpretation of SDMT for determination of soil profiling,
shear wave velocity, constrained modulus and horizontal stress
index, which when, combined with lab data, allowed a better
understanding of the buildings response during the earthquake.
4.2. Field condition and/or site performance
Haza-Rozier et al. study the behavior of a soil foundation
improved by rigid columns to support wind turbines. This
foundation was fixed on a rigid slab, lying on a granular layer,
improved by 84 rigid columns. The authors monitored the
structure behavior during excavation, machine construction, and
over a period of time for the working service of the wind
turbine. They observed that the working platform induced an
important confinement of the columns heads with subsequent
small levels of displacement.
Svinkins paper discusses the controversial and
contradictory evaluations of ground vibrations from pile driving
theories. He pointed out that pile driving is a powerful and
wide-spread source of construction vibrations which may
detrimentally affect adjacent or remote structures. The paper
thus presented several issues in the assessment of ground
vibrations generated by pile driving.
The paper from Matei et al. presents a case history with
the use of hydro test results for designing steel tanks on
improved ground with 660 stone columns. The authors
described the conducted hydro tests as part of a technical
monitoring assessment from all elements of the tank structure.
The paper presents and discusses all experimental data and
states that they could be wisely used to improve the tank design.
Jeon and Mimura present elasto-viscoplastic FEM analyses
to assess the long-term deformation of a reclaimed island over a
Pleistocene foundation from the adjacent construction of an
offshore (twin) airport. It is a numerical modeling paper where
simulation was compared to instrumentation results. The
authors introduced the concept of mass permeability to model
the excess pore water pressure dissipation and concluded that it
functioned well to assess the long-term deformation of the
foundation, including the interactive construction behavior.
Chous et al. paper discusses survey results of damaged
areas after a flood disaster caused by the 2009 Morakot
Typhoon in Taiwan. A comprehensive site survey was
conducted after the flood disaster and ten failure mechanisms
were identified depending on the different geological
environments. The paper presented the site survey observations,
analyzed the causes and mechanisms of failures, and drafted
strategies and suggestions for the restoration projects.
The paper from Lin et al. uses a multi-scale sediment
monitoring device to assess the remediation effectiveness on a
watershed reservoir after sedimentation processes were
originated by the same typhoon cited on Chou et al. It is stated
that it caused unprecedented landslide and sediment-related
disasters in mountain areas of the Tsengwen reservoir
watershed, drastically reducing its storage capacity. Hence, the
paper describes the method and how to systematically study and
analyze soil erosion and landslide areas with the aid of sediment
accumulation trapping dams and aforementioned device.
Al-Saudi et al. is another paper that deals with a non-text
book type geomaterial: gypseous soils, another problematic
soil given its intrinsic characteristics. According to the authors,
it covers about 20 to 30 % of total Iraq area. An important
characteristic of this soil is the collapsibility, a sudden and large
volumetric strain when exposed to water. Proposals for soil
treatment are presented, focusing on the control of settlement by
reducing or even preventing humidity changes within the soil
foundation.
Shulyatievs et al. paper presents a case study related to the
construction of the Okhta-center high-rise tower in St.
Petersburg. Static load tests on real scale barrette pile types
were carried out to adjust the design soil parameters. The paper
also presents a comparison between the derived bearing
capacity values and those from Russian and foreign building
codes. The authors concluded that pile tests are an effective way
to calibrate design parameters for usage in real case designs.
The paper from Chen et al. presents a generalized
(dimensional analysis type) solution to be used into
underground geological-mechanical interaction excavation
problems. The model groups the geological characteristics into
three categories: brittle (rock-like), ductile (soil-like), and
brittle-ductile (gravel-like), with respect to thrust and force
cuttings. Two case histories are presented to validate the
approach to assess the efficiency of a tunnel cutting machine.
Bellatos et al. paper presents a case study to discuss the
quality control of Cutter Soil Mixing (CSM), i.e., a relatively
new deep mixing method suitable for various types of ground
improvement. The materials and the testing program were
described in the paper. The obtained results under an innovative
experimental apparatus underline the influence of the physical,
and chemical, characteristics of the natural soil on the strength
gain of the stabilized materials.
5. SOIL AND ROCK PROPERTIES
In this Conference Session, fifteen papers were selected to be
part of this main topic where seven of them presented new
theoretical advances as a major subtopic and eight dealt with the
evaluation of geotechnical parameters.
5.1. New theoretical advances
The paper from Baud et al. discusses stress-strain hyperbolic
curves obtained with a self-boring Mnard PMT test. The
authors determined E-moduli values by assimilating the
pressure-volume plot of a Mnard PMT to a 2
nd
degree
hyperbolic arc. The self-boring Mnard PMT tests were carried
out using a self-bored steel slotted tube implemented either by
the STAF
technique, or by the ROTOSTAF
method. The
authors derived the hyperbolic best fit of the plotted readings to
obtain an original equation of the radial borehole expansion, =
f(G
0
, p
o
, p
LM
, P
L
). After that, they derived the tangent modulus
G
t
for each reading and the corresponding G
t
/G
0
ratio as a
476
th
th
6
function of , and similarly the secant modulus G
s
. The authors
affirmed that their variation can be well compared with those
given by the usual hyperbolic model, except for cases with very
small initial strains. They concluded that the results are similar
to those obtained by laboratory tests and geophysical surveys.
Benzs et al. paper presents the recent developments
achieved on the Panda 3
dynamic penetrometer to improve its

use for site characterization. This new improvement is
schematically represented in Figure 9 including a typical test
result. According to the authors the load-settlement
p
-s
p
curve
can be derived from the measurement and decoupling of sonic
waves created by each impact of the penetrometer, which allows
the determination of the strength and deformation parameters.
The paper presented calibration chamber test data for two
different soils to validate the given results. It was observed a
good repeatability and sensibility to the soil conditions. The
authors compared the results with those obtained by triaxial and
oedometer tests and also found a good agreement for sands.
This new test is now currently used in the field to improve the
derivation of geotechnical soil parameters via site derived load-
settlement Panda curves.

Figure 9. Schematic representation of Panda 3
dynamic penetrometer
with a typical test result (Benz et al).
The paper from Nishimura et al. presents the use of the
Swedish Weight Sounding (SWS) test with the objective of
making a diagnosis of man made earth-fills, hence increasing
their lifetime especially because their shear strength is
generally required for investigations with this scope. The study
is justified by the existence of several earth-fill dams for farm
ponds in Japan, with some of them under final life stages.
Although the strength can also be predicted by the SPT N-
values, the authors used the SWS test as a simple method for
obtaining the spatial distribution of the N-values in short
interval exams. The paper also presented an indicator simulation
(geostatistical) method to interpolate the spatial distribution of
derived N-values. The results are used to determine degraded
regions within existing embankments. The shear strength
parameter was derived through the empirical correlation with
the N-values, and the reliability analysis of the embankments
was conducted considering the variability of the internal friction
angle of the material.
The paper from Poulsen et al. shows how a change in cone
penetration rate affects all cone penetration measurements in a
silty soil. The authors emphasized the fact that for the standard
rate of penetration (20 mm/s) it is generally accepted that
undrained penetration occurs in clay, while it is drained in
sands. Data from 15 field cone penetration tests with varying
penetration rates were conducted at a sandy silt test site. Figure
10 depicts the pore pressure and cone resistance at depths
ranging from 4.5 to 11.4 m for CPTs conducted with variable
penetration rates (60 and 0.5 mm/s can be observed). The CPT
conducted with a penetration rate of 0.5 mm/s corresponds to
fully drained penetration conditions, since the measured pore
pressure is close to u
0
. On the other hand, the CPT conducted
with a penetration rate of 60 mm/s corresponds to undrained or
partially drained conditions. The authors did not observe any
correlation between sleeve friction and cone penetration rates.
They concluded that a correlation between the penetration rate,
the cone resistance, and the derived excess pore pressure, do
exist. They have also suggested an approach to determine when
the penetration is partially drained or not, and how to convert it
into a fully drained or undrained condition, hence changing
derived geotechnical parameters.

Figure 10. a) Comparison of the pore pressure and b) cone resistance
carried out with penetration rates of 60 and 0.5 mm/s, with 3 CPTs test
for each rate (Poulsen et al.).
Galaa et al. present a paper describing a methodology for
establishing more representative design values for the hydraulic
conductivity (K) of glacial deposits, particularly when
performing large scale subsurface investigations for tunnels.
They justify their study given the known glacial deposit
heterogeneities and the difficulties to determine proper design
values for K. The authors affirm that conventional pumping
tests can not provide reliable design parameters due to their
small zone of influence, and inherent variable nature of glacial
deposits. Hence, the paper describes a subsurface investigation
which involved 400 boreholes, 88 slug tests and 16 pumping
tests. The authors established a correlation between K from the
field tests (K
field
) and K calculated by the Kozeny-Carman
formula (K
KC
). They observed that the Kozeny-Carmen formula
with the incorporation of a site specific correlation factor
predicted K values ranging between 1/3 to 3 times the K
field

values. The calculated and measured K values were used to
form a statistical analysis of this parameter, and to provide a
more reliable design number for dewatering problems.
Phoon & Ching present a paper using a statistical approach
for a better interpretation of the geotechnical data when
considering soil variability. The paper presented the concept of
a virtual site with the purpose of emulating site investigation
efforts as realistically as possible. The authors affirmed that in
the present time, it is still not possible to emulate every aspect
of a real site deposit. So, the scope was to reproduce the
information content arising from a typical mix of laboratory and
field tests conducted at a site with the aim of estimating
undrained shear strengths (s
u
) for clays and friction angles (')
for sands. However, the development of a virtual site does not
replace the site investigation need, but it quantifies the
uncertainty in the derived s
u
and design values by
incorporating into the analyses the effect of either higher quality
or larger numbers of testing results.
Motaghedi et al. present a new analytical method to predict
cohesion (c) and friction angle () using q
c
, u and f
s
from the
piezocone test, considering the bearing capacity mechanism of
failure at the cone tip and a direct shear failure along the
penetrometer sleeve. The authors state that one of the
advantages of this method is the improvement of the accuracy in
the case of (eventually) using erroneous data related to all three
outputs from the CPTu test. The paper presented laboratory test
results, together with two sets of nonlinear equations derived by
the proposed approach and existing correlations for both c and
parameters. The authors state that the obtained by current
techniques is relatively higher than real measured values.
However, when adopting the advocated method, the
comparisons indicate a good consistency with lower scatter.
477
th
7
5.2. Parameter evaluation
The paper from Tumay et al. discusses the challenge for the
effective identification of organic content in the soil based on
traditional CPT and CPTU methodologies. It is very important
to overcome this interpretation limitation since the cone is a
popular and handy tool for subsurface investigations and soil
characterization. The paper presents a comprehensive
CPT/CPTU-based organic content identification method using a
probabilistic soil classification system. The paper describes the
probabilistic method, which employs a non-traditional modeling
approach that takes the uncertainty of the correlation between
soil composition and soil behavior into account. The authors
affirmed that the use of the compositional soil classification (U)
and in-situ behavior (V) indexes for organic profiling improves
the capability of determining organic material at any given
depth. A detailed description of the proposed methodology and
the discussion of its effective application are included in the
paper.
Mulabdics paper presents the use of penetration testing
devices, including the CPT and SDMT, for site characterization
of a compacted earth dam. This is a case study of a small earth
dam for which the remediation work was necessary given
construction errors and the possible damage to the earth
structure during the filling stage of the reservoir. The site
investigation campaign consisted of drilling boreholes and
carrying out in-situ tests (4 CPTs and 3 SDMTs) along the crest
of the dam, complemented with laboratory tests. The paper
focused on assessing the potential of these in-situ tests in
describing physical and mechanical properties of the compacted
(man-made) clay strata, since the traditional interpretation
methods were developed for natural soils. The authors
concluded that both CPT and SDMT clearly detected the
inhomogeneous clay conditions. They also showed remarkable
repeatability and proved to be valuable tools in characterizing
the embankment quality, both in terms of non homogeneity and
of physical and mechanical properties.
Zabielska-Adamska & Sulewska present the use of both
static (classic) and dynamic CBR methods to establish
relationships between the bearing ratio and degree of
compaction of fly ash. The objective was the use of the
compaction degree, and also the California Bearing Ratio, as an
indicator of the soil bearing capacity in compacted material. The
dynamic CBR test is described in the paper, where fly ash
samples were compacted by the standard and modified Proctor
methods without soaking to replicate field conditions during
earth structure construction. Test results indicate that both the
dynamic CBR as well as the classic CBR are closely connected
with the characteristics of compaction, and can therefore be
used to assess the compaction of fly ash and cohesive soils. The
authors suggested that the dynamic CBR test should be widely
used as an alternative way to the classical method of quality
control to assess the subgrade capacity of the soil.
The paper from Chapuis discusses scale effects in the
permeability of sandy aquifers. The authors initial hypothesis is
that the large-scale tests are more likely to meet preferential
flow paths, so yielding larger K values than small-scale tests,
which may be viewed as some sort of scale effect. In the paper,
the small scale was simulated via lab soil samples, the middle
scale from field permeability tests, and the large scale with site
pumping tests. The paper presents and discusses some few real
case studies, observing that for all of them the K distributions
provided consistent images of the aquifers. It was finally
concluded that scale effect was not of importance for the test
interpretation in such phreatic deposits.
Mlynareks et al. paper discusses the interrelationship
between deformation moduli from CPTU and SDMT tests in
overconsolidated soils. The authors point out that glaciations in
Poland overconsolidated its deep soil layers. So, it is imperative
to take it into account in calculations of differential settlements
of structures. The paper presents deformation characteristics
estimated from CPTU and SDMT tests in clays, and focuses on
a method to identify soil preconsolidation and to establish
relationships between deformation moduli derived from CPTU
and SDMT tools. The authors concluded that the simultaneous
use of CPTU and SDMT provides a continuous picture of the
changes in stiffness of heterogeneous subsoil. They emphasized
the need for establishing specific calibration functions for each
soil type, which may be a useful tool in the construction of a
model for the subsoils rigidity based on G
0
or M
0
values.
Lius et al. paper reports the practice and development of
the piezocone test in the geotechnical engineering field of
China. In this paper, the history and current development status
of CPT and CPTu in China practice were systematically
presented. The most used (standard) cone has the 10 cm
2
tip
area, but both 15 and 20 cm
2
CPT probes are frequently used in
China. The relationship between international standardized
CPTu and Chinas CPT is based on a large data bank of testing
results related to a great number of soils. The paper presents a
comparison review of the soil characterization methods in
China, including the determination of stress history,
deformation, consolidation and permeability characteristics.
The paper from Espinace et al. presents their 10 years of
experience on the use of Panda
penetrometer to assess the

stability of Chileans tailings dams. The authors report around
40 cases of mechanical instability from tailing dams in Chile,
which were mainly due to liquefaction, slipping of banks, or
settlement. The paper presents the main results that have
allowed the proposition of a new methodology to control and
diagnose tailing dams. It is based on in-situ determination of the
geomechanical parameters (internal friction angle and density
index) using the Panda
penetrometer in order to characterize

the constituent materials and their variability. The authors also
pointed out that this methodology allows taking into account the
variability concept for stability and liquefaction risk studies
when using a probabilistic approach.
Hamza & Shahiens paper studies the compressibility
parameters of Egyptian cohesive soils via piezocone tests. The
major objective was to provide additional data on drained
compressibility parameters, focusing on the constrained
modulus (M
o
) and on the overconsolidation ratio (OCR) for
cohesive soils from geotechnical investigations at seven major
sites of the Nile delta river deposit in Egypt. Enhanced
propositions to estimate the OCR and M
o
for the studied clays
are presented, allowing settlement analyses to be done with the
proposed equations. The authors believe that the presented data
and correlations are a valuable contribution, since it improves
the current state of the art in estimating the compressibility
parameters of sedimentary soils with the CPTU test.
6. FINAL REMARKS
Site characterization using in-situ testing techniques has
considerably changed in the last two decades along with the
rapid transformation and advances of the technology, either by
the development of newer and economical electronic devices
operated by laptop computers or by new mathematical and
software approaches based on multi-variable, statistical or
probabilistic calculations. Besides of such remarkable
accomplishments, the traditional old fashion (past century.)
laboratory and site investigation methods are still widely in use,
sometimes as the preferential or unique available method. It was
clear from aforementioned review that, on the 21
st
century, the
proper site investigation, material characterization and soil
behavior prediction for the geotechnical design cannot solely
rely in one isolated test technique, or on simple local
unadjusted correlations that are probably not universally valid.
Higher sensorial levels of testing tools and combined
investigation procedures are surely now available that can be
478
th
th
8
applied on a more regular basis, especially for large or
important jobs. Improved interpretation methods or background
geotechnical theories, advanced computer software codes, and
more prominent hardware equipment, should further be
explored in research as well as in practical in-situ testing
settings. This is a challenge for the overall community as a
whole, from practitioners to academicians, manufacturers, and
designers. Nevertheless it can not be accomplished without a
strong change in mentality from the geotechnical field itself,
moving from a priced-based design to perhaps a more
expensive and sound quality-based criteria.
The papers presented in this Conference Session show how
in-situ testing technology has developed, matured, and used to
tackle several geotechnical problems of difficult order, for
instance from the uncertainty in site characterization and
understanding of different geomaterials, to the challenging task
of retrieving high quality soil samples in a difficult
environment. Sometimes, in standard project cases, only
traditional tests were required and used for the site
characterization. However, in more complex or ambivalent
conditions, the usage of todays available technological
advances was surely an asset for the design. Although not
directly mentioned throughout this review, the human factor,
i.e., the good education based on solid concepts of the
geotechnical area together with the access to a free flow of
technical information and knowledge, will be the key factor for
the transformation of our field, and the future society, as we all
envisage with rational use of resources and technology, self-
sustained projects, quality based design and environmentally
safe site procedures.
7. ACKNOWLEDGEMENTS
The authors acknowledge the hard work of some of the Ph.D.
candidates from the Geotechnical Graduation Program of the
University of Braslia during the early stages of paper
compilation and description. Therefore they are indebted to the
work of the engineers Osvaldo Neto, Edgar Rincon and Raydel
Lorenzo. Additionally, this report would not be possible without
the use of the language skills from engineer Adrien Treguer,
native in French, who is a student from Clermont-Ferrand
University and fortunately happens to be at the moment in an
undergrad exchange program with the University of Brasilia.
8. REFERENCES
Al-Saoudi N.K.S.; Al-Khafaji A.N., Al-Saoudi N.K.S. Challenging
problems of gypseous soils in Iraq.
Amoroso S.; Totani F., Totani G. Site characterization by seismic
dilatometer (SDMT): the Justice Court of Chieti.
Baud J.P., Gambin M. Dtermination du coefficient rhologique de
Mnard dans le diagramme Pressiorama.
Baud J.P.; Gambin M., Schlosser F. Courbes hyperboliques contrainte
dformation au pressiomtre Mnard autofor.
Bellato D.; Simonini P.; Grisolia M.; Leder E., Marzano I.P. Quality
control of Cutter Soil Mixing (CSM) technology A case study.
Benz M. A.; Escobar E.; Gourvs R.; Haddani Y.; Breul P., Bacconnet
C. Mesures dynamiques lors du battage pntromtrique
Dtermination de la courbe charge enfoncement dynamique en
pointe.
Cao L.F.; Peaker S.M., Ahmad S. Performance of a deep excavation in
downtown Toronto.
Chapuis R.P. Permeability scale effects in sandy aquifers: a few case
studies.
Chen L.; Chen Y.C.; Chen W.C., Liu H.W. A study of cuttability
Indices for tunnel penetration.
Chou J. C.; Huang C. R., Shou K. J. Survey results of damaged areas in
flood disaster of Typhoon Morakot and suggestions for restoration
projects.
Espinace R. A.; Villavicencio G. A.; Palma J.; Breul P.; Bacconnet C.;
Benz M.A .N, Gourvs R. Stability of Chileans tailings dams with
the Panda penetrometer. Experiences of the last 10

years.
Fenton G.A.; Hicks M.A. Site sampling: assessing residual uncertainty.
Frost J. D., Martinez A. Multi-sleeve axial-torsional-piezo friction
penetration system for subsurface characterization.
Galaa A.; Manzari M., Hamilton B. Hydraulic properties of glacial
deposits based on large scale site investigation.
Giacheti H. L.; Pedrini R. A. A., Rocha B. P. The seismic SPT test in a
tropical soil and the G
0
/N ratio.
Hamza M., Shahien M. Compressibility parameters of cohesive soils
From piezocone.
Haza-Rozier E.; Vinceslas G.; Le Kouby A., Crochemore O.
Comportement de la structure de sol amlior par inclusions
rigides, supportant une olienne.
Hokmabadi A.S.; Fatahi B., Samali B. Seismic response of
superstructure on soft soil considering soil-pile-structure
interaction.
Ivi T.; Ortolan ., Kavur B. Applicability of the RNK-method for
geotechnical 3D-modelling in soft rocks.
Jacquard C.; Rispal M.; Puech A.; Geisler J.; Durand F.; Cour F.;
Burlon S., Reiffsteck P. Une nouvelle sonde permettant de mesurer
sans extrapoler la pression limite pressiomtrique des sols.
Jeon B.G., Mimura M. Long-term Deformation of the reclaimed
pleistocene foundation of the offshore twin airport.
Kayser M., Gabr M. Assessment of scour potential using in-situ jetting
device.
Kim A. R.; Cho G.C.; Kwon T. H., Chang I. H. Practical reviews on
CO2
sequestration in Korean sedimentary basins and geophysical
responses of CO
2
-injected sediments.
Lin B.S.; Ho H.C.; Hsiao C.Y.; Keck J.; Chen C.Y.; Chi S.Y.; Chien
Y.D., Tsai M.F. Using multi-scale sediment monitoring techniques
to evaluate remediation effectiveness of the Tsengwen Reservoir
watershed after sediment disasters induced by Typhoon Morakot.
Matei L.; Mihaljevi I.; Grget G., Kvasnika P. The use of hydro test
results for design of steel tanks on stone column improved ground -
a case history.
Mynarek Zb.; Gogolik S., Sanglerat G. Interrelationship between
deformation moduli from CPTU and SDMT tests for
overconsolidated clays.
Monnet J. Le Gomcamtre, un nouvel essai in-situ adapt la mesure
des caractristiques hydro-mcaniques du sol.
Motaghedi H.; Eslami A., Shakeran M. Analytical approach for
determining soil shear strength parameters from CPT & CPTu data.
Mulabdic M. Use of penetration testing for determination of soil
properties in earth dam.
Nishimura S.; Shuku T., Suzuki M. Diagnosis of earth-fills and
reliability-based design.
Phoon K.K., Ching J. Construction of virtual sites for reliability-based
design.
Poulsen R.; Nielsen B. N., Ibsen L. B. Correlation between cone
penetration rate and measured cone penetration parameters in silty
soils.
Rito F., Emura T. Sampling method and pore water pressure
measurement in the great depth (-400m).
Serratrice J.F. Une mthode de classification de la sensibilit des sols au
moyen du pizocne.
Shulyatiev O.; Dzagov A.; Bokov I., Shuliatev S. Correction of soil
design parameters for the calculation of the foundation based on the
results of barrettes static load test.
Somasundaram S.; Khilnani K.; Shenthan T., Irvine J. Characterization
and modeling settlement of deep inert debris fills.
Songyu. L; Guojun. C; Anand J. P., Yanjun D. Practice and
development of the piezocone penetration test (CPTu) in
geotechnical engineering of China.
Steenfelt J.S.; Yding S.; Rosborg A; Hansen J.G., Yu R. Site
characterization of the HZM immersed tunnel.
Svinkin M.R. Controversial and contradictory evaluations in analysis of
ground vibrations from pile driving.
Tmay M. T.; Hatipkarasulu Y.; Marx E. R., Cotton B. CPT/PCPT-
based organic material profiling.
Yasufuku N.; Ochiai H., Maeda Y. Geotechnical challenge for total cost
reduction related to construction of connecting bridge with pile
foundations.
Zabielska-Adamska K., Sulewska M.J. CBR as a method of
embankment compaction assessment.
479
Challenging Problems of Gypseous Soils in Iraq
Des problmes difficiles des sols gypseux en Irak
Al- Saoudi N.K.S.
University of Technology-Baghdad-Iraq
Al- Khafaji A.N., Al- Mosawi M.J.
University of Baghdad Baghdad- Iraq
ABSTRACT: Gypseous soils are classified as one of the problematic soils due to their complex and unpredictable behaviour.
They exist in many parts of the world, concentrated mainly in arid and semi-arid regions. In Iraq gypseous soils cover about 20 to
30 % of its total area concentrated primarily on the west desert and extended to the southern parts and directed towards south
west. Gypsum soils experience sudden collapse upon exposure to water, losses of serviceability of many structures were observed
in different parts of Iraq. Extensive research was made in Iraq to investigate and understand the behavior of Gypsum soils and to
set safety limits for the collapse and suggest practical precautions during construction. The enormous amount of data collected
from different research sources revealed wide spectrum of information covering the overall performance of Gypsum soils under
different environmental and climate conditions The present paper focuses on the main geotechnical properties of gypseous soils
and their effect on the collapsible mode of failure, some practical solutions are also proposed that provide safety precautions
RSUM : Les sols gypseux sont classs comme des sols problmatiques cause de leurs comportements complexes et imprvi-
sibles. Ils existent dans plusieurs rgions du monde, principalement dans des rgions aride et semi-aride. En Irak, les sols gypseux
couvrent entre 20 et 30 % du pays et sont principalement concentrs dans le dessert de louest et stendent vers les rgions du sud
et orientes vers le sud-ouest. Les sols gypseux seffondrent soudainement lorsquils sont soumis leau et beaucoup de structures
inutilisables ont t observes dans diffrentes zones en Irak. Des recherches approfondies ont t menes en Irak pour tudier et
comprendre le comportement des sols gypseux afin de dterminer les limites avant leffondrement et de suggrer des prcautions
concrtes lors de la construction. Lnorme quantit de donnes recueillies auprs de diffrentes sources a rvl un large spectre
dinformations couvrant lensemble des performances des sols gypseux sous diffrentes conditions environnementales et climati-
ques. Le prsent document se concentre sur les principales proprits gotechniques des sols gypseux et sur leurs effets sur les
croulements de structure, enfin quelques solutions pratiques sont aussi dveloppes pour proposer des mesures de scurit.

KEYWORDS: Gypsum,collapsibility, Gypseous Soils, Problematic Soils

1. INTRODUCTION
Gypseous soils are one of the most complex materials that
challange the geotechnical engineers. Structures or dams
founded on gypseous soil may experience unpredictible defor-
mations that ultimatley may cause catostrophic failure. In iraq it
has been reported that several structures have experienced dif-
ferent patterns of cracks and uneven deformations generated
primearly from the exposion of the supporting gypseous soils to
water. It is a well known fact that gypseous soils demonstrate
high bearing capacity and very low compressibility when they
are in the dry state. On the contrary a sudden collapsibile behav-
iour was reported when the gypseous soils are exposed to wa-
ter.The collapsibility of gypseous soils results from the direct
contact of water. The dissolution of different types of salts con-
tained inside the mass of gypseous soil will generate new pores
inside the soil skeleton and loosen the cementing bonds between
the particles. This process creates a meta stable structure that fa-
cilitates the sliding of particles into a more dense state. The rate
of dissolution of gypsum depends primarily on environmental
changes in moisture content generating from fluctuation of
ground water table and /or surface water, climate changes typi-
cally temperature, permeability and state of flow conditions in
addition to the type and content of gypsum.
During the last three decads many attempts were made in
Iraq through intensive research programs set in many institution
to investigate and underestand the behaviour of gypseous soils
under various enviromental and loading conditions. The first ob-
jective of these research programs was to determine the physical
proerties of the gypseous soils and to check whether staandard
tests can be used and if not what modifications are required. Fol-
lowing that several attempts were made to determine the geo-
technical properties such as compressibility, collapsibility and
shear strength parameters under various flow and environmental
conditions and loading conditions. The tests were performed us-
ing triaxial and Rowe cells allowing soaking and leaching of the
soil samples. Plate load tests were also performed under differ-
ent soaking periods to moniter the generated deformation with
time. Numerical techniques were also used to simulate the diso-
lution process of gypsum under soaking and leaching conditions.
The abundant amount of data obtained from the lengthy re-
search programs revealed in many cases contradicting results
due to the complexity of the gypseous soils. So no regid conclu-
sions are yet been drawn.
The paper sheds the light on the distribution of gypsum in
Iraq. A summary of main physical and geotechnical properties
with emphises on the collapsibility is presented and a some re-
midied are proposed.
480
th
2. DISTRIBUTION OF GYPSEOUS SOILS IN IRAQ
Gypseous soils exist mainly in arid and semi arid regions,
concentrates in contenents like africa,central and souther asia.
Iraq is among the contries of south asia where gypsum covers
about 12 % of its total area. (FAO 1990), although more recent
study (Ismail 1994) reported that gypseous soils cover 31.7 % of
the total area of iraq.the first map demonstrating the distribution
of gypsum in iraq was presented by (Buringh 1960) indicating
five zones as shown in figure 1. The primary gypsum is located
in the extreme north area between tigres and euphraties rivers.
The second zone where primary gypsum mixed with limstone
located below and parrallel to the euphraties river extending
from the west desert to the south. The secondary gypsum is
identified in two areas, one in the north below the first zone and
one in the south west. The fourth zone is gypsiferous alluvium
extens from the north in a narrow band and gradually widened
towards the south. The fifth zone representing the non
gypsiferous soil, mainly limestone is identified in two ares one
in the north east and the other in the west desert.

Figure 1 first map of distribution of gypsum in iraq
A more refined map exhibiting the distribution of gypsum in
iraq was presented by (Al-Barrazanji 1973). He investigated
thoroughyl the type and gypsum content in different parts of
iraqand proposed the map shown in figure 2.Six zones are
distiguished according to their origin and gypsum content.Zone
one of slightly gypseous over gypsum bedrock denoted by nar-
row parallel lines taking the shape of a triangle in the upper
north of Iraq. The second zone is of moderately to highly gyp-
seous soils over gypsum and anhydrate rock denoted by wider
parallel lines, located in the north part between the Tigress and
Euphrates rivers. Zone three is gypsum desert denoted by a mesh
of small squares, located between zones one and two in the
north. Zone four contains highly gypsiferous soils on Pleistocene
terraces covering two narrow strips on the left and right of Ti-
gress River denoted by moderately dense dots. The fifth zone is
non to slightly gypseous soils denoted by parallel hashes, ex-
tends from the upper mid third of Iraq up to the Kuwaiti borders
in the south. The sixth zone is moderately to highly gypsiferous
soil associated with lime denoted but heavily condensed dots,
covering the west jazeria. The two maps comply each other in
most of their subdivisions with slight divergence in others, al-
though different terminologies have been used. Based on figure
two, if the soil in zone four is considered as non gypseous soil
that does not possess any hazardous impact then most likely 50
to 60 % of the totalarea of Iraq is covered with active gupsum.
This indicates how serious the challanges are for geotechnical
engineers when dealing with such unpredictible material.

Slightly over gypsum bedrock
Moderately to highly gypseous soils
over gypsum and anhydrate rock
Gypsum desert
Highly gypsiferous soils on Pleistocene ter-
Non to slightly gypseous soil
Moderately to highly gypsiferous associated with lime
Figure 2 Distribution of gypsum in Iraq (Al Barazanji 1973)

3. PROPERTIES OF GYPSEOUS SOILS

The Physical, chemical and geotecnical properties of gypseous
soils collected from different parts of Iraq are summarized and
discussed below.

3.1. Physical properties

The physical properties of natural gypseous soils varies consid-
erably with the amount and type of gypsum soil in addition to
the texture and constituents of the soil

3.1.1. Specific gravity

(Schultz and Cleaves 1955) reported that the specific gravity of
gypsum ranges between 2.31 to 2.33, increased to 2.95 for anhy-
drite type. Thus any increase in gypsum content of specific
gravity less than 2.33 will lead to a decrease in specific gravity
of thr soil. (Saleam 1988,Nashat 1990 and Al- Mufty 1997)
reached to the same finding unless the gypsum of the unhydrated
type.

3.1.2. Maximum dry unit weight

The results of maximum dry unit weight showed contradicting
relationship with gypsum content.(Khattab 1988 and AlDulaimy
1989) found that the dry unit weight increases with increasing
gypsum content up to a certain limit followed by a gradual drop.
On the other hand (Subhi 1987 and others) reported a decrease in
481
dry unit weight with increasing gypsum content. Further
more,(Al Heeti 1990) showed an increase in dry unit weight with
increasing gypsum content. This descrepency may be due to the
type of gypsum ( hydriate or anhydriate), type of soil and range
of gypsum content considered in the investigation.

3.1.3. Soil constiuents and texture

The samples of gypseous soils collected from different parts of
Iraq showed that gypsum exists primearly in Sandy soil and silty
sand and less in silty clay or clayey silt. The presence of appre-
ciable amount of gypsum creats problems in determining the
constituents of the soil. During sedementation test disolution of
gypsum will occur causing the flucculation of silt and clay parti-
cles.Pretreatment with water was suggested by (Al-Khashab
1981 and Mohammed 1993). EDTA was suggested by ( AL-
Khuzaie 1985 and others). Most of the treated agents used cause
distruction of bonds and most likely revealed an unreliable
amount of constituents.

3.2. Chemical properties
3.2.1 Chemical compsition of gypsum
Pure chemical proportions of gypsum as reported by (Nashat
1990) are
20.9 % combined with water
46.6 % sulphur trioxide SO
3

32.5 % calcium oxid CaO
(Majeed 2000) observed that the alkalinity increases with in-
creasing gypsum content. On the other hand the electrical con-
ductivity, cation exchange capacity and exchangeble sodium
percentage decrease with increasing gypsum content.
3.2.2. Solubility of gypsum
The most effective parameter in the general behaviour of gyp-
sous soils is the solubility. Gypsum is classified as a moderate
soluable salt. The solubility of the hydrated type in pure water is
2g/l ( Hesse 1971). Some higher values, 2.41g/l and 2.6g/l , were
reported for Iraqi gypseous soils ( Sirwan etal 1989, Seleam etal
1988)

3.2.3 Rate of dissolution of gypsum

The rate of dissolution of gypsum is responsible for the devel-
opment of cavities and sinkholes. It is very complex to be evalu-
ated as it is affected by many environmental conditions such as
temperature,source of water,time, concentration of sodium chlo-
rid and calicum sulphate etc.

3.3. Geotechnical properties
The geotechnical prperties of gypseous soils cover, compressi-
bility, collapsibility, permeability and shear strength parameters
( c and )

3.3.1. Compressibility

More than ten researchers have investigated the influence of
gypsum on the copressibility characterstics.(Al-Khashab 1981
and many others) reported a decrease in the copression index
with increasing gypsum content. It is hard to judje about the con-
tradicting results as many parameters such as the placement con-
ditions, degree of disturbance, and testing methodology. Similar
contradicting results were reported for the recompression index.
Most of the researchers demonstrated an increase in the secon-
dary compression index with increasing gypsum content. This
phenomenon is attributed to the contieous dissoltion process of
gypsum with timeas reprted by (Saleam 1988 and Nashat 1990).
The same researchers and many others showed that the coeffi-
cient of consolidation remains unchanged with increasing gyp-
sum content.

3.3.2. Collapsibility

Gypseous soils are distinguished by their collapsible behaviour
upon wetting. The term collapse potential is used to classify the
hazarduse state of collapsibility. ( Jennings and Knight 1957)
proposed a double oedometer collapse test to predict the col-
lapsibility of the foundation soil. Two identical samples are
tested, one at natural water content and the other after submerge-
ing in water for one day.The collapse potential C.P. is defined as

C.P. = e / 1+ e
o
(1)

Where
e is the difference in void raio of the two samples at a specific
stress
e
o
is the natural void ratio

The severity according to the collapse potential is shown in table
1.

Table 1. Collapse identification ( Jennings and Knight 1975)
Severity No prob-
lem
Moderate Trouble Severe Very
Severe
C.P. % 0-1 1-5 5-10 10-20 > 20

(Saleem 1988, Nashat 1990 and many others), found that the
collapse potential under a constant stress of 200kN/m
2
increases
with increasing gypsum content. The gypsum content of the
tested samples ranged between 20 - 80% revealed a moderate
type of 4 % maximum collapse potential.

3.3.3. Moduluse of deformation

Al Khafaji etal 2009 investigated the deformability of gypseous
soils through plate load tests performed on natural and soaked
soils. The tests were performed on two sites GP-GM soil and
SM soil and socking period extended to 7 -11 days under 300
mm head of water. All types of stiffness moduli were calculated,
the initial tangent moduluse, the permissible secant modulus at
half the yeild, the yeild secant modulus at the yeild and the yeild
tangent modulus after the yeild. The outcomes revealed that
soaking decresed the stiffness moduli in the range of 2 to 5 folds
for GP-GM soil and from 2 to 3.5 for SM soil. The field tests
highlights on the hazardius degree of constructing structures on
gypsious soils without awarness of the expected generated set-
tlements that may result from the contamination of water.

3.3.4. Hydraulic conductivity

Hydraulic conductivity or coefficient of permeability of gyp-
seous soils is very hard to predict. Standard constant head test on
sandy gypseous soils does not reveal reliable results as the gyp-
sum disolves during flow creating more free space for the soil
particles to reorient themselves to a closer state of packing, caus-
ing a suddent fluctuation of rate of flow during test This phe-
nomenon is very difficult to evaluate as the dissolution process
is influenced by many factores like type and amout of gypsum,
hydralic gradient, initial placement of soil sample. Attempts
were made to perform leaching permeability tests under different
stress levels using Rowe cell (Al-Kaisi 1997 and many others).
(Al- Qaissi 2001 and many others) using triaxial permeability
leaching apparatus cited that the variation in hydraulic gradient
combined with diffusion of gypson encountered serious difficul-
ties in predicted reliable values of the coefficient of permeabil-
ity.

482
th
4. PROPOSED REMEDIES FOR GYPSEOUS SOILS

The invenitable geotechnical problems associated with the aban-
dance of gypseous soils in Iraq create real challanging issues.
Based on that proposals were made for treatment of soils focus-
sing on controlling the settlement and reducing the coefficient of
permeability or preventing any contact of water between the
foundation soil and any source of water. All the proposed treat-
ments are based on elemt tests and not verified by field applica-
tions

4.1. Chemical treatment

The treatment materials proposed are basically cement, lime and
petroleuum products

4.1.1. Treatment with cement

(Khattab1986) reported that sulphate resisting cement improved
the unconfined compression strength of granular gypsified soil,
but a substaintial amount of reduction in strength and stiffness
upon immersion in water.

4.1.2. Treatment with lime

Al-Obaidy 1992 and Al-Zory 1993 showed that mixing 5 -7 %
lime with gypsous soil of 43 % gypsum content exhibited an in-
crease in strength and high resistance to leaching. It is reported
that the soil became practically impermeabile after 28 days cur-
ing.

4.1.3. Treatment with petroleum products

Various types of petroleum products such as kerosene, automo-
bile oil ,fuel oil and bitumenous materials such as S-125 and R-
250 were proposed as improvement agents for gypseous soils.
(Saleam 1988) found that treating soil of gypsum content be-
tween 40-50 % with kerosene caused a decrease in compressi-
bility and permeability by delaying the removal of gypsum.(Al-
Aqaby 2001) observed a reduction in cohesion of soil of gypsum
content between 30 -67 % upon immersion in water or kerosene.
The angle of internal friction was reduced by 6 degrees upon
soaking in kerosene.
(Al-Kaisi 1997) found that 4 % automobile oil caused a re-
duction in the coefficient of permeability by not less than ten
folds.
(AlHassany 2001)perfromed consolidation tests on two
samples of gypsum content 26 % and 51% treated with fuel oil.
The fuel oil tends to to fill the pores of soil and prevent water
perculation and hence reduce the permeability. The presence of
fuel oil also reduced the copmressibility and collapsibility.
Bitumenous materials S-125 and R-250,emulsified asphalt,
Cut-Back MC-30 were used by (Al-Morshedy 2001 and many
others). Gypseous soils treated with one of the above materials
showed reduction in coefficient of permability as well as com-
pressibility and collapsibility.

4.2. Physical treatment

(Al-Khafaji 1997) developed simple and quick equations for es-
timating the optimum water content and maximum dry unit
weight to control field compaction of soils with gupsum content
ranging between 0.5 -50 %.

5. CONCLUSIONS

Inspite of the abandant data collected concerning the
the geotechnical properties of gypsefeous soils and the attempts
to investigate and understand the behaviour of gypseous soils
under different stresses and environmental conditions.The
challanges still exist due to the scarceyt and complexity of such
natural material. No real firm solution or a general improvement
technique can be proposed. It is advised that geotechnical engi-
neers must investigate each case seperatly depending on the type
of structure, characteristics of site, environmental conditions
coupled with the engineering judgement of the consultant.

6. REFERENCES

FAO. 1990 . Management of gypsiferous soils. Food and Agricultural
Organization of United Nations Rome. Internet
http//fao.org/docrep/to323e/ro323e03.htm
Ismail H.N. 1994. The use of gypseous soils. Symposium on Gypsifer-
ous soils and their effect on structures. NCCL Baghdad. Iraq
Buringh P. 1960. Soils and soil conditions in Iraq. Ministry of Agricul-
ture . Baghdad. Iraq
Al-Barazanji A.F. 1973. Gypsiferous soils in Iraq. PhD. Dissertation.
Ghent University. Belgium.
Shultz J.R. and Cleaves A.B. Geology in engineering.John Wiley and
sons . New York
Saleam S.N.1988. Geotechnical characteristics of gypseois sandy soil in-
cluding the effect of contamination with some oil products. MSc.
Theisis . University of Technology. Baghdad.Iraq.
Nashat I.H. 1990. Engineering characteristics of some gypseous soils in
Iraq. PhD. Thesis . University of Baghdad. Iraq
Al- Mufty A.A. 1997. Effect of gypsum dissolution on the mechanical
behaviour of gypseous soils. PhD. Thesis . University of Baghdad.
Iraq
Khattab S.A. 1986. Effect of gypsum on strength of cement treated
granular soil and untreated soil. MSc. Thesis . university of Mosul.
Iraq
Al-Dilaimy F.H. 1989. Effect of gypsum content on strength and defor-
mation of remolded clayey soil. MSc.Thesis University of
Salahddin. Iraq
Subhi H.M. 1978 The properties of salt contaminated soils and their in-
fluence on the performance of rocks in Iraq. PhD. Thesis Queen
Mart College. University of London
Al-Heeti 1990. The engineering properties of compacted gypsified soil.
MSc. Thesis. University of Baghdad. Iraq
Al- Khashab M.N. 1981 Investigation of foundation soil behaviour of
Qadisiah site Mosul. MSc. Thesis. University of Mosul.
Mohammed R.K. 1993 Effect of wetting and drying of engineering char-
acteristics of gypseous soils. MSc. Thesis University of technology.
Baghdad. Iraq.
Al-Khuzaie H.M.A.1985 The effect of leaching on the engineering prop-
erties of Al-Jezirah soil.. MSc. Thesis . university of Mosul. Iraq
Majeed A.H. 2000. Data base for gypseous soils. PhD. Thesis University
of Baghdad. Iraq.
HesseP.R. 1971. A textbook of soil chemical analysis. Chemical publish-
ing co..Inc.New York pp520
Jennings J.E. and Knight K.1957. The additional settlement of founda-
tionsandy subsoil on wetting. Proceeding 4th Int. Conf. Soil me-
chanics and foundation engineering . vol.1. pp316-319
Al- Khafaji A.N, Al-Mosawi M.J., Khorshid N.S. and Al-Obaid B.M.
2007 Proceeding of the 17th ICSMGE Alexandia .Eygept pp 727-
729.
Al- Khafaji A.N. Densification of gypseous soil by compaction. Sympo-
sium on ground improvement geosystems. London
Al-Morshedy A.D. 2001 The use of cutback MC-30 for controlling the
collapsibility of gypseous soils.MSc.thesis. University of Technol-
ogy. Baghdad.Iraq.
483

Site characterization by seismic dilatometer (SDMT): the Justice Court of
Chieti
Caractrisation du site par dilatomtre sismique (SDMT): la Cour de justice de
Chieti
Amoroso S., Totani F., Totani G.
University of LAquila, Italy

ABSTRACT: A detailed investigation of several seismic dilatometer (SDMT) tests was performed in 2011 on Chieti hill to restore the
Justice Court, an historical building damaged by the April 6, 2009 LAquila earthquake. Moreover, boreholes were carried out to
investigate foundation base level and cyclic simple shear tests with double sample were realized to analyze the seismic site response.
The paper illustrates the potential of the seismic dilatometer to efficaciously approach a geotechnical problem by the interpretation
of SDMT parameters, as the shear wave velocity V
S
, the constrained modulus M and the horizontal stress index Kd. Fi- nally, the
paper combines SDMT results with laboratory data to analyze the site response of the Justice Court.

RSUM : Une tude dtaille de plusieurs sismiques dilatomtre (SDMT) tests a t ralise en 2011 sur la colline de Chieti pour
restaurer la Cour de justice, un btiment historique endommag par le tremblement de terre qui a eu lieu le Avril 6 2009 L'Aquila . En
outre, des sondages ont t effectus pour tudier le niveau de base de fondation et cycliques essais de cisaillement simple avec
chantillonnage double ont t ralises pour analyser la rponse sismique du site. Cet article montre efficacement le potentiel de la
dilatomtre sismique l'approche d'un problme gotechnique par l'interprtation des paramtres SDMTs, comme la vitesse de
l'onde de cisaillement V
S
, le module M et l'indice de contrainte horizontale Kd. Enfin, le document combine les rsultats SDMT aux
donnes de laboratoire pour analyser la rponse du site de la Cour de justice.

KEYWORDS: seismic dilatometer, horizontal stress index, shear wave velocity, site response analysis, local site effetcs.

1 INTRODUCTION

The April 6, 2009 LAquila (Italy) earthquake (M
W
= 6.3)
caused heavy damages not only in the city of LAquila basin but
also in few cities, as Chieti, approximately 100 km far from the
epicenter. In this respect, a detailed investigation of several
seismic dilatometer (SDMT) tests (Marchetti et al., 2008) in
virgin soils and inside boreholes backfilled with sand (Totani et
al. 2009), foundation boreholes and cyclic laboratory tests were
performed in 2011 on Chieti hill to restore the Justice Court, an
historical building damaged by the above mentioned earth-
quake. The geotechnical campaign allowed to characterize the
subsoil, to investigate foundation base level and to analyze the
seismic site response of this construction. In particular, the pa-
per illustrates the potential of the seismic dilatometer to effica-
ciously approach a geotechnical problem by the interpretation of
SDMT parameters, as the shear wave velocity V
S
, the constrai-
ned modulus M and the horizontal stress index Kd, even combi-
nig SDMT results with laboratory data for the evaluation of the
local site effects (e.g. topography, soil conditions) with modo-
dimensional (1D) and bidimensional (2D) seismic site response
analyses.

2 GEOTECHNICAL INVESTIGATION ON CHIETI HILL

A detailed investigation of eleven SDMT tests, six in virgin
soil, each 10-20 m in depth, and five inside boreholes backfilled
with sand, each 30-50 m in depth, were performed in 2011 on
Chieti hill to restore the Justice Court, an historical building
damaged by the April 6, 2009 LAquila earthquake. Moreover,
boreholes were carried out to investigate foundation base level
and cyclic simple shear tests with double sample were realized
to analyze the seismic site response. The historical centre was
built on sandy and arenaria deposits (45 m in depth), while
moving towards the bottom of the slope the colluvial cover start
to emerge over the OC silty clay, as shown in Figure 1 together
with the shear wave velocity V
S
profiles.
Figure 2 emphasizes the main reason of the damage due to the
seismic action on the construction. As shown by the inspection,
the structure appears to be considerably fissured in its Southern
part, while in the Northern area it seems to be intact (Figure 2a).
This aspect can be justified referring to the four SDMTs per-
formed along the perimeter of the building (Figure 2b). SDMT1
and SDMT4, as well as SDMT3 and SDMT2 profiles, can be
coupled. In fact, in the Northern part of the Justice Court the
constrained modulus M reaches on average values over 100
MPa and the horizontal stress index Kd indicates OC soils. In-
stead, in the Southern area, until about 8.00 m in depth, M as-
sumes very low values (under 50 MPa) and Kd is about equal to
2 and thus Kd individuates NC layers (TC16, 2001). In both the
cases V
S
appears less sensible to the stress history and the stiff-
ness of the deposits compared to M and Kd. In addition, the
boreholes on the foundations illustrate that in the Southern part
484
th

the base level is at about 4.90 m in depth. Then, before inter-
cepting sandy and arenaria deposits, it was found a landfill layer
from 4.90 to 8.00 m. In terms of stiffness it means that in the
Northern portion of the structure the foundations stand on soil
with higher mechanical properties compared the soil in the
Southern part. A possible solution to restore the historical build-
ing is to improve the stiffness of the soils placed in the Southern
portion, without acting on masonry foundations. This achieve-
ment could be realized for example, using, from 4.90 to 8.00 m
in depth, special injections at low pressure, able to penetrate and
mix with the existing soil structure.

Figure 1. Geotechnical cross section with V
S
profiles.

Figure 2. (a) Justice Court: site investigation by Seismic Dilatometer; (b) SDMT results: M, Kd and V
S
profiles.

son was carried out between the Northern portion (absence of
filling material) and the Southern portion (presence of filling
3 SITE RESPONSE ANALYSIS

Numerical analyses of seismic site response were carried out us-
ing the computer codes EERA (Bardet et al. 2000), a monodi-
mensional linear equivalent model, and QUAD4M (Hudson et
al. 1994), a bidimensional linear equivalent model. that consid-
ers a cross section of 3.5 km of width, with 5860 elements and
5844 joints.
The evaluation of the local site effects (e.g. topography, soil
conditions) plays an important role in the non-uniform amplifi-
cation response obtained at different sites (Paolucci 2002).
In order to compare the 1D and 2D analyses, the 1D elastic re-
sponse spectrum were multiplied by the topographic amplifica-
tion factor, assumed equal to 1.2 (CEN 2003). Both the analyses
were performed on the top of Chieti hill, in correspondence of
Southern portion of the Justice Court. Moreover, a 1D compari-
material) to evaluate the site effects due to the different me-
chanical behaviour of the upper 8 meters.

3.1 Input ground motions

For the numerical analyses two natural accelerograms, applied
on the bedrock, were selected as input ground motions. Both the
accelerograms were chosen from the software REXEL (Smerz-
ini et al. 2012) and the Italian Accelerometric Archive ITACA
(Working Group ITACA 2010).
The first accelerogram UM_EW is the strong motion recorded
at the Assisi station (Italy) during the September 26, 1997 Um-
bria-Marche (UM) earthquake (Mw = 6, on outcrop, normal
fault, site-source distance 20 km), scaled, according to CEN
(2003), to a peak ground acceleration of 0.164g, for a return pe-
riod TR = 475 years and a soil type, for the site of Chieti. The
485

3)
N
o
r
m
a
l
i
z
e
d

s
h
e
r

m
o
d
u
l
u
s

G
/
G
0
D
a
m
p
i
n
g

R
a
t
i
o

D
(
%
)
S
p
e
c
t
r
a
l

e
r
a
t
i
o
n
S
a
(
g
)

a
c
c
e
l

second accelerogram VN_NS is the strong motion recorded at
the Cascia station (Italy) during the September 19, 1979 Val
Nerina (VN) earthquake (Mw = 5.8, on outcrop, normal fault,
site-source distance 9 km), scaled to the same peak ground
acceleration of UM earthquake.
second accelerogram VN_NS is the strong motion recorded at
the Cascia station (Italy) during the September 19, 1979 Val
Nerina (VN) earthquake (Mw = 5.8, on outcrop, normal fault,
site-source distance 9 km), scaled to the same peak ground
acceleration of UM earthquake.

3.2 Geotechnical model 3.2 Geotechnical model

The geotechnical model of Chieti hill, used in the numerical
analyses, is illustrated in Figure 1 and 3 and Table 1, by includ-
ing the soil and dynamics parameters (unit weight , Poisson
coefficient , shear wave velocity V
S
, stiffness decay curves
G/G
0
and damping D curves).
The geotechnical model of Chieti hill, used in the numerical
analyses, is illustrated in Figure 1 and 3 and Table 1, by includ-
ing the soil and dynamics parameters (unit weight , Poisson
coefficient , shear wave velocity V
S
, stiffness decay curves
G/G
0
and damping D curves).
an average value interpolated from the experimental relation-
ship Crespellani et al. (1989) and SDMT profiles.
an average value interpolated from the experimental relation-
ship Crespellani et al. (1989) and SDMT profiles.
The site campaign of the Justice Court had provided only a cy-
clic simple shear tests with double sample in OC silty clay. In
this respect, the following reference laboratory curves were as-
sumed to evaluate the non-linear and dissipative soil behaviour:
Anh Dan et al. (2001) for filling material, Marcellini et al.
(1995) for sand and sandstone, MSAQ Working Group (2010)
for silty clay colluvial cover. The bedrock has G/G
0
- and D
linear behaviour.
The site campaign of the Justice Court had provided only a cy-
clic simple shear tests with double sample in OC silty clay. In
this respect, the following reference laboratory curves were as-
sumed to evaluate the non-linear and dissipative soil behaviour:
Anh Dan et al. (2001) for filling material, Marcellini et al.
(1995) for sand and sandstone, MSAQ Working Group (2010)
for silty clay colluvial cover. The bedrock has G/G
0
- and D
linear behaviour.

3.3 Results 3.3 Results

1D numerical analyses of seismic site response were carried out 1D numerical analyses of seismic site response were carried out
11

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

00

1,2 - Fillin g materi al 1,2 - Fillin g materi al
4,5 - San d , San dstone 4,5 - San d , San dstone
6 - OC Silty clay 6 - OC Silty clay
3 - Silty clay co lluvial cover 3 - Silty clay co lluvial cover
1, 2 - Filling material 1, 2 - Filling material
4,5 - San d , San dstone 4,5 - San d , San dstone
18
by considering the Northern portion and the Southern portion.
18
by considering the Northern portion and the Southern portion.
16
The spectral accelerations (Figure 4) highlight the different me-
16
The spectral accelerations (Figure 4) highlight the different me-
14 chanical behaviour of the upper 8 meters: the Southern portion 14 chanical behaviour of the upper 8 meters: the Southern portion
12 shows pronounced amplifications for a period of 0.1-0.2 s, that 12 shows pronounced amplifications for a period of 0.1-0.2 s, that
10 is the fundamental period of the filling material, while the spec- 10 is the fundamental period of the filling material, while the spec-
8
tral accelerations of Northern portion appear lower.
8
tral accelerations of Northern portion appear lower.
6
2D numerical analyses of seismic site response were performed
6
2D numerical analyses of seismic site response were performed
4
by considering in the Southern portion two point, A and B, 20
4
by considering in the Southern portion two point, A and B, 20
2
m far from each other, shown in Figure 3. The peak ground ac-
celerations (Figure 5) doesnt appear influenced by spatial posi-
2
m far from each other, shown in Figure 3. The peak ground ac-
celerations (Figure 5) doesnt appear influenced by spatial posi-
00
0.0001 0.001 0.01 0.1 1 0.0001 0.001 0.01 0.1 1
Shear Strain (%) Shear Strain (%)

Figure 3. Stiffness decay curves G/G
0
and damping D curves of
Chieti
hill for numerical analyses.
Figure 3. Stiffness decay curves G/G
0
and damping D curves of
Chieti
hill for numerical analyses.

Table 1. Geotechnical model of Chieti hill for numerical analyses. Table 1. Geotechnical model of Chieti hill for numerical analyses.
tion and input ground motion, even thought the analyses con- tion and input ground motion, even thought the analyses con-
sider only two time histories. In addition, the spectral
accelerations emphasize the site effect due to the topography:
point A, closer than point B to the hillside, shows higher ampli-
fications for a period of 0.2-0.4 s, compared to the ones of point
B .
sider only two time histories. In addition, the spectral
accelerations emphasize the site effect due to the topography:
point A, closer than point B to the hillside, shows higher ampli-
fications for a period of 0.2-0.4 s, compared to the ones of point
B .
The comparison of the average results from 1D and 2D numeri- The comparison of the average results from 1D and 2D numeri-

Layer Layer

Material Material

V
S
V
S
cal analyses in the Southern portion of the Justice Court (Figure cal analyses in the Southern portion of the Justice Court (Figure
6) illustrates that the 1D peak ground accelerations are higher

6) illustrates that the 1D peak ground accelerations are higher

(kN/m (m/s)
than the ones evaluated from 2D analyses, probably due to the
1 Filling material 17.1 0.30 220

2
Filling material 17.1 0.30 440

3
Silty clay colluvial cover 18.7 0.45 280
higher sensitivity of 1D model to stratigraphic effects. In addi-
tion, 2D method shows local site effects mainly due to topogra-
phy for a period of 0.3-0.4 s, that is the fundamental period of
the Justice Court.

4
Sand, sandstone 20.6 0.30 580

5
Sand, sandstone 20.6 0.30 870
1.8

1.6

UM_EWSouthernportion 1D
VN_NS Southernportion 1D
UM_EWNorthern portion 1D

6
OC silty clay 20.2 0.45 600

7
OC silty clay 20.2 0.45 800

8
Bedrock 21.0 0.30 1000

9
Bedrock 22.0 0.30 1300
1.4

1.2

1.0

0.

8
0.6
VN_NS Northern portion 1D
Average Southernportion 1D
Average Northernportion 1D

On the top of the hill, in correspondence of the Justice Court, the
subsoil was modelled by considering in the upper 8 m filling
material in the Southern portion of the Justice Court and sand
and sandstone in the Northern portion, sand and sandstone be-
tween 8 m and 42 m of depth, OC silty clay between 42 m and
342 m of depth and the bedrock beyond 342 m of depth, while on
the hillside the model reflects the silty clay colluvial cover in the
upper 15 m up to the OC silty clay layer. In the upper 50
meters the V
S
profile was defined as an average of SDMT pro-
files, while in the lower OC silty clay V
S
was estimated by using
0.4

0.

2
0.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Period T (s)
Figure 4. Spectral accelerations form 1D analyses.
486
th

1.8

1.6

S
p
e
c
t
r
a
l

a
c
c
e
l
e
r
a
t
i
o
n
S
a
(
g
)

S
p
e
c
t
r
a
l

a
c
c
e
l
e
r
a
t
i
o
n
S
a
(
g
)

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.0

1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.0

UM_EWSouthe rnportion2D point A
VN_NS Southernportion 2D point A
UM_EWSouthe rnportion2D point
B VN_NS Southernportion2D point
B Average Southe rnportion2D
point A Average Southe rn portion
2D point B

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Period T (s)
Figure 5. Spectral accelerations from 2D analyses.

Average Southernportion 1D
Average Southernportion 2D pointA
Average Southernportion2D point B

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Period T (s)
Bardet J.P., Ichii K., Linn and C.H. 2000. EERA A Computer
Program for Equivalent-linear Earthquake site Response Analyses
of Layered Soil Deposits. University of Southern California.
CEN, EuropeanCommittee for Standardisation (2003) Eurocode 8: de-
sign provisions for earthquake resistance of structures, Part 1.1: ge-
neral rules, seismic actions and rules for buildings, prEN 1998-1
Crespellani T., Ghinelli A. and Vannucchi G. 1989. An evaluation of
the dynamic shear modulus of a cohesive deposit near Florence, I-
taly. Proc. XII ICSMFE, Rio de Janeiro.
Hudson, M., Idriss, I.M., and Beikae, M. 1994. QUAD4M: A Computer
Program to Evaluate the Seismic Response of Soil Structures using
Finite Element Procedures and Incorporating a Compliant Base.
Center for Geotechnical Modeling, Dep. of Civil & Env.
Engng, University of California, Davis.
Marcellini A., Bard P.Y., Vinale F., Bousquet J.C., Chetrit D., De-
schamps A., Marcellini A., Iannaccone G., Romeo R.W., Silvestri
F., Bard P.Y., Improta L., Meneroud J.P., Mouroux P., Mancuso C.,
Rippa F., Simonelli A.L., Soddu P., Tento A. and Vinale F. 1995.
The Benevento Seismic Risk Project. I- Seismotectonic and Geote-
chnical Background. Proc. 5th International Conference on Seismic
Zonation, Nice, France 1: 802- 809.
Marchetti S., Monaco P., Totani G. and Marchetti D. 2008. In Situ Tests
by Seismic Dilatometer (SDMT). In J.E. Laier, D.K. Crapps
& M.H. Hussein (eds), From Research to Practice in
Geotechnical Engineering, Geotechnical Special Publication No.
180: 292311. ASCE.
MSAQ Working Group. 2010. Microzonazione sismica per la ricostru-
zione dellarea aquilana. Regione AbruzzoDipartimento della
Protezione Civile, LAquila, 3 vol. & Cd-rom (in Italian).
Paolucci R. (2002). Amplification of earthquake ground motion by ste-
ep topographic irregularities. Earthquake Engineering and Structu-
ral Dynamics, 31: 1831-1853.
Smerzini C., Galasso C., Iervolino I. and Paolucci R. 2012. Engineering
ground motion selection based on displacement-spectrum compati-
bility. Proc. 15th World Conference on Earthquake
Engineering, Lisbon, Portugal, September 24-28, 2012.
TC16. 2001. The DMT in Soil Investigations. A Report by the ISSMGE
Committee TC16. May 2001, 41 pp. Reprint in R.A. Failmezger &
J.B. Anderson (eds), Flat Dilatometer Testing, Proc. 2nd Int. Conf.
on the Flat Dilatometer, Washington D.C.: 748.
Totani G., Monaco P., Marchetti S. and Marchetti D. 2009. Vs measu-
rements by Seismic Dilatometer (SDMT) in non-penetrable soils. In
M. Hamza et al. (eds), Proc. 17th Int. Conf. on Soil Mechanics and
Figure 6. Comparison between 1D and 2D spectral accelerations.

4 CONCLUSION

The paper illustrates the potential of the seismic dilatometer to
efficaciously approach a geotechnical problem by means of the
results analyses. While V
S
appears less sensible to both the
stress history and the deposits stiffness, M gives precious in-
formation on soil stiffness, while Kd provides for important de-
tails about the deposits overconsolidation. Combining SDMT
results with laboratory data it has been possible to evaluate the
the local site effects by means of 1D and 2D seismic site res-
ponse analyses of the Justice Court. These numerical analyses
indicates that in complex stratigraphic and topographic condi-
tions, it appear appropriate to combine 1D and 2D methods.

5 ACKNOWLEDGEMENTS

This study was founded by Provincia di Chieti and Studio Prof.
Marchetti s.r.l.

6 REFERENCES

Anh Dan,L.Q., Koseki,J. and Tatsuoka,F. 2001. Viscous deformation in
triaxial compression of a dense well-graded gravel and its model
simulation. In Tatsuoka et al. (eds) Advanced Laboratory Stress- Strain
Testing of Geomaterials, Balkema, pp.187-194.
Geotechnical Engineering, Alexandria, 2: 977980, IOS Press.
Working Group ITACA. 2010. Data Base of the Italian strong motion
records: http://itaca.mi.ingv.it
487
th
1
Dtermination du coefficient rhologique de Mnard dans le diagramme
Pressiorama
.
Obtaining the Mnard Rheological Factor in a Pressiorama
Diagram.
Baud J.-P.
Eurogo, France
Gambin M.
Apago, France
RSUM: Louis Mnard a dfini le coefficient rhologique laide du rapport adimensionnel des deux caractristiques classiques
de lessai pressiomtrique E
M
/p*
LM
. La connaissance de ce rapport est un lment essentiel de la validit des calculs de dformation
du sol au contact de toutes sortes de fondations. Sa valeur dpend simultanment de la qualit de ralisation du forage
pressiomtrique, et de la proportion entre cohsion et frottement dans la rsistance du sol soumis lessai, cest--dire de sa nature.
Dans le but de complter par un axe des valeurs de le diagramme Pressiorama dsormais prsent dans certaines normes
dapplication de lEurocode 7, nous nous sommes talonns sur des essais pressiomtriques autofors par la technique STAF dans des
sols divers allant de sols trs mous des rochers trs massifs. En effet ces essais, de trs bonne qualit, permettent dobtenir des
valeurs E
M
/p*
LM
allant de 4 pour les sols purement granulaires plus de 100 pour les sols cohrents trs consolids et les roches. Les
auteurs proposent ici, partir de ces donnes, une expression, la fois chiffre et graphique, de la valeur du coefficient rhologique ,
base sur 3 paramtres E
M
, p
LM
et p
0
- ce dernier estim ou, mieux, mesur et donc variable selon la profondeur de lessai dans le sol.
ABSTRACT: In the early years of the development of his direct method of design based on PMT results, Louis Mnard introduced
a rheological factor based on the ratio E
M
/p*
LM
for each test. The knowledge of this factor is necessary to estimate settlement and
horizontal displacement of all sorts of foundations. Its value is a function of both the quality of the borehole and the ratio between soil
friction and cohesion, i.e. its nature. In order to complete the Pressiorama diagram with a values axis, the authors used a
calibration mostly based on so-called self-bored PMT tests performed with the STAF technique, in various soil types from soft clay to
rock. These very good quality tests permit obtaining E
M
/p*
LM
values from 4 in granular soils to 100 in highly consolidated soils and
rock. From these data, the authors propose an evaluation of this rheological factor only based on the values of 3 parameters, namely,
E
M
, p*
LM
and the earth pressure at rest p
0
, either estimated, or, much better, measured during the early part of the test. Results are
given under the shape of an equation and graphically on the Pressiorama diagram.
MOTS CLS : Pressiomtre, autoforage, classification des sols, coefficient rhologique .
KEYWORDS: Mnard pressuremeter, self-boring, soil classification, rheological factor.
1 INTRODUCTION
Le rapport adimensionnel E
M
/p*
LM
des deux caractristiques
classiques de lessai pressiomtrique Mnard est un facteur
complexe et puissant, qui dpend simultanment de la qualit de
ralisation du forage pressiomtrique, et de la proportion entre
cohsion et frottement dans le comportement du sol soumis
lessai, cest--dire de sa nature. Son utilisation par Mnard
pour dfinir le coefficient rhologique est un lment
essentiel de la validit des calculs de dformation du sol par les
mthodes pressiomtriques.
Au cours des annes rcentes, en raison de laccroissement
des essais pressiomtriques produits par des oprateurs
manquant de formation et de matrise des techniques de forage
les mieux adaptes chaque type de sol, techniques spcifiques
cet essai mais remarque valable aussi pour tous les essais
gotechniques, de nombreux utilisateurs ont dcel une
distorsion entre les prvisions de tassement par la mthode
pressiomtrique et les dformations rellement observes sur les
ouvrage construits. La quasi-totalit des pratiques de forage
entranant le remaniement des parois de forage pressiomtrique
allant toujours dans le sens dune diminution parfois dramatique
des modules mesurs, les prvisions de tassement qui en ont t
dduites deviennent notoirement pessimistes, ceci alors mme
que la mthode de calcul, confirme par les normes et
rglementations nationales et europennes, a montr depuis
longtemps sa fiabilit lorsquelle est applique des donnes de
terrain acquises dans des conditions de qualit normale.
(Baguelin et al. 1978).
Avec lmergence de codes de calcul aux lments finis, de
nombreux ingnieurs ont pens quil tait possible, voire
ncessaire, de dlaisser la clbre rgle souvent nomme T-0
(Mnard & Rousseau, 1962) pour le calcul du tassement partir
dun profil de modules pressiomtriques, au profit dune
modlisation numrique complexe tenant compte de la
gomtrie de linterface sol-structure, et de traiter le problme
de la dformation en appliquant ce modle des lois de
comportement bases sur llasticit linaire, laide dune
corrlation entre le module pressiomtrique et un module
dYoung.
Cette approche nous semble voue lchec, comme chaque
fois que lingnieur croit pouvoir fait fi de la mthode
exprimentale et la remplacer par des calculs que lon prend
pour rigoureux raison de leur complexit (Briaud & Gibbens
1994, Gambin 2003, Gambin 2010). Notre approche de ce
problme majeur relatif la crdibilit des prvisions de
dformation faites par lingnierie gotechnique, se distingue de
cette tendance aux modlisations complexes, et vise plutt
assurer ce qui fait loriginalit et la cause du succs de la
mthode pressiomtrique, cest dire atteindre rapidement un
488
th
th
2
rsultat par un calcul simple et fond sur une connaissance la
plus parfaite possible de la loi contrainte-dformation de la
sollicitation pressiomtrique du sol vierge. Dans ce sens, les
essais autofors dans la plus large gamme de sols possible sont
une bonne rponse (Arsonnet et al., 2005), car ils permettent
la fois la mesure du module initial G
0,
de la loi de dcroissance
du module avec la contrainte (Baud & Gambin, 2005 ; Baud &
Gambin, 2008 ; Baud et al., 2012 et 2013), enfin lvaluation du
coefficient o (Baud, 2005 ; Baud & Gambin, 2012).
2 PARAMETRES PRESSIOMETRIQUES NECESSAIRES
ET SUFFISANTS A LA DETERMINATION DE o.
2.1 Module E
M
, pression limite p*
LM
et confinement de
lessai.
La premire prsentation du coefficient rhologique o en
tableau double entre, en fonction du rapport E
M
/p*
LM
et de
ltat de consolidation du sol, qui avait t donne dans la notice
D60 sur les rgles dexploitation des techniques
pressiomtriques et dexploitation des rsultats obtenus pur le
calcul des fondations (Mnard, 1968) a t reconduite telle
quelle dans les rglementations franaises puis europennes.
Mnard prvoyait pourtant de rviser le tableau en fonction de
rsultats dessais exprimentaux , raison principale pour
laquelle les valeurs de o taient donnes sous forme de fractions
unitaires. Mais le principe tait tabli : o est croissant quand
E
M
/p*
LM
est croissant, depuis le comportement sableux jusquau
comportement argileux, et croissant galement lorsque le sol
passe de la sous-consolidation ou de laltration la sur-
consolidation.
Pour suivre cette rgle usuelle, des lignes disovaleurs des
valeurs fractionnaires de o peuvent tre traces directement
dans le diagramme bi-logarithmique Pressiorama (Baud,
2005) en log (p*
LM
), log (E
M
/p*
LM
), tel quil a t dit en
annexe des normes NF P94-261 et NF P94-262 (Fig. 1), la
valeur 1 formant la limite suprieur du diagramme et la valeur
, la plus faible du tableau de Mnard, tant affecte au
rapports E
M
/p*
LM
trs faibles, correspondant soit des sables et
graviers, soit des sols trs remanis, remaniement naturel in
situ, ou bien li une mauvaise excution du forage
pressiomtrique.
La principale raison de tracer des droites en coordonnes bi-
logarithmiques, tait que lon ne dispose absolument pas, dans
le rfrentiel des confrontations entre mthode pressiomtrique
et fondations instrumentes, de donnes suffisantes pour
proposer des courbes plus sophistiques. Ces droites sont donc
des constructions mathmatiques simples, base sur
lhypothse :
( )n
m
n
k
LM
LM
M
* p .
* p
E

1
|
.
|
\
|
= o
(1)
Sous cette forme en effet, les coefficients m et n (en
exposants) et le facteur k, tous sans dimensions prennent
empiriquement les valeurs ncessaires pour assurer le trac
dsir : m dtermine langle des droites iso-o, n dtermine
lcartement entre les valeurs, et k dtermine, pour un couple de
valeurs (m, n) donn, la position de la ligne maximale o = 1.
Les valeurs de o ainsi proposes peuvent tre rendues assez
conformes au tableau double entre de Mnard pour des essais
profondeur moyenne de quelques mtres utiliss pour des
fondations superficielles.
Mais on voit assez vite que si les valeurs de E
M
/p*
LM
dcrivent toujours assez bien la nature du sol, sableux,
intermdiaire ou argileux, qui peut tre connu par ailleurs au
moment du forage pressiomtrique ou, mieux par un sondage
Figure 1 Valeurs de o dans le diagramme Pressiorama tel quil peut
apparatre dans les normes NF-P94-261 et NF-P94-262 (en projet).
carott au voisinage, la position des lignes disovaleurs de o par
rapport p*
LM
nest plus compatible ds lors que lon considre
des essais raliss des profondeurs importantes : ceci est li au
fait que ce nest pas la valeurs absolue de p*
LM
qui doit tre
prise en compte, mais la valeur relative p*
LM
/p
0
, sans
dimension, qui tient compte du confinement de lessai par la
pression horizontale des terres au niveau o elle est mesure.
On est ainsi conduit la formulation suivante, dans laquelle
k
E
, m et n jouent le mme rle.
n
m
E
n
k
|
.
|
\
|
|
.
|
\
|
=
0
LM
LM
M
p
* p
.
* p
E

1
o
(2)
Le nouveau diagramme [log (p* /p
0
), log (E
M
/p*
LM
)] qui en
itue une prsentation peu
intuitive, essentiellement parce que le gotechnicien est trs
habitu ce que reprsente la pression limite, directement
proportionnelle la rsistance du sol, alors que la grandeur sans
dimension p*
LM
/p
0
est difficile saisir. Elle reprsente en
quelque sorte le degr de rsistance du sol soumis lessai par
rapport une augmentation normale de rsistance attendue
croissante en fonction de la profondeur.
LM
rsulte nest pas trac ici. Il const
Les coefficients m et n peuvent prendre en thorie une large
gamme de valeurs, mais en pratique seule une faible gamme de
variations laisse les droites reprsentatives toutes prsentes et
tales dans le champ de vision du diagramme. Nous proposons
de les arrter m=0,5 et n=2, et avec ce couple de coefficients,
k
E
prendra une valeur comprise entre 3 et 5 pour que o reste
compatible avec lusage. Nous avons retenu ici la valeur entire
k
E
=4, do :
489
th
3
4
1
2
1
4
|
.
|
\
|
|
.
|
\
|
=
0
LM
LM
M
p
* p
.
* p
E
o
(3)
2.2 Quelle est la relation possible entre E Young et E
M
?
Tout dabord, il parat peu judicieux de comparer le module
d'Young mesur par traction sur des solides polycristallins
(barres mtalliques) o la dformation est linaire jusqu' la fin
de la phase strictement lastique, et le module de dformation
des sols, variable avec lintensit de la contrainte, cette
dformation tant de nature hyperbolique tout au long de
lapplication de la charge.
Cest la raison quexpose Mnard, dans son article
fondateur de Sols-Soils n1 (Mnard & Rousseau, 1962), pour
crer la notion de coefficient rhologique o pour lequel il
propose simultanment la gamme des valeurs fractionnaires
dans diffrents sols, et une expression base sur le module
altern Ea :
|.| \| v + ( 21
(
= o a E
E |.| \| v +
(
(
= o 21
a E
E
(4)
E tant aujourdhui not E
M
, et v tant une valeur faible
compare , dpendante de facteurs secondaires . Mnard
pensait approcher avec Ea la valeur du module de micro-
dformations, not alors Ec ; on saccorde plutt aujourdhui
voir dans Ea une approche dun module lastique E
Y
.
Paraphrasant Mnard on peut crire ;
E
Y
= E
M
/o
n
(4b)
avec nO2, soit en ngligeant les facteurs secondaires :
E
Y
= E
M
/o (4c)

Pour des raisons non explicites, lhabitude avait t prise en
France par les utilisateurs de rsultats pressiomtriques, de
tronquer cette proposition et de retenir Ey = E
M
/o pour des
estimations du module dYoung qui ne donnent pas satisfaction.
Il y a eu sans doute attraction et confusion lie au fait que
Mnard a plus tard galement indiqu que o = E
M
/E+, E+
module de dformation du sol dans un champ quasi-
isotrope , plutt assimilable donc un module de type
domtrique et non un module dYoung. Des utilisateurs de
longue date des mthodes pressiomtriques avaient gard
lusage de cette relation entre Ea et E
M
pour donner une
estimation de o partir dessais cycliques (R Heintz, 2012).
En identifiant o entre les relations (3) et (4), il vient :

(5)
Cette relation remarquable qui limine E
M
et o et tablit une
relation directe entre Ey et p*
LM
, dcoule directement du choix
fait ci-dessus pour les coefficients m et n. En effet, quelle que
soient les valeurs adoptes pour ces coefficients, il se maintient
toujours une forte corrlation entre un paramtre de rupture,
p*
LM
, et un module dYoung dfinissant une relation linaire
lastique, donc constituant une corde sur la courbe
pressiomtrique, entre son origine (p = p
0
) et un point situ vers
le milieu de lintervalle p
0
- p
LM
cest--dire proche de la
pression de fluage de lessai.
La mise en vidence de cette relation confirme bien la
dfinition initiale de Mnard et lexpression qui en est propose
ici. En effet, dans un modle (q,c) ramenant le comportement du
sol une phase lastique linaire borne par un critre de
rupture fixe, celui-ci impose bien un rapport unique entre le
module E et la valeur choisie pour le dviateur q.
3 PROPOSITION DUN NOUVEAU CADRE POUR LE
DIAGRAMME PRESSIORAMA
.
En vue de dterminer la valeur du coefficient rhologique o
pour chaque essai pressiomtrique dont on connait normalement
et simultanment la pression limite, le module pressiomtrique,
et la profondeur permettant destimer p
0
, ou mieux la valeur
mesure de p
0
, il est possible de proposer une faon diffrente
de placer les rsultats dessais, dans un nouvel abaque construit
de la faon suivante (Fig. 2):
- en abscisse, le coefficient o , en chelle logarithmique
et en valeurs dcroissant de gauche droite.
- en ordonne, plac sur o = 1, le module pressiomtrique
relatif E
M
/p
0
, qui est donc un nombre sans dimension,
en chelle logarithmique et en valeurs croissantes vers
le bas.
- laxe des pressions limites relatives p*
LM
/p
0
vient se
placer en oblique des deux axes, avec un angle variable
selon les rapports dchelles.
- laxe des rapports E
M
/p*
LM
est alors conjugu et
orthogonal laxe p*
LM
/p
0
.
Chaque essai pressiomtrique est reprsent par un point
unique au croisement de ses 4 caractristiques.
Limit vers le haut par la ligne E
M
/p*
LM
= 3 au-del de
laquelle on ne doit pas trouver de matriau naturel ou fabriqu,
labaque est un triangle rectangle englobant tous les types de
sols, roches et matriaux fabriqus. La base, que lon tronque
plus ou moins tt selon que lon sintresse plus la mcanique
des roches, ou celle des sols, ou au domaine intermdiaire,
reprsente les matriaux ciments. Les sols trs mous, les vases
et boues sont dans la pointe galement tronque. Les sols
habituels de la gotechnique sont entre ces deux extrmes, et
sont ici qualifis dans un quadrillage de 3 fois 3 cases, N1
N9, dont les matriaux sont identifis dans la lgende de la
figure.
Lexprience rduite dessais pressiomtriques que nous
avons personnellement dans le rocher franc (Baud & Gambin,
2011 et 2012) nous a permis de confirmer les zones N10, pour
les graves et roches trs fractures, N11, pour les roches
tendres, fractures ou altres et N12 pour les roches trs
dures. Elles sont galement en accord avec des tudes
antrieures sur les roches (Failmezger et al., 2005).
4 CONCLUSIONS
Cette tude a permis de montrer quil est possible dintroduire
un axe des o comme abscisse de notre diagramme
Pressiorama
, lordonne tant E
M
/p
0
, et den graduer la valeur
en fonction des rsultats obtenus au pressiomtre, ce qui ntait
pas vident a priori. (Fig. 2).
Deux autres axes apparaissent transversalement aux
coordonnes cartsiennes : la pression limite relative p*
LM
/p0 et
le rapport bien connu E
M
/p*
LM
. Labscisse o est galement
norme arithmtiquement en dfinissant le comportement
pressiomtrique du sol par un indice de granularit g =
[2.Ln(o)]/k
E
, proportionnel un angle de frottement dductible
directement de lessai :
M
= arctan (g).
Nous ne considrons pas notre travail comme termin, car
il est ncessaire quil soit confront de nombreux rsultats
dessais pressiomtriques dans les roches en particulier. Et il est
possible que cette confrontation, ainsi que des tudes de
corrlations entre E
Y
et p*
LM
, conduisent une volution de
notre schma dont les bases paraissent cependant bien acquises.
2
1
|
.
|
1

\
|
0
LM
LM
Y
p
* p
. 6 =
* p
E
490
th
th
4
Figure 2 Diagramme Pressiorama en coordonnes bilogarithmiques
adimensionnelles [ , module relatif E
M
/p
0
].
5 REFERENCES.
Arsonnet, G., Baud, J.-P., Gambin, M.P. 2005. Ralisation du forage
pour essais pressiomtriques par un systme de tube fendu auto-
for (SFAF), Actes du Symp. Intern. ISP5 PRESSIO 2005, sous la
direction de Gambin, M., Magnan, J.-P., & Mestat., Paris, 22-24,
Aot. 2005, Vol.1, Paris: Presses des Ponts.
Baguelin F., Jezequel J.F. Shields D.H. 1978. The Pressuremeter and
foundation engineering. Trans Tech Publications, Clausthal,
Germany, chap. 3 pp 284-291.
Baud, J.-P, Analyse des rsultats pressiomtriques Mnard dans un
diagramme spectral [log (p
LM
), log (E
M
/p
LM
)] et utilisation des
regroupements statistiques dans la modlisation dun site, Actes du
Symp. Intern. ISP5 PRESSIO 2005, Paris, sous la direction de
Gambin, M., Magnan, J.-P., & Mestat, P. 22-24 Aot. 2005, Vol.1,
Paris: Presses des Ponts.
Baud, J.-P, Gambin, M. P. 2005. Dduction dune loi de rponse
hyperbolique unique par complilation de courbes pressiomtriques
dans un sol de lithologie homogne Actes du Symp. Intern. ISP5
PRESSIO 2005, sous la direction de Gambin, M., Magnan, J.-P., &
Mestat, P., Paris, 22-24, Aot. 2005, Vol.1, Paris: Presses des
Ponts.
Baud, J.-P, Gambin, M. P. 2008. Homogenising MPM tests curves by
using a hyperbolic model, in Huang, A.-B., & Mayne, P. W. (eds)
Geotechnical an Geophysical Site Characterization, Proc. ISC3
Taipei, Taiwan, 1-4 April 2008, London: Taylor & Francis
Baud, J.-P, Gambin, M. P. 2011, Classification des sols et des roches
partir dessais dexpansion cylindrique en haute pression, C. R. du
15me Congrs Europen de Mcanique des Sols et de
Gotechnique, edited by A. Anagnostopoulos, M. Patchakis,
C.Tsatsanifos, ISO Press, Amsterdam .
Baud, J.-P, Gambin, M. P. 2012. 50 MPa Mnard PMTs help linking
Soil and Rock Classifications. in A. Anagnostopoulos, ed.,
Geotechnical & Geological Engineering Journal, Special Issue on
Hard Soils and Weak Rocks, Springer Verlag, Berlin.
Baud, J.-P, Gambin, M. Schlosser F. 2012. Stress-strain hyperbolic
curves with Mnard PMTs in R.Q. Coutinho ed., Geotechnical and
Geophysical Site Characterization, Proc. ISC4 Porto de Galinhas,
Brazil, Sept.18-21, 2012, London: Taylor & Francis
Baud, J.-P, Gambin, M. Schlosser F. 2013. La courbe contrainte-
dformation au pressiomtre Mnard Actes du 18me CIMSG,
Paris, 1-5 sept.
Briaud, J.-L., & Gibbens, R. 1994, Test and Prediction Results for Five
Spread Footings on Sand ASCE Geotechnical Specification
Publication No.41
Failmezger, R., Zdinak, A., Darden, J., Fahs, R. 2005, Use of Rock
Pressuremeter for Deep Foundation Design Actes du Symp. Intern.
ISP5 PRESSIO 2005, sous la direction de Gambin, M., Magnan,
J.-P., & Mestat, P., Paris, 22-24, Aot. 2005, Vol.1, Paris: Presses
des Ponts.
Gambin, M. 2003. Etude lmentaire dun mythe. FONSUP 2003,
Symposium International sur les fondations superficielles, Paris, 5-
7 novembre 2003, p.251-254, J.P. Magnan & N. Droniuc d.,
Presses de lENPC, Paris.
Gambin M. 2010. Les thories et leur volution face la ralit en
Gotechnique. VIIme Confrence Coulomb, Comit Franais de
Mcanique des Sols, Paris.
Heintz R. 2011, Communication personnelle. Eurasol, Luxembourg.
Mnard L. Rousseau J. 1962. Lvaluation des tassements, tendances
nouvelles. Sols-Soils, N1, Paris.
Mnard L. 1968. Rgles dexploitation des techniques pressiomtriques
et dexploitation des rsultats obtenus pur le calcul des fondations. (en
anglais, trad. Franaise TLM notice D60), Sols-Soils, N26 Paris.
491
Courbes hyperboliques contrainte-dformation au pressiomtre Mnard autofor
Stress-Strain Hyperbolic Curves Obtained With a Selfboring Mnard PMT
Baud J.-P.
Eurogo, Avrainville, France
Gambin M.
Apago, Paris, France
Schlosser F.
cole des Ponts Paris-Tech, Navier-CERMES, Marne-la-Valle, France
RSUM : On prsente ici les derniers rsultats de nos recherches sur le module de dformation du sol dtermin en assimilant un
arc d hyperbole la courbe pression-volume obtenue dans un essai au pressiomtre Mnard. Lessai est ralis en utilisant une cavit
obtenue par un systme de tube fendu auto-for, soit en roto-percussion, le STAF, soit en rotation seule, le ROTOSTAF.
Lajustement des points de mesure sur une branche dhyperbole du second degr permet dobtenir une expression analytique originale
de la dforme sous la forme = f (G
0
, p
o
, p
LM
, P
L
) dans laquelle P
L
reprsente labscisse de la pression limite vraie prise comme
asymptote. Pour chaque essai, il est alors facile dobtenir la variation, calcule partir des points de mesure, du module tangent Gt et
du rapport Gt/G0 en fonction de , et pareillement le module scant Gs. Cette variation peut tre compare avec celle donne par le
modle hyperbolique courant : la concordance est bonne, sauf pour les trs faibles dformations initiales o les courbes obtenues
partir des points dessais donnent des modules sensiblement plus levs. Ces rsultats sont en bon accord avec les rsultats dessais en
laboratoire et en gophysique. Ils mettent en vidence une dcroissance typique des modules lorsque la dformation saccroit.
ABSTRACT : The present stage of our research work on soil E-moduli values are submitted here. These values are obtained by
assimilating the pressure-volume plot of a Mnard PMT to a 2nd degree hyperbole arc. The tests were performed using a self-bored
steel slotted tube implemented either by the STAF technique involving a drifter and a full-face bit, or the ROTOSTAF method
with a drag bit protruding from the tube outlet. Getting the hyperbolic best fit of the plotted readings makes it possible to obtain an
original equation of the radial borehole expansion as = f (G
0
, p
o
, p
LM
, P
L
) in which P
L
is the true limit pressure value of the
vertical pressure asymptote. Then, it is easy to derive the tangent modulus G
t
for each reading and the corresponding G
t
/G
0
ratio as a
function of , and similarly the secant modulus G
s
. Their variation can be compared with those given by the usual hyperbolic model: a
very good agreement is obtained, except for very small initial strains where the readings plot yields moduli sensibly higher in value.
These results are similar to those obtained by laboratory tests and in geophysical surveys. They exemplify the typical decrease of the
deformation modulus when the stress or the strain increases.
MOTS-CLS : Pressiomtre, autoforage, modle hyperbolique,
KEYWORDS: Mnard Pressuremeter, self-boring, hyperbolic soil model.
1 CONDITIONS DOBTENTION DUNE EXPANSION
PRESSIOMETRIQUE QUASI VIERGE.
Lanalyse expose ici est applique des essais
pressiomtriques raliss dans les conditions traditionnelles
de chargement par paliers de lessai pressiomtrique Mnard
(NF P91-110-1 et Pr EN-IS0 22476-4), avec des mthodes et
matriels de forage et dessai permettant de rduire autant que
possible les effets perturbateurs pour que le sol, avant le dpart
de lessai, ne soit ni dcomprim ni comprim.
1.1. Essais pressiomtriques non remanis : ncessit de
lautoforage.
Depuis les annes 1970, la mise en place du pressiomtre par
autoforage a t utilis dans les sols mous, suivi par des essais
en dformations contrles (Baguelin et al., 1978, Mair &
Wood, 1987). Progressivement, les essais en auto-forage ont t
essentiellement limits pour tester des sols sous-consolids,
supposs dans les conditions idales de dmarrage p
o
, en vue
dobtenir des relations contrainte-dformation sous cycles de
petites dformations, usuellement de moins de 20%
damplitude. De cette faon, plusieurs modules de sol pouvaient
tre obtenus, mais aucune pression limite ntait recherche
(Clarke & Gambin, 1998).
Plus rcemment, une technique dauto-forage a t propose
(Arsonnet et al., 2005) pour raliser des essais pressiomtriques
Mnard. Cette mthode, appele STAF, consiste enfoncer
un tubage continu, muni dun tube fendu au voisinage de sa
base, laide dun taillant dbordant en roto-percussion
lextrmit dun train de tiges centr. Il tait ainsi possible
dobtenir un essai en auto-forage de trs bonne qualit en petites
dformations, en raison de la rduction de la dcompression du
sol durant la cration de la cavit, tout en permettant de raliser
des essais jusqu de grandes dformations rendant possible
datteindre la pression limite conventionnelle (Baud & Gambin,
2005).
Avec cette conception, le STAF fut utilis avec succs
dans les sols cohsifs mous et moyennement compacts, et dans
les matriaux granulaires compacts, mais peu dexemples
pouvaient tre proposs dans les argiles raides et les sols
marneux, la vitesse davancement de lappareillage restant
rduit par rapport aux mthodes traditionnelles de pr-forage,
telles que la tarire continue ou les outils dents avec injection
de boue.
1.2. Essais autofors des sols meubles aux roches tendres par la
technique Rotostaf.
Pour rsoudre ce problme du faible avancement du STAF
dans les sols raides ou trs compacts, lutilisation de la rotation
simple dun train de tige a t rsolu laide dune tte de
forage hydraulique spcifique combinant une faible vitesse de
492
th
rotation du tube extrieur avec une bonne rapidit du vibro-
marteau agissant sur le taillant dbordant.
Avec cette conception, appele ROTOSTAF, les essais
peuvent tre raliss dans des sols tels que les argiles
compactes, les marnes et mme des couches calcaires.
1.3 Rgulation et conduite automatise dun essai par le
pressiomtre GoPac
Lusage dun contrleur pression-volume (CPV) traditionnel en
pression contrle est certes possible dans un tube fendu
descendu par le STAF. Mais de meilleurs rsultats peuvent
tre atteints avec un nouveau type de CPV, GeoPac, intgrant
le logiciel de traitement Govision. Au lieu de travailler au gaz
comprim, ce pressiomtre comprend un piston motoris
extrmement prcis qui permet de raliser les essais normaliss
en paliers de pression, avec une prcision volumtrique de 10
-3
cm
3
, soit une prcision sur la dformation radiale moyenne de
10
-5
.
Actuellement, dans la mise au point de ce CPV, au moins
quatre avances fondamentales dans la procdure de lessai ont
t recherches :
- une stricte compensation automatique de la rsistance
propre de la membrane des sondes, ce qui permet une correction
de la pression vraie sur la paroi du forage tout instant
- une automatisation complte de la procdure dessai : le
systme assist par ordinateur dtermine les modules du sol
partir des trois premiers pas de pression, et ajuste la procdure
dessai de manire mener bien lessai avec un nombre de
paliers de pression optimis. Loprateur conserve cependant
toujours la possibilit de travailler en semi-automatique.
- le calibrage automatique du volume initial de la sonde avec
le volume rel du trou auto-for, la pression pour ce premier
palier de lessai tant mis en quilibre avec la pression des
terres au repos. Ce procd dans le systme volume en
fonction de la pression rappelle celui du lift-off dans le
systme pression en fonction du dplacement des sondes du
pressiomtre auto-for SBP avec capteurs mcaniques.
- enregistrement des lectures (p,V) avec une trs grande
prcision, la courbe ne prsentant pas de point dinflexion.
Le logiciel Govision traite automatiquement les donnes
reues dun Gopac, avec sur option de lutilisateur drivation
des courbes de modules tangent et scant depuis p
o
comme dans
lexemple donn ci-aprs.
1.4 Un exemple dessai
Sur lessai ci-contre (Fig.1), on remarquera que la pression du
premier point de lecture est acceptable comme p
o
, avec un
volume de contact exactement gal au volume annulaire entre la
sonde dilatable et le tube fendu, et que lessai est rgul jusqu
un volume de prs de 1000 cm
3
soit un dplacement de la paroi
du forage de lordre de 12 mm. Lerreur moyenne rpartie entre
les points mesurs et le modle hyperbolique est ici de 1,8 cm
3
(soit environ 25m).
Dautres caractristiques de lessai sont galement
remarquables et observes aussi sur de nombreux essais
autofors. La premire est que la courbe de fluage prsente deux
cassures : lune la pression pf
1
prise ici pour le calcul du
module pressiomtrique, lautre une pression pf
2
nettement
plus leve (Monnet et Khlif, 1994). Une autre est que le
modle hyperbolique permet de tracer automatiquement les
courbes de dcroissance rgulire du module scant Es et de
module tangent Et, qui recouvrent bien les mmes courbes
calcules sur les points dessais, ceux-ci montrant dans les trs
faibles dformations initiales un module nettement plus lev,
ici presque doubl, que nous chercherons expliquer.
Figure 1. Rsultat dun essai 18 m de profondeur dans une argile
sableuse raide (Cnomanien de la bordure Ouest du Bassin Parisien),
autofor par Rotostaf, sonde diamtre 44 mm cellule de mesure de 37
cm dans le tube fendu de 63 mm de diamtre, essai pilot par le CPV
GoPac. et rsultats tracs par Govision
2. MODLE HYPERBOLIQUE DE LEXPANSION DUNE
CAVIT CYLINDRIQUE
2.1. De la double hyperbole au modle hyperbolique
Trs tt dans la pratique de lessai pressiomtrique, Louis
Mnard avait dfini la pression limite p
LM
, notion ne avec
lessai, au doublement du volume de la cavit de forage initiale,
en sachant que cette convention nimpliquait pas que la
dformation correspondante soit matriellement atteinte par les
sondes. La socit Mnard et les concessionnaires pionniers ont
ds lors propos successivement de nombreuses mthodes
dextrapolation de la courbe vers la pression limite : courbe
inverse, mthode des volumes relatifs, coordonnes log-log,
courbe Leme, dont la convergence vers une pression limite
unique ntait pas vidente (Baguelin et al. 1978). La double
hyperbole est lune de ces mthodes dextrapolation, drivant
du dessin des essais en (P, 1/V) (dHemricourt 2005).
La mthode en double hyperbole a t formalise et
programme (Baud et al., 1992) et constitue un modle
dcrivant bien la forme en S des essais en prforage, lie
aux vicissitudes de diamtre de forage, du temps dattente et de
dcompression entre forage et essai et du choix de paliers
initiaux infrieurs la pression des terres p
o
avant forage, par un
oprateur qui en principe ignore la valeur de p
o
et de p
LM
, sujets
de la mesure, et qui il est demand de prvoir avant la mesure
un rsultat final divis en n paliers gaux.
Il apparat rapidement que lapplication aux essais
autofors tel que celui de la figure1 simplifie plus ou moins
radicalement la modlisation, en rduisant la premire
hyperbole au rle de facteur secondaire. Pour un essai dbutant
par un palier de pression corrige lgrement suprieure ou
idalement gale la pression des terres au repos p
o
sans
dcompression ni refoulement du sol avant lessai, les points
dessai dcrivent une simple hyperbole, de la forme
- p A
A
p A A V
4
3
2
.
1
(1)
o V est le volume mesur au-del du volume de la sonde au
repos V
o
et p la pression dessai corrige de ltalonnage et du
calibrage. Pour le pressiomtre Mnard lusage est de nommer
le V mesur par la simple notation V .
Il est possible dexprimer ces donnes dessai en fonction
de la dformation dite circonfrentielle
c
:
1
0
0
0

V V
V V
a
a a
P
p
c
(2)
493
o a est le rayon du forage en expansion, a
0
le rayon initial au
moment du contact sonde-sol la pression p
o
, V
p
le volume de
la sonde au repos, AV
0
volume de contact entre la sonde et le sol
vierge correspondant a
0
.
Lessai pressiomtrique ne mesurant quun dplacement la
paroi de la cavit, c
c
est la seule dformation relative qui peut
tre ainsi dduite de lessai et dans la suite du texte nous
dsignerons simplement
- la dformation pressiomtrique la paroi par c, sans
indice ;
- les pressions nettes la paroi, aprs dduction de la
pression horizontal du sol au repos p* = p p
o
Lhyperbole ainsi ajuste sur les donnes dessai est de la
forme
*
4
3 *
2 1
.
- p C
C
p C C + + = c
(3)
Nous rappelons rapidement les tapes conduisant de cet
ajustement mathmatique sur les points de mesure aux
paramtres du modle hyperbolique prsent ici :
- C
1
est homogne une dformation, C
2
une unit de
contrainte ayant le rle la dimension dun module, C
3
et C
4
des contraintes (pressions). On montre
facilement (Baud & Gambin, 2005, 2008) que C
4
est
lordonne de lasymptote verticale pour c
=
, note p
L
par les Anglo-Saxons, et ici p*
L
, et que les 3 autres
paramtres ne sont pas indpendants et se rduisent 2 :
une dformation c
0
et un module E
0
qui est la pente de
lasymptote oblique :
*
p
*
L 0
0
*
0
- p p

E
p
*
L

+ = c

(4)
- Le module de cisaillement du sol pour les dformations
infinitsimales partir de p
o
, not G
0
est un des
paramtres du modle (Baud et al., 2012), comme on
peut le voir en construisant hors de toute rfrence
une base exprimentale une hyperbole passant par le
point (p
o
, 0), ayant une pente 2G
0
ce point initial et
admettant une asymptote verticale p*
L
:
* *
* *
0
*
1
1
. 2
L
c
p p
p p
G
p
+
= .
(5)
Cette expression ncessite pour tre dtermine de
connatre un point quelconque de la courbe (p
x
, e
x
)
dfinissant comme paramtre complmentaire une
pression nette p*
c
telle que :
( )
* * * *
0
2
* *
*
. . . 2
.
x L x L x
x L
c
p p p p G
p p
p

=
c
(6)
- En complment de G
0
et p*
L
, un seul autre paramtre est
donc ncessaire la dfinition complte du modle, soit
p
c
, soit c
0
:Erreur ! Signet non dfini.
* *
*
.
0
*
*
0
0
0
. 2
1
p p
p
p
p G
L
L
L

|
|
.
|
\
|
+ + =
c
c
c c
(7)
- Les expressions (5) et (7) sont quivalentes ds que lon
dtermine le mme point complmentaire sur la courbe.
2.2. Rle de la limite conventionnelle de lessai
Le choix dune valeur de rupture conventionnelle pour lessai,
p*
LM
, a t dict par la ncessit pragmatique de dduire de
cette caractristique globale du sol au niveau de lessai des
rgles de dimensionnement la rupture ralistes, et
indpendantes de la recherche dune pression limite vraie
p*
L
correspondant une dformation infinie. On peut remarquer
quelle signifie, pour les essais en forage calibrs de diamtre
60 63 mm (2 pouces ) qui sont devenus la pratique et la
norme de lessai, un dplacement absolu de la paroi de 13 mm
environ ( pouce, ou 20% du diamtre).
Lexprience de lutilisation de sondes de diamtres trs
diffrents dmontre que cette convention nest quivalente au
doublement du volume de la cavit que par le hasard historique
du choix pa r Mnard de sondes de 55 63 mm, comme les plus
pratiques et les plus rpandues (en ralit, jusqu 76 mm ou
3 pouces si on considre la pratique du pressiomtre Mnard au
niveau mondial). On constate que des sondes de diamtres trs
diffrents conduisent une pression limite quivalente p*
LM
pour des taux dexpansion qui ne correspondent pas au
doublement de volume de leur cavit, mais un mme
dplacement absolu de la paroi du forage de lordre de 13 mm :
- Dans les petits diamtres, les sondes de diamtres
22 mm et 32 mm dites minipressiomtre , de volume
240 cm
3
au repos, restent dans le domaine pseudo-
lastique ou proche de la pression de fluage lorsquelles
ont doubl de volume, et ncessitent une expansion
jusqu 350 400 cm
3
pour montrer une rupture franche
du sol ; ce volume, atteint sans difficult grce
llancement important de ces sondes, correspond un
dplacement de la paroi de 11 13 mm.
- Dans les diamtres plus importants, les utilisateurs des
sondes autoforeuses de type PAF76, de diamtre
140 mm, ont montr que la rupture tait amorce ds le
dbut de lessai, et ont fix exprimentalement une
quivalence avec la pression limite Mnard pour une
pression p
20
dtermine par une dformation diamtrale
de 20%, soit un dplacement absolu de 14 mm.
(Baguelin et al. 1978).
Ds lors, nous proposons dutiliser comme convention pour
le calcul de la pression limite p
LM
pour tout essai de chargement
radial la dformation conventionnelle c = Ar/R
ref
et non plus
c = Ar/r
0
, ce qui rend c dpendant dune longueur absolue
R
ref
=13 mm dont la signification reste rechercher, mais
indpendant de la sonde utilise et du diamtre du forage.
2.3. Expression de la pression limite conventionnelle
Le modle de comportement de sol hyperbolique de type
lasto-plastique avec crouissage dit Hardening soil model
de Plaxis, B.V. est bien connu. Ce modle utilise dans le repre
(c
1
, q) o c
1
est la dformation axiale des essais triaxiaux, et le
dviateur q = o
1
- o
3
, une courbe hyperbolique passant par
lorigine et dasymptote horizontale q
a
. Par analogie avec ce
modle dfinissant un module E
50
correspondant la
dformation acquise pour la moiti du dviateur de contrainte
de rupture, soit q
f
, nous dfinissons sur la courbe
pressiomtrique le module de cisaillement scant G
M
atteint la
moiti de la pression limite p*
L
(
1
). Soit :
M
L
M
p
G
c . 4
*
=
(8)
En portant cette valeur c
M
dans (7), on obtient

G G
p
M
L

|
|
.
|
\
|
=
1 1
2
0
*
0

(9)
Do une expression du modle pressiomtrique :
( )
0
*
* *
2
*
0
. 2
1 1
2
1
G
p
p p
p
G G
L
M
+
|
|
.
|
\
|
|
|
.
|
\
|
= c
(10)
Pour obtenir la prsentation de lexpression (10), nous avons
choisi un module la moiti de la rupture vraie ,
asymptotique, et non la moiti de la rupture conventionnelle. Ce
second choix aboutit galement une expression c = (p)
dtermine par 3 paramtres physiques (G
0
, p*
LM
et un module
scant G
M
), mais moins simple. En pratique, ces deux modules
G
M
et G
M
sont videmment trs peu diffrents, puisque lon

1
N.B. Les modles hyperboliques hardening soil et pressiomtrique
ntant pas dans les mmes coordonnes de contraintes et dformations,
il ne sagit ici que dune analogie.
494
th
constate sur les essais autofors que le rapport p*
LM
/p*
L
,
analogue au rapport R
f
= q
f
/q
a
, est de lordre de 0,7 0,9.
4. RFRENCES.
La pression limite conventionnelle p*
LM
est celle qui
correspond c = 1, ce qui conduit pour p*
LM
une simple
quation du second degr :
AFNOR, 2012, Pr EN ISO 22476-4,. Reconnaissance et essais
gotechniques Essais en place - Partie 4: Essai au pressiomtre
Mnard.
0 1
. 2
1
. 2
1
* *
0
*
2
*
0
=
|
|
.
|
\
|
+ +
|
|
.
|
\
|
L LM
L
p p
G
p
p
G G
LM
M
(11)
Arsonnet, G., Baud, J.-P., Gambin, M. P. 2005. Ralisation du forage
pour essais pressiomtriques par un systme de tube fendu auto-
for (STAF), in ISP5 PRESSIO 2005, Actes Symp. Intern. Paris,
Gambin, M. P., Magnan, J.-P., Mestat, P. (eds) 22-24 aot, 2005,
Paris: Presses des Ponts. Vol.1 pp 31-45.. dont p*
LM
est lunique solution positive.
Baguelin F., Jezequel J.F., Shields D.H. 1978. The Pressuremeter and
Foundation Engineering. Trans Tech Publications, Clausthal,
Germany, pp 425-439.
2.4. Expression des modules rduits G/G
0
.
Baud, J.-P, Gambin, M. P. 2005. Dduction dune loi de rponse
hyperbolique unique par compilation de courbes pressiomtriques
dans un sol de lithologie homogne, in ISP5 PRESSIO 2005,
Actes Symp. Intern., Gambin, M. P., Magnan, J.-P., Mestat, P. (eds)
Paris, 22-24 aot, 2005, Vol.1 Paris: Presses des Ponts pp 175-186.
Baud, J.-P, Gambin, M. P. 2008. Homogenising MPM Tests Curves by
Using a Hyperbolic Model, in Huang, A.-B., & Mayne, P. W. (eds)
Geotechnical and Geophysical Site Characterization, Proc. ISC3
Taiwan, 1-4 April 2008, London: Taylor & Francis
Baud, J.-P, Gambin, M. P., Schlosser F. 2012. Stress-strain Hyperbolic
Curves with Mnard PMTs, in R.Q. Coutinho (ed.), 4th Int. Conf.
on Geotechnical and Geophysical Site Characterization (ISC'4), P.
de Galinhas, Brazil, 18-21 sept. 2012. London: Taylor & Francis
Baud J.-P., Gambin M., Uprichard S. 1992. Modeling and automatic
analysis of a Mnard pressuremeter test. Gotechnique et
Informatique, Presses des Ponts, Paris.pp 25-32.
Briaud, J.-L., Hossein K. et BarfknechtJ, .2003. Mthode de
dtermination de la courbe charge-tassemennt pour les fondations
superficielles dans les sables. Presses des Ponts, Rev. Fr. de
Gotechnique N105, pp 29-39.
Figure 2. Module tangent rduit Gt/G
0
en fonction de la dformation
pressiomtrique radiale pour la gamme des valeurs du rapport E
M
/p*
LM
.
Briaud, J.-L., 2006. The preboring pressuremeter, some contributions, in
ISP5 PRESSIO 2005, Actes Symp. Intern., Gambin, M. P.,
Magnan, J.-P., Mestat, P. (eds) Paris, 22-24 aot, 2005, Vol.2
Paris: Presses des Ponts pp 103-124.
Les expressions (10) et (11) permettent de calculer les modules
de cisaillement tangent Gt et scant Gs, normaliss par leur
rapport au module tangent initial G
0
, et de visualiser lvolution
de ces modules en fonction de la dformation, pour la gamme
des valeurs possibles pour les rapports G
0
/p*
LM
ou p*
L
/p*
LM
,
rapports caractristiques du type de comportement du sol. En
figure 2 est prsente la mme volution du module tangent
rduit Gt/G
0
en fonction des valeurs de E
M
/p*
LM
, plus familires
aux utilisateurs du pressiomtre.
Clarke B. G., Gambin, M. 1998 Pressuremeter Testing in Onshore
Ground Investigations: A report by the ISSMGE Committee TC16,
Geotechnical Site Investigation, Proc. First Int. Conf. on Site
Characterization (ISC1), P. K. Roberts on & P. W. Mayne eds.,
Vol.2, A. A. Balkema, Rotterdam, pp 1429-1468.
dHemricourt J. 2005. Linterprtation de lessai pressiomtrique : de la
courbe inverse la double hyperbole. in ISP5 PRESSIO 2005,
Actes Symp. Intern. Paris, Gambin, M. P., Magnan, J.-P., Mestat, P.
(eds) 22-24, 2005, Paris: Presses des Ponts. Vol.1 pp 319-328.
Mair, R. J., Wood, D. M. 1987. Pressuremeter Testing, CIRIA Series,
London: Butterworths
Monnet J. et Khlif J. 1994 Etude thorique de lquilibre lasto-
plastique dun sol pulvrulent autour du pressiomtre. Presses des
Ponts, Rev. Fr. de Gotechnique N65.
3. CONCLUSION
La possibilit existe avec la mthode du tube fendu autofor
(forage STAF) et lutilisation dun pressiomtre qui pilote et
rgule lessai automatiquement (pressiomtre Gopac), de
raliser dans des sols varis et offrant la plus large gamme de
rsistances, des essais pressiomtriques autofors, fournissant
instantanment (logiciel industriel Govision) non seulement
la courbe pressiomtrique "classique", mais aussi ses drives
en module tangent et scant depuis une origine p
o
. Elle ouvre
des perspectives de dveloppements nouveaux dont nous
navons pu donner ici quun aperu. En particulier nous pensons
possible de traduire trs facilement les paramtres des courbes
dessais ncessaires au dimensionnement direct de fondations
par la mthode de J.-L Briaud (2003, 2006).
495
Quality control of Cutter Soil Mixing (CSM) technology a case study
Contrle de la qualit des la technologie Cutter Soil Mixing (CSM) une tude de cas
Bellato D., Simonini P.
University of Padua - Department of Civil, Environmental and Architectural Engineering
Grisolia M., Leder E., Marzano I.P.
Sapienza University of Rome - Department of Civil, Environmental and Architectural Engineering

ABSTRACT: The Cutter Soil Mixing (CSM) is a relatively new Deep Mixing (DM) method that offers versatile construction
solutions suitable for various types of ground improvement. Besides the many advantages compared to the most common DM
methods, CSM has a high level of process control. Quality control and quality assurance (QC/QA) procedures are essential aspects of
each DM project, and a successful treatment is related closely to the professional ability to control and verify the DM construction.
This paper presents the results of laboratory tests carried out on wet grab samples collected from a CSM construction site
characterized by the presence of sandy soil. Similar soil-binder mixture were then produced and tested in the laboratory accordingly,
using the same binder adopted for the in situ panel construction and the sandy soil taken directly from the jobsite. A comparison
between the results obtained by UC tests carried out on the wet grab and the laboratory mixed samples is also presented. The results
obtained using an innovative experimental apparatus underline the influence of the physical and chemical characteristics of the natural
soil on the strength gain of the stabilized material.
RSUM : Le Cutter Soil Mixing (CSM), appartenant des mthodes Deep Mixing, est une technique rcente qui offre des solutions
constructives adapts diffrents types d'amlioration du sol. En plus des nombreux avantages sur les mthodes les plus courantes, le
CSM a un niveau lev de contrle de processus. Les procdures de contrle et d'assurance de la qualit (QC/QA) sont des aspects
essentiels du projet, et le succs du traitement est troitement lie la capacit de contrler la phase d'excution. Cet article prsente
les rsultats de tests de laboratoire effectus sur des chantillons prlevs wet grab d'un site CSM caractris par la prsence d'un
sol sableux. Semblables sol-liant mlanges ont ensuite t fabriqus et tests dans le laboratoire avec le mme liant adopt pour la
construction des panneaux in situ et le sol sablonneux prises directement partir du site. En particulier, il est prsent une
comparaison entre les rsultats obtenus avec des essais de compression simple, effectue sur des chantillons prlevs wet grab sur
le site et prouvettes raliss en laboratoire. Les rsultats obtenus par l'utilisation d'un appareil exprimental innovateur ont soulign
l'influence des caractristiques physiques et chimiques du sol naturel sur l'augmentation de la rsistance du matriau stabilis.
KEYWORDS: deep mixing, cutter soil mixing, sandy soil, unconfined compressive strength.

1 INTRODUCTION
The Cutter Soil Mixing (CSM) offers numerous advantages
over the more traditional methods of mixing soils using
standard rotary tools (Fiorotto et al. 2005), being equipped with
two sets of cutting wheels rotating around horizontal axes
producing treated soil panels of rectangular shape.
Several successful applications in different geotechnical
contexts for various engineering purposes have been recently
documented by Gerressen and Vohs (2012).
The Quality Control/Quality Assurance (QC/QA) programs
have the objective to ensure the compliance between the actual
field performance and the design requirements, therefore special
attention is required.
Due to the significant uncertainties related to the site
activity, most of the mix design and mixing procedure
calibration is performed in the laboratory.
In order to develop a tool for an effective comparison
between laboratory and field values, a specific CSM jobsite
located in the city of Zandvoort (NED) has been selected. The
subsoil condition is characterized by the presence of sandy soil.
Despite the fact that higher performance are usually obtained
in the laboratory (Porbaha et al. 2000), the comparison between
strength tests on wet-grab samples and laboratory specimens
have shown sometimes opposite outcomes (Bellato et al. 2012).
The mechanical properties of in-situ improved soil may be
found larger than that of laboratory specimen when using
cement slurry (wet method) to stabilize loose sandy ground due
to water drainage (Yoshimura et al., 2009).
Three types of water drainage may occur during soil mixing
operations: potential expulsion of part of pore-water contained
in the original soil by the injection of the cement slurry;
bleeding of the soil-binder mixture, i.e. drainage of water due to
sedimentation processes; possible drainage towards the
surrounding soil layer of part of the water in the mixture due to
consolidation under the effective overburden pressure.
In this paper the effect of water drainage was investigated
trough an original laboratory experimental apparatus.
To assess the influence of the granular soil type on test
results, the analysis were replicated on a different marine sand.
Moreover, important considerations regarding the significant
influence of the physical and chemical characteristics of the
natural soil on the strength gain of the stabilized sands are
presented and discussed.
2 SITE DESCRIPTION
A requalification activity was planned in Zandvoort, a small
village next to the North Sea coast at about 30 km west of
Amsterdam. Preliminary geotechnical ground investigations
were performed in the jobsite area. The results show a relatively
uniform sand profile characterized by the prevalence of a
medium to fine sand, generally of medium density, whose grain
size distribution is reported in Figure 1. The groundwater level
ranges around 2.5 m below the ground surface.

496
th

Figure 1. Grain size distribution of the Zandvoort and Serapo sands .

To construct the 11.5 m deep CSM panels (2.4 x 0.55 = 1.32
m
2
sectional area) the 1-Phase system was chosen, therefore the
grout was injected on both downward and upward stroke. The
grout composition adopted for the panel production was
characterized by a water-to-cement ratio w/c = 1.12 and a binder
factor = 509 kg/m
3
of natural soil. The cement used was a
special composite cement especially produced for ground
improvement applications.
After mixing, several wet-grab samples were collected from
the fresh panels at about 2,0 m from the ground level and
immediately sealed into watertight tins (inner diameter of 98
mm and height of 113 mm).
3 MATERIALS AND TESTING PROGRAMME
The wet-grab samples collected from the site (in the following
referred as SWGS) were cured under controlled condition
(room temperature of about 20C and at a relative humidity >
95%.) for 40 and 125 days in order to measure also the time
influence on the unconfined compressive strength (UCS) of the
treated soil. Before testing, the wet-grab samples were cored to
provide specimens of 37 and 54 mm in diameter with an aspect
ratio of 2. Finally, the specimens were trimmed to regularize the
bases and wrapped with plastic film to prevent moisture loss.
Laboratory soil-binder mixtures were prepared at the same
grout/sand ratio used on site, according to the treatment
parameters evaluated from the elaborated machine production
data. The grout and the soil were first prepared separately and
then mixed together for 10 minutes using a high power mixer to
produce the stabilized soil, according to the recommendations
for laboratory mixed specimens provided by the Japanese
Geotechnical Society (JGS0821-2000).
The stabilized soil was then poured into plastic moulds 50
mm in diameter and 100 mm in height using the No
Compaction technique (simply consisted in filling the mold) to
realize the laboratory mixed specimens (referred as LS).
Past experiences of sandy soil stabilization (Yoshimura et al,
2009, Grisolia et al, 2010, Bellato et al., 2012) showed the
following occurrences related to water drainage conditions:
The physical properties (water content and wet density)
of sandy soil collected from the site, especially when taken
below the groundwater table, typically are different from the
initial in-situ conditions, due to the loss of fine particles and
water during sampling and transportation to the laboratory;
Bleeding, i.e. separation of water from the soil-binder
mixture, generally occurs immediately after the mixing
process in the bowl and causes the sedimentation of some
amount of cement at the surface;
Every molded sample usually shows the occurrence of
bleeding phenomena, that inevitably leads to a reduction in
the specimens height;
In addition, when the mixture is taken from the bowl for
molding operations, separation among constituent materials
may be observed. This further increases the variability in
terms of amounts of binder, water and sand of the samples.
Moreover, during in-situ soil treatments, some water
drainage may also arise depending on the type of mixing
procedure adopted and the specific subsoil conditions. In
particular, sedimentation mechanisms in the liquid soil-binder
slurry mixture may develop just after the passage of the
mixing tools and some amount of water can be radially
drained away into the surrounding permeable sandy layers
(Yoshimura et al, 2009).
To simulate the effects of water drainage on the mechanical
properties of stabilized soils in the laboratory an original
experimental set up was designed and used (Figure 2).
The apparatus was essentially composed of a watertight
container in which a cylindrical sand core, reproducing the site
conditions, is placed and surrounded by a gravel filter, with
installed an open pipe for water level control (Figure 2a).
A cylindrical cavity was then prepared and filled with the
stabilized soil just after the mixing operations (Figure 2b). After
a time span equal to that adopted on site before sampling, a
laboratory wet-grab specimen was retrieved (Figure 2c and 2d).
The two types of specimens, i.e. laboratory (LS) and
laboratory wet-grab (LWGS), were cured at 20C and at 95%
relative humidity in curing tanks and removed from the moulds
just before the test.
In order to investigate the influence of the sand type and
mineralogy on the performance of the stabilized material, a
marine soil namely Serapo Sand (Figure 1) was also used to
prepare laboratory and laboratory wet-grab specimens.
The experimental investigation mainly consisted of
unconfined compression tests. The specimens were tested at
different curing times, ranging from 7 to 125 days.
To evaluate the influence of the physical and chemical
characteristics of the natural soils (Zandvoort and Serapo sands)
SEM (Scanning Electron Microscope) and EDS (Energy
Dispersive Spectroscopy) analysis were carried out.
A CamScan MX2500 electron microscope, equipped with a
EDAX EDS (energy dispersive X-ray spectrometer) system was
used to determine both the morphology and chemical
composition of the grains. Two small samples for both sands
were first oven dried at 40C for 24 h and then coated with a
layer of carbon using an high-vacuum evaporative coater to
prevent the accumulation of electrostatic charges at the surface
during irradiation.
4 RESULTS AND DISCUSSION
The results of the unconfined compression tests performed on
the three series of samples (SWGS, LS, LWGS) are presented in

Figure 2. Experimental set-up for laboratory wet grab specimens: a)
cavity preparation, b) mixture pouring, c,d) specimen retrieval.
497
Figure 3.
From this figure it clearly appears that higher strength was
provided by the tests carried out on the SWGS, for which UCS
has been found to range between 12 and 16 MPa at 40 curing
days.
UCS obtained from the LS is lower at any curing time
investigated. In particular the UCS was found to be about 7.0
MPa at 40 curing days.
From the same figure, it also appears that UCS of LWGS
approaches the field values.
These results underlines the effectiveness of the
experimental set up in simulating the real field conditions, and
emphasizes the significant effect of drainage conditions, which
increase the UCS of about 1,9 times at 40 curing day.
To evaluate the influence of the type of sand, and, therefore,
of the related drainage effect on strength properties, the same
experimental procedure for sample preparation was replicated
on Serapo sand.
The results of Figure 4 confirm also for this kind of sand an
increment, even though less significant, of the UCS due to the
drainage effect. The increment was about 40% at 40 curing days
for the LWGS specimens with respect to the classical LS.
It is important to note (Figure 4) that similar UCS at 40
curing days was obtained from the LS of both Zandvoort and
Serapo sands (prepared according to JGS0821-2000). This was
expected since the two sands presents similar grain size
distributions.
The results obtained from the newly developed experimental
apparatus show that the type of sand and the corresponding
water drainage effect may greatly influence the mechanical
properties of the stabilized sandy soils.
0
5
10
15
20
25
0 25 50 75 100 125 15
U
n
c
o
n
f
i
n
e
d

c
o
m
p
r
e
s
s
i
v
e

s
t
r
e
n
g
t
h
,

U
C
S

(
M
p
a
)
Curing time, tcur (days)
To investigate in more details the reason of this particular
outcome, mineralogical and microstructural tests were
performed on the two types of sand.
The SEM and EDS analyses results are shown in Figure 5
and 6.
Figure 5 presents two backscattered electron (BE) images of
two different sand grains: the grain on the left referring to
Zandvoort sand, whereas that on the right to Serapo sand.
Generally, both sands are predominantly composed of quartz
minerals, but in the Serapo sand a significant portion of
carbonate particles is present (Figure 6).
In addition, it is easily detectable the more irregular and
angular morphology of quartz grains of Zandvoort sand with
respect to the more rounded, sub-angular carbonate grain of
Serapo sand.
To quantify the degree of angularity different methods have
been proposed in the literature (de Santiago et al., 2008).
Among them, the procedure based on the Fourier descriptors
(Bowman et al, 2001) is one of the most diffuse recent
approaches.
0
SWGS Zandvoort
LS Zandvoort
LWGS Zandvoort

Figure 3. UC test results on Zandvoort sand specimens

Figure 7. Lower order Fourier Descriptors obtained from the two
types of sands.
The boundary of the particle is circumnavigated in the
complex plane at a constant speed. The step size is selected so
that the circumnavigation takes 2 and the number of steps is
2
k
. The complex function presented in Eq. (1) allows to
determine the aforementioned Fourier descriptors

2 /
1 2 /
2
exp
N
N n
n m m
M
m n i
Z y i x

(1)

where x, y are the coordinates of the particle boundary, N is the

Figure 5. Comparison between SE images of a Zandvoort (on the left)
and Serapo sand grain (on the right)

Figure 6. Chemical compositions from EDS analysis performed on a
Zandvoort (on the left) and a Serapo sand grains (on the right).
0
5
10
15
20
25
0 25 50 75 100 125 15
U
n
c
o
n
f
i
n
e
d

c
o
m
p
r
e
s
s
i
v
e

s
t
r
e
n
g
t
h
,

U
C
S

(
M
p
a
)
Curing time, tcur (days)
0
LS Zandvoort
LWGS Zandvoort
LS Serapo
LWGS Serapo

Figure 4. UC test results on Zandvoort and Serapo specimens
498
th
total number of descriptors, n is the descriptor number, M is the
total number of points describing the boundary, m is the index
number of a point on the boundary, Z
n
is the Fourier descriptor
and i is the imaginary number.
Each Fourier descriptor, especially those of the lower order,
are associated to specific and morphological features of the
particle shape.
The average shape descriptors obtained for a reasonable
number of grains taken from each sand sample are shown in
Figure 7.
A clear more unevenness in the boundary of the Zandvoort
grains can be recognized due to the higher contribution of
higher order Fourier descriptors to the shape morphology.

5 CONCLUSIONS
The calibration of relationships between real and laboratory
scale treatment may support soil mixing QC/QA procedures.
In sandy soil, laboratory specimens tests results may be
lower than that obtained by wet grab samples due to water loss
during in situ mixing operations.
The results show that quite a good match may be achieved
by simulating in laboratory the in situ water drainage.
The results obtained from the newly developed experimental
apparatus show that the kind of sand may greatly influence the
water drainage effect on the mechanical properties of the
stabilized soil.
The different degree of angularity of the grains and the
different nature of the minerals composing the two sands
considered in this study should be considered as relevant factors
affecting the performance of the stabilized soil, as well as the
grain size distribution.
Further study are needed to validate and extend the results
and findings described in this case history.
To simulate in situ condition it is also necessary to carefully
take into account other possible factors such as: mixing energy,
use of compressed air, molding technique and curing conditions.
6 AKNOWLEDGEMENT
The authors would like to thank the Hoffman Group and the
Bauer Group, especially Mr. Franz Werner Gerressen & Mr.
Thomas Vohs for kindly providing the construction data for this
study.
The authors wish to thank Mr. Fabrizio Tocci for his help in
conducting the laboratory test at the Department of Civil,
Environmental and Architectural Engineering of Sapienza
University of Rome.

7 REFERENCES
Bellato D., Simonini P., Marzano I.P., Leder E., Grisolia M.,Vohs T.,
Gerrresen F.W., 2012. Mechanical and physical properties of a
CSM cut-off/retaining wall. International Conference on Ground
Improvement and Ground Control (ICGI 2012), University of
Wollongong, Australia. ISBN 978-981-07-3561-6.
Bowman, E.T., Soga, K., and Drummond, T.W., 2001. Particle shape
characterization using Fourier descriptor analysis. Gotechnique, 51
(6), pp. 545-554.
De Santiago, C., Santana, M., and Manzanas, J., 2008. Digital Image
processing and Fourier descriptors analysis of the porosity in
various volcanic rocks. Proc. Of the International Geotechal
Conference - Development of urban areas and geotechnical
engineering, S. Petersburg (Russia), Vol. 2, pp. 449-454.
EN 14679, 2005. Execution of special geotechnical works - Deep
mixing, CEN - Technical Committee CEN/TC 288.
Fiorotto, R., Schpf, M., and Sttzer,E., 2005. Cutter Soil Mixing
(CSM) - An innovation in Soil mixing for creating Cut-off and
Retaining walls. Proceedings 16th ICSMGE, 15 sept. 2005, Osaka
(Japan), pp. 1185-1188.
Gerressen, F.-W. and Vohs, T., 2012. CSM - Cutter Soil Mixing -
Worldwide experiences of a young soil mixing method in soft soils.
Proceedings 4th International Conference on Grouting and Deep
Mixing, New Orleans (USA).
Grisolia M., Marzano I.P., De Lentinis D., Leder E. 2010.
Performance of CSM stabilised soils in geotechnically complex
formations. Proc. Geotechnical Challenges in Megacities, Moscow
(Russia). ISBN 978-5-9902005-2-4.
Grisolia M., Kitazume, M., Leder E., Marzano I.P., Morikawa Y. 2012.
Laboratory study on the applicability of molding procedures for the
preparation of cement stabilised specimens International
Symposium & short courses on Recent Research, Advances &
Execution Aspects of ground improvement works, Brussels.
JGS 0821-00 2000. Practice for Making and Curing Stabilised Soil
Specimens Without Compaction (Translated version). Geotechnical
Test Procedure and Commentary, Japanese Geotechnical Society.
Kitazume, M., 2005. State of Practice Report: Field and laboratory
investigation, properties of binders and stabilised soils.
Procceedings International Conference on Deep Mixing Best
Practice and Recent Advances. Swedish Deep Stabilization
Research Centre, Stockholm (Sweden). Vol. 2, pp. 660-684.
Larsson, S., 2001. Binder distribution in lime-cement columns.
Proceedings of the ICE - Ground Improvement. Vol. 5, No. 3, pp.
111-122.
Marzano I.P., Leder E., Grisolia M., Danisi C. 2012. Laboratory study
on the molding techniques for QC/QA process of a Deep Mixing
work. 3
rd
International Conference on New Developments in Soil
Mechanics and Geotechnical Engineering, Near East University,
Nicosia, North Cyprus. ISBN 975-8359-28-2.
Porbaha, A., Shibuya, S., and Kishida, T., 2000. State of the art in deep
mixing technology. Part 3: geomaterial charachterization.
Proceedings of the ICE - Ground Improvement. 4 (3), pp. 91-110.
Yoshimura K., Mochizuki H., Kon N., Saito S., Suzuki Y., Sugiyama
T., Takahashi, S., 2009. New Procedure for Making Specimens in
Laboratory Mix Test for Sandy Soil Improved by Cement Slurry.
International Symposium on Deep Mixing & Admixture
Stabilization, Okinawa, Japan.
499
Mesures dynamiques lors du battage pntromtrique Dtermination de la courbe
charge-enfoncement dynamique en pointe
Dynamic measurements of the penetration test Determination of the tips dynamic load-
penetration curve
Benz M.A., Escobar E., Gourvs R., Haddani Y.
Sol-Solution Gotechnique Rseaux, Riom, France
Breul P., Bacconnet C.
Institut Pascal-PolytechClermont-Ferrand Universit Blaise Pascal, Clermont-Ferrand, France
RSUM : Dans cet article, nous prsentons les rsultats des rcents dveloppements raliss sur lessai de pntration Panda 3 en
vue de permettre denrichir son exploitation. A partir de la mesure et du dcouplage des ondes cres suite limpact sur lappareil,
nous pouvons obtenir pour chaque coup la courbe charge enfoncement
p
-s
p
permettant de dterminer des paramtres de rsistance et
de dformation mis en jeu pendant lenfoncement de la pointe. Une campagne dessais au laboratoire dans une chambre de calibration
pour deux sols a t mene afin de valider les rsultats obtenus par lexploitation de la courbe
p
-s
p
. Les rsultats obtenus sont
confronts aux rsultats obtenus laide dessais triaxiaux et domtriques.
ABSTRACT: In this paper, we present the results of the recent developments done on the Panda 3 dynamic penetrometer aiming at
improving its use will be presented. From measurement and decoupling of waves created by the impact on the penetrometer, we can
obtain for each blow the load-settlement
p
-s
p
curve allowing determination of the strength and deformation parameters brought into
play during the cone penetration. A series of tests in a calibration chamber for two soils were conducted to validate the results
obtained by the exploitation of the
p
-s
p
curve. The obtained results are compared with the results obtained using the triaxial and
oedometer test.
MOTS-CLS : caractrisation des sols, pntromtre dynamique, Panda 3, propagation dondes, courbe charge-enfoncement.
KEYWORDS : soil characterization, dynamic penetrometer, Panda 3, wave propagation, load-penetration curve.
1 INTRODUCTION
En reconnaissance de sols, du fait des contraintes des essais de
laboratoire (cot, chantillonnage, transport), lutilisation
dessais in-situ est une pratique trs rpandue. Parmi ceux-ci,
les pntromtres dynamiques sont les plus utiliss dans le
monde et sont intressants pour ltude du comportement
dynamique des sols (Tokimatsu 1988) et ce bien que pour la
plupart des ingnieurs, leur caractre dynamique soit considr
comme un dsavantage. Toutefois, les pntromtres
dynamiques ne permettent dobtenir quune seule information
sur le sol : la rsistance de pointe ; et au contraire des
pntromtres statiques qui sont devenus des outils trs
sophistiqus grce lincorporation de diffrents capteurs dans
les pointes, les pntromtres dynamiques sont rests loigns
de ces avances et demeurent dune technicit ancienne. Par
ailleurs, le battage pntromtrique a longtemps t expliqu
par la thorie des chocs de Newton, bien que lon sache que ce
problme ne peut tre rsolu avec la seule application de cette
thorie. A lheure actuelle, on sait que le battage
pntromtrique est mieux reprsent par la thorie de transport
des ondes o le transfert dnergie se fait sous forme dune onde
de compression qui parcourt le pntromtre aprs chaque
impact (Smith 1962, Aussedat 1970).
Cest dans ce cadre et sur la base du pntromtre
PANDA dvelopp depuis plus de vingt ans (Gourvs 1995)
que nous avons conu et dvelopp un pntromtre et un
procd de mesure permettant damliorer linformation
obtenue lors dun sondage : Le PANDA 3 (Benz et al. 2010).
Le principe consiste mesurer et dcoupler les ondes cres
par limpact du marteau sur la tte de lappareil et calculer
ensuite la force, lacclration et la vitesse subis en pointe pour
nous permettre de tracer la courbe charge-enfoncement
p
-s
p
pour chaque coup fourni lors du battage (figure 1).
Figure 1. Principe de lessai Panda 3 (c.f. Benz, 2009).
Lexploitation de la courbe permet de dterminer des
paramtres de rsistance et de dformation du sol mis en jeu
lors de la pntration de la pointe tels que la clrit des ondes,
le module pntromtrique et lamortissement de Smith.
2 PRINCIPE DU PANDA 3
Le principe de lessai est simple : au cours du battage on vient
mesurer dans les tiges, au voisinage de lenclume, les variations
de dformation (x,t) et/ou dacclration a(x,t) entranes par
londe de compression cre par limpact. En effet, quand le
marteau de masse M anim dune vitesse v
m
heurte la tte du
pntromtre, une onde de compression u(x,t) est engendre
dans celui-ci et se propage une vitesse constante c
t
vers le
cne. Lorsque u(x,t) arrive linterface cne/sol, une partie de
celle-ci est utilise pour dformer le sol et une autre partie est
rflchie vers le haut. La propagation de u(x,t) dans les tiges est
dcrite par la quation (1) et sa solution gnrale correspond
la superposition de deux ondes, u
d
et u
r
, descendante et
remontante (quation 2). Lors de son parcours u(x,t) entrane
dans tout point x des tiges des variations de dformation (x,t) et
500
th
de vitesse particulaire v(x,t) reprsentes par la superposition
des ondes lmentaires.
2
2
2
2
2
) , ( ) , (
x
t x u
c
t
t x u
t
c
c
=
c
c
(1)
) / ( ) / ( ) , (
t r t d
c x t u c x t u t x u + + = (2)
) / ( ) / ( ) , (
t r t d
c x t c x t t x + + = c c c (3)
) / ( ) / ( ) , (
t r t d
c x t v c x t v t x v + + = (4)
Il est possible de montrer que lexpression (4) peut tre
exprime en fonction des ondes de dformation
d
et
r
.
| ) / ( ) / ( ) , (
t r t d t
c x t c x t c t x v + = | c c (5)
La connaissance des ondes
d
(t) et
r
(t) permet ainsi de
dcrire entirement le phnomne dynamique du battage
pntromtrique en tout point x le long des tiges. Dans la
pratique, cette connaissance peut se faire laide de mesures
ralises par le biais de jauges de dformation et/ou
dacclromtres. Toutefois, dans les enregistrements raliss
lors du battage, ces ondes se trouvent souvent imbriques les
unes avec les autres et il devient ncessaire de les dcoupler.
2.1 Dcouplage dondes et construction de la courbe charge-
enfoncement en pointe
Diffrentes mthodes peuvent tre employes pour dcoupler
les ondes
d
(t) et
r
(t) partir des enregistrements raliss.
Celles-ci diffrent suivant le type de mesures (dformation,
acclration), suivant la quantit (1 n) ainsi que suivants les
conditions aux limites imposes. Toutefois, il a t montr que
la mthode propose par (Casem et al. 2003) est celle qui
sadapte le mieux au cas du battage pntromtrique. A partir
des enregistrements
A
(t) et v
A
(t) raliss dans un point A, les
ondes
d
(t) et
r
(t) sont dcouples daprs :
(
+ =
(
=
t
A
A r
t
A
A d
c
t v
t t et
c
t v
t t
) (
) (
2
1
) (
) (
) (
2
1
) ( c c c c
(6)
En supposant les efforts externes nuls le long de tiges, la
connaissance de
d
(t) et
r
(t) permet de calculer les signaux de
force F
N
(t) et de vitesse v
N
(t) pour tout point N situ en dessous
du point de mesure A, notamment dans la pointe, selon la
solution propose par (Karlsson et al. 1989).
| |
| ) ( ) (
2
) ( ) (
2
1
) (
) 1 ( 1 ) 1 ( 1
) 1 ( 1 ) 1 ( 1

A A + +
A + A + =
n n N n n N
n
n n N n n N N
t t v t t v
Z
t t F t t F t F
|
(7)
| |
| ) ( ) (
2
1
) ( ) (
2
1
) (
) 1 ( 1 ) 1 ( 1
) 1 ( 1 ) 1 ( 1

A A + +
A + A + =
n n N n n N
n
n n N n n N N
t t F t t F
Z
t t v t t v t v
|
(8)
avec t
n-(n-1)
=(x
n-1
-x
n
)/c
n
et Z
n
=E
n
A
n
/c
n
limpdance mcanique
de la section n dfinie par son module de Young E
n
, sa section
A
n
et la clrit de londe c
n
. Ainsi, partir des enregistrements
A
(t) et a
A
(t) il est possible de calculer les signaux de force F
p
(t),
vitesse v
p
(t) et enfoncement s
p
(t) en pointe et donc de tracer la
courbe o
p
-s
p
pour chaque coup de marteau fourni lors du
battage. De nombreux tests ont t raliss pour valider la
faisabilit dun tel essai. Dans la figure 2 on prsente un
exemple de courbes obtenues pour deux types de sol. On peut
remarquer que ces courbes sont rptitives pour un mme
matriau et varient selon la nature du milieu auscult permettant
didentifier des comportements diffrents.
2.2 Exploitation de la courbe charge-enfoncement o
p
-s
p

Pour interprter la courbe o
p
-s
p
il a t propos une
mthodologie fonde sur les travaux de (Smith 1962).
Figure 2. Exemple de courbes charge-enfoncement pntromtriques
obtenues avec le Panda 3 pour deux types de sol (c.f. Benz 2009).
Pour chaque impact, on calcule des paramtres de rsistance,
damortissement, de dformation ainsi que de clrit dondes
dans le sol.
2.2.1 Paramtres de rsistance
En supposant que la contrainte en pointe qd(t) est la
rsultante des composantes statiques R
s
(obissant une loi
lastoplastique parfaite) et dynamiques R
d
(t) (proportionnelle
la vitesse denfoncement v
p
(t)); on dtermine la valeur de R
s
en
admettant que lorsque v
p
(t) est nulle la composante dynamique
R
d
(t) sannule et R
s
est donc gale qd(t).
Figure 3. (a) Modle de Smith et (b) exploitation de la courbe o
p
-s
p
Les valeurs de R
d
(t) et du coefficient damortissement de
Smith J
s
sont dtermins dans lintervalle denfoncement [s
e
;
s
max
], avec s
e
et s
max
les enfoncements lastique et maximal, en
crivant que R
d
(t) = q
d
(t)-R
s
et J
s
=R
d
(t)/(R
s
v
p
(t))(figure 3.b).
2.2.2 Paramtres de dformation
Une fois lenfoncement maximal s
max
atteint, on admet que
sol et pntromtre se mettent vibrer ensemble dans un rgime
pseudo-lastique. Dans cette partie de la courbe o
p
-s
p
, deux
modules sont ainsi dfinis : un module de dchargement E
d
P3
(droite AB) et un module de rechargement E
r
P3
(droite BC)
(figure 3.b). En assimilant la pointe une petite plaque
encastre lintrieur dun massif lastique semi-infini, on
calcule la valeur de E
d,r
p3
en appliquant la quation de
Boussinesq (9) propose par (Arbaoui 2006).
M
p
p
d r d
p
k
d
s
q
E
1
4
) 1 (
2 '
3
t
v
A
A
=
(9)
avec suppos gal 0,33, d
p
le diamtre de la pointe et k
M
le
coefficient dencastrement de mindlin.
2.2.3 Clrit des ondes c
p
et c
s

La clrit des ondes de compression c
p
et de cisaillement c
s
dans le sol est calcule par le biais des polaires de choc
prconise par (Aussedat 1970). Pour chaque impact on mesure
les pics des ondes descendantes et remontantes dans un espace
de temps to+2L
t
/c
t
nous permettant de calculer la valeur de c
p
(Benz 2009). La valeur de c
s
est calcule daprs lexpression
(10) en supposant la valeur de gale 0,33.
501
Tableau 1 Caractristiques des matriaux et des prouvettes.
Caractristiques Sable dAllier (GTR : B1 - USCS : SP) Argile de Laschamps (GTR : A2 - USCS : ML)
T
max
/2mm/80m/IP 2,73mm / 90,3% / 1,14% / NP 0,08mm / 99,3% / 96,3% / 15,1% (W
L
:42,7% )
OPN W
OPN
17,07 kN/m
3
7% 18,08 kN/m
3
15,8%
prouvettes 1 2 3 4 1 2 3 4 5 6
- w% 0,8 0,8 14,6 14,6 0,6 0,6 0,6 15,38 15,61 19,15
- s kN/m
3
16,1 16,83 16,26 16,52 15,73 16,69 17,43 16,65 17,43 16,72
- %OPN D.R% 94 - 28 99 - 57 95 - 35 97 - 44 87 92 96 92 96 93
- qd(MPa) Pda2 2,1 4,5 2,4 4,6 9,8 33,9 43 1,9 2,8 1,2
- E
triaxial
MPa 2,9 11,2 4,1 6,3 - - - - - -
- E
oed
MPa (*) 28,1 97,6 37 57,6 19,6 38,1 63,5 24 36 17,6
) 1 ( 2 ) 2 1 (
p s
c c (10)
Enfin, lissue dun sondage Panda 3, on trace en fonction
de la profondeur z les pntrogrammes de : rsistance de pointe
q
d
, clrit des ondes c
p
et c
s
, module pntromtrique E
d
p3
et du
coefficient damortissement J
s
(i.e. figure 6)
3 ESSAI EN CHAMBRE DE CALIBRATION
Une srie dessais a t ralise dans une chambre de calibration
compose dun cylindre en acier de 400mm de diamtre et
810mm de hauteur (figure 4). Lobjectif tait dune part de
valider les rsultats obtenus par lexploitation de la courbe
p
-s
p
du Panda 3, dautre part de vrifier leur sensibilit ltat du
sol et enfin de les confronter aux rsultats obtenus par le biais
dessais classiques (domtre, triaxial). Deux sols ont t
employs : un sable dAllier et une argile de Laschamps.
3.1 Ralisation des prouvettes et des essais
Diverses prouvettes ont t ralises en faisant varier la
densit
s
et la teneur en eau w pour chaque sol tudi (tableau
1). Le compactage des prouvettes est ralis laide dun vrin
quip de diffrents capteurs suivant la procdure dcrite par
(Chaigneau 2001). Lors du compactage, des essais de
chargement domtriques taient raliss selon la procdure
prconise par (Gorena 2007). Lexploitation des courbes
domtriques a permis de dterminer les valeurs E
oed
prsents
dans le tableau 1. Une fois les prouvettes ralises, trois essais
de pntration taient raliss avec le Panda 2 et le Panda 3 et
dont au moins 1 avec ce dernier (figure 4.c).
Figure 4. (a) essais de chargement domtriques lors du compactage,
(b) prouvette compacte et (c) ralisation des essais Panda 3.
Des essais triaxiaux ont t galement raliss pour le sable
dAllier. Les chantillons, conus mme densit que les
prouvettes auscultes, ont t cisaills dans un chemin triaxial
pression de confinement gale 50kPa et les rsultats obtenus
(module tangent initial E
triaxial
) sont prsents dans le tableau 1.
3.2 Rsultats
Au total, une dizaine dessais Panda 3 ont t raliss, un
pour chaque prouvette (tableau 1). Pour chaque essai et donc
pour chaque impact ralis lors du battage, la courbe
p
-s
p
a t
dtermine. Un exemple des courbes obtenues lors dun essai
pour quatre prouvettes diffrentes est prsent dans la figure 5.
Dans lexemple, les chelles de charge
p
et denfoncement s
p
sont normalises par rapport la charge
p-max
et
lenfoncement s
p-max
maximal mesurs pour chaque coup trac.
Figure 5. Courbes
p
-s
p
obtenues pour du Sable dallier (a)
s
: 16,26kN/m
3
,
w% :14,6, (b)
s
: 16,83kN/m
3
, w% :0,8 et pour de lArgile de Laschamps
(c)
s
:16,72kN/m
3
, w% :19,15 et (d)
s
:17,43kN/m
3
, w% :0,6
(b) sable dallier
dense
(a) sable dallier
moyennement dense
(d) argile de Laschamps
dense
(c) argile de Laschamps
moyennement dense
A partir des courbes
p
-s
p
exposes, on peut remarquer que
celles-ci sont caractristiques et rptitives pour chaque sol
auscult. De mme, lallure des courbes est sensible ltat du
sol. Dans le cas du sable dAllier, on constate que la courbe
p
-
s
p
devient presque asymptotique avec laugmentation de la
densit et que le retour lastique augmente aussi (figure 5.a,b).
Quant largile, on peut remarquer que les courbes
p
-s
p
sont
sensibles ltat du sol. Lorsque le sol est trs humide (figure
5.c) la contrainte augmente trs rapidement jusquau pic
p-max
,
puis elle chute mme vitesse vers la valeur rsiduelle. Cela
nest pas le cas lorsque le sol est sec et plus dense (figure 5.d)
Pour chaque prouvette, lensemble de courbes
p-max
est
exploite automatiquement selon la procdure cite dans 2.2.
Les paramtres calculs sont tracs sous forme de
pntrogrammes, tel que montr dans la figure 6. Dans
lexemple prsent, on compare les pntrogrammes obtenus
pour deux prouvettes de sable dAllier diffrents tats de
densit (D1 et D2). De manire gnrale, on peut constater que
lensemble des paramtres est sensible lvolution de la
densit du milieu.
A partir des pntrogrammes obtenus nous avons calcul la
valeur moyenne pour chaque paramtre issu de lexploitation
des courbes
p
-s
p
(tableau 2). On peut remarquer que ceux-ci
varient en fonction de la nature et de ltat du sol auscult. De
mme, lordre de grandeur des paramtres calculs, tel que la
502
th
clrit des ondes c
p
et c
s
, a une bonne correspondance avec ceux proposs dans la littrature (Sharour et Gourvs, 2005).
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0.0 0.2 0.4 0.6 0.8 1.0
Coefficientd'amortissementJs
Js D1 Js D2
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
10 100
Module,E
d
p3 (MPa)
EdP3 D1 Edp3 D2
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
10 100 1000
Clritdesondes (m/s)
Cp D1 Cs D1
Cp D2 Cs D2
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0.1 1 10 100
P
r
o
f
o
n
d
e
u
r
,
z
(
m
)
Rsistancedepointe,(MPa)
Pda2 D1 Pda3 D1
Pda2 D2 Pda3 D2
*Pda2 : valeurs
obtenus avecla
formuledes
Hollandais selon
(NF 94105).
Figure 6. Rsultats Panda 3 obtenus en chambre de calibration pour un sable dAllier sec 2 tats de densit diffrents D1, D2 (prouvettes 1 et 2).
Tableau 2 Synthse des rsultats obtenus laide du Panda 3 dans chambre de calibration
Sable dAllier (GTR : B1 - USCS : SP) Argile de Laschamps (GTR : A2 - USCS : ML)
prouvettes 1 2 3 4 1 2 3 4 5 6
w% 0,8 0,8 14,6 14,6 0,6 0,6 0,6 15,38 15,61 19,15
s kN/m
3
16,1 16,83 16,26 16,52 15,73 16,69 17,43 16,65 17,43 16,72
Rsultats Panda 3 (valeurs moyennes calcules pour chaque prouvette)
- qd
P3
(MPa) 2,3 11 1,6 2,6 8 32 65 1,5 3 4
- c
p
(m/s) 620 1504 383 504 2380 6151 5775 597 989 1081
- c
s
(m/s) 298 708 184 242 1145 2955 2773 286 475 519
- E
d
P3
(MPa) 37 120 41 65 60 130 221 31 62 64
- Js (Ns/m) 0,26 0,09 0,15 0,15 0,63 0,65 0,70 0,43 0,56 0,75
Par ailleurs, nous avons confront les valeurs moyennes des
modules E
d
P3
obtenus pour chaque prouvette avec ceux obtenus
par le biais des essais triaxiaux (cas du sable) et domtriques
(figure 7). Dans le cas du sable, on peut constater quil existe
une trs bonne corrlation entre les modules E
d
P3
et les modules
triaxiaux E
triaxial
et domtriques E
oed
(figure 7.a). Toutefois, la
corrlation E
d
P3
-E
oed
est moins bonne pour le cas de largile
(figure 7.b). Cela peut sexpliquer par le caractre dynamique
du chargement pntromtrique et par laugmentation des
pressions interstitielles au sein du milieu pendant lenfoncement
du cne. De mme, les valeurs du module E
oed
obtenues lors du
compactage des prouvettes peuvent tre perturbes par la
cration des surpressions interstitielles lors du chargement.
Figure 7. Corrlation entre les modules E
d
P3
et les modules triaxiaux E
triaxial
et
domtriques E
oed
pour (a) Sable dAllier et (b) Argile de Laschamps.
4 CONCLUSION
Lessai au pntromtre dynamique bien que largement
utilis travers le monde souffrait du peu de dveloppements
raliss pour permettre damliorer la qualit des mesures
effectues et enrichir son exploitation. Cet article a prsent les
dveloppements rcents raliss sur le pntromtre Panda 3,
qui permettent partir de la mesure et du dcouplage des ondes
cres par limpact sur lappareil, dobtenir pour chaque coup
une courbe charge-enfoncement
p
-s
p
du sol test.
Lexploitation de cette courbe permet de dterminer des
paramtres de rsistance (rsistance de pointe), de dformation
(module dynamique), des caractristiques damortissement et de
clrit des sols auscults en fonction de la profondeur tout au
long du sondage. Les tests raliss en chambre de calibration
ont montr la bonne rptabilit des mesures ainsi que leur
sensibilit aux conditions du sol (tat de serrage et tat
hydrique) et leur bonne adquation avec les valeurs de la
littrature. Des tudes comparatives entre les modules obtenus
au pntromtre et ceux obtenus partir dessais de chargement
domtrique ou triaxial ont montr une bonne corrlation pour
le sable. Cet outil est maintenant oprationnel in situ et des
travaux complmentaires sont actuellement mens en vue
dobtenir une meilleure interprtation des paramtres extraits
partir de la courbe charge/enfoncement.
5 REFERENCES
Tokimatsu, K., (1988). Penetration tests for dynamic problems, Proc.,
ISOPT1, 1, pp. 177-136.
Gourvs R, Barjot R (1995). Le pntromtre dynamique PANDA,
Proc. of ECSMFE, Copenhagen, Denmark, 1995, p 83- 88.
Benz, M.A. (2009). Mesures dynamiques lors du battage du
pntromtre Panda 2. Thse de lUniversit Blaise Pascal,
Clermont-Fd, (2009).
Casem, D., Fourney, W. et Chang, P. (2003), Wave separation in
viscoelastic pressure bars using single-point measurements of strain
and velocity, Polymer testing 22, 2003, pp 155-164.
Chaigneau, L. (2001). Caractrisation des milieux granulaires de surface
laide dun pntromtre. Thse de lUniversit Blaise Pascal,
Clermont-Fd (2001).
Karlsson L.G., Lundberg B, Sundin K.G. (1989), Experimental study of
a percussive process for rock fragmentation, Int J Rock Mech Min
Sci Geomech, 1989, pp.45-50.
Smith, E.A.L. (1962), Pile-Driving Analysis by the Wave Equation,
ASCE. Paper No. 3306, Volume 127, Partie I, 1962, pp 1145-1193.
Arbaoui, H., Gourvs, R., Bressolette, Ph., Bod, L. (2006), Mesure de
la dformabilit des sols in situ laide dun essai de chargement
statique dune pointe pntromtrique, Canadian geotechnical
journal, vol. 43, 2006, pp. 355-369.
Sharour, I et Gourvs R (2005) Rconnaissance des terrains in situ. Ed.
Hermes Lavoisier.191pp. 2005
Aussedat G. (1970). Sollicitations rapides des sols, Thse de doctorat,
Facult de sciences de lUniversit de Grenoble.
Gorena A. (2007). Mesure des proprits de dformabilit de sols de
rfrence, Mmoire dIngnieur CUST, Juin 2007.
Benz, M.A., Gourvs, R. et Haddani, Y. (2010). Dtermination de la
courbe charge enfoncement dynamique en pointe pntromtrique
par dcouplage des ondes. JNGG 2010, Grenoble 7-9 Juillet 2010,
France Tome 1, pp17-24.
(a) Sable dAllier (b) Argile de Laschamps
503
th
1
Performance of a deep excavation in downtown Toronto
Performance d'une excavation profonde au centre-ville de Toronto
Cao L.F., Peaker S.M., Ahmad S.
SPL Consultants Limited, Ontario, Canada
ABSTRACT: This paper presents field measurements of soldier pile walls installed in the clayey soils and shaly rock in downtown
Toronto. The method of deducing wall bending moments from the inclinometer measurements was evaluated and discussed. Back-
analysis using a finite element program has been carried out to evaluate the shoring wall performance as well as the creep behaviour
of the shaly rock. Recommendation for the design of soldier pile walls in the similar soils and bedrock conditions were provided.
RSUM : Cet article prsente des mesures de terrain pour des murs de pieux soldats installs dans les sols argileux et le roc shaleux
du centre-ville de Toronto. La mthode donnant les moments de flexion du mur partir des mesures inclinomtriques a t value et
discute. Une analyse a t effectue avec un programme d'lments finis pour valuer la performance du mur tay ainsi que le
comportement en fluage du shale. Des recommandations, pour la conception des murs de pieux soldats dans des conditions
semblables de sols et de roc, ont t fournies.
KEYWORDS: deep excavation, field measurement, inclinometer, bending moment, finite element, time-dependent deformation
1 INTRODUCTION
Underground structures such as basements and subway have to
go deeper today than in the past due to limited space in densely
populated urban environments. As deep excavations induce
large stress and strain, underground structures and the adjacent
structures/utilities will confront risks of being damaged. As the
soil/rock stress-strain behaviour is non-linear and affected by
many factors, it is difficult to predict the ground movement
induced by excavation. In practices, field measurements are
widely used to monitor soil/rock behaviour and to control
ground movement.
This paper presents a case study of a deep excavation in
downtown Toronto. Soldier piles with tiebacks were used to
support the excavation. Two inclinometers and one hundred and
twenty seven reflective targets were installed to monitor the
movements of the shoring walls during and after excavation.
The inclinometer measurements have been used to deduce the
wall bending moments. A finite element program has been
carried out to evaluate the performance of the shoring walls. It
is found that the total stress analysis leads a good prediction of
wall deflections during the excavation, whereas the effective
stress analysis is required to model the behaviour of shoring
walls after excavation. The back-analysis also shows the
evidence of the creep movement of the shaly rock.
2 GROUND CONDITION AND TEMPORARY SUPPORT
SYSTEM
The site is located at 352 Front Street West in Toronto, Ontario.
Field investigation with drilled boreholes revealed that the site
stratigraphy was made up of about 1 m thick, compact sand to
gravel fill with asphalt surface overlying 3 to 4 m thick, firm to
hard clayey silt fill over 2 to 5 m thick, stiff to very stiff clayey
silt till. Both clayey fill and till are low plasticity soils. Georgian
Bay formation of shale and limestone/siltstone was encountered
at 9 to 9.5 m below existing ground surface. The groundwater
table was about 5 m below grade.
The Georgian Bay formation is generally massive shaly rock
with widely spaced jointing and sub-horizontal bedding planes.
The influence of sedimentary shaly bedrock formations on the
engineering performance of underground structures in Southern
Ontario was summarized by Lo (1989). The shaly bedrock
formations are subjected to high in-situ horizontal stresses with
typical coefficient of lateral earth pressure K
o
of 4 or greater.
Upon relief of the high residual horizontal stresses, time-
dependent, creep-like deformations take place. These time-
dependent deformations that are highly stress dependent, persist
well beyond the initial elastic deformations and generally
exceed the magnitude of the elastic movements.
Soldier piles of steel H-beam W610x82 at 3.05 m spacing
with wood lagging were employed to support an approximately
14 m deep excavation in which 9.3 m excavation was inside
overburden soils and 4.7 m excavation inside the bedrock. The
soldier piles were installed typically 16 m below the existing
ground surface in 910 mm diameter drilling holes. The drilling
holes were backfilled by 0.4 MPa concrete with the exception at
the pile toe, where 20 MPa concrete was used to support the pile
toe. Two layers of tiebacks were installed at approximately 3.3
and 8.3 m below the existing ground surface, respectively to
support the soldier pile walls during excavation. The tiebacks
were installed within 150 mm dia. cased boreholes and bonded
in bedrock. Each tieback was made up of 6 to 7 numbers of 15
mm strand tendons. The upper and lower tiebacks were installed
at 45
o
and 25
o
to the horizontal direction, respectively. The bond
length of the upper tiebacks was typically 5 m and the free
length 9.4 m. The bond length of the lower tiebacks was
typically 3 m and the free length 3.9 m. The tiebacks were
generally post-grouted the day after they were installed. The
typical design loads for the upper and lower tiebacks were 1000
and 800 kN respectively. Figure 1 shows outlook of soldier
piles with wood lagging supported by tiebacks.
Two performance tests for the tiebacks were conducted up
to 138% and 200% of the design load, respectively. The test
loads were maintained for 0.5 to 1 hour and the tests met the
PTI criteria (PTI, 1996). Proof tests were carried out for all
504
th
th
International Conference on
2
Soil Mechanics and Geotechnical Engineering, Paris 2013
0
5
10
15
20
25
30
-15 -10 -5 0 5 10 15
D
e
p
t
h

(
m
)
Lat eral Deflection (mm)
Inclinometer after upper tieback installed
Reflective target after upper tieback installed
Inclinometer after lower tieback installed
Reflective target after lower tieback installed
Inclinometer 1 d after excavation to bottom
Reflective target 1 d after excavatio to bottom
Inclinometer 11 m after excavation to bottom
Reflective target 11 m after excavation to bottom
Excavation
Level
Bedrock Surface
Tieback
tiebacks. The test load was 133% of the design load and
maintained for 10 minutes. All tiebacks except three tiebacks
met the PTI criteria. The three tiebacks could not reach the test
load due to the broken wires. A lower design load was used for
the three tiebacks. Detailed discussions on the tiebacks are
presented by Cao and Peaker (2011).
Figure 1. Outlook of soldier piles with wood lagging supported by
tiebacks
0
5
10
15
20
25
-600 -400 -200 0 200 400 600
D
e
p
t
h

(
m
)
Bending Moment (kN/m)
Inclinometer after upper tieback installed
1 d after excavation to bottom
11 m after excavation to bottom
Excavation Level
Bedrock Surface
Two inclinometers were installed inside the soldier pile walls
during the pile installation. The inclinometers were monitored
during and after the excavation. Figure 2 shows the monitoring
results of one inclinometer including the reading taken after
upper and lower tieback installations, 1 day after the excavation
to bottom, and 11 months after the excavation. The lateral
deflections measured by reflective targets installed at the top of
soldier piles are also shown in Figure 2. The measurements of
reflective targets are consistent with the inclinometer
measurements.
3 BENDING MOMENT FROM WALL INCLINOMETER
MEASUREMENTS
The inclinometer measurements have been used to estimate wall
bending moments by some researchers (Poh et al. 1999). The in-
wall inclinometers provide a direct measurement of the rotation.
These measurements can be subsequently converted into wall
deflections along the wall. The wall curvatures can be derived
from the wall deflection data. The second differential equations
of the wall deflection will give the along the wall. The
bending moment M can be computed from using the
following equation (West, 1993)

M = I = KI(d
2
y/dx
2
) (1)
where E is the elastic modulus of the wall, I is the inertia
moment of the wall, y is the lateral deflection of the wall and x
is the distance along the wall.
Using Microsoft Excel spreadsheet, the inclinometer
measurements were fitted with a sixth- degree polynomial and
double differentiation of this polynomial gave . The coefficient
of determination value obtained during the curve fitting ranged
from 0.98 to 0.99, indicating minimal error during the process
of curve fitting. The Youngs moduli of 0.4 MPa concrete and
H-beam W610x82 were taken as 2.8 GPa and 200 GPa,
respectively. The sum of concrete EI and H-beam EI was used
in the calculation of the bending moment. Figure 3 shows the
bending moments deduced from the wall inclinometer
measurements. Higher bending moments were observed at the
locations of tiebacks. However, significant high values of
bending moments were obtained near the ground surface, which
is against the typical distribution of bending moment along a
cantilever beam. This could be an error inducted in the double
differentiation of the wall deflection. Further study using a
Figure 2. Lateral deflection of soldier pile wall

Figure 3. Wall bending moments deduced from inclinometer
measurements
higher degree polynomial and a defined boundary condition is
required.
4 FINITE ELEMENT BACK-ANALYSIS
The finite element program Phase 2 (version 8.0) was used in
the back-analysis. The program can be used to simulate
excavation in soil and rock under plane strain condition. Six-
node triangle elements were used to model the soil and bedrock
media. The soldier pile wall and tiebacks were modelled by
505
th
3
structural beam elements. The analysis modelled a half width of
the excavation where the right-hand boundary of the mesh
represented the line of symmetry at the centre line of
excavation. The finite mesh was 140 m long and 84 m deep.
The half width of the excavation was 20m. The bottom
boundary was strained from both vertical and horizontal
movements. The left-hand and right-hand boundaries were free
to move in the vertical direction.
The soil and bedrock profiles used in the analysis were
based on borehole logs. The groundwater level was taken at 5 m
below existing ground surface at the initial stage. During the
excavation, the groundwater level was assumed to be drawn
down to the excavation level at the excavated side.
For the surface sandy fill, the Youngs modulus E of 25 MPa
and the friction angle of 30
o
were assumed. For the clayey
soils, the undrained shear strength s
u
was estimated from 6N,
where N is the blow counts of the standard penetration testing.
The residual s
u
was taken as 50% of the initial s
u
. The undrained
E of clayey soil was estimated from 1500s
u
for the native low
plasticity clayey silt till and 500s
u
for the clayey silt fill,
respectively. The soil Poissons ratio was taken as 0.3. The unit
weight was obtained from available laboratory testing data.
Mohr Coulomb failure criterion was used for soils. The soil
properties used in the analysis are shown in Table 1.
Table 1. Soil parameters used in the finite element analysis
Depth
(m)
Type
(kN/m
3
)
E
(MPa)
s
u
(kPa)
K
o
0 1
Sandy
fill
20 25 0 30 0.5
1
4.5
Clayey
fill
20 30 60 - 0.75
4.5
9.3
Clayey
till
21 225 150 - 0.75
0
5
10
15
20
25
30
-15 -10 -5 0 5 10 15
D
e
p
t
h

(
m
)
Inclinometer readings after upper tieback installed
Computered deflections after upper tieback installed
Inclinometer readings after excavation to bottom
Computered deflections after excavation to bottom
Excavation Level
Bedrock Surface
Note: K
o
is the coefficient of lateral earth pressure (total stress)
For the jointed shaly bedrock, the generalized Hoek-Brown
constitutive model was used. The following parameters were
used to generate the generalized Hoek-Brown rock-mass
strength criterion: (1) The geological strength index was taken
as 60 for sound bedrock, respecting blocky to very blocky, good
to fair joint surface, and 30 for weathered bedrock, respecting
blocky/disturbed/seamy joint surface; (2) The intact rock
constant was taken as 8 for sound bedrock (highest value for
shale) and 4 for weather bedrock (lowest value for shale); (3)
The disturbance factor was taken as 0 for excellent quality
controlled excavation; and (4) The modulus ratio was taken as
250 for bedrock (highest value for shale) and 150 for weather
bedrock (lowest value for shale). The intact compressive
strength was obtained from available results of rock point load
testing and unconfined compressive testing. The rock Poissons
ratio was taken as 0.15. Based on the above assumptions, the
obtained strength parameters for the generalized Hoek-Browns
model are summarized in Table 2.
Table 2. Rock parameters used in the finite element analysis
Depth
(m)
Type

(kN/m
3
)
E
(MPa)
m
b
s a K
o
9.3
11.3
Weathered
bedrock
25 244 0.3 0.004 0.52 2
>11.3
Sound
bedrock
26 3072 1.3 0.004 0.51 4
Note: m
b
, s and a are parameters used in generalized Hoek-Browns
model
The soldier pile wall was modelled as reinforced concrete
with W610x82 at spacing of 3.05 m. The equivalent thickness
of 0.4 MPa concrete was taken as 0.2 m and the Youngs
modulus was 2.8 GPa. The concrete compressive and tensile
strengths were taken as 400 kPa and 40 kPa, respectively. The
compressive and tensile strengths of W610x82 were taken as
345 MPa. The Poissons ratio for steel and concrete was taken
as 0.2. The equivalent bolt diameters for the upper and lower
tiebacks were taken as 32 mm and 24.5 mm, respectively. The
Youngs modulus of tiebacks was taken as 200 GPa. The bond
shear stiffness was taken as 6000 kN/m/m based on the tieback
proof test results. The bond lengths of the upper and lower
tiebacks were taken as 5 m and 3 m, respectively. The spacing
of tiebacks was taken as 3.05m.
The measured and computed wall deflections after the
installation of upper tiebacks and the excavation just to the
bottom are shown in Figure 4. The computed wall deflections
are in a good agreement with the inclinometer measurements,
indicating that the in-put parameters used in the analysis are
reasonable.
The computed bending moments for the excavation just to
bottom are compared with those deduced from the inclinometer
measurements as shown in Figure 5. The bending moments
deduced from the inclinometer measurements are comparable
with the computed except near the ground surface where
significant high values deduced from the inclinometer
measurements. Ignoring the high bending moments near the
ground surface, the bending moments deduced from the
inclinometer measurements can be used for the checking of the
capacity of the soldier piles.
The inclinometer measurements show that up to 7 mm lateral
movement was developed after the excavation to bottom as
shown in Figure 2. This could be due to three possible reasons:
(1) the consolidation of clayey soil; (2) de-stressing of tiebacks;
and (3) time-dependent deformation of the shaly rock upon
relief of the initial high horizontal stresses. The first two
possible reasons have been studied in the finite element analysis
using the effective parameters and reduced modulus for the
Figure 4. Measured and computed wall deflections
506
th
th
4
5 CONCLUSIONS
0
5
10
15
2
2
-600 -400 -200 0 200 400 600
D
e
p
t
h

(
m
)
Bending Moment (kN/m)
In-wall inclinometer has an importation role in the monitoring
of shoring walls during and after excavation. Wall bending
moments can be estimated from the inclinometer measurements
except near the ground surface where the bending moments are
overestimated probably due to the error in double differentiation
of the wall deflection.
Both total and effective stress analyses should be used for
the design of shoring walls installed in the clayey soils. As
supported by the finite element back-analysis, the clayey soils
behave as undrained during excavation and as drained after
excavation. The finite element analysis confirmed that the time-
dependent deformations of shale rock during 11 months after
excavation could be up to 5 mm for 4.5 m excavation in the
shaly rock. The time-dependent deformation should be
considered in the shoring wall design.
0
5
Deducted from inclinometer readings (1 d after
excavation to bottom)
Computed (1 d after excavation to bottom)
Excavation Level
Bedrock surface
6 REFERENCES
Cao L.F. and Peaker S.M. 2011. Use of tieback in Southern Ontario.
64
th
Canadian Geotechnical Conference, Toronto, Paper 175
Lo K.Y. 1989. Recent advances in design and evaluation of
performance of underground structures in rocks. Tunnelling and
Underground Technology 27, 461-477.
0
5
10
15
20
25
-15 -10 -5 0 5 10 15
D
e
p
t
h

(
m
)
Inclinometer reading (11 m after
Computed deflections (11 m after
Excavation Level
Bedrock Surface
Tieback
Figure 5. Deduced and computed wall bending moments
clayey soils, and reduced prestressing for tiebacks. Figure 6
show the comparison of the computed and measured lateral
deflections. In this analysis, the E was taken as 80% of the
initial E; the effective and cohesion for clayey fill were taken
as 30
o
and 3 kPa, respectively; the effective and cohesion for
clayey till were taken as 32
o
and 7.5 kPa, respectively; and the
lower tiebacks were assumed to be de-stressed to 50% of the
initial prestressing. The computed lateral deflections within the
upper portion of the overburden are in a good agreement with
the inclinometer reading. However, there is a difference of up to
5 mm between the computed and measured deflections in the
lower portion of the overburden and the bedrock, which should
belong to the time-dependent deformation of the shaly rock as
the de-stressing of tiebacks or the consolidation of clayey soils
could not lead such deformation.
Poh T.Y., Goh A.T., Wong K.S., Wong I.H. and Poh K.B. 1999.
Determination of bending moments in diaphragm wall. 5
th
International Symposium on Field Measurements in Geomechanics,
Singapore, 229-234.
PTI 1996. Recommendations for prestressed rock and soil anchors.
Post-Tension Institute, Phoenix.
West H.H. 1993. Fundamentals of structural analysis. John Wiley &
Sons, Inc.

Figure 6. Measured and computed wall deflections (11 months after
excavation)
507
Permeability scale effect in sandy aquifers: a few case studies
Effet dchelle et permabilit des aquifres sableux : quelques tudes de cas
Chapuis R.P.
cole Polytechnique, Montral, QC, Canada
ABSTRACT: In sandy aquifers, stratification results in a range of values for the hydraulic conductivity K, which can be evaluated at
three scales. Since large-scale tests are more likely to meet preferential flow paths, they are also likely to yield larger K values than
small-scale tests, which may be viewed as a scale effect. The small scale is that of soil samples: their quality must be assessed and
their grain size distribution analyzed to check for mixes of sub-layers, before using reliable methods to predict the K values. The
middle scale is that of field permeability tests for which it is important to respect the standards and perform verifications. The large
scale is that of pumping tests. The paper presents a few case studies of sandy aquifers. Their stratification led to unimodal or
multimodal grain size distributions. For all cases, the K distributions provided consistent images of the sandy aquifers. It was then
concluded that, after a quality control of data and interpretations, there was no scale effect in the aquifers.
RSUM : Dans les aquifres sableux, la stratification donne une gamme de valeurs pour la conductivit hydraulique K qui peut tre
value trois chelles. Les essais grande chelle ayant plus de chances de tester des zones dcoulement prfrentiel, ils ont aussi
plus de chances de donner des valeurs leves de K que les essais petite chelle, ce qui peut tre vu comme un effet dchelle. La
petite chelle est celle des chantillons : leur qualit doit tre value et leur granulomtrie analyse pour dtecter les mlanges de
strates, avant dutiliser des mthodes fiables de prdiction de K. Lchelle moyenne est celle des essais de permabilit in situ pour
lesquels on doit respecter les normes et faire des vrifications. La grande chelle est celle des essais de pompage. Larticle prsente
des tudes de cas daquifres sableux. Leur stratification a donn des granulomtries unimodales ou multimodales. Pour tous les cas,
les distributions de K ont fourni des images cohrentes des aquifres sableux. On a conclu, aprs un contrle de qualit des donnes et
des interprtations, quil ny avait pas deffet dchelle dans ces aquifres.
KEYWORDS: aquifer, grain size distribution, monitoring well, permeability test, pumping test, scale effect

1 INTRODUCTION
In sandy aquifers, groundwater seepage is controlled by
stratification, with coarse size sediments deposited at high water
velocities and small size sediments settling at low water
velocities, or in temporary ponds. Many methods can be used to
assess the hydraulic conductivity, K, which can vary over orders
of magnitude. It is often believed that since large-scale tests
involve large volumes, which are more likely to meet
preferential flow paths, they are likely to yield larger K values
than small-scale tests (Bradbury and Muldoon 1990; Rovey
1998; Rovey and Niemman 1998). Thus, there should be a scale
effect for the K value, some increase with the tested volume.
There is no consensus about this scale effect. Many studies
tried to check or challenge theoretical opinions. They differed
about testing techniques, investigated scales, and geologic
media. Alas, the quality of each K value usually was not
questioned even if poor quality data and interpretation are
known to yield an artificial scale effect. Regrettably, the quality
control of groundwater parameters, which must be methodically
completed for engineered facilities, is not always done (Chapuis
1995). This paper examines quality control issues with data and
interpretation, in order to exclude artificial scale effects.
The idea of scale effect was rejected by Butler and Healey
(1998). They argued that scale effect results from artifacts
linked to incomplete well development and low-K skins around
well screens, but they did not study what produce a positive or
negative skin. These skin phenomena and their effects on the
apparent K value being related to safety issues, they are more
studied in geotechnique (Chapuis and Chenaf 2010) than in
geosciences.
Moreover, many studies have not examined how incorrect
interpretation methods for slug tests and pumping tests can yield
artificial scale effects. However, the quality control of slug test
methods has been largely investigated in geotechnique (Chapuis
et al. 1981; Chapuis 1988, 1998, 1999, 2001; Chapuis and
Chenaf 2002, 2003). For pumping tests in unconfined aquifers,
the large-scale K values obtained were shown to be incorrect if
the interpretation was performed using current methods for
unsteady-state (Akindunni and Gilham 1992).
Therefore, when studying scale effect, some caution must be
observed to avoid using scale effect as a final excuse, or as a
fudge factor, when the heterogeneity of the tested material could
have been more thoroughly investigated and when errors
involved in sampling, testing and interpretation methods could
have been taken into account.
Note that properly taking into account scale effect is
important for numerical analyses, since an aquifer numerical
model cannot be as detailed as the physical reality. Most often,
the grids of numerical models cannot contain enough elements
to model the detail of real features. This is why up-scaling
techniques are needed to define some equivalent K value for
grid elements (Renard and de Marsily 1997; Zhang et al. 2011).
In this paper, the results of three sites are briefly examined.
The small scale, about 10
-3
m
3
, is that of samples recovered in
boreholes for which the K value was evaluated using predictive
methods. The middle scale, about 1 m
3
, is that of field
permeability tests in monitoring wells. The large scale, about
10
3
m
3
, is that of pumping tests. Now, the problems linked to
the collected data at three scales in sandy aquifers are examined
in detail, starting with the soil samples taken in boreholes.
2 SMALL-SCALE K VALUES (SAMPLES)
Many soil samples can be taken in boreholes, usually with a
split spoon. Quality issues relative to soil sampling have been
the topic of many geotechnical researches. Five sample classes
are defined by considering the relationships between sampling
508
th
methods, quality of sample and quality of laboratory tests. All
borehole samples in sandy aquifers are of class-3 or class-4
quality. For information, the class-4 quality is obtained with the
hollow stem auger, rotary, percussion, cable tool and sonic
drilling methods (Baldwin and Gosling 2009). These methods
strongly influence not only the quality of samples, but also the
quality of permeability tests, and the quality of the MW
installation (Chapuis and Sabourin 1989; Chesnaux et al. 2006;
Chesnaux and Chapuis 2007). In sandy aquifers, a tube sampler
with a clear plastic liner can be used. This tool does not provide
class-1 or -2 samples. It roughly preserves the grain size
distribution curve (GSDC), with major margin disturbance
(thick-walled sampling) plus some mixing between adjacent
sub-layers. It does not preserve the water content w, void ratio e,
and K in situ values. For that reason, this sampler provides
class-4 samples, and not intact ones as claimed in a few papers.
Several methods can be used to predict the K value of a soil
sample. Chapuis (2012a) listed 45 methods and assessed their
capacity against large data sets for laboratory permeability tests
performed on homogenized fully saturated specimens. All tests
were not plagued by one of the 14 most frequent mistakes when
performing such tests. For sandy aquifers, the in situ porosity n
can be assessed using the method of Chapuis (2012b) and the K
values can be predicted with the method of Chapuis (2004),
which yields good predictions for natural soils in the ranges
0.003 d
10
3 mm and 0.3 e 1. The range for the effective
diameter d
10
was recently extended up to 150 mm (Ct et al.
2011; Chapuis et al. 2012).
If the soil sample is homogenous, its GSDC is smooth. This
is not the case for most borehole samples in sandy aquifers.
Therefore, when studying the GSDCs, caution must be taken to
avoid confusing homogenous samples (single layer) with those
made by mixing 2 or 3 small layers. The analysis proceeds with
a modal decomposition (Chapuis 2010; Chapuis et al. 2013),
which provides the GSDC and percentage of each layer in the
composite sample. The equivalent horizontal K value (stratified
sample) is then obtained using the composition rule.
3 MEDIUM-SCALE K VALUES (SLUG TESTS IN MWS)
The middle scale, about 1 m
3
, is that of permeability tests (slug
tests) performed in monitoring wells.
It is important to use the standard methods to interpret the
slug test data. In Canada, CAN/BNQ 2501-135 is the standard
for an overdamped response (CAN/BNQ 1988, 2008), but there
is no standard for an underdamped response. ASTM, however,
has standards for the underdamped response (ASTM 2012a) and
for the critically damped response (ASTM 2012b).
For overdamped slug tests, the velocity graph method helps
to establish the correct piezometric level (PL) and K value for
the test. It also helps to detect several phenomena during the
test. Even if the aquifer is unconfined, and even if the MW is
correctly installed, there are several reasons why the test data
must be corrected by a systematic error on the assumed PL, of a
few centimetres (Chapuis 2009a, b). The velocity graph gets rid
of any systematic error, which may be due to incorrect
calibration of a pressure transducer (PT), waiting time, PT line
slippage, piezometric modification, faulty MW installation, and
unknown PL. However, it cannot make a distinction between
these six errors.
For underdamped slug tests, it is preferable to fit the test data
using a least squares method, instead of a visual fit, and the
verification of three physical conditions must be done for each
tests, otherwise large errors can be made (Chapuis 2012c).
4 LARGE-SCALE K VALUES (PUMPING TESTS)
For the large scale of pumping tests, about 10
3
m
3
, precautions
must be taken when installing the pumping well and MWs, and
also when interpreting the pumping test data. The common
theories for unsteady-state are based on some wishful thinking
about drainage, unsaturated seepage and a misleading concept
of specific yield (e.g., Akindunni and Gilham 1992; Chapuis et
al. 2005a). For MWs, it is commonly admitted that two thirds of
them are improperly installed (Nielsen and Schalla 2005).
5 THE SITES
5.1 The Lachenaie site
The site is located 50 km north-east of Montreal. The sand
unconfined aquifer has been used for field training and research.
The GSDCs could be correctly fitted using a unimodal
lognormal distribution. The little variability for the mean and
the standard deviation indicate homogeneity (Fig. 1). For the
pumping test, the steady-state drawdown data were used, the
interpretation methods being proven to be reliable (Chapuis et
al. 2005a, b). In this aquifer, the average K values at the three
scales are very close, and thus there is no scale effect (Fig. 2).

0
50
100
150
200
250
300
350
-2.0 -1.0 0.0 1.0 2.0
log (mean size) and log (std dev.)
m
e
a
n

s
a
m
p
l
i
n
g

d
e
p
t
h

(
c
m
)
mean
sdt dev.

Figure 1. Lachenaie: modal decomposition of the sand GSDCs.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1.E-03 1.E-02 1.E-01 1.E+00
x = K (cm/s)
%

o
f

c
a
s
e
s

l
o
w
e
r

t
h
a
n

x
pred. K (small scale)
K tests in MWs
lognormal best fit MWs
data pumping

Figure 2. Lachenaie: comparison of the K values obtained at the small,
medium, and large scales.
5.2 The Shannon site
Shannon is a small town about 30 km north-west of Quebec
City. For the TCE-contamination case, a lot of information was
given in the defendants' expert reports, but without a quality
control, which led to contradictions. The quality control and a
synthesis were done in Chapuis (2009c, 2010, 2013a, b). There
were about 1000 MWs for this huge contamination case.
The aquifer stratification could be considered or not when
analyzing the GSDCs. When it was not, the distribution of
509
predicted K values could not explain the high large-scale K
values of pumping tests (Fig. 3). When it was, after using a
modal decomposition of each GSDC, the distribution of
predicted K values yielded a large-scale K value very close to
that of pumping tests (Chapuis 2010, 2013b).
For the slug tests in MWs, Chapuis (2010) showed that the
defendants expert reports gave K values that were obtained
without following the standards and without making the
required verifications. They were about three times smaller than
the K values obtained when following the standards and making
the verifications. When the standards were not respected, the
distribution of the slug test K values could not explain the large-
scale K values of pumping tests (Fig. 4). When the standards
were followed, the slug test K distribution yielded a large-scale
K value very close to that of pumping tests (Fig. 4).

0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1.E-05 1.E-04 1.E-03 1.E-02
x = estimated K (m/s)
%

o
f

c
a
s
e
s

l
o
w
e
r

t
h
a
n

x
homogeneous
stratified
pumping wells
usual range
for
pumping
#
homogeneous
stratified

Figure 3. Shannon: K values predicted using the GSDCs, assuming
either homogeneous or stratified samples (modal decomposition) and
large scale pumping tests.

0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1.E-06 1.E-05 1.E-04 1.E-03 1.E-02
x = slug test K (m/s)
%

o
f

c
a
s
e
s

l
o
w
e
r

t
h
a
n

x
not following standards
following standards
pumping wells
usual range for
pumping wells
NFS
FS

Figure 4. Shannon: K values obtained with slug tests in monitoring
wells (following or not the standards), and large scale pumping tests.
How to perform the modal decomposition of a GSDC, and
that of predicted or measured K values, is explained elsewhere
(Chapuis 2013b; Chapuis et al. 2013). These papers also explain
how to predict, for a K distribution, the large-scale K value
which would be given by a pumping test, in order to logically
compare the data at the three scales. A closed-form equation is
also provided for the soil specific surface, more general than
that of Chapuis and Lgar (1992). According to the detailed
study following the quality control for the Shannon aquifer, all
K distributions provided a coherent image of the hydraulic
properties in the aquifer. Therefore, there was no scale effect.
5.3 The Sorel site
The Sorel site, 100 km north-east of Montreal, has been used for
many years for field training of students in groundwater
engineering and geophysics. The site is part of the floodplain at
the confluence of the Richelieu River and the St-Lawrence
River. Down to about 5 m deep, the stratigraphy includes many
layers of fine sand (deposited in low velocity water) and silty
clay (deposited in ponds). Over 300 soil samples were
recovered in over 40 boreholes.
The soil samples provided clearly bimodal GSDCs and K
values (Chapuis et al. 2013). The split-spoon sampler could
recover 30 or more individual layers of silty clay and fine sand,
which were uniform in color. The GSDC modal decomposition
provided results such as those of Fig. 5 for a few boreholes in
the vicinity of the pumping well. The fine sand and silty clay
were fairly homogeneous (Fig. 5). According to the modal
decompositions, the portion between 1.9 and 3.1 m deep had
more clayey silt than the upper and lower portions. The screens
of the pumping well and nearby MWs were installed in the
portion between 3.1 and 4.4 m this confined aquifer.
The horizontal K distribution curve was obtained from the
modal decomposition of GSDCs and the K composition rule.
The predicted K distribution was in good agreement with the
pumping test K values, whereas the slug test K values were
somewhat below the pumping test K values (Chapuis et al.
2013). Due to the fine stratification of fine sand and silty clay
sub-layers, the development of monitoring wells was not
effective. Therefore, the slug tests have slightly underestimated
the horizontal medium-scale K value due to smearing between
layers during drilling and MW installation. Therefore, there was
no scale effect for the Sorel highly stratified aquifer.

0
1
2
3
4
5
6
-9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4
log of the means and standard deviations
d
e
p
t
h

(
m
)
fine sand, mean fine sand, std dev
clayey silt, mean clayey silt, std dev

Figure 5. GSDCs modal decomposition for stratified samples of Sorel,
showing fairly homogeneous layers of fine sand and clayey silt.
6 CONCLUSIONS
This paper studies the permeability of sandy aquifers at three
scales. The aquifers are stratified or not, which leads to
multimodal or unimodal distributions for grain size distribution
curves. The small scale is that of soil samples: their quality
must be assessed and their GSDC analyzed to check for mixes
of sub-layers before using reliable methods to predict the K
values. The middle scale is that of field permeability tests for
510
th
which it is important to respect the standards and perform
verifications. The large scale is that of pumping tests, which
must be interpreted for steady-state.
The results presented here have shown that, when
stratification is adequately considered, slug tests are interpreted
according to standards, and the resulting K distributions are
taken into account, the conclusion is that there is no scale effect.
Therefore, using a quality control approach for analyzing the
GSDCs and interpreting field test data is essential for cross-
checks, and for avoiding the creation of artificial scale effects.
7 ACKNOWLEDGEMENTS
The author thanks the National Research Council of Canada,
BFI Ltd (Lachenaie), the Rgie de leau de Sorel-Tracy and
Aquatech (Sorel), the Shannon Citizens Committee, Charles
Veilleux and Associates, and the FARC (Shannon).
8 REFERENCES
Akindunni F.F. and Gillham R.W. 1992. Unsaturated and saturated flow
in response to pumping of an unconfined aquifer: Numerical
investigation of delayed drainage. Ground Water 30, 873884.
ASTM 2012a. D5785: Standard test method (analytical procedure) for
determining transmissivity of confined nonleaky aquifers by
underdamped well response to instantaneous change in head (slug
test). Annual CDs of standards, vol. 04.09, ASTM, West
Conshohocken, Penn.
ASTM 2012b. D5881: Standard test method (analytical procedure) for
determining transmissivity of confined nonleaky aquifers by
critically damped well response to instantaneous change in head
(slug test). Annual CDs of standards, vol. 04.09, ASTM, West
Conshohocken, Penn.
Baldwin M. and Gosling D. 2009. BS EN ISO 22475-1: Implications for
geotechnical sampling in the UK. Ground Engineering, August
2009, 2831.
Bradbury K.R. and Muldoon M.A. 1990. Hydraulic conductivity
determinations in unlithified glacial and fluvial materials. ASTM
STP 1053, 138151.
Butler J.J. and Healey J.M. 1998. Relationship between pumping test
and slug-test parameters: scale effect or artefacts. Ground Water
36(2), 305313.
CAN/BNQ 1988. Canadian Standard CAN/BNQ 2501-135/1988: Soils
determination of permeability by the Lefranc method.
CAN/BNQ 2008. Canadian Standard CAN/BNQ 2501-135/2008: Soils
determination of permeability by the Lefranc method.
Chapuis R.P. 1995. Controlling the quality of ground water parameters:
some examples. Can Geotech J 32(1), 172177.
Chapuis R.P. 1988. Determining whether wells and piezometers give
water levels or piezometric levels. In Ground Water Contamination:
Field Methods, ASTM STP 963, 162171
Chapuis R.P. 1998. Overdamped slug test in monitoring wells: Review
of interpretation methods with mathematical, physical, and
numerical analysis of storativity influence. Can Geotech J 35(5),
697719.
Chapuis R.P. 1999. Borehole variable-head permeability tests in
compacted clay liners and covers. Can Geotech J 36(1), 3951.
Chapuis R.P. 2001. Extracting the local piezometric level and hydraulic
conductivity from tests in driven flush-joint casings. Geotech
Testing J 24(2), 209219.
Chapuis R.P. 2004. Predicting the saturated hydraulic conductivity of
sand and gravel using effective diameter and void ratio. Can
Geotech J 41(5), 787795.
Chapuis R.P. 2009a. Permeability or hydraulic conductivity tests in a
monitoring well: Why are piezometric level corrections required?
Geotech News 27(2), 4649.
Chapuis R.P. 2009b. Interpreting slug tests with large data sets. Geotech
Testing J 32(2), 139146.
Chapuis R.P. 2009c. Recours collectif - Rsidents de Shannon -
Expertise sur les conditions hydrogologiques - Rapport
prliminaire en 3 volumes, 970 p.
Chapuis R.P. 2010. Recours collectif - Rsidents de Shannon
Expertise sur les conditions hydrogologiques - Rapport
dexpertise, 156 p.
Chapuis R.P. 2012a. Predicting the saturated hydraulic conductivity of
soils: A review. Bull Eng Geology Envir 71(3), 401434.
Chapuis R.P. 2012b. Estimating the in situ porosity of sandy soils
sampled in boreholes. Engng Geology 141142, 5764.
Chapuis R.P. 2012c. Improved curve fitting methods for underdamped
slug tests. Geotech Testing J 35(5), 752761.
Chapuis R.P. 2013a. TCE-contaminated groundwater in Shannon,
Quebec: 2. Evaluating the hydraulic conductivity with permeability
tests in observation wells. Bull Eng Geol Env, submitted
Chapuis R.P. 2013b. TCE-contaminated groundwater in Shannon,
Quebec: 3. Evaluating the hydraulic conductivity at three scales.
Bull Eng Geol Env, submitted
Chapuis R.P. and Chenaf D. 2002. Slug tests in a confined aquifer:
Experimental results in a large soil tank and numerical modeling.
Can Geotech J 39(1), 1421.
Chapuis R.P. and Chenaf D. 2003. Variable-head field permeability
tests in driven casings: Physical and numerical modeling. Geotech
Testing J 26(3), 245256.
Chapuis R.P. and Chenaf D. 2010. Driven field permeameters: Rein-
venting the wheel? Geotechnical News 28(1), 3742.
Chapuis R.P. and Lgar P.P. 1992. A simple method for determining
the surface area of fine aggregates and fillers in bituminous
mixtures. In Effects of Aggregates and Mineral Fillers on Asphalt
Mixture Performance, ASTM STP 1147, 177186.
Chapuis R.P. and Sabourin L. 1989. Effects of installation of
piezometers and wells on groundwater characteristics and
measurements. Can Geotech J 26(4), 604613.
Chapuis R.P., Par J.J., and Lavalle J.G. 1981. Essais de permabilit
niveau variable. Proc. 10th ICSMFE, Stockholm, Balkema, Vol. 1,
401406.
Chapuis R.P., Chenaf D., Acevedo N., Marcotte D. and Chouteau M.
2005a. Unusual drawdown curves for a pumping test in an
unconfined aquifer at Lachenaie, Quebec: Field data and numerical
modeling. Can Geotech J 42, 11331144.
Chapuis R.P., Dallaire V., Marcotte D., Chouteau M., Acevedo N. and
Gagnon F. 2005b. Evaluating the hydraulic conductivity at three
different scales within an unconfined aquifer at Lachenaie, Quebec.
Can Geotech J 42, 12121220.
Chapuis R.P., Weber S. and Duhaime F. 2012. Intrinsic permeability of
materials ranging from sand to rock-fill using natural air convection
tests: Discussion. Can Geotech J 49(11), 13191322.
Chapuis R.P., Dallaire V. and Saucier A. 2013. Getting information
from modal decomposition of grain size distribution curves.
Geotech Testing J, submitted.
Chesnaux R. and Chapuis R.P. 2007. Detecting and quantifying leakage
through defective borehole seals: A new methodology and
laboratory verification. Geotech Test J 30(1), 1724.
Chesnaux R., Chapuis R.P. and Molson J.W. 2006. A new method to
characterize hydraulic short-circuits in defective borehole seals.
Ground Water 44(5), 676681.
Ct J., Fillion M.H. and Konrad J.M. 2011. Intrinsic permeability of
materials ranging from sand to rock-fill using natural air convection
tests. Can Geotech J 48, 679690.
Nielsen D.M. and Schalla R. 2005. Design and installation of ground-
water monitoring wells. Chapter 10, Practical Handbook of
Environmental Site Characterization and GroundWater
Monitoring, 2nd edition, CRC Taylor & Francis.
Renard P. and de Marsily G. 1997. Calculating equivalent permeability:
a review. Adv Water Resources 20(5-6), 253278.
Rovey C.W. II 1998. Digital simulation of the scale effect in hydraulic
conductivity. Hydrogeology, 6(2), 216225.
Rovey C.W. II and Niemann W.L. 1998. Wellskins and slug tests:
wheres the bias? J Hydrology 243(1-2), 120132.
Zhang Y., Liu B.Z. and Gable C.W. 2011. Homogenization of hydraulic
conductivity for hierarchical sedimentary deposits at multiple
scales. Transp Porous Med 87, 717737.
511
A Study of Cuttability Indices for Tunnel Penetration
tude sur les indices daptitude la coupe pour la pntration de tunnels
Chen L.-H.
Dept. of Civil Engineering, National Taipei University of Technology, Taipei, Taiwan
Chen Y.-C., Chen W.-C., Liu H.-W.
Dept. of Construction Engineering, National Taiwan University of Science & Technology, Taipei, Taiwan
ABSTRACT: To speed up construction of mass transit subway and the popularity rate of sewage, Taiwan's underground excavation
works, especially for mechanical cutting cases, show an ascendant tendency. This study presents a generalized solution for
underground geological-mechanical interaction. By using dimensional analysis, this model generalizes geological characteristics
grouped into three categories: (1) brittle (rock-like), (2) the ductile (soil-like), and (3) brittle-ductile (gravel-like) type with respect to
two cutting forces: (1) thrust and (2) torque to evaluate their excavation/penetration rate. Furthermore, the leading cuttability indices
can be obtained to enable to assess the underground excavation. Meanwhile, in-situ experimental results from shield tunneling and
pipe jacking construction were used to examine this model and it showed a nice agreement between both. From this analytical
approach, a proposed oval-shaped cutting ellipsoid, including its center (O), area (A), and long/short axis (ax/by, or ay/bx), can be
used not only to estimate the functionality and efficiency of cutting machine adopted for tunnel project, but also to offer a warning
information for inadequate cutting strategy.
RSUM: Pour acclrer la construction du systme de transport en commun souterrain et des eaux uses, les travaux d'excavation
souterrains de Tawan, en particulier les coupes mcaniques, montrent une tendance ascendante. Cette tude prsente une solution
gnralise pour linteraction gologique-mcanique souterraine. En utilisant l'analyse dimensionnelle, ce modle gnralise les
caractristiques gologiques regroupes en trois catgories: (1) fragile (comme la roche), (2) l'ductile (comme le sol), et (3) fragile-
ductile (comme le gravier) en respectant deux types de coupe: (1) la pousse et (2) torsion pour valuer leur taux dexcavation /
pntration. Par ailleurs, les indexes daptitude la coupe peuvent tre obtenue pour permettre d'valuer l'excavation souterraine.
Pendant ce temps, les rsultats exprimentaux in-situ de bouclier tunnel et de la construction de tuyau de fonage ont t utiliss pour
examiner ce modle et ceux-ci concordent. A partir de cette approche analytique, une proposition de ellipsode de coupe de forme
ovale " (comprenant son centre (O), sa surface (A) et ses axes longs et courts (ax/by, ou ay/ bx) ) peut tre utilis non seulement pour
estimer le bon fonctionnement et l'efficacit de la machine de dcoupage adopt pour le projet de tunnel, mais aussi pour fournir une
alerte propos dune stratgie de coupe inadapte.
KEYWORDS: Generalized cutting mechanism , Thrust, Pipe jacking, Cuttability indices

1 INTRODUCTION
1.1 Multi-scale underground cutting project
Recently, the construction projects increase the cases of
underground tunneling by mechanical cutting such as tunnel in
the mountain, mass rapid transportation system in the city and
sewer system, etc. There are different types of cutting methods
including TBM, shield tunnel (ST), as well as pipe jacking (PJ)
with various sizes corresponding to different geological
conditions (see Figure 1). This study presents a normalized
evaluation to meet the multi-scale underground cutting projects
so that all of the in-situ data can be collected and compared with
each other.

D
S
= 2920 mm
DT = 11740 mm
DP = 1500 mm

Figure 1. Different size of cutting machines ranged from 11740 to 1500
mm in diameter.
2 CONCEPTUAL MODEL
2.1 Indentation-typed fracture mechanism
Based upon normal indentation fracture in a Mohr-
Coulomb material, Huang 2000 proposed a conceptual
model as follows:
512
th
*( 1)/ *( 1)/
(1 )
p p d d
k k k k

+ +
+ =
c
(1)
where is a key dimensionless factor, which is a function of
wedge angle of cutter , elastic constants and plastic strength
parameters.
( ) ( )
*
* * u i
= E, or G , q , , , , u | | | o
(2)
and is defined as dimensionless elasto-plastic (E-P) radius
while * reach critical value where brittle fracture occurs on this
E-P interface. Thrust force, therefore, can be estimated using the
indentation pressure P and indentation force F as seen in Figure
2 schematically.
p p
n (K 1) / K p
*
p p
(n 1) K
P 1
1
q K 1 K n

+

=
`
+

)

(3)
n 1
i
d
F (3 n) P( )
tan
= t
(4)

d
Intrinsic
flaw
Core
Plastic zone
Elastic zone
o
|
2a
* r
Indenter
F
E-P
interface
, max x
o
X
y

Figure 2. Schematic normal indentation fracture
2.2 Generalized trust system
This study presents a generalized trust system of cutter head
globally by taking each different types of individual cutters into
account locally with respect to different methods (TBM, ST,
and PJ) and geological conditions (rock, soil, & gravel). Figure
3a and 3b show the total trust force F, which is consists of front
resistance F
f
and lateral resistances F
p
including both F
p,m
for
machine itself and Fp,p for pipes.
f p,m p,
F F F F = + +
p
(5)
p, p
F
p,m
F
f
F
F
p
F

Figure 3a. Trust force system

Figure 3b. Trust force system with inclined angle
Therefore, Ff (kN) is estimated from different types (nj) of
cutters fj and water/earth pressure Ps (kN/m2):
m
f j j s
j 1
F n f P
=
A = +
(6)
where A is cross area of cutter head (m
2
).
This paper presents an analytical estimation to deal with
different mechanical cutting methods (tunnel boring machine,
shield tunnel and pipe jacking), construction types (earth
pressure balance, slurry pressure balance, thick-mud), and
geological conditions (soil, gravel and rock) by generalizing
their total thrust system. The straight-line thrust is calculated for
either wedge- or conical-typed cutters of tunneling machine. In
this generalized work, the upper bound and lower bound of trust
are highlighted for the warning situations for risk assessment.
3 CASE STUDY
3.1 Case I: Taoyuan tunneling project in Taiwan
In addition, the in-situ data of trust in shield machine
(Taoyuan tunneling project) is presented to confirm with. It
depicts a favorable agreement for the estimation of thrust in this
study as shown in Figure 4 (cutter head), and Figure 5 (results)
with respect to normal cutting as well as abnormal conditions
(point a and b shown in Fig.5) once the in-situ data out of the
theoretical boundaries.

Figure 4. Cutting head in field for shield tunnel project
513

4 CONCLUSIONS
The results shows that the total thrust for upper bounds and
lower bounds are: (401%, 37.8%) and (258.2%, 31.7%)
compared with normal condition in gravel and weathered
sandstone cases respectively, which normalized boundary
values are (1310-4 ~ 8210-4) and (0.97 ~ 4.98) for cutting-
head resistance respectively. It also found that the cutting-head
resistance take about 28% of the total trust resistance
(F=4773kN) in the gravel case by taking cutters forces into
account.
5 REFERENCES
Figure 5. Theoretical upper and lower bounds associated with data
curve in field for shield tunnel project (vertical axis: trust in kN and
horizontal axis: rate of penetration in m)
3.2 Case II: Pipe jacking project in Taiwan
Another case study is presented for pipe jecking tunnel
shown in Fig. 6.

Balci, C., "Correlation of rock cutting tests with eld performance of a
TBM in a highly fractured rock formation: A case study in
Kozyatagi-Kadikoy metro tunnel, Turkey", Tunnelling and
Underground Space Technology, Vol.24, 2009, pp. 423 - 435.
Chen, L. H., Failure of Rock under Normal Wedge Indentation, Ph. D.
Thesis, University of Minnesota, U.S.A., 2001.
Huang, H., Detournay, E., and Alehossein, H., "Analytical Model for
the Indentation of Rocks by Blunt Tools," Rock Mechanics and
Rock Engineering, Vol.33, No.4, 2000, pp. 267 - 284.

Farrokh, E. and Rostami, J., "Correlation of Tunnel Convergence with
TBM Operational Parameters and Chip Size in the Ghomroud
Tunnel, Iran," Tunnelling and Underground Space Technology,
Vol.23, 2008, pp.700 - 710.

Figure 6. Cutting head in field for pipe jacking project
Unlike a flat data curve in field for the case of shield tunnel,
the in-situ data curve for pipe jacking method in Fig. 7 increases
in trust (vertical axis) with the increase of rate of penetration
(horizontal axis) due to the lateral resistance is proportional to
the pipe length. In this cutting case of sewer system, there is no
abnormal excavation situation such that the data curve does not
reach the theoretical boundaries.

Figure 7. Theoretical upper and lower bounds with in-situ data curve for
pipe jacking project of sewer system (vertical axis: trust in kN and
horizontal axis: rate of penetration in m)

515
th
1
Survey results of damaged areas in flood disaster of typhoon Morakot
and suggestions for restoration projects
Rsultats des investigations sur les zones ravages et inondes par le typhon Morakot,
propositions de projets de restauration
Chou J.C., Huang C.R.
Sinotech Engineering Consultants, Taipei, Taiwan
Shou K.J.
National Chung Hsing University, Taichung, Taiwan
ABSTRACT: Typhoon Morakot attacked Taiwan on August 8th, 2009 and caused heavy rainfall in Taiwan. The precipitation from
August 5th to 10th, 2009 in the survey station of Ali Mountain was about 3049.5mm and the precipitation of 24-hour and 48-hour
period in the south of Taiwan was close to the maximum observed precipitation in the world. This heavy rainfall caused many natural
disasters including landslides, floods and debris flows. These natural disasters severely damaged the middle, south and east of Taiwan
and cost huge property and life losses. Sinotech Engineering Consultants conducted a comprehensive survey after the flood disaster of
Typhoon Morakot (1) to collect information of damaged areas, (2) to analyze the causes and mechanisms of failures and (3) to draft
strategies and suggestions for restoration projects and future projects. Survey areas included the drainage basins of Chenyulan Creek,
Laonong River, Cishan River and Ailiao River, the watershed of Nahua reservoir and Taitung area. Depending on different geological
environments, the survey team concluded ten mechanisms causing failures. This article presents survey observations, discussions of
failures in flood disaster and provides strategies and suggestions for restoration projects and future projects.
RSUM : Le typhoon Morakot a atteint Taiwan le 8 aot 2009 accompagn de fortes pluies. Les prcipitations du 5 aot au 10 aot
2009, mesures la station dAli Mountain, ont t de 3 049,5 mm, entranant dans le sud de Taiwan des niveaux de prcipitation en
24 heures et 48 heures proches des maximums de prcipitations connues dans le monde. Les fortes pluies ont caus de nombreux
cataclysmes naturels tels que glissements de terrains, inondations et charriage de dbris, qui ont ravag le Centre, le Sud et lEst de
Taiwan entranant des pertes en vie humaines. Par la suite, le bureau dingnierie Sinotech a ralis des tudes dtailles afin de
dresser un bilan exhaustif des dgts des inondations causes par le typhon Morakot : (1) recensement et investigation des zones
ravages ; (2) analyse des causes et des mcanismes des dsordres, (3) bauches de stratgies et solutions pour les travaux de
restaurations et les projets futurs. Les zones dinvestigations comprenaient les bassins versants de Chenyulan Creek, Laonong River,
Cishan River et Ailiao River, le bassin hydrographique du rservoir Nahua et de la zone de Taitung. Suivant les diffrents
environnements gologiques identifis, les investigations ont conclu dix mcanismes lorigine des dsordres. Cet article prsente
les rsultats des investigations sur sites, ltablissement des mcanismes des dsordres lis aux inondations qui en suivirent ainsi que
les solutions de restaurations et les projets correspondants.
KEYWORDS: Typhoon Morakot, Landslides, Debris Flows
1 INTRODUCTION
Typhoon Morakot attacked Taiwan on August 8th, 2009
causing heavy rainfall in Taiwan. The precipitation from August
5th to 10th, 2009 in the survey station of Ali Mountain was
about 3049.5mm and the precipitation of 24-hour and 48-hour
period in the south of Taiwan was close to the maximum
observed precipitation in the world. Figure 1 shows the
maximum accumulated precipitation in 24 hours in the south of
Taiwan. This heavy rainfall caused floods, many natural
disasters and severely damaged mountain areas of the middle,
south and east of Taiwan. All these floods and disasters are
called 88 Flood. 88 Flood caused 643 deaths, 60 missing,
2,555 injured and 16.5 billion NT dollars economy lost which
makes 88 Flood the worst natural disaster since Chi-Chi
Earthquake happened on September 21, 1999. The most severe
disaster is the disaster in Xiaolin village. The village was
destroyed by a catastrophic debris flow during Typhoon
Morakot. Over 500 residents were buried alive and 350 houses
were damaged.
Right after 88 Flood, Sinotech Engineering Consultants
conducted a comprehensive survey on slopelands, roads,
bridges and hydraulic facilities in the drainage basins of
Chenyulan Creek, Laonong River, Cishan River and Ailiao
River, the watershed of Nahua reservoir and Taitung area (See
Figure 2). Purposes of this survey are (1) to collect information
of damaged areas, (2) to analyze the causes and mechanisms of
failures and (3) to draft strategies and suggestions to restoration
projects and future projects.
RainfallIsoline
Figure 1. Maximum accumulated precipitation in 24 hours in the south
of Taiwan (Precipitation in mm).
516
th
th
2
Drainage basins of
Chenyulan Creek
Drainage basins
of Cishan River
Watershed of
Nahua reservoir
Drainage basins
of Laonong River
Drainage basins
of Ailiao River Taitung Area
Drainage
Basin
County Line
Survey Area
Figure 2. Survey area conducted by Sinotech Engineering Consultants
survey team.
2 CAUSES AND MECHANISMS OF FAILURES
Flood and heavy rain are two main reasons causing slopeland
and road failures. Depending on different geological
environments, the survey team concluded ten mechanisms of
slopeland and road failures. Table 1 listed these mechanisms
and numbers of failures caused by each mechanism.
Figure 3 to Figure 12 show failures caused by Mechanism
No.1 to No.10. One mechanism should be noticed is Complex
Failure Mechanism. Complex Failure Mechanism means that
slopeland or road failures were caused by more than one
mechanism simultaneously. Damages caused by Complex
Failure Mechanism were usually severe and massive. Figure 12
shows one failure site (County Road 64 in Taitung County)
damaged by Complex Failure Mechanism which included
Erosion of Bottom Roadside Slope and River Channel Erosion
of Roadbed mechanisms.
Table 1. Failure Mechanisms and Number of Failures
No. Failure Mechanism
# of
Failures
1 Erosion of Bottom Roadside Slope 35
2 Debris Flow 33
3 Erosion of Top Roadside Slope 27
4 Shallow Slope Failure 22
5 River Channel Erosion of Roadbed 21
6 Dip slope 4
7 Circular Failure of Slope 4
8 Barrier Lake 2
9 Deep Sliding Failure of Colluvium 2
10 Complex Failure Mechanism 15
Figure 3. Failure caused by Erosion of Bottom Roadside Slope
Mechanism (Provincial Highway 21 (228K+500)).
Figure 4. Failure caused by Debris Flow (County Road 179 in Nantou
County).
Figure 5. Failure caused by Erosion of Top Roadside Slope (County
Road 60 in Nantou County).
Figure 6. Failure caused by Shallow Slope Failure (River in Tao Yuan
District in Kaohsiung City).
517
th
3
Chenyulen
Creek
Figure 7. Failure caused by River Channel Erosion of Roadbed
(Provincial Highway 21).
DipSlope
Figure 8. Failure caused by Dip slope (County Road 179-1 in Nantou
County).
CircularFailure
Figure 9. Failure caused by Circular Failure of Slope (At 3K of
Road 179 in Nantou County).
County
ColluviumFailure
Area
Figure 11. Failure caused by Deep Sliding Failure of Colluvium (Xin Fa
Village in Kaohsiung City).
Erosionof Top
Roadside Slope
50m
205m
RiverChannel
Erosionof Roadbed
Figure 12. Failure caused by Complex Failure Mechanism (County
Road 64 in Taitung County).
res and Mechanism No.2, No.4 and
No
RESTORATION PROJECTS AND FUTURE PROJECTS
rom
In general, Mechanism No.1, No.3 and No.5 were main
mechanisms of road failu
.5 were main mechanisms of slopeland failures. These five
mechanisms caused most of failures in 88 Flood.
3 STRATEGIES AND SUGGESTIONS TO
From survey observations and analyses, strategies to prevent
future slopeland and road failures can be outlined starting f
following aspects: (1) debris flow control and river remediation,
(2) proper treatment of landslides, (3) soil and water
conservation of slopeland and (4) soil and water conservation of
road.
Furthermore, improvements of road designs can be done to
avoid slopeland and road failures in the design phase: (1)
co
Figure 10. Failure caused by Barrier Lake (County Road 179-1 in
Nantou County).
nsider and prevent all possible failure mechanisms, (2) use
rock shed in potential rock fall area (see Figure 13), (3) from
survey observations, the tunnel and bridge are the best solution
for roads in potential debris flow area (see Figure 14), (4) use
deep foundation (e.g. pile foundation) as retaining wall at
bottom roadside slope retaining wall to avoid erosion. If the
roadbed can be protected from erosion, it is easier and faster for
future road restorations. (5) avoid area where river channel
erosion occurs and (6) install proper water drainage system for
slopelands (see Figure 15).
518
th
th
4
4 CONCLUSIONS
Because of the global climate change, natural disasters caused
by the extreme weather become more frequently in last few
years. The magnitude of these disasters and the damage caused
by these disasters become greater and greater and are greater
than what we experienced in the past. Therefore, in future
engineering projects, engineers should consider uncertainties
caused by the extreme weather in their designs and understand
limitations of engineering techniques. In addition, engineers
should try to avoid unnecessary development and construction
in environmental sensitive area.
5 ACKNOWLEDGEMENTS
Authors would like to thank Taiwan Geotechnical Society (TGS)
for giving this great opportunity to present our work in the 18
th
International Conference on Soil Mechanics and Geotechnical
Engineering. Authors would also like to thank engineers who
involved in this survey project and supports from Sinotech
Engineering Consultants.
Figure 13. Rock shed used to prevent landslides and rock fall area.
6 REFERENCES
SINOTECH Engineering Consultants, 2010. Survey results of flood
disaster of typhoon Morakot and suggestions to restoration projects,
SINOTECH, Taipei, 334p
Figure 14. Bridge used to avoid damages from Debris Flow in Wanrung
Township, Hualien County (Provincial Highway No. 16).
CreekorRiver
Drainage
Pipe
Drainage
Gallery
Slip
Surface
Road
Building
Ditch
Water
TableLevel
DewateredWater
TableLevel
Drainage
Well
Figure 15. Different types of water drainage systems used in slopelands.
519
Stability of chileans tailings dams with the Panda penetrometer. Experiences of
the last 10
th
Dix ans dtudes de la stabilit des barrages de rsidus miniers chiliens laide du
pntromtre Panda
Espinace R., Villavicencio G., Palma J.
Grupo de Geotecnia. Escuela de Ingeniera en Construccin. Pontificia Universidad Catlica de Valparaso, Chile.
Geotecnia Ambiental, Chile.
Breul P., Bacconnet C.
Institut Pascal PolytechClermont-Ferrand. Universit Blaise Pascal, Clermont-Ferrand, France.
Benz M.A., Gourvs R.
Sol-Solution Gotechnique Rseaux, Riom, France.
ABSTRACT: In Chile, since the beginning of the 20
th
century, about 40 cases of mechanical instability of the tailing dams have been
reported mainly due to liquefaction, slipping of banks or settlement. In order to solve this problem, a scientific and technological
cooperation has been established in 2001 between the geotechnical of groups at the Catholic University of Valparaso (Chile) and
Blaise Pascal University Clermont-Ferrand (France) whit the support of two companies, Sol-Solution in France and Geotecnia-
Ambiental in Chile. This article presents the main results that have allowed to propose a methodology for control and diagnosing of
tailing dams and its application in the medium mining sector. It is based on in-situ determination of geomechanical parameters
(internal friction angle and density index) using the Panda lightweigh penetrometer in order to characterize the constituent materials,
the variability of these materials and their implementation in the works. Finally, this methodology allows taking into account this
variability in the study of stability and the risk of liquefaction of these structures in a probabilistic approach.
RSUM: Au Chili, depuis le dbut du XX
e
sicle, environ 40 cas d'instabilit mcanique de ces dpts, principalement par
liqufaction, glissement des talus et tassements, ont t rapports. Cest dans ce contexte et pour apporter une rponse ce problme,
quune coopration scientifique et technologique a t tablie en 2001 entre les groupes de gotechnique de lUniversit Catholique
de Valparaiso (Chili) et de luniverist Blaise Pascal Clermont-Ferrand (France), avec le soutien des entreprises Sol-Solution (France)
et Geotecnia Ambiental (Chili). Cet article prsente les principaux rsultats qui ont permis de proposer une mthodologie pour le
contrle et le diagnostic des barrages de rsidus miniers de relave ainsi que son application au secteur de lindustrie minire moyenne.
Elle sappuie sur la dtermination in situ des paramtres gomcaniques (angle de frottement interne et densit relative) laide du
pntromtre Panda en vue de caractriser les matriaux constituants, de la variabilit de ces matriaux et de leur mise en uvre au
sein des ouvrages. Finalement, cette mthodologie permet de prendre en compte cette variabilit pour ltude de la stabilit et du
risque de liqufaction de ces ouvrages dans une approche probabiliste.
KEYWORDS: soils and site investigation, structures in seismic areas.

1 INTRODUCTION
Mine tailings are frequently stored in dams. This is the case for
copper for which the coarse fraction (fine sands) of the tailings
form the body of the dams, while the fine saturated fraction
(sludge and silts) is poured by cycloning into the reservoirs of
the dams thus formed.
Chile has a very large number of tailings dams built in this
way. Due to the construction methods and materials used, these
dams comprise failure mechanisms such as loss of stability,
liquefaction, and internal and external erosion leading to major
risks for the populations and their environments. Such risks are
highlighted by the accidents that have occurred around the
world and recently in the case of failures occurring during the
earthquake of 27 february 2010 in Chile, with fatal
consequences (Dobry and Alvarez 1967, ICOLD 2001, GEER
2010). In order to manage these risks, it appears necessary to
employ a probabilistic approach to predict their behaviour
during construction and after closing. However, applying such
an approach in practice at present is limited by the difficulty of
managing the data (random variables and stochastic fields) to be
introduced in the reliability calculations for the limit conditions
involved and conditioned by the relevance of the probability
models chosen to represent the variability of tailings dam
properties (Villavicencio et al. 2011). This is the reason why,
this article presents an approach of estimating calculation
parameters (friction angle and density index ID%) governing
the stability of these dams, and its variability from dynamic
penetration tests. Then models are proposed for all dams
composed of the same mine tailings types, making it possible to
link a probability law to the calculation parameters and ID%.
This method, applied to Chilean dams constructed from copper
mine tailings, proposes a single model for all tailings dams so as
to associate a probability law to the and ID%.

2 ESTIMATION OF THE DENSITY INDEX (ID%) AND
THE FRICTION ANGLE ()
1.1 The objective
In mine tailings with non plastic fine particles (size < 80 m)
ID% and are very important parameters, related to the in situ
penetration strength (N, qd, qc, etc), the input parameter of
static and dynamic stability models and for the evaluation of the
liquefaction (Troncoso 1986). These parameters are greatly
influenced by the origin and mineralogy of the particles, by the
physical characteristics and state of arrangement of the grains
determined by the state of compacting and by the extent of
stresses in-situ (Bolton 1986).
The methods used to implement mine tailings lead to the
prevalence of stratified internal structures that can be
heterogeneous. This can result in variations of resistance
properties, especially and ID%, as a function of depth. Thus
it is important to estimate the values and variability of these
520
th
parameters. To do this, we propose an estimation method based
on measuring the dynamic cone resistance (qd) that can be
relatively easily measured on this type of structure.
1.2 Normalisation of qd
Estimating ID% and by using empirical and semi-empirical
relations, first implies normalising qd at a reference stress
corresponding to atmospheric pressure (pa), using the following
equation 1.

q 1 N C qd qd with
c '
v a q p C (1)

where: qd
N1
is the dimensionless normalised dynamic cone
resistance, qd is the dynamic cone resistance, pa is the
atmospheric pressure,
v
is the effective vertical stress, c is
the normalisation coefficient (0.5 to 0.75).

According to Moss et al. (2006), this reference stress value is
considered as reasonable if the depth/stress relation is taken into
account. According to Salgado et al. (1997) and Moss et al.
(2006), the normalisation coefficient is not only linked to the
intrinsic properties of the soil such as the type of grain and the
physical characteristics of the material (mineralogy,
granulometry, particle shape and texture characteristics), lateral
pressure (K
o
), compressibility, cementation, resistance to
crushing of the particles, etc.
1.3 Experimental approach
Our study is based on the use of cone penetration resistances
(qd) obtained by using the Panda test. The Panda device is a
manual light dynamic penetrometer with variable energy and a
small cone section (2.0 or 4.0 cm
2
) (Gourvs et al. 1997, Benz
2009). The Panda provides the cone resistance qd of the soil as
a function of depth, and is capable of performing a large number
of in situ tests thanks to its small size and its quick
implementation. This device can operate until 6.0 (m) in depth
and for materials having particles size lower than 50.0 (mm).

Table 1. Geotechnical properties of mine tailings. Values and statistical
nalyses of experimental data from three representative tailings dams. a
No. 1 No. 2 No. 3
Geo.
Prop
Av. CV Av. CV Av. CV
s
3.09 4.6 3.36 8.0 3.1 2.2
D
50
0.13 19.0 0.11 15.2 0.25 8.7
F.C 28.0 28.7 33 26.3 17 10.0
IP 0 0 0 0 0 0
dmax
18.2 6.2 20.8 8.0 18.5 2.3
d
17.5 6.6 20.1 8.2 18.1 2.9
w
nat
11.0 22.3 3.3 43.1 7.5 27.3
qd 4.8 50.6 2.87 45.9 1.95 52.8
N
60
22 62.5 12 58.8 - -
s: specific weight (kN/m
3
), D
50
: median diameter (mm), F.C: percentage
of fines less than 80 (m), IP: plasticity index (%), dmax: Proctor dry
density (kN/m
3
), d: dry density in situ (kN/m
3
), w
nat
: water content in-situ
(%), qd: cone resistance PANDA test (Mpa), N
60
: corrected penetration
resistance index, Av: average, CV: coefficient of variation (%).
A serie of Panda tests have been performed on the mine
tailings coming from three dams studied, under controlled
laboratory conditions in a calibration chamber. The following
procedure was used:
a) Determination of the physical characteristics of 3 samples of
mine tailings of copper sulphates (Table 1).

b) Performing dynamic cone resistance tests in a calibration
mould for different states of density to obtain the relation d/qd
(calibration curve). A logarithmic relation can be observed, in
agreement with previous results (Chaigneau et al. 2000) for this
type of material. Figure 1 gives the calibration curves d/qd
obtained for dams No. 1, No. 2 and No. 3.
c) Normalisation of qd at atmospheric pressure (equation 1).

d = 1,0811ln(qd) + 15,983
R = 0,9948
d = 0,8693ln(qd) + 15,552
R = 0,9755
d = 1,086ln(qd) + 15,543
R = 0,9883
Figure 1. Relation d/qd for tailings dams No. 1, No. 2 and No. 3 in the
study.
1.3.1 Relation ID% = f (qd
N1
)
The equivalence between the state of density (% Optimum
Proctor Normal) and ID% was estimated for each calibration
test. On the basis of the normalised cone resistance (qd
N1
), and
by considering the classification modified by Skempton (1986)
and adapted by Villavicencio (2009), we estimated ID%
associated with each degree of compaction (table 2).

Table 2. Estimation of the state of compaction and associated
echanical behaviour for silty sands. Villavicencio (2009). m
qd
N1
ID%
State of
compaction
Mechanical
behaviour
Liquefaction
potential
0 17 0 15 Very low Contractant High
17 69 15 55 Low Contractant High
69 82 55 60 Average
Contractant
/Limit
Limit
82 162 60 80 Dense Dilatant Null
162 326 80 100 Very dense Dilatant Null

Studies conducted by Troncoso (1986) have concluded that
for mine tailings with a percentage of fines around 15% , with
confining stresses between 50 kPa and 350 kPa, ID% below
50%-60% is an indicator of contractancy. Under this condition,
if the material is saturated or partially saturated, under seismic
conditions, the risk of liquefaction is real. On the other hand,
the material will tend to a dilatant behaviour for a relative
density over these values. Verdugo (1997) have conducted an
analysis of the variation of the minimum and maximum
densities (Vibratory and Proctor compaction) both with mine
tailings and similar soils (sands and silts) with different
percentage of fines. They conclude that in situ ID% of 60% is a
very reasonable compaction value with a satisfactory
mechanical behaviour (dilatancy) in structures that allow certain
degree of deformation such as the tailing dams.
An empirical model was adapted by using a simple
regression on all the pairs of experimental data (qd
N1
, ID%) for
the three samples of mine tailings. Since we consider that mine
tailings can be globally classified in a single geotechnical class,
it is possible to estimate ID% as a function of the resistance
qd
N1
by a single relation. The model used is given by the
following equation:
4 . 65 ln 5 . 28 %
1

N
qd ID

with 10.0 qd
N1
326.0 (2)

14,0
5
0
5
0
5
0
5
0
5
19,0
0,0 2,0 4,0 6,0 8,0 10,0 12,0 14,0 16,0 18,0
D
r
y

d
e
n
s
i
t
y

(
K
N
/
m
3
)
Dynamic cone resistance, qd (MPa)
14,
15,
15,
16,
16,
17,
17,
18,
18,
Tailings Dam No 1
Tailings Dam No 2
Tailings Dam No 3
521
Figure 2 shows that the results of the model are very close
to the experimental results. In addition, the relation proposed by
Tatsuoka et al. (1990) was used by replacing qc
N1
by qd
N1
.

Figure 2. The experimental points, relations proposed and references for
estimating the ID% of mine tailings as a function of qd
N1
.
1.3.2 Relation = f (qd
N1
)
Since we considered that mine tailings can be globally classified
within one geotechnical class, it is possible to estimate as a
function of the resistance of qd
N1
by a single relation. To do
this, a regression analysis was performed on all the pairs of
experimental data (qd
N1
, ) obtained during the calibration
tests, for the three samples of mine tailings (figure 3). The
model used is given by the following equation:

N1
qd ln 5.54 14.79 ' with 10.0 qd
N1
280.0 (3)

As it can be seen on figure 3, the results of the model are
very close to the experimental results. In addition, the relation
proposed by Daz and Rodrguez-Roa (2007) was used by
replacing qc
N1
by qd
N1
.

Figure 3. Experimental points, proposed and bibliographic relations for
estimating of mine tailings as a function of the qd
N1
.
This result is in full agreement with the works already
carried out on the correlation between qc and qd obtained with a
Panda penetrometer. Indeed, it has been proven (Chaigneau et
al. 2000, Lepetit 2002) that in the case of sands and silty sands,
the average value obtained for the ratio qd/qc is equal to 1.03.
More recent research performed by Rahim et al. (2004)
confirmed the relation between qd and qc. Their results obtained
for granular soils have been demonstrated experimentally and
analytically on the basis of the cylindrical cavity expansion
theory and that of cavitation collapse.
The resistance qd obtained with a light Panda penetrometer
can therefore be assimilated with qc. In conclusion, in the case
of mine tailings:
(1) density index (ID%) and effective friction angle () can
be deduced very precisely from the normalised cone penetration
resistance qd
N1
by a two single relations,
(2) relation qd
N1
= qc
N1
is very well validated which allows
using either static or dynamic penetrometers according to need.
1.4 Application case: density index (ID%) and associated
mechanical behavior
On the basis of equation 2, it is possible to estimate the
profiles of the density index (ID%) as a function of depth from
the penetrometric tests performed in situ. The adaptation of the
correlation presented in table 2, allows estimating the
mechanical behaviour of mine tailings as a function of ID%. At
global scale (measurements processed at the scale of the tailings
dam by using the ID% distribution obtained from all the
penetration tests performed), the distribution of all these ID%
values for each dam can be adjusted by a normal law (Figure 4).

Figure 4. Density function of ID%. Proposed relation for tailings dams
No. 1, No. 2 and No. 3 in the study.

At global scale, the density function makes it possible to
obtain a global idea of the mechanical behaviour of the mine
tailings stored, by considering the limit value of ID%, which
permits classifying contractant or dilatant behaviour and
associate in a qualitative way the liquefaction potential. As an
example, table 3 presents a probabilistic analysis in global scale
of the variability of ID% and the mechanical behaviour for the
tailing dam No.1.

Table 3. Density index (ID%) and associated mechanical behaviour.
nalysis at global scale. Tailings dam No. 1. A
ID% Analysis of the mechanical behavior
Av.
C.V
%
ID%
% of
values
State of
compaction
Mechanical
behaviour
Liquefaction
potential
< 55 58 Low Contractant High
55 60 13
Average
density
Contractant
/Limit
Limit
52 28.3
60 100 29
Dense to
very dense
Dilatant Null

At a local scale (measurements processed at the scale of
each penetration test, by using the ID% distribution), the
distribution of all these ID% values can then also be adjusted by
a normal law (figures 5a, 5b). The so-obtained results are
consistent with the compaction test performed during the
construction of the three tailings dams.
The results are similar for the three dams, they show that a
local test can be used to estimate ID% for each penetration test,
with sufficient precision provided that the calibration tests have
been carried out on the material characteristics of the dam at the
scale of the structure concerned. The variability of ID% and the
soil mechanical behaviour associated, allows to estimate in a
first stage, the liquefaction potential of tailings dams in both
scales, global and local, and identify in a local scale the zones
with lower strengths through a layer by layer penetration test
(Figure 6).
The evaluation of the risk of liquefaction has been
expressed in an equation formulated by Seed and Idriss (1981).
The classical method compares locally the ratio of the cyclic
resistance of the soil (CRR) with the ratio of the cyclic shearing
stress ratio (CSR) stemming from seismic stress. The notion of
liquefaction potential is therefore linked to the fact that ratio
522
th
CRR/CSR is lower than unity. It is widely accepted that
estimating the cyclic resistance ratio (CRR) can be estimated on
the basis of dynamic and static penetration tests (Robertson and
Wride 1998, Boulanger 2004 and Idriss, etc.).

a) b)
Figure 5. a) The breakdown into layers and density index (ID%). b) The
distribution of Density Index (ID%). Test No. 1. Tailings dam No. 1.

Figure 6. Example of the factor of safety (F.S) profile. Test No. 1.
Tailings dam No. 1.

3 CONCLUSIONS
To predict the behaviour of mine tailings dams in view to
managing the risks inherent to them, it appears necessary to
carry out a probabilistic approach However, in practice
implementing this type of approach is limited by the difficulty
of managing the data to be used in reliability calculations for the
limit conditions concerned. This article proposed a method for
estimating in situ the density index (ID%) and the effective
friction angle () and its variability, making it possible to carry
out a probabilistic study of these structures. A single model was
proposed for all the mine tailings dams in Chile, in view to
linking a probability law to ID% and the .
A method was proposed that takes into account the spatial
variability of data for performing a reliability calculation of
liquefaction potential, which is the main cause for the failure of
this type of structure. On the basis of the results obtained, we
showed that the method proposed for estimating liquefaction
potential permits evaluating the probability of triggering this
phenomenon. Estimating the reliability of a dam in relation to
the limit states of static and dynamic stability demonstrates the
advantages and applicability of the approach, by using the
variability of the geotechnical characteristics of mine tailings
and resistance to penetration (qd
N1
) in particular.

4 ACKNOWLEDGEMENTS
Fundings for the work described in this paper was provided by
the research department of the Pontifical Catholic University of
Valparaiso Chile. This article was developed with the important
collaboration from the Professor, Mr Pierre Foray, Laboratory
3S-R, Institut National Polytechnique de Grenoble, France.
2 REFERENCES
Benz M.A. 2009. Mesures dynamiques lors du battage du pntromtre
Panda 2. Ph. D. Thesis, Blaise Pascal-Clermont II Univ, France.
Bolton M. 1986. The strength and dilatancy of sands. Gotechnique 36
(1), 65-78.
Boulanger R. and Idriss I.M. 2004. State normalization of penetration
resistance and the effect of overburden stress on liquefaction
resistance. Proceedings 11
th
SDEE and 3
rd
ICEGE, Berkeley, CA,
484- 491.
Chaigneau L. Bacconnet C. and Gourvs R. 2000. Penetration test
coupled with geotechnical classification for compacting control. An
International Conference on Geotechnical & Geological
Engineering, GeoEng2000, Melbourne, Australia
Daz E. and Rodrguez-Roa F. 2007. Ensayos in-situ en Arenas. VI
Chilean Congress of Geotechnical Engineering. Chilean Society of
Geotechnics. Univeridad Catlica de Santiago. Chile, November,
28-30.
Scale
Global
Local
Index Density (ID%)
D
e
n
s
i
t
y

F
u
n
c
t
i
o
n
0 10 20 30 40 50 60 70 80 90 100
0
0,01
0,02
0,03
0,04
Dobry R. and Alvarez L. 1967. Seismic failures in chilean tailings
dams. J. Soil Mech. & Foundation Eng. ASCE, SM6 (93), 237-260.
ICOLD. 2001. Tailings dams. Risk of dangerous occurrences. Lessons
learnt from practical experiences. Bulletin N 121. UNEP, DTIE
and ICOLD, Paris.
GEER (Geo-Engineering Extreme Events Reconnaissance Association)
2010. Dams, levees, and mine tailings dams. Turning disaster in
knowledge: geo-engineering reconnaissance of the 2010 Maule,
Chile Earthquake. J. Bray and D.Frost, Eds., 204-226.
Gourvs R. Oudjehane F. and Zhou S. 1997. The in situ characterization
of the mechanical properties of granular media with the help of
penetrometer. Proceedings of 3
rd
International Conference on
Micromechanics of Granular Media, Powders and Grains, Duram,
USA, 57-60.
Lepetit, L. 2002. Etude dune mthode de diagnostic de digues avec
prise en compte du risque de liqufaction. Thesis, Blaise Pascal-
Clermont II Univ, France
Moss R.E. Seed R.B. Kayen R.E. Stewart J.P. and Der Kiureghian A
2006. CPT-Based probabilistic assessment of seismic soil
liquefaction initiation. PEER Report 2005/15.
Rahim A. Prasad SN, and George K.P. 2004. Dynamic cone penetration
resistance of soils-theory and evaluation. Proceedings of the Geo-
Trans 2004 Conference, Los Angeles, California.
Robertson P.K. and Wride C.E. 1998. Evaluating Cyclic Liquefaction
Potential Using The Cone Penetration Test. Canadian Geotechnical
Journal, 35 (3). 442-459.
Salgado R. Boulanger R. and Mitchell J. 1997. Lateral effects on CPT
liquefaction resistance correlations. J. of Geotechnical and
Geoenvironmental Engineering, ASCE, 123 (8). 726-735.
Seed H.B. and Idriss I.M. 1981. Evaluation of liquefaction potential of
sand deposits based on observations and performance in previous
earthquakes. In Situ Testing to Evaluate Liquefaction
Susceptibility, ASCE Annual Convention, St. Louis.
Skempton S.M. 1986. Standard penetration test procedures and the
effects in sands of overburden pressure, relative density, particle
size, aging and overconsolidation. Geotechnique 36 (3). 425-447.
Tatsuoka F. Zhou S. Sato T. and Shibuya S. 1990. Evaluation method of
liquefaction potencial and its application. Report on Seismic
Hazards on the Ground in Urban Areas. Tokyo. 75-109.
Troncoso J. 1986. Envejecimiento y estabilidad ssmica de un depsito
de residuos minerales en condicin de abandono. ISSN-0716-0348.
(22), 147-158.
Verdugo R. 1997. Compactacin de Relaves. IV Chilean Congress of
Geotechnical Engineering. Chilean Society of Geotechnics,
Santiago. Universidad Federico Santa Mara. Chile. October, 29-4.
Villavicencio G. 2009. Mthodologie pour evaluer la stabilite des
barrages de rsidus miniers. Ph. D. Thesis, Blaise Pascal-Clermont
II Univ, France.
Villavicencio G. Bacconnet C. Breul P. Boissier D. and Espinace R.
2011. Estimation of the Variability of Tailings Dams Properties in
Order to Perform Probabilistic Assessment. Geotechnical and
Geological Engineering. 29 (6). 1073-1084.
523
Site Sampling: Assessing Residual Uncertainty
chantillonnage du site : valuation de l'incertitude rsiduelle
Fenton G.A.
Faculty of Civil Engineering and Geosciences, Delft University of Technology, Delft, The Netherlands
Department of Engineering Mathematics, Dalhousie University, Halifax, Nova Scotia, Canada
Hicks M.A.
Faculty of Civil Engineering and Geosciences, Delft University of Technology, Delft, The Netherlands
ABSTRACT: Geotechnical design is plagued by the uncertainty associated with site characterization. Common questions are How
many samples should be taken? and How do these samples reduce my uncertainty? Of considerable interest is the question What
site sampling plan will give the best cost to effectiveness ratio? This papers looks specifically at the effect of the number of samples
on residual uncertainty. The results can be used to quantitatively select the required number of samples needed to achieve a target
maximum residual uncertainty level. To study this problem, a square domain is selected (the site) and a stationary Gaussian random
field is simulated within the domain (the random soil properties). The random field is sampled at a series of locations and a trend is
estimated from the samples. The trend is then removed from the random field and the residual random field is statistically analyzed to
determine various measures of the effectiveness of the sampling scheme. These measures include: 1) the variance of the residual field
average (i.e. does the estimate represent the average?), 2) the residual standard deviation (i.e. how much residual uncertainty
remains?), and 3) the residual correlation length (i.e. how does trend removal affect the perceived correlation lengths?).
RSUM : Le design gotechnique est traditionnellement affect par des incertitudes associes la caractrisation du site. Les
questions les plus courantes sont : combien dchantillons devraient tre prlevs ? Comment ces chantillons peuvent rduire mon
incertitude ? Un des intrts les plus importants vient de cette question. Quel plan dchantillonnage du site donnera le meilleur
coefficient defficacit? Cet article examine spcifiquement leffet du nombre dchantillons sur des incertitudes rsiduelles. Les
rsultats peuvent tre utiliss pour quantifier et slectionner le nombre demand dchantillons ncessaires pour atteindre un objectif
dincertitude maximal avec le niveau rsiduel. Pour tudier ce problme, un domaine carr est slectionn (le site) et un champ
gaussien alatoire stationnaire est simul dans le domaine (les proprits du sol alatoires). Le champ alatoire est chantillonn une
srie demplacements et une tendance a t estime partir de lchantillon. La tendance retire du champ alatoire et le champ
rsiduel alatoire est statistiquement analyses afin de dterminer les mesures diverses de lefficacit du plan dchantillonnage. Ces
mesures comprennent : 1) la variance de la moyenne de champ rsiduel, cest dire comment la tendance estime reprsentent la
moyenne relle sur le terrain ? 2) lcart type rsiduel, cest--dire quel degr dincertitude rsiduelle demeure, et 3) la valeur
longueur rsiduelle de corrlation, cest--dire comment la suppression tendance affecte les longueurs de corrlation ?.
KEYWORDS: geotechnical design, site characterization, residual uncertainty, sampling, required number of samples, sampling plans.
1 INTRODUCTION
Site characterization is clearly an essential component of any
geotechnical design and a great deal of effort has been devoted
over recent decades on how to best perform such a
characterization. How many samples should be taken? How
should these samples be used in the design process?
The ground is one of the most complex of engineering
materials, and yet is the most fundamental, in all senses of the
word. While steel, concrete, and wood, for example, have fairly
well established and relatively small uncertainties, the ground
can vary by orders of magnitude from site to site, and even
within a site.
As a result of the large uncertainty in the ground, all
geotechnical designs must start with a geotechnical
investigation so that the best nominal or characteristic
ground parameters can be used in the design process.
Traditionally, the intensity of the site investigation has not been
particularly important, so long as a reasonable estimate of the
characteristic design values can be estimated. However, recent
impetus has been towards providing reasonable estimates of the
reliability of designed geotechnical systems. In order to do so
the ground used to provide the geotechnical resistance needs to
be properly evaluated, in both the mean and the covariance.
In this paper, the ability of a soil sampling scheme to predict
the actual mean, variance, and correlation length of the soil at a
site is investigated. A key question is how does the number of
samples affect the accuracy of the estimate? Or, put another
way, how many samples are required to achieve a certain
desired accuracy? The answer is found by considering a square
site and using random field simulation to generate realizations
of the soil properties over the site, sampling each realization,
and then comparing the estimated mean, variance, and
correlation length to the true values. The goal here is to
investigate the discrepancies between the estimated statistics
and the true local statistics, the latter obtained by sampling the
field at all locations. Note that the local statistics will differ
from the population parameters, (mean), (standard
deviation), and (correlation length), which are used by the
random field generator, due to the fact that the local statistics
are derived from a single realization. In detail, the soil is
represented by a stationary Gaussian random field, x X , at
spatial position , which is simulated within the domain and
sampled at
x
s
n locations. The samples are then used to estimate a
mean trend, x , which can then be compared to the field
realization to assess its ability to represent the actual mean
trend. Defining the residual to be
( ) ( ) ( ) X X
r
x x x (1)
then
x is a good estimate of the mean trend if is
generally small. If the site is sampled at all locations, then
r
X
x
can be taken to be equal to X x , in the event that a pointwise
trend is assumed (as in Kriging), in which case 0
r
X x
everywhere. Sampling at all locations is the best case since
there is then minimum residual uncertainty (zero in the case of
Kriging).
524
th
Sampling at all locations is, of course, prohibitively
expensive and would also change the resulting field properties
while measuring them (see, e.g., Heisenberg, 1927). In practice,
soil properties are estimated from a relatively small number of
samples so that will only ever approximate in some
way (i.e., via a trend).
( )
x ( ) X x
In assessing the ability of ( )
x to represent ( ) X x , it will
also be useful to consider the average residual over the domain,
( ) ( ) ( )
1
1 1
r r i
n
i
D D
X d X
D D n

=
= =

}
x x x x
i
( (2)
where D is the edge dimension of the D D
( )
square domain.
The domain is broken up into cells in the simulation,
resulting in the summation form on the right, in which is the
location of the center of the th cell.
n
i
x
i
The agreement between and will be determined
here by considering three measures; 1) the standard deviation of
the residual field average,
( ) x X x
r
(i.e., how well does the estimated
trend represent the actual field average?), 2) the standard
deviation of the residual,
r
(i.e. how much residual
uncertainty remains?), and 3) the residual correlation length (i.e.
how does the trend removal affect the perceived correlation
lengths?).
X
Five sampling schemes are considered in the paper, ranging
from a single sample taken at the field midpoint to nine samples
taken over a 3 x 3 array at the quarter points of the field. In
some cases a further maximum' sampling scheme is performed,
where every point in the field is sampled, to see what the
maximum attainable uncertainty reduction is.
For each sampling scheme, three types of trend removal are
performed; a) removing the constant sample mean, b) removing
a bilinear trend surface which is fit to the sample, and c)
removing a Kriged surface fit to the sample. The residual
statistics are determined by Monte Carlo simulation, with 2000
realizations for each case, where the field is discretized into 128
x 128 cells and the random fields generated using the Local
Average Subdivision method (Fenton and Vanmarcke, 1990).
2 RESULTS
Consider first the average of the residual,
r
, given by Eq. 2. It
can be shown that the mean of
r
is zero, so that a measure of
how accurately represents ( )
x ( ) X x can be obtained by
looking at the standard deviation of
r
small values of this
standard deviation imply that ( )
x remains close to the field
average. Figure 1 illustrates how the standard deviation of
r
,
normalized by dividing by the standard deviation of the random
field value, , in the th cell (referred to as
cell
( X x )
i
i o ), varies
as a function of the number of samples taken from the domain,
s
n , and the normalized correlation length, / D u . Note that if
only one sample is taken at the midpoint of the domain,
s
1 n = ,
then a bilinear trend cannot be fit to the sample, nor is a Kriged
surface removal attempted. Thus, parts b and c in Figure 1 do
not have a curve corresponding to
s
. In all plots it is
apparent that as the number of samples increases, the accuracy
improves (in agreement with the findings of Lloret-Cabot, et al.,
2012). It can be seen, however, that for
s
to 9, there is
very little difference between the detrending methods, so far as
the field average is concerned. It is to be noted that the field
average is a constant, not a trend, so it is not expected that the
bilinear and Kriged surface trends will do any better than the
sample mean, when compared to the field average.
1
3 n =
n =
Figure 1. Standard deviation of the field average residual (eq. 2),
normalized by the standard deviation of X, versus normalized
correlation length.
In all cases in Figure 1, the agreement between ( )
x and
( ) X x improves as the correlation length increases. This is
because the field becomes increasingly smooth, or flat, as the
correlation length increases, so that all trends considered
become closer to the flatter ( ) X x .
525
Figure 2. Standard deviation of the residual (eq. 1), normalized by the
standard deviation of X, versus normalized correlation length.
A possibly better measure of how well represents the
field is obtained by considering the standard deviation of the
residual,
r
(see eq. 1), directly. This measure will include
the effects of trend removal and is illustrated in Figure 2, again
with the standard deviation of the residual,
r
( )
x
( ) X x
o , divided by the
standard deviation of ,
cell
X o . In detail, the standard deviation
of the residual is estimated as the square root of the variance,
( ) ( )
1
2
2
1
1
n
i
i
r i
X
n
o
=

x x
for each realization. The value of
r
o used in Figure 2 is
averaged over all realizations. As in Figure 1, the 1
s
n = case
only appears in Figure 2a, since bilinear trend and Kriging
surfaces are not well defined for only one sample point.
However, Figures 2a and b now include a limiting case where
the entire simulation has been sampled (
s
all), representing
the best site knowledge possible. This case was not included in
Figure 1 since, when all values are sampled,
n =
0
r
= , that is, the
average residual is zero. In Figure 1, this would have
corresponded to a horizontal line at zero standard deviation. In
Figure 2, the
s
n = all case corresponds to the classical case
where both the estimated mean (trend) and the variance are
computed from the same set of observations. As the correlation
length decreases, these observations become increasingly
independent, and the estimated standard deviation approaches
the true standard deviation, so that
r cell
1. / 0 o o as seen in
Figures 2 a and b when
s
n = all. In Figure 2 c, the case
s
n =
all is not included in the Kriging surface case since, when the
entire field is sampled, the residual is zero with zero variability,
and so the curve corresponding to this case lies at zero.
As in Figure 1, Figure 2 also shows that the ability of
( )
x to represent ( ) X x improves as the correlation length
increases, for all of the trends considered. In the limit, as
/ D u , all random fields become uniform (under the
assumed finite variance correlation structure), random from
realization to realization, but constant within each realization. In
this limiting case, the sample perfectly predicts the uniform
field, and the residual becomes zero everywhere so that 0
r
o = .
It is apparent in Figure 2 that all curves are heading towards 0,
as / D u .
One of the perhaps surprising results of Figure 2 is that the
removal of a bilinear trend is not generally as good as the
removal of the constant sample mean at smaller correlation
lengths, and especially at a lower number of samples. The
reason for this becomes apparent when, for example, the case
where 3
s
n = is considered. If the correlation length is small,
then the three samples will be largely independent, and the
resulting fitted bilinear plane could (and often does) end up with
quite an unrepresentative slope, leading to a high variability in
the residual. Even when 9
s
n = the residual variability is higher
at low correlation lengths than seen using the constant sample
mean. At low correlation lengths, the Kriging surface performs
about the same as the constant sample mean.
At large correlation lengths, e.g. / 1 D 0 u = , the bilinear
trend performs better than the constant sample mean for all
s
n
except 3
s
n = , where the relative standard deviation is 0.35
versus 0.32 for the constant sample mean. For higher number of
samples, the relative standard deviation using the bilinear trend
is 0.25, versus 0.31 for the constant sample mean. The Kriged
surface performs the best out of the three methods (relative
standard deviation of 0.30) when the number of samples is 3,
and about the same as the bilinear trend for higher numbers of
samples.
The last measure of the quality of the trend type used
considered in this paper is how well the estimated correlation
length agrees with the actual correlation length, Figure 3. Once
( )
x has been established from the soil samples, the
correlation length is estimated here using the following steps;
1. for each direction through the soil domain, 1, 2 i = ,
2. estimate the semi-variogram along all lines through the
domain in direction i using the entire ( ) X x field,
r
3. average the semi-variograms obtained in step 2 to obtain
the final semi-variogram estimate in direction i ,
4. fit a theoretical semi-variogram, having parameter u
(correlation length), to the semi-variogram estimated in
step 3 by minimizing the sum of squared errors (i.e.
regression).
( (3)
526
th
In general, when D the estimated correlation length is
overestimated, and often considerably overestimated, especially
when the actual correlation length is small. This occurs because
errors between the estimated trend (of any of the three types)
and actual bilinear field trend (bilinear because correlation is a
measure of the degree of linear dependence between random
variables) are perceived in the estimation process to be caused
by a strong lingering correlation (and not by an error in the
original trend estimate) hence a longer correlation length is
estimated to account for the evident residual trend.
Of the three trend types considered, the best is the constant
sample mean and the worst is the bilinear trend (except when
s
n all). The Kriged surface is slightly worse than the constant
sample mean. For example, when and 9
s
n / 0.0 D 5 , then
5 6 / .
r
, 10.0, and 6.3 for the constant sample mean,
bilinear trend, and Kriging surface, respectively. It should be
noted that the best performer, the constant sample mean, may be
so only because the simulated field is assumed stationary (i.e.
constant mean).
At the other end of the plot, where D , the correlation
length is underestimated ( 1 /
r
). In general, this is because
the removal of a trend in a strongly correlated field is also
removing the evidence of the strong correlation (strong
correlation is evidenced by a trend having little variation off the
trend) resulting in a residual field without strong correlation
hence a small correlation length. Of the three trend types
considered the best performer at the large correlation length end
is again the constant sample mean. For example, when
9
s
n and / 1 D 0 , then
r
/ 0.08 , 0.05, and 0.06 for the
constant sample mean, bilinear trend, and Kriging surface,
respectively.
3 CONCLUSIONS
There is no difference between the accuracies of the trend type
selected when matching the trend to the field average,
r
. As
expected, the accuracy improves as the number of samples and
the correlation length increase. If a target standard
deviation,
r
, equal to 20% of the random field standard

deviation,
cell
, is desired, then only one sample is required if
/ D 10 , while 9 or more samples are required if / 1 D .
In general, if the correlation length is small, the most
accurate approach is to use a constant sample mean, which
shows the best general results for all three measures of accuracy
considered in this paper. Kriging is almost identical, only losing
out slightly when considering the residual estimated correlation
length. At the other end of the scale, when the correlation length
is large, the bilinear trend is more accurate with respect to the
residual standard deviation than is the constant sample mean, as
expected.
In the absence of knowledge about the actual correlation
length, it appears that the Kriging surface removal, although not
generally the best in any one measure, is very competitive and is
certainly a good overall choice.
4 REFERENCES
Fenton G.A. and Vanmarcke, E.H. 1990. Simulation of Random Fields
via Local Average Subdivision, ASCE Journal of Engineering
Mechanics, 116(8), 1733 1749.
Figure 3. Estimated correlation length of the residual, normalized by the
point correlation length, versus normalized actual correlation length.
The correlation length estimated from the residual,
r
, will
agree with the actual correlation length used in the simulation,
, when the ratio
r
1 / . It can be immediately seen in
Figure 3 that this only occurs in general when the entire field is
sampled and the correlation length is relatively small (i.e.
significantly less than D ). That is, when the entire field is
sampled (
s
all), so that the sample average is equal to the
actual field average, the estimated correlation length becomes
equal to the actual correlation length when the samples are
relatively independent (small
n
).
Heisenberg W. 1927. ber den anschaulichen Inhalt der
quantentheoretischen Kinematik und Mechanik, Zeitschrift fr
Physik, 43(3-4), 172 198.
Lloret-Cabot, M., Hicks, M.A., and Van Den Eijnden, A.P. 2012.
Investigation of the reduction in uncertainty due to soil variability
when conditioning a random field using Kriging, Gotechnique
Letters, 2, 123 127.
527
Multi-Sleeve Axial-Torsional-Piezo Friction Penetration System for Subsurface
Characterization
Systme de pntromtre friction axial-torsional-piezomtrique manchons multiples
pour la reconnaissance des sols superficiels
Frost J. D., Martinez A.
Georgia Institute of Technology
ABSTRACT: The multi-sleeve penetration system is an in-situ testing device that is derived from the cone penetration test. It
incorporates a series of friction sleeves with varying surface texture along with a series of pore pressure sensors, in addition to the
standard smooth friction sleeve and pore pressure sensor located directly behind the tip in the conventional CPT device. The
multiple measurements made with this device allow it to provide new insight into soil type and stratigraphic variations as well as
in-situ shear strengths as a function of sleeve texture height. This paper describes a third generation version of this device that
incorporates torsional load sensing capabilities in addition to the standard axial load sensing capabilities. In this manner, the
effects of different vertical and horizontal stress states on measured sleeve stresses can be explored. This multi-sleeve technology
offers benefits over devices which are used to measure the mechanical response of soils.
RSUM : Le systme multi-manchon de pntration est un dispositif de test in situ qui est drive partir de l'essai de
pntration de cne. Il comporte une srie de manchons de friction avec plus ou moins de surface le long d'une srie de capteurs
de pression de pore, en plus de la douille de friction lisse et standard de capteur de pression de pore situ directement derrire
l'extrmit dans le dispositif de CPT classique. Les multiples mesures effectues avec cet appareil permettent d'apporter un
nouvel clairage sur le type de sol et les variations stratigraphiques ainsi que in situ la rsistance au cisaillement en fonction de la
hauteur de la texture manche. Cet article dcrit une version de troisime gnration de ce dispositif qui intgre la charge de
torsion capacits de dtection, en plus de la charge axiale norme capacits de dtection. De cette manire, les effets des diffrents
tats de contraintes verticales et horizontales sur les contraintes manches mesures peuvent tre explores. Cette technologie
multi-douille offre des avantages par rapport d'autres appareils qui sont utiliss pour mesurer la rponse mcanique des sols.
1 INTRODUCTION
The general trend followed for in-situ site characterization
practice has been to utilize devices that incorporate only one
sensor of a given type to measure desired engineering
properties. While a number of different sensor types may be
incorporated into a single device, they typically measure
different properties and then rely on empirical correlations
to predict engineering properties. The primary reason for
this single sensor approach has been historical precedent as
opposed to any compelling technical limitations. While this
approach has proven to yield generally acceptable results for
many projects, opportunities remain to improve practice. For
example, as the complexity and uniqueness of investigation
projects increase, the merit of conventional single sensor in-
situ tools decreases. Hence, recent efforts have sought to
develop new tools for subsurface characterization studies
configured with multiple sensors, which have the ability of
providing more reliable information as part of more detailed
investigations.
As noted above, invasive site characterization tools have
traditionally followed the approach of using single-sensor
configurations. An example is the cone penetration test
(CPT). The CPT measures, as a minimum, the penetration
resistance of a conical tip inserted into the ground, the
frictional force that the soil exerts on a smooth sleeve
located just above the cone tip, and the pore pressure
(assuming the pores are fluid filled) recorded at a location
also typically close to the penetrating tip as the probe is
inserted into the subsurface. Such an in-situ tool can provide
a robust set of data in the sense that it measures the bearing
and frictional resistances of the soil being tested. However,
one shortcoming is that it only measures the frictional
response of the soil when sheared against a surface of fixed
and specified low roughness. Studies by Frost and DeJong
(2005) have shown that friction measurements of soil
against smooth surfaces are more indicative of soil particle
sliding along the surface and not of shearing against the
sleeve surface. A more robust characterization of interface
strength can be achieved when the soil is sheared against a
range of surfaces of different roughnesses (DeJong et al.,
2001).
2 MULTI-SENSOR IN-SITU TOOLS
Among the new generations of more specialized in-situ tools
that exploit the multiple sensor approach are the multi
sleeve penetrometer attachments developed at the Georgia
Institute of Technology (DeJong, 2001; DeJong and Frost,
2002; Hebeler, 2005; Hebeler and Frost, 2006; Frost et al.,
2012). These attachments are designed to be used behind a
regular 15cm
2
CPT, or as a stand-alone device behind an
instrumented tip. The first and second generation devices
were described in detail by DeJong and Frost (2002) and
Hebeler and Frost (2006), respectively, and are briefly
summarized below. The third generation device is under
development and is introduced herein.
528
th
2.1 First Generation: Multi-sleeve Friction Attachment
(MFA). The first generation of multi sensor attachments
deviates from the standard CPT in that the MFA is capable
of measuring four different friction sleeve stresses in
addition to the standard CPT measurements (q
t
, f
s
, u
2
). Each
sleeve position offers the possibility of being equipped with
a sleeve of different roughness, with the intention of
inducing different degrees of shearing in the soil. Figure 1
shows a schematic of the MFA. According to studies
conducted by Frost and DeJong (2005), the standard smooth
CPT friction sleeve measurement is more indicative of soil
sliding against the sleeve as opposed to shearing against the
soil. The reason is that the conventional CPT friction sleeves
are manufactured with an intentionally smooth surface. As a
consequence of the MFAs multi-sensor configuration, the
device is able to determine the end bearing capacity of the
soil and the relationship between interface shear strength
resistance and surface texture in a single sounding. The
important relationship between interface shear strength and
surface roughness was originally identified through
laboratory tests by Uesugi and Kishida (1986).
2.2 Second Generation: Multi-sleeve Piezo-Friction
Attachment (MPFA). The second generation of multiple
sensor devices offers the ability to directly measure the
interface response over a range of counterface profiles,
while simultaneously measuring the excess pore water
pressure ahead of and after each friction sleeve as the device
is advanced into the subsurface. This is achieved by means
of its four independent load cells attached to the textured
sleeves and five independent dynamic pore pressure sensors.
C
o
n
v
e
n
t
i
o
n
a
l

1
5

c
m
2
C
P
T

M
o
d
u
l
e
M
u
l
t
i
-
F
r
i
c
t
i
o
n

S
l
e
e
v
e

P
e
n
e
t
r
o
m
e
t
e
r

A
t
t
a
c
h
m
e
n
t
q
c
f
s
f
s
#1
u
2
f
s
#2
f
s
#3
f
s
#4
Attachment
Digital
Housing
Attachment
Digital
Board
Digital
Housing
Digital
Board
Dual Axis
Inclinometer
Friction Sleeve
Tip
Pore Pressure
6
1

c
m
1
0
9

c
m
Friction Sleeve
Friction Sleeve
Friction Sleeve
Friction Sleeve
Mandrel
Mandrel
Mandrel
Mandrel
4.37 cm
C
o
n
v
e
n
t
i
o
n
a
l

1
5

c
m
2
C
P
T

M
o
d
u
l
e
M
u
l
t
i
-
F
r
i
c
t
i
o
n

S
l
e
e
v
e

P
e
n
e
t
r
o
m
e
t
e
r

A
t
t
a
c
h
m
e
n
t
q
c
f
s
f
s
#1
u
2
f
s
#2
f
s
#3
f
s
#4
Attachment
Digital
Housing
Attachment
Digital
Board
Digital
Housing
Digital
Board
Dual Axis
Inclinometer
Friction Sleeve
Tip
Pore Pressure
6
1

c
m
1
0
9

c
m
Friction Sleeve Friction Sleeve
Mandrel Mandrel
Mandrel Mandrel
Mandrel Mandrel
Mandrel Mandrel
4.37 cm
Figure 1. Schematic of the multi-sleeve friction penetrometer along
with a standard CPT module.
The coupling of axial load and pore pressure sensors gives
the MPFA the ability to provide a direct measure of pore
water pressure generation due to shearing against surfaces of
different roughnesses. Several advantages offered by the
MPFA are the ability to consider the measured interface
response data within an effective stress framework which is
useful for applications such as liquefaction as well as
strength degradation, flow and consolidation characteristics
along the penetrometers shaft, more detailed data for
improved stratigraphy profiling, and the ability to
distinguish between drained, undrained and partially drained
conditions at the various sensor locations (Hebeler, 2005).
Figure 2 shows a schematic of the MPFA. Examples of the
unique insights resulting from the multi-sleeve sensor
technology include in-situ determination of the relationship
between interface friction and sleeve surface roughness
(Figure 3) and soil classification using interface behavior
(Figure 4).
3 PROPOSED SITE CHARACTERIZATION TOOL
3.1. Third Generation: Multi-sleeve Piezo-Friction-Torsion
Attachment (MPFTA). The third generation of multi-sensor
devices being developed at the Georgia Institute of
Technology incorporates both axial and torsional shear as
well as pore pressure sensing capabilities.
f
s
f
a1
u
2
f
a2
f
a3
f
a4
Attachment
Digital
Housing
Digital Housing
u
a1
u
a0
u
a4
u
a3
u
a2
q
c
Friction Sleeve
Pore Pressure
Tip Load
Dual Axis
Inclinometer
Digital
Board
Attachment
Digital
Boards
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Mandrel
Mandrel
Mandrel
Mandrel
Mandrel
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Attachment
Sleeve
Mandrel
Replaceable
Attachment
Friction
Sleeve
Attachment
Individual
Load Cell
Attachment
Individual
Piezo Sensor
(0.67)
(0.81)
(0.88)
(1.07)
(1.14)
(1.33)
(1.40)
(1.59)
(1.66)
f
s
f
a1
u
2
f
a2
f
a3
f
a4
Attachment
Digital
Housing
Digital Housing
u
a1
u
a0
u
a4
u
a3
u
a2
q
c
Friction Sleeve
Pore Pressure
Tip Load
Dual Axis
Inclinometer
Digital
Board
Attachment
Digital
Boards
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Friction
Sleeve
Mandrel Mandrel
Mandrel Mandrel
Mandrel Mandrel
Mandrel Mandrel
Mandrel Mandrel
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Piezo
Sensor
Attachment
Sleeve
Mandrel
Attachment
Sleeve
Mandrel
Replaceable
Attachment
Friction
Sleeve
Replaceable
Attachment
Friction
Sleeve
Attachment
Individual
Load Cell
Attachment
Individual
Load Cell
Attachment
Individual
Piezo Sensor
Attachment
Individual
Piezo Sensor
(0.67)
(0.81)
(0.88)
(1.07)
(1.14)
(1.33)
(1.40)
(1.59)
(1.66)
(a) (b)
Figure 2. Schematic of the multi-piezo-sleeve friction penetrometer
along with a standard CPT module (a) schematic - brackets indicate
sensor offset from tip in meters and (b) piezo friction sleeve
mandrel design detail.
Its dimensions and external characteristics, with and without
pore pressure sensing capabilities, are similar to the MFA
and MPFA shown in Figures 1 and 2, respectively. However,
the new concept incorporated into the device consists of a
dual load-torsion cell being installed in each sleeve module
529
and location, with the goal of measuring both axial and
torsional shear responses of the soil throughout the same
sounding. In this manner, the effects of special variability
(vertical and horizontal) will be eliminated and more
detailed information about the soils anisotropy and state of
stress can be provided. The proposed texture of the
MPFTAs friction sleeves is the same to the texture of the
MFA and MPFAs sleeves as shown in Figure 5. The
friction sleeve texture pattern consists of machined diamond
shaped features with a height that typically ranges from 0.25
to 2 mm in order to induce different degrees of shearing.
The configuration of penetration angle, diamond width,
diagonal spacing, texture slope and areas with no textural
features ensures that shearing is induced with the soil and
prevents clogging of the textural features.
Normalized Roughness, R
n
* (=R
max
(diamond height)/D
50
(sand))
0 2 4 6 8 10
I
n
t
e
r
f
a
c
e

S
l
e
e
v
e

S
t
r
e
s
s

(
k
P
a
)
0
100
200
300
400
Values from Individual Sleeve Tests (Soundings S17-S22)
S31 - CPT - Smooth - 30H.125S3 - 30H.25S3 - 30H.5S3
S32 - CPT - Smooth - 30H.5S3 - Smooth - 30H1S3
S33 - CPT - Smooth - 30H1S3 - Smooth - 30H2S3
Figure 3. Relationship between surface roughness and interface
friction determined using multi-sleeve technology.
Figure 4. Soil Classification Chart based on multi-sleeve data.
3.1 Comparison to Existing In-Situ Testing Systems. The
MPFTA device has relatively little in common with the in-
situ shear vane test. The shear vane is typically used to
characterize the response of soft clays. The reason is that
stiffer soils can compromise the structural integrity of the
shear vane, resulting in blade bending. It is considered that
this is not a limiting factor for the MPFTAs frictional
elements because of their different design and thus stiffer
configuration. It is important to note that the MPFTAs
intent is to the surface interface strength of the soil in the
axial and radial directions, while the shear vanes intent is to
measure the soils undrained shear strength. Finally, as
shown by Chandler (1988), different diameter sizes can
impose strain-rate effects; however since the diameter of the
MPFTA device is constant and only the height of the
diamond texture elements changes, the results of the
MPFTA will not need to be corrected for this and other
potential geometry effects.
Figure 5. MFA and MFPAs friction sleeves with increasing
diamond height (from left to right)
3.2 Sleeve Locking Mechanism. For the MFA and MPFA
devices, the axial force is derived from measurements using
a series of bonded strain gauges configured as the four-
branches of a wheatstone bridge. Application of the soil
shear force on the textured sleeves brings them into contact
with a shoulder and the resulting change in length of the
bonded strain gauges changes the output of the Wheatstone
bridge. In order to measure the torque applied when the
sleeve is rotated, the sleeve is temporarily fixed to the core
of the mandrel by an electromagnet which prevents rotation
of the sleeve and instead induces changes in resistance of a
set of orthogonally bonded strain gauges also configured as
the branches of a Wheatstone bridge. Given the magnitude
of the forces on even the most heavily textured sleeves,
relatively low currents are required to lock the
electromagnets and thus sleeves during torsional testing. A
sketch and photograph showing the axial and torsional load
application modes for the new device are shown in Figure 6.
Final designs of the actual combined axial-torsional cell are
being completed. Once measurements at a given sounding
elevation are completed, the electromagnets are turned off
and the penetration of the device and recording of axial
loads is continued. In many instances, the device will be
advanced so that a sleeve is advanced to the same elevation
that the adjacent preceding sleeve was located at in a
previous torsional test so that successive torsional test
measurements are made at the same elevations with sleeves
of increasing texture height. This eliminates the need to
account for lateral and vertical variability since successive
tests are performed on the same material.
530
th
4 CONCLUSIONS
Significant advances have been made in the last four
decades in the design, use and interpretation of the results
from penetrometer devices used for subsurface
characterization. Similarly, over the past two decades,
significant new insight has been developed into the role of
surface roughness on the behavior of geotechnical interfaces.
An emerging family of innovative devices has been
developed in the last decade that leverages the advances in
performance of penetrometer devices with the new
understanding of interface behavior to produce multi-sleeve
devices that allow for direct in-situ determination of the
relationship between surface roughness and interface shear
as well as the development of shear induced pore pressures
when surfaces of various roughness are sheared against soils.
A unique aspect of this family of devices is the use of
multiple friction sleeves with surface of different roughness
in the same sounding so that the effects of material
variability can be isolated and/or eliminated. To date, all
these devices rely on response of soils during axial
penetration.
Figure 6. Schematic and photograph of multi-sleeve piezo- friction-
torque penetrometer showing load application modes.
This paper describes the development of a new device that
embodies the attributes of the existing multi-sleeve devices
but incorporates the ability to also conduct torsional friction
penetrometer tests in the same sounding. In contrast to
existing vane shear type devices which involve the
application of a torsional force to a rigid central shaft and
measure the resistance to rotation of a set of blades in a soil,
the new device enables measurement of torsional resistance
with the same textured sleeves used in the axial stage of the
test. This is possible through the use of an innovative
electro-mechanical system that allows independent
measurements of axial and torsional resistance of the sleeves
of the penetrometer device. The availability of
complimentary axial and torsional shear forces along with
the associated pore pressures generated by friction sleeves of
different surface roughness represents a potential disruptive
technology in the in-situ characterization of soil properties
ranging from soil type to soil strength and deformation
properties to assessment of the in-situ state of stress and
associated parameters such as the in-situ stress ratio.
Significant opportunities exist for dramatic advances in
subsurface investigation. Single sensor historical precedent
has guided the design and configuration of in-situ devices,
there are clear advantages to developing multiple sensor
systems for future investigation studies.
The recent development of various configurations of multi-
sensor systems including the previously presented
generation 1 MFA and generation 2 MPFA as well as the
proposed generation 3 MPFTA device introduced herein
represent a significant departure from traditional practice.
Amongst the benefits of the latter device are:
i) Up to sixteen independent measures of interface shear
ranging from smooth surface sliding to textured surface soil
shearing can be realized in a single sounding in contrast to
the three measures possible with conventional cone
penetration systems.
ii) The effects of different vertical and horizontal stress
states on measured sleeve stresses can be accounted for by
means of the axial and torsional interface stress response.
iii) The procedure for performing combined axial-torsional-
piezo penetration testing of the subsurface using the
proposed MPFTA system involves a series of steps which
allows them to be readily controlled from a remote location
and to be performed using robotic systems.
5 ACKNOWLEDGEMENTS
The initial device development reported herein was funded
in part by a grant from the US National Science Foundation
to the Georgia Tech Research Corporation under Contract #
CMS 9978630.
6 REFERENCES
Chandler, R.J. (1998). The in-situ measurements of the undrained
shear strength using the field vane, Vane Shear Strength
Testing in Soils. Field and Laboratory Studies. A.F. Richards
(ed.), ASTM STP 1014, ASTM, Philadelphia, pp. 13-44.
DeJong, J.T. (2001). Investigation of Particulate-Continuum
Interface Mechanics and Their Assessment Through a Multi-
Friction Sleeve Penetrometer Attachment, PhD Dissertation,
Georgia Institute of Technology, Atlanta, May, 360 pp.
DeJong, J.T. and Frost, J.D. (2002). A Multi-Friction Sleeve
Attachment for the Cone Penetrometer, ASTM Geotechnical
Testing Journal, 25, No. 2, pp. 111-127.
DeJong, J.T., Frost, J.D., and Cargill, P.E. (2001). Effect of
Surface Texturing on CPT Friction Sleeve Measurements.
Journal of Geotechnical and Geoenvironmental Engineering,
127, No. 2, pp. 158-168.
Frost, J.D., and DeJong, J.T. (2005) In Situ Assessment of the Role
of Surface Roughness on Interface Response, Journal of
Geotechnical and Geoenvironmental Engineering, 131, No. 4,
pp. 498-511.
Frost, J.D., Hebeler, G.L., and Martinez, A., (2012), Cyclic Multi-
piezo-friction Sleeve Penetrometer Testing for Liquefaction
Assessment, Proceedings of 4
th
International Conference
(ISC4) on Geotechnical and Geophysical Site
Characterization, Pernambuco, Brazil, Vol. 1, pp. 629-636.
Hebeler, G.L. (2005). Multi Scale Investigations of Interface
Behavior. PhD Dissertation. Georgia Institute of Technology,
Atlanta, August, 772 pp.
Hebeler, G.L., and Frost, J.D., (2006), A Multi Piezo-Friction
Attachment for Penetration Testing, Proceedings of ASCE
Geo-Institute Congress: Geotechnical Engineering in the
Information Technology Age, Atlanta, CD ROM.
Uesugi, M. and Kishida, H., (1986) Frictional Resistance at Yield
Between Dry Sand and Mild Steel. Soils and Foundations,
Vol. 26, No. 4, pp. 139-149.
531
Hydraulic Properties of Glacial Deposits Based on Large Scale Site Investigation
Les proprits hydrauliques des dpts glaciaires bases sur une enqute de chantier grande
chelle
Galaa A., Manzari M.
Coffey Geotechnics, Toronto, Ontario, Canada
Hamilton B.
CH2M Hill, Toronto, Ontario, Canada
ABSTRACT: Glacial deposits by nature comprise variable soil types in relatively short distances. Hydraulic conductivity (K) is the
most important parameter in design of construction dewatering for underground structures. However, determination of proper design
values for K is not an easy task. Due to the inherent variable nature of the glacial deposits, even conventional pumping tests may not
provide reliable design parameter due to its smaller zone of influence compared to that of the actual dewatering for a structure. This
paper describes the methodology created for establishing more representative design values for hydraulic conductivity of glacial
deposits during a large scale subsurface investigation for planned tunnels. The subsurface investigation involved 400 boreholes,
including 88 slug tests and 16 pumping tests. A relation was established between K obtained from the field tests (K
field
) and K
calculated by applying Kozeny-Carman formula (K
KC
). Subsequently, the calibrated K-C formula was applied to 1,200 grain size
analyses conducted on various soil types. The calculated and measured K were used to form statistical analysis of the parameter and
provide more reliable design values for dewatering.
RSUM : Les dpts glaciaires comprennent des sols variables travers des distances relativement courtes. La conductivit
hydraulique (K) est le paramtre le plus important qui est ncessaire durant la construction des structures souterraines. Cependant, la
dtermination des valeurs de calcul appropries pour K n'est pas une tche facile. cause de la nature variable des dpts glaciaires,
mme les essais de pompage peut-tre ne fourniront pas des rsultats fiables pour une bonne conception pour une bonne conception
parce que les structures dshydrats ont une plus grande zone d'influence. Ce document dcrit la mthodologie cre pour tablir les
paramtres de conception plus reprsentatives au cours d'une enqute de chantier grande chelle pour les tunnels de mtro prvues.
L'tude a port sur 16 essais de pompage avec des puits d'observation associs, et 88 essais de conductivit hydraulique. Une relation
a t tablie entre K obtenue partir des essais sur le terrain (K
field
) et K calcul en appliquant la formule de Kozeny-Carman (K
KC
).
Par la suite, la formule de K-C calibre a t applique des analyses granulomtriques effectue 1200 chantillons. Les valeurs de K
calcules et mesures ont t utilises pour former une analyse statistique, et pour fournir des valeurs plus fiable.
KEYWORDS: Kozeny-Carman formula, hydraulic conductivity, Glacial Tills, dewatering.

1 INTRODUCTION
The Greater Toronto and Hamilton Area (GTHA), located in
southern Ontario, is Canadas largest and fastest growing urban
region. The Government of Ontario Province through its
transportation authority known as Metrolinx, has embarked in a
massive transportation plan called The Big Move, which is a
25-year, $50 billion plan that will transform regional
transportation across the GTHA. The Eglinton Scarborough
Crosstown (ESC) Light Rail Project is part of that Big Move
program. The ESC is a 19-kilometre light rail transit line (LRT)
that will run along Eglinton Avenue, connecting west to east of
the city. Eleven kilometers of the alignment will be tunneled
underground, crossing well established urban areas which are
densely populated and congested. The tunnel construction is
divided in two contract packages: West Twin Tunnels
Construction and East Twin Tunnels Construction, with Yonge
Street the dividing limit. Dewatering operations will be required
for a total of twenty four structures along the tunnel alignment:
sixteen cross passages, four launch and exit shafts, and six
emergency exit buildings.
In order to meet a very tight schedule while properly
managing subsurface risk and support the design of the tunnel,
an aggressive multi-phase geotechnical investigation program
was undertaken. The geotechnical investigation for the west and
east tunnel contracts was conducted during a two-stage program
between 2010 to mid-2012; which followed by a
hydrogeological study for each section. In summary, about four
hundred (400) shallow and deep sampled boreholes were
advanced including three hundred (300) monitoring wells along
the subject alignment to obtain information regarding the
subsurface stratigraphy and groundwater conditions.
Furthermore, eighty eight (88) slug tests and sixteen (16)
pumping tests (150 mm O.D.) were completed as part of the site
specific hydrogeological study. At the time of preparation of
this paper, only the results of eight (8) pumping tests for the
west tunnels are available and used in analyses.
Due to projects very tight schedule and ongoing progress of
design, the proposed locations of some structures were revised
after completion of the pumping tests. Furthermore, it was not
practical to conduct the pumping tests for all of the structures.
Innovative techniques were developed and used to establish
more representative design value of hydraulic conductivity
while not having pumping test at exact location of each
structure and also consider the inherent variable nature of the
glacial deposits. This paper describes the methodology
developed and summarizes the range of hydraulic conductivity
for various types of glacial deposits obtained from this large
scale subsurface investigation which is generally more refined
than older published range for the same deposits.
2 GEOLOGY SETTING
A detailed regional description of the Quaternary geology of the
project area can be found in the Ontario Geological Survey Map
(Sharpe, 1980). The soil deposits in the project area are result of
glacial depositional systems that took place during various
glacial periods. From the published geological data, the GTHA
experienced three glacial and two interglacial periods. This
532
th
fluctuating glacial advance and retreat produced a complex
distribution of over-consolidated glacial till layers, separated by
interstadial and interglacial stratified deposits of glacio-
lacustrine plastic silt/clays and non-plastic silt/sands.
The subsurface overburden encountered during the site
investigation were initially classified into 17 different soil types
(Types 1 through 17). The soil classification system followed
the modified version of Unified Soil Classification System.
Identification of soil origin as till was based on their
heterogeneous structure, the relatively broad grain size
distributions and the documented local geology. Many of the
different soil types demonstrate relatively comparable
engineering characteristics and may possibly have similar
geological origin. Consequently, the various soil types were
consolidated into six engineering classes (Classes A through F).
The six soil classes are as follows:

- Class A: Fill and Topsoil
- Class B: Interstadial Sand to Gravel
- Class C: Interstadial Silt to Sand
- Class D : Non-Plastic Till
- Class E : Plastic Glacio-lacustrine
- Class F : Plastic Till

Class B was divided into two subclasses based on the
percentage of silt and clay particles (<75 m). Sandy soils with
less than 20% silt and clay particles were grouped under Class
B2,3,4 and the rest (> 20% silty and clay) under Class B5,6.
3 ESTABLISHING HYDRAULIC CONDUCTIVITY
Glacial deposits by nature comprise of variable soils types in
relatively short distances. Due to the inherent variable nature of
the glacial deposits at project area, conventional filed pumping
tests may not provide fully reliable results for a proper
dewatering calculation as the zone of influence of a pump test
may only extend a few tens of meters. On the other hand, the
actual dewatering volume of a structure is affected by the
characteristics of surrounding soil within a few hundreds of
meters. Furthermore, the pumping tests were not necessarily at
the exact location of some structures.
It became necessary to complement the hydraulic
conductivity values obtained through field testing in order to
expand the test results to a larger domain or be able to focus on
any specific area. It was decided to use the available semi-
empirical methods/formulae in literature to complement
hydraulic conductivity values obtained through filed testing
with predicted values based on index properties such as grain
size distributions, pore size distributions and/or specific surface.
The following sections will outline the procedure followed to
predict hydraulic conductivities and provide design parameters.
3.1 Kozeny-Carman formula
Since Kozeny (1927) introduced his theory for a series of
capillary tubes and Carman (1938 and 1956) followed this work
and provided formulations that takes into the account the
tortuosity of the flow path of a fluid in a porous medium. The
following formula presented by Carman was then referred to as
the Kozeny-Carman (K-C) formula (Carrier, 2003).
Details of the formula can be found in the subject references.
In summary, the hydraulic conductivity of the soil can be
estimated as follows:
| | ( ) )] 1 /( )[ / 1 ( )} /( { / % 100 10 99 . 1
3 2
2
5 . 0 5 . 0 4
e e SF D D f K
si li i
+ =
(1)

Where, e is the void ratio; SF is a shape factor; f
i
is the
fraction of particles between two sieves (%), denoting the larger
sieve with (l) and the smaller one as (s) in, and D
ave-i
=
(D
li
D
si
)
0.5
is the average particle size, in cm, between two sieve
sizes
.
.
The Kozeny-Carman formula takes into account specific
surface area of full range of particle sizes and soil void ratio
which leads to better accuracy than the famous Hazen formula
(Lambe and Whitman 1969) in predicting the hydraulic
conductivity for a wide range of soils. Notwithstanding the
above, the application of K-C formula is constrained by almost
the same limitations as Hazen (Carrier 2003). Such constrains,
as discussed below, arise when dealing with soils at the
extremes of any spectrum such as the grain size, particle size
distribution, particle shape, and particles orientation
(anisotropy).
The formula does not account for the electrochemical forces
between particles and particles and water which disqualify the
formula from being applied to clayey soils. In addition, the
formula assumes laminar flow, which may not be satisfied in
gravels and gravelly sands. The formula does not produce a
close estimate to the specific surface area of particles with
extreme shapes such as platy or flakey particles. Therefore, the
K-C formula may not be applicable in these cases or can be
applied after replacing the calculated specific surface area by
the measured value. Also, K-C formula does not account for soil
anisotropy which is more pronounced in natural deposits than
for laboratory constructed samples.
Locat et al (1984) measured the specific surface area (S) for
several clays and found that clays with low plasticity (8 < PI <
15) have S between 23 and 30 m
2
/kg and is independent of the
percentage of soil finer than 2 m. Chapuis and Aubertin (2003)
picked a constant number between 23 and 30 m
2
/kg as an
estimate for S of the soil fraction finer than 2 m and calculated
S for the fraction coarser than 2 m as per original K-C
formula. Consequently, the results of these hybrid methods in
using K-C formula were in good agreement with measured
hydraulic conductivities in laboratory for clayey soils with
PI<15. In this study, the approach proposed by Chapuis and
Aubertin (2003) was followed for plastic glacial tills with PI
less than 15. However, the effect of weathering and factures in
the upper portion of the clayey till deposits must be considered
in any assessment (McKay, 1993; Hendry, 1982).
3.2 Site specific correlation factor for K-C formula
This section outlines the work completed in the field to obtain
in-situ hydraulic conductive (K) for the different soil classes
and explains the approach followed to establish site specific
correlation factor for using K-C formula.
Hydraulic conductivities for each soil class were measured in
the field by a combination of pumping tests and/or falling or
rising head slug tests. The results of 8 pumping tests with
associated observation wells and 88 slug tests conducted along
the tunnel alignment, distributed among six soil classes are used
in this study. The number of the field tests performed on the
aquifers materials was greater than those performed on the
other soil types. However, a significant number of the tests were
performed on both plastic and non-plastic tills.
One grain size distribution analysis was conducted, as
minimum, on the soil samples recovered from within the screen
interval of the 88 slug tests and pumping tests with associated
observation wells. These grain size distributions were
determined by undertaking sieve analysis, in accordance to
ASTM C136-06, and the hydrometer test, in accordance to
ASTM D422-63. These grain size distributions analyses were
used to calculate K based on the K-C formula. After excluding
the tests for samples with PI >15 and/or field test conducted in
the clayey till deposits with obvious signs of weathering and
fracture, K-C formula was applied to about 80 grain size
analyses that were screened as suitable (not within the
limitations of the formula) and correspond with K obtained
from field tests. As a result, for every in-situ measured K in the
field (K
field
) there is a corresponding predicted K from applying
KC formula to the grain size analysis associated with the screen
interval (K
KC
), as shown in Figure 1.
533

1.E07
1.E06
1.E05
1.E04
1.E03
1.E02
1.E01
1.E+00
1.E07 1.E06 1.E05 1.E04 1.E03 1.E02 1.E01 1.E+00
K
(
F
i
e
l
d
)
(
c
m
/
s
e
c
)
K(KozenyCarman)(cm/sec)
Correlationline
K(field)=3K(KC)
K(field)=1/3K(KC)

Figure 1 In-situ measured field hydraulic conductivity versus calculated
by Kozeny-Carmen Formula (K
KC
vs. K
field
)
The dashed line represents the equality line and the solid
black line represents the site specific correlation line which has
a slope shown in equation (2).

log K
field
= 1.03log K
KC
(2)

The grey lines in Figure 1 represent the boundaries that
encompass 90% of the data points. These lines have the same
slope as the correlation line with 0.5 offset in the log-log scale.
This indicates that Kozeny-Carmen formula with incorporation
of the site specific correlation factor of 1.03 (equation (2))
predicts a K value ranging between 1/3 to 3 times the in-situ
measured field hydraulic conductivity (K
field
) for the glacial
deposits in this specific site. These conclusions are comparable
to the margin obtained from laboratory permeability test results
shown by Chapuis (2002) and Chapuis and Aubertin (2003).
3.3 Overall hydraulic conductivity for each soil class
Hydraulic conductivity (K) values for each soil class of glacial
deposits were calculated using the K-C formula as per method
described in the previous sections for about 1,200 grain size
analyses conducted on various soil types along the alignment.
Equation (2) is then used to correct K
KC
assuming that 90% of
the predicted values fall between 1/3 to 3 times the actual K in
the field. The statistical parameters were calculated for the
corrected K
KC
obtained for each soil class in conjunction with
the K values directly obtained from field tests (slug and
pumping tests). The statistical distribution of K for each soil
class is plotted in histograms as shown in Figure 2a to 2e.
The K values obtained from the field tests conducted in the
plastic till deposits (Class F) with obvious signs of weathering
and fracture has also been added to the calculated K values and
other field measurement results which all together included in
the statistical distribution of K for Class F (Figure 2a).
Generally, the higher end of the K distribution in Figure 2a is
associated with the field measured hydraulic conductivity in the
fractured plastic till. This is in conformance with the finding of
other studies in similar soil condition (e.g., DAstous 1989,
Ruland 1991). Although, some of the slug tests conducted on
this fractured zone were as low as the results typically
associated with soil matrix values; which could by the results of
the smeared zone tend to form around augered boreholes.

Figure 2a to 2e Statistical distribution of hydraulic conductivity for
various soil classes of glacial deposits obtained from the investigation.

The K values for Class B2,3,4 (interstadial sand with less
than 20% fines) fit a bimodal distribution (Figure 2e). Further
review of the resutls indicated that the higher peak (10
-2
cm/s) is
associated to sand with leass than about 10% fine; while the rest
of the class resutled to the lower peak.
3.4 Design hydraulic conductivity for structures
The zone of influence for 72 hours pumping tests ranged from
15 m to less than 100 m, depending on the location. On the
534
th
other hand, the zone of influence for actual dewatering volume
of the structures would be a few hundreds of meters and
therefore, the dewatering volume would be affected by the
characteristics of surrounding soil within this larger zone. In
order to assess the reliability of the pumping test results for
dewatering calculation, the uniformity of the soil within the
dewatering zone was verified using the correlation described in
the previous sections.
For each structure location, a zone of influence of 350 m
radius is assumed. Corrected K
KC
in conjunction with K values
directly obtained from field tests (slug and pumping tests)
within the assumed zones around each structure were pulled out
of the overall data available. Subsequently, the statistical
distributions of K-values for every soil class encountered within
the dewatering zone were prepared for each structure. Examples
of the cumulative distributions are shown in Figure 3a and 3b
for Structure No.1 and No.2.
Based on the localized distribution of the K-values for each
structure, the pumping tests results for some structures fall
within 70 percentile or higher; on the other hand, the results for
other structures could be as low as 20 to 50 percentile.
A detailed review of the results and interpretive subsurface
profile showed that generally when the zone of the influence of
the pumping tests was small, the K obtained from pumping test
tends to be on the lower side of the cumulative distribution.
This has also been augmented where random presence of
pockets/seams of Class C soil within Class B deposits has
dominant effect on pumping test results. The design K-value for
dewatering calculation has been selected based on the result of
the localized distribution of the K-values prepared for each
structure. Two examples are shown in Figure 3.
4 CONCLUSION
Glacial deposits comprise of variable soil types in relatively
short distances. Conventional pumping tests may not provide
fully reliable results for a proper dewatering calculation as the
zone of influence of a pump test may only extend tens of meters
while the actual dewatering volume of a structure is affected by
the characteristics of surrounding soil within hundreds of
meters. Presence of pockets/seams with higher silt content
within sandy deposits has dominant effect on pumping test
results. Smaller the zone of influence of the pumping tests, K
obtained from the test tends to be on the lower side of the
cumulative distribution for the dewatering zone of influence.
The pumping test results for some structures could be as low as
20 to 50 percentile of accumulative distribution. It is imperative
to assess the reliability of the pumping test results for
dewatering calculation in the variable glacial deposits;
particularly when the zone of the influence of the pumping tests
is relatively small.
The Kozeny-Carman formula takes into account specific
surface area of full range of particle sizes and soil void ratio and
proven to provide reliable predictions of K for wide range of
soils. Based on the results of this large scale investigation,
Kozeny-Carmen formula with incorporation of the site specific
correlation factor, predicts K values ranging between 1/3 to 3
times the in-situ hydraulic conductivity (K
field
) for the glacial
deposits. This provides a powerful tool in verifying the
reliability of pumping test results in glacial deposits. However,
careful consideration must be given to proper interpretation of
the field test results and applicability of the formula to site
conditions.
It also should be noted that K of weathered zone of clayey
deposits is controlled by flow through the fractures. The field K
measured in this zone could be up to a few orders of magnitude
greater than the clay matrix. Field measurements in this zone
may also be sensitive to smearing during the installation of
piezometers. Physical scale of field measurements may strongly
influence the resulting hydraulic conductivity values.

Figure 3 Localized distribution of hydraulic conductivity for (a) Class
B5,6 in Structure No.1 and Class B2,3,4 Structure No.2.
5 ACKNOWLEDGEMENTS
The authors would like express their gratitude to Metrolinx for
authorizing the preparation of this paper.
6 REFERENCES
Carman, P. C. 1938. The determination of the specific surface of
powders. J. Soc. Chem. Ind. Trans. 57, 225.
Carman, P. C. 1956. Flow of gases through porous media, Butterworths
Scientific Publications. London.
Carrier, W. D. 2003. Goodbye, Hazen; Hello, Kozeny-Carman. Journal
of Geotechnical and Geoenvironmental Engineering, 129(11),
1054-1056
Chapuis, R.P. and Aubertin, M. 2003. On the use of the Kozeny
Carman equation to predict the hydraulic conductivity of a soil.
Canadian Geotechnical Journal, 40(3), 616-628.
Chapuis R.P. 2002. The 2000 R.M. Hardy Lecture: Full-scale hydraulic
performance of soilbentonite and compacted clay liners. Canadian
Geotechnical Journal, 39(2), 417-439.
DAstous et al. 1989. Fracture effects in the shallow groundwater zone
in weather Sarnia-area clay. Canadian Geotechnical Journal, 26,
43-56.
Hendry, M. J. 1982. Hydraulic conductivity of a glacial till in Alberta.
Ground Water 20(2), 162-169.
Kozeny, J. 1927. Ueber kapillare Leitung des Wassers im Boden. Wien,
Akad. Wiss. 136 (2a), 271.
Lambe, T. W., and Whitman, R. V. (1969). Soil mechanics. Wiley, New
York.
Locat, J., Lefebvre, G., Ballivy, G. 1984. Mineralogy, chemistry, and
physical properties interrelationships of some sensitive clays from
Eastern Canada. Canadian Geotechnical Journal, 21 (3), 530-540.
McKay L., Cherry J., and Gillham R 1993. Field Experiments in a
Fractured Clay Till. Water Resources Research, 29(4), 1149-1162.
Ruland W. et al. 1991. The depth of active groundwater flow in a clayey
till plain in southwestern Ontario. Ground Water 29(3), 405-417
Sharpe 1980. Quaternary Geology Series, Quaternary Geology
Toronto and Surrounding Area, Southern Ontario. Ontario
Geological Survey Map 2204.
535
The seismic SPT test in a tropical soil and the G
0
/N ratio
L'essai SPT sismique pour le sol tropicaux et la relation G
0
/N
Giacheti H.L., Pedrini R.A.A.
Universidade Estadual Paulista, Departamento de Engenharia Civil e Ambiental, Bauru SP Brazil
B. P. Rocha
Universidade de So Paulo, Departamento de Geotecnia, So Carlos SP Brazil
ABSTRACT: The seismic SPT, a test which associates the up-hole technique to the SPT, is briefly described. The maximum shear
modulus (G
0
) can be determined together with the N value with this hybrid test. Seismic (Cross-hole, Down-hole and SCPT) and SPT
test data for a Brazilian tropical sandy soil are presented and discussed emphasizing the advantage of using the interrelationship
between the small strain stiffness (G
0
) and an ultimate strength (N value) to identify different soil behavior. A seismic SPT test was
carried out in this research site and the G
0
/N ratio is discussed as an interesting index to help characterize tropical soils, similar to
what has been suggested for the Go/q
c
ratio determined in a single test.
RSUM : Le SPT sismique, qui associe le up-hole au SPT est brivement dcrit. Le module de cisaillement maximale (G
0
) peut tre
dtermin avec la valeur N de ce test hybride. Des donnes sismiques (Cross-hole, Down-hole and SCPT) et SPT pour un sol sableux
tropical du Brsil sont prsentes et discutes soulignant l'avantage d'utiliser la corrlation entre (G
0
) et une rsistance la rupture
(valeur N) afin d'identifier le comportement de diffrents sols. Un essai SPT sismique a t ralis dans le site exprimental et la
relation G
0
/N est discute comme un indice intressant pour aider caractriser les sols tropicaux, de la mme faon que ce qui a t
propos pour le rapport G
0
/q
c
mesur dans un essai unique.
KEYWORDS: In situ testing, SPT, seismic, up-hole, tropical soil, G
0
/N ratio.
1 INTRODUCTION
Site characterization can be defined as the process of identifying
the geometry of relatively homogeneous zones and developing
index, strength and stiffness properties for the soils within these
zones. Some in situ testing can be used as an alternative to the
traditional approach of drilling, sampling and laboratory tests.
Combining stratigraphic logging with a specific measurement in
a in situ test is a modern approach for site characterization.
Some authors have shown that it is possible to incorporate
the measurement of shear wave velocities using the SPT blow
by the up-hole technique. This hybrid test is known as the
seismic SPT (S-SPT), which combines stratigraphic logging,
estimative of geotechnical parameters and determining small
strain stiffness (Go) in one single test similarly to the SCPT.
This paper briefly describes a system to carry out the S-SPT
test and the approach to interpret the seismic data. It also
discusses the applicability of the interrelationship between (Go)
and N value to identify unusual soil behavior based on the tests
carried out in a research sites located in the city of Bauru, inland
of So Paulo State, Brazil emphasizing the advantage of using
the S-SPT test for this approach.
2 BACKGROUND
2.1 Tropical Soils
Tropical soils are formed predominantly by chemical alteration
of the rock, and they are considered a non-textbook type
geomaterial because their peculiar behaviors that cannot be
explained by the principles of classical soil mechanics.
The term tropical soil includes both lateritic and saprolitic
soils. Saprolitic soils are necessarily residual and retain the
macro fabric of the parent rock. Lateritic soils can be either
residual or transported and are distinguished by the occurrence
of the laterization process, which is an enriching of a soil with
iron and aluminum and their associated oxides, bonding a
highly porous structure. Saprolitic soil has structural or
chemical bonding retained from the parent rock. The
contribution of this cementation to the soil stiffness depends on
the strain level the soil will experience. Differences between the
mechanical behaviors of the mature (lateritic) and young
(saprolitic) soils have been reported for both natural and
compacted conditions.
2.2 Go/q
c
Ratio
The pore pressure measurements cannot always be considered
useful to allow an adequate classification of tropical soil based
on CPTU data. The small strain stiffness (Go) and cone tip
resistance (qc) ratio has been suggested as an additional
information for classifying different soil types, especially to
identify soils with unusual compressibility. Schnaid et al (2004)
suggested that the ratio Go/qc provides a measure of the ratio of
the elastic stiffness to ultimate strength and may therefore be
expected to increase with sand age and cementation, primarily
because the effect of these on Go are stronger than on qc. They
proposed a chart and boundaries by correlating Go/qc versus
normalized type resistance (qc
1
). This chart can be used to
evaluate the possible effects of stress history, degree of
cementation and ageing for a given profile. Three lines divide
upper and lower bounds for cemented and uncemented sands.
Giacheti & De Mio (2008) presented SCPT test results from
three tropical research sites in the State of So Paulo, Brazil and
plotted all the data in the Schnaid et al (2004) chart as shown in
Figure 1. The authors pointed out that the SCPT test allows
calculating Go/qc ratio simplifying interpretation and reducing
site variability. The SCPT data interpretation indicated that the
bonded structure of tropical soils gives Go/qc ratios that are
systematically higher than those measured in cohesionless soils.
Rocha B.P.
536
th
The results are in agreement with the propositions of Schnaid et
al. (2004). They also observed that lateritic soils tends achieve a
higher Go/qc ratio than the saprolitic soils.
Figure 1. Relationship between Go and qc (Giacheti & De Mio, 2008).
2.3 Go/N Ratio
Schnaid et al. (2004) suggested that the N values from SPT test
can also be combined with Go, using the Go/N ratio, to help
assessing the presence of bonding structure. This approach is
presented in Figure 2. Viana da Fonseca & Coutinho (2008)
included data from experimental sites from Portugal in this
figure. These authors pointed out that the bonded structure have
a marked effect on the behavior of residual soils, with a Go/N
values considerably higher than those observed in cohesionless
materials. Lines are also shown in Figure 3 to define the upper
and lower bounds for cemented and uncemented sands.
Similarly to what has been presented by Giacheti & De Mio
(2008) for tropical soils based on SCPT data, the
interrelationship between small strain stiffness (Go) and N value
could be used to identify different soil behavior using the
seismic SPT similarly to the SCPT.
Figure 2. Relationship between Go and N
60
(Schnaid et al 2004,
completed by Viana da Fonseca & Coutinho, 2008).
3 STUDIED SITE
3.1 The site
The Unesp experimental research site is located inland of the
State of So Paulo, Brazil, in the city of Bauru (Figure 3).
Several site characterization campaigns including SPT, DMT,
PMT, CPT, SCPT, cross-hole and down-hole tests were
previously carried out at the site. A sample pit was excavated to
retrieve disturbed and undisturbed soil blocks to be tested in the
laboratory to characterize the soils and to determine
geotechnical properties.
The subsoil is a sandy soil where the top 13 m has lateritic
soil behavior (LA) overlaying a soil of non-lateritic behavior
(NG) derived from weathering of Sandstone rock. The MCT
Classification System proposed by Nogami & Villibor (1981)
for tropical soils was used to define and classify the soils with
regards to its lateritic behavior.
Figure 3. Bauru city, where the studied site is located.
3.2 SPT and seismic testing data
The typical soil profile for the studied site was defined based on
the SPT tests and it is presented in Figure 4.a, together with N
values correct by 60% efficiency (N
60
) for all SPT tests (Figure
4.b). The shear wave velocities (Vs) were determined with
cross-hole, down-hole and SCPT tests (Figure 4.c). Total mass
densities were obtained from undisturbed soil samples collected
in a sample pit excavated at the site. They were used to
calculate Go values based on Elastic Theory and the data are
presented in Figure 4.d.
An average Go/N
60
ratio for every one meter depth was
calculated, so the Go
avr
/N
60avr
values versus depth are presented
in Figure 4.e. The criteria to calculate this ratio was averaging
Go and N
60
from all the tests and after that calculating the
average ratio with the closest depth from Go and N
60
.
Site variability can be assessed based on N
60
and Vs values
and these data indicate that the site is quite variable. Giacheti at
al (2003) discussed variability for this site based on several CPT
tests. They also concluded that the site is variable and test data
can be affected by suction and cementation. The authors pointed
out that the SCPT1 shows the presence of a region with low qc
and high R
f
between around 10 and 16 m depth. These data are
quite different from those recorded with the SCPT2 test, so Vs
values were not considered to calculate Go for this portion of
the soil profile for the SCPT1 test. This variation is probably
related to the morphogenetic and pedogenetic processes and
probably reflects different degrees of cementation in the profile.
N
60
0 10 20 30 40
D
e
p
t
h

(
m
)
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
LA'
NA'
SPT profile
1
2
3
1-Red clayey fine sand
2-Red clayey silty fine sand
(b)
(a)
SM - SC
G
o
(MPa)
0 100 200 300 400 0 10 20 30 40 50 60
CH1
SCPT1
SCPT2
DH
Go
avr
/ N
60avr
(d) (e)
V
s
(m/s)
0 200 400 600
(c)
CH1
SCPT1
SCPT2
DH
Figure 4. SPT and seismic testing data and Go/N
60
for the studied site.
3.3 The Go/N ratio
It can be observed in Figure 4.e that the average Go/N
60
tends
do decrease with depth, with an average value equal to 35
537
between 1 and 6 m depth, 23 between 6 to 13 m depth and 10
below 13 m depth. These results indicate that Go/N
60
ratio is
higher in the lateritic soil layer (1 to 13 m depth) and tends to
decrease as the residual soil is less developed.
The average Go and N
60
values for the study site were
plotted in the Go/N
60
versus (N
1
)
60
chart (Figure 5). Almost all
the data points are in the upper bound for cemented sands which
indicates that the bonded structure of tropical sandy soils
produces Go/N
60
that are systematically higher than those
measured in cohesionless soils. It is also interesting to note in
Figures 4.e that the lateritic soils (G
0
/N
60
=35 to the upper
portion and 23 to the lower portion) present a higher
cementation than the saprolitic soils (average G
0
/N
60
=10).
These results are similar to what had been presented by Giacheti
& De Mio (2008) based on Go/qc from SCPT test (Figure 1)
and indicate the use of the ratio between the small strain
stiffness (Go) and an ultimate strength (N or qc) to identify
unusual soil behavior and degree of evolution of residual soils.
Normalized (N
1
)
60
1 10 100
R
a
t
i
o

(
G
o
/
p
a
)
/
N
6
0
10
100
1000
Unaged
uncemented sands
Upper bound
(cemented geomaterials)
Lower bound
Lateritic Soils
Saprolitic Soils
60
4 THE SEISMIC SPT TEST
4.1 Principle
It is possible to incorporate the shear wave velocity (Vs)
measurements during the SPT test applying the up-hole
technique. This approach has been used in the past and it is
recently presented in detail by Bang & Kim (2007). This hybrid
test allows measuring the SPT N value together with Vs (so Go)
at the same time and in the same borehole. For each sampler
depth (usually at every meter) a seismic wave is generated and
it can be recorded on the ground surface. A schematic
representation of the S-SPT test is show in Figure 6.
Case with
geophones
H1
H2
H3
Hi
DAQ
System
Manual SPT Equipment
Trigger &
Anvil
L1
L2
L3
Li
1
2
3
i
Figure 6. Schematic representation of an S-SPT test and a seismic
refracted path (adapted from Bang & Kim, 2007 by Pedrini et al, 2012).
The test equipment is the same currently used for the SPT
test. An arrangement of transducers (usually geophones) placed
in appropriate boxes on the ground surface, a triggering system
and the seismic source, which is the SPT sampler itself, are
added for the seismic SPT test.
4.2 Vs from the S-SPT test
Determining Vs from the S-SPT test data is not straightforward.
Bang & Kim (2007) described two methods: DTR (delay time
between serial receivers) and DTS (delay time between serial
sources). Pedrini (2012) suggested using the DTS method. In
this method, the time interval of the S waves arrival for each
sample depth in which the test was carried out is determined
identifying the exact moment of the first arrival time plotting
the wave receptions generated at different depths. Figure 7
shows a typical wave recordings profile as well as the point of
the first S wave arrival. Another important aspect is the
geometry. Bang & Kim (2007) recommend that Snells Law
(the refraction and reflection during the propagation of waves in
stratified layers of different densities) should be taken into
account when determining the refracted wave path.
Figure 7. Profile of seismic wave and the identification of the common
arrival point of the S waves (Pedrini et al, 2012).
The refracted ray pathway calculated based on Snells Law
depends on various wave velocities and it can be determined by
considering two conditions: the Snells law and a geometrical
criteria. The following assumptions must be done: 1) each
sample layer is equal to the depth where the SPT test was
carried; 2) each layer is homogeneous and the propagated wave
velocity is assumed constant in each layer as show in Figure 6.
An iterative method must be used to solve the equation system
and determine the length (L) that the wave propagates in each
soil layer. Details can be found in Bang & Kim (2007).
4.3 The S-SPT equipment
The system for carrying out S-SPT tests and the method of
analysis were implemented and described by Pedrini (2012).
The main characteristics of this system are presented by Pedrini
et al (2012) and will be summarized herein.
Bang & Kim (2007) used the drop of the SPT weight as the
source to generate waves while Pedrini (2012) used a 2 kg
sledgehammer. The triggering device was digital, with one
terminal (positive or negative) fitted into the anvil head and the
other attached to the sledgehammer.
Two geophones were installed inside of six boxes placed on
the ground, one vertical and other horizontal oriented in a radial
pattern. A National Instruments, model NI-USB-6353, data
acquisition system was used. It has 16 bits resolution, 32 single
ended channels and 16 differential channels, a digital and
analogue trigger and a receiving rate of 1.25 ms/s. Software in
the Labview and Matlab platforms were developed by Pedrini
(2012) to trigger, capture the waves, signal processing,
represent the traces, analyzing the recorded data and calculating
the velocities.
538
th
4.4 The S-SPT test procedure
Normalized (N
1
)
60
1 10 100
R
a
t
i
o

(
G
o
/
p
a
)
/
N
6
0
10
100
1000
Unaged
uncemented sands
Upper bound
Lower bound
Lateritic Soils
Saprolitic Soils
An S-SPT test was carried out using this system in the studied
site. Seismic data were recorded from waves generated every
one meter depth up to 21 m, right after the N SPT measurement
using the equipment described in the previous sub-item. A six
box arrangement was placed on the ground surface after
removing the top soil to enable better coupling. The distance
between each box (which contains a pair of geophone) was 1.5
m and they were all placed between 4.5 m to 12.0 m away from
the test borehole.
4.5 The S-SPT testing data
The N
60
values measured during the S-SPT test carried out at
the studied site are presented in Figures 9.b. This hybrid test
allowed determined Vs simultaneously to N every 1 m interval
(Figure 8.c) for calculating Go (Figure 8.d). The Go/N
60
values
versus depth are also presented in Figure 8.e for the studied site
with no averaging.
60
for the seismic SPT test.
0 10 20 30 40
D
e
p
t
h

(
m
)
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
0 200 400 600
Vs (m/s) Go (MPa) Go / N
60
(b)
(c)
(d)
N
60
(a)
0 100 200 300 400
(c)
0 10 20 30 40 50 60
LA'
NA'
1
2
3
2-Red clayey silty fine sand
SM - SC
Vs - 6,0m
Vs - 8,0m
Vs - 10,0m
Vs - 12,0m
Vs avr
SPT profile
(d) (e)
5 CONCLUSIONS
It was observed that the average ratio Go/N from several SPT
and seismic tests carried out in the studied site was higher in the
lateritic soil than in the saprolitic soil, particularly in the top
layer. The seismic SPT test was used to derive Go/N values in
the same site. Similar results were achieved with this single test,
which allows determining both parameters simultaneous,
reducing the effects of site variability. Relating an elastic
stiffness (Go) to an ultimate strength (N value) is an interesting
approach to help identify tropicals soils since the low strain
modulus from seismic tests reflects the weakly cemented
structure of lateritic soils while the SPT sampler penetration
brakes down all cementation. The preliminary results from the
seismic SPT test indicate that this hybrid test opens up new
possibilities for geotechnical site characterization of tropical
soils, based on the relationship Go/N, which is similar to the
Go/qc ratio in the SCPT test.
6 ACKNOWLEDGEMENTS
The authors gratefully acknowledge FAPESP (State of So
Paulo Research Foundation) and CNPq (National Council for
Scientific and Technological Development). Figure 8. S-SPT testing data and Go/N
60
4.6 The Go/N ratio
The Go/N
60
profile (Figure 8.e) obtained from the S-SPT test
data are similar to what was found when averaging all SPT and
seismic test data (Figure 4.e) for the top lateritic layer (1 to 6 m
depth) with a lower average Go/N
60
equal to 27, a bit lower than
what was previously found, 35. In the lower part of the lateritic
layer (6 to 13 m depth) it was found an average Go/N
60
equal to
14, also lower than what was previously found (23) and the
same average value for the saprolitic layer.
7 REFERENCES
Bang, E. S. & Kim, D.S. 2007. Evaluation of shear wave velocity
prole using SPT based up-hole method, Soil Dynamics and
Earthquake Engineering 27, p. 741-758.
Giacheti, H. L. & De Mio, G. 2008. Seismic cone penetration tests on
tropical soils and the ratio Go/qc. 3
rd
Geotechnical and Geophysical
Site Characterization Conference, ISC3, Taiwain, v.1. p. 1289-
1295.
These data were also plotted in the Go/N
60
versus (N
1
)
60
chart as shown in Figure 9. All the data points are in the upper
bound for cemented sands. In this case the difference between
lateritic and saprolitic soils is not so clear, just the upper portion
of the lateritic layer reflects a higher degree of cementation. Soil
variability in this particular site probably related to the
morphogenetic and pedogenetic processes, already pointed out
by Giacheti et al (2003) and Giacheti & De Mio (2008) could
explain the observed differences.
Giacheti, H. L.; Peixoto, A. S. P. & Marques, M. E. M. 2003. Cone
Penetration Testing on Brazilian Tropical Soils. XII Panamerican
Conference on Soil Mechanics and Geotechnical Engineering,
Cambridge/USA, v.1. p. 397-402.
Nogami, J. S. & Villibor, D. F. 1981. Uma nova classificao de solos
para finalidades rodovirias, Simpsio Brasileiro de Solos Tropicais
em Engenharia, Brasil, V. 1, p. 30-41.
Pedrini, R. A. A. 2012. Desenvolvimento de sistema para realizao de
ssmica up-hole em conjunto com sondagem SPT. M.Sc. thesis,
FEB/Unesp. Bauru/SP, Brazil.
Pedrini, R. A. A. & Giacheti, H. L. 2012. The seismic SPT to determine
the maximum shear modulus, 4
th
Geotechnical and Geophysical
Site Characterization Conference, ISC4, Brazil, CD-Rom.
Schnaid, F.; Lehane, B. & Fahey, M. 2004. In situ test characterization
of unusual geomaterials. 2
nd
Geotechnical and Geophysical Site
Characterization Conference, ISC2, Portugal, v. 1. p. 49-74.
Viana da Fonseca. A. & Coutinho, R. Q. 2008. Characterization of
residual soils, 3
rd
Geotech. and Geoph. Site Characterization Conf.,
ISC3, Taiwain, v. 1. p. 195-249.
539
Compressibility Parameters of Cohesive Soils From Piezocone
Paramtres de compressibilit de sols cohsifs au pizocone
Hamza M.
Faculty of Engineering, Suez Canal University & Chairman of Hamza Associates, Egypt
Shahien M.
Faculty of Engineering, Tanta University, Egypt
ABSTRACT: Drained compressibilty parameters for cohesive soils can be determined by carrying out one dimensional consolidation
tests on undisturbed samples. The compressibility parameters include the compression and recompression indices,
overconsolidation ratio and coefficient of consolidation. Some of these parameters or in other forms have been already correlated in
the literature to results of piezocone. The aim of this paper is to provide additional data on drained compressibility parameters,
focusing on constrained modulus and overconsolidation ratio, for cohesive soils from geotechnical investigations in seven major sites
of river Nile Delta deposits in Egypt where piezocone CPTU data are also available. The results of consolidation tests are used to
evaluate and modify the available correlations(s) with CPTU data. It is believed that the data and analysis in this paper shall be a
valuable contribution to the literature by providing a better ground for improving the current state of the art of estimating the
compressibility parameters from the CPTU data.
RSUM : Les paramtres de compressibilit draine pour les sols cohrents peuvent tre dtermins en excutant un test de
consolidation unidimensionelle sur les chantillons intacts . Ces paramtres incluent les indices de compression et de
recompression, le taux de surconsolidation et le coefficient de consolidation. Certains de ces paramtres ont dj t corrls dans la
bibliographie aux rsultats du pizocone. L'objectif de cet article est de fournir des donnes supplmentaires sur les paramtres de
compressibilit draine en se concentrant sur le module contraint et sur le taux de surconsolidation pour des sols cohrents tudis
dans sept sites majeurs des dpts du Delta de Nil en Egypte, o des donnes de CPTU sont aussi disponibles. Les rsultats dessais
de consolidation sont utiliss pour valuer et modifier les corrlations disponibles avec les donnes de CPTU. On estime que les
donnes et l'analyse prsentes ici seront une contribution valable la bibliographie en fournissant de meilleurs fondements pour
amliorer ltat de lart actuel concernant l'estimation des paramtres de compressibilit partir de donnes de CPTU.
KEYWORDS: constrained modulus, overconsolidation ratio, sample quality designation, piezocone, clay

1 INTRODUCTION
Drained compressibilty parameters for cohesive soils are useful
in; a) carrying out long term settlement analysis, b) providing
key parameters for analysis and design of ground improvement,
and c) profiling undrained shear strength parameters with the
aid of other insitu field investigation equipments such as field
vane and piezocone.
Drained compressibility parameters for cohesive soils can be
determined from End of Primary (EOP) void ratio versus
effective stress relationship that results from carrying out
incremental load one dimensional consolidation tests on
undisturbed samples. The drained compressibility parameters
include the compression and recompression indices,
overconsolidation ratio and coefficient of consolidation. These
parameters can be influenced with variable degrees by quality
of samples used in the tests. (Jamiolkowski et al., 1985 and
Terzaghi et al., 1996). Empirical correlations to estimate these
parameters or equivalent in other forms, from insitu tests such
as piezocone are available in the literature (e.g. Jamiolkowski et
al, 1985, Lunne et al., 1997 and Mayne, 2009). Availability of
such correlations provides a great aid for geotechnical engineers
to estimate such parameters in continuous profiles for a site in
relatively short period of time and perform fewer consolidations
tests for confirmation. However, estimating drained parameters
from undrained piezocone test results could be complicated and
sometimes may have various degrees of uncertainties (Lunne
etl. 1997). Therefore, there is a need for continuous feed of data
from local experiences to confirm, validate, and even modify
the existing correlations.
The aim of this paper is to provide additional data on both
constrained modulus and overconsolidation ratio as determined
from oedometer consolidation tests on undisturbed samples of
cohesive soils and CPTU data from seven sites from the Nile
Ddelta deposits. The authors believe that the addition of the data
presented in this paper to the literature provides a better ground
for improving the current state of the art of estimating drained
compressibility parameters from the CPTU data. With such
belief, the data are used to evaluate and modify the available
correlations.

2 INVESTIGATED SITES
Comprehensive geotechnical investigation campaigns were
carried out in seven sites of major projects along the north coast
and within the Delta of the Nile River of Egypt. The seven sites
provide full coverage of the Nile Delta deposits starting from
Idku at west of the Nile Delta, to Metobus within the Nile Delta,
to Damietta, to El-Gamil and Port Said further east of the Delta.
Three of these sites were reported in Hight et al. (2000), Hamza
et al. (2002), (2003) and (2005). The seven sites were used by
Hamza and Shahien (2009) to investigate the correlations of
estimating the efective stress friction angle from piezocone data.
The stratifications of the sites are shown in Fig. (1).
The statification of the sites consists of silty sand top layer
over very soft to medium stiff clay layer over sand over stiff to
hard clay. The thickness of the soft clay layer tends to thicken
as moving from west to east of the Delta (Hamza et al., 2005).
540
th
Natural Water Content, %
10 100
C
o
m
p
r
e
s
s
i
o
n

I
n
d
e
x
,

C
c
0.1
1 Idku
Metobus
Dammietta 3
Dammietta 4
Port Said 2
El-Gamil
Dammietta 2
C
c
=
0
.1
w
C
c
=
0
.0
0
5
w
C
c
=
0
.0
0
3
w
C
c
=4x10
-6
w
2.85
Volumetric Strain at '
vo
,
vo
, %
0 1 2 3 4 5 6 7 8 9 10
O
v
e
r
c
o
n
s
o
l
i
d
a
t
i
o
n

R
a
t
i
o
,

O
C
R
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Idku
Metobus
Dammietta 3
Dammietta 4
Port Said 2
El-Gamil
Dammietta 2
SQD Scale
A B C D E
D
e
p
t
h
,

m
0
10
20
30
40
50
60
Silty Sand
Soft to
Firm Clay
Sand
with Silt
Occasionaly
interbeded by
Hard Clay
ELGamil
Silty
Sand
Alternating
Soft to
Med. Stiff
Clay,
Organic Silt
& Silty Sand
Sand
with
Silt
Idku
Limemud
Silty
Sand
Soft to
Med. Stiff to
Very Stiff
Clay,
Interbeded
with Very
Stiff Peat
Sand
with
Silt
Metobus
Soft to
Firm Clay
Port Said 2
Soft Clay
Silt/
Silty Sand/
Clay
Stiff Clay
Silty Sand
Sand - Occas.
Silt and Clay
Silt
Firm to
Stiff Clay
Silty Sand
Stiff Clay
D
e
p
t
h
,

m
0
10
20
30
40
50
60
Silty
Sand
Soft
to
Firm
Clay
Silty
Sand
Damietta 2
Very
Soft to
Medium
Stiff
Clay
Silty
Sand
Stiff
to
Hard
Clay
Silty
Clayey
Sand
Damietta 4
Hard
Clay
Very
Soft to
Medium
Stiff
Clay
Silty
Clayey
Sand
Stiff
to
Hard
Clay
Silty
Clayey
Sand
Damietta 3
Hard
Clay
Silty
Clayey
Sand

Figure 1. Stratigraphy of the soil formations in the seven sites.
3 COMPRESSIBILITY PARAMETERS FROM
OEDOMETER TESTS
3.1. General
The results of total 125 consolidation tests were used in this
study. The tests were carried out on clay undisturbed samples
that were collected by means of stainless steel thin wall Shelby
tubes with cutting edge sharpened to approximately 5
o
.
Incremental loading procedure was utilized with a load
increment ratio of 2. End Of Primary (EOP) consolidation was
determined for each load increment using the Taylor method.
EOP void ratio versus logarithm of effective vertical pressure
(e-log
v
) curves were plotted for each test.
3.2. Overconsolidation Ratio
The overconsolidation ratio, OCR, is defined as the ratio
between the preconsolidation or yield pressure,
p
, to in situ
effective overburden pressure,
vo
. The
p
is the pressure that
distinguishes between low compressibility in the recompression
range and the high compressibility in the compression range.
There are several mechanisms for a deposit to demonstrate a
p

(Jamiolkowski et al., 1985 and Mayne et al., 2009). Those
mechanisms include; decrease in vertical effective stress,
freeze-thaw cycles, repeated wetting-drying, tidal cycles,
earthquake loading, desiccation, aging, cementation or
geotechnical bonding. The decrease in effective stress could be
caused by; mechanical removal of overburden, overburden
erosion, rise in sea level, increased groundwater elevations,
glaciation, and mass wasting. The conventional and most
common Casagrande method is used to determine
p
from the
EOP e-log
v
curves from the Oedometer tests carried out.
Sample quality was evaluated on the basis of the magnitude
of the volumetric strains,
vo
, during reconsolidation to
vo
in
oedometer tests as suggested by Andresen and Kolstad (1979).
The Sample Quality Designation (SQD) scale using
vo

suggested by Andresen and Kolstad (1979) and modified by
Terzaghi et al. (1996) is used in this paper. Figure (2) shows the
OCR values in this study versus
vo
. Shown also on the plot, is
the above mentioned SQD scale. The scale suggests that the
majority of samples have quality B to C. Such sample qualities
correspond to verbal scale of very good to good samples.
The OCR values for the clay are in the range of 1 to 2. It
should be noted that OCR values might be influenced by sample
disturbance. As sample disturbance increases (i.e.
vo
increases),
the OCR value decreases due to the de-structuring of the samples
during sampling. One possible major source for sample
disturbance in Nile Delta deposits is the natural gas exsolution in
the pore water (Hight et al., 2000). The OCR values, for the very
few tests, that are less than 1 were corrected to 1 for use in
evaluations and correlations developed in this study.

Fig. 2 Overconsolidation ratio (OCR) versus
vo
as a measure of SQD
3.3. Compression Indices and Moduli
The compression, C
c
, and re-compression, C
r
, indices were
calculated for each test as the slopes of the e-log
v
curve in
the normally consolidated and the re-compression ranges,
respectively. The recompression index, C
r
, was calculated as the
average slope of the unloading-reloading cycle of e-log
v

curve between vertical effective stress value of twice of the
preconsolidation pressure,
p
, and effective overburden
pressure,
vo
or the average slope of the unloading curve from
consolidation pressure of 3200 kPa.
Compression index values in this study are plotted in Figure (3)
versus natural water content, the Terzaghi et al. (1996) plot for
filling and reference. The water content is a major variable as it
reflects how much water held in the deposit to be squeezed out
upon the increase in effective stress. As expected, the data show a
band that compares relatively well with data from all over the
world as collected originally by Terzaghi et al. (1996). The overall
average of ratio of re-compression to compression indices C
r
/C
c
is
calculated to be about 0.1.

Figure.3 Data of this study on the compression index versus natural water content
Terzaghi et al (1996) relationship
Constrained modulus is another form of compressibility
parameter instead of the recompression or compression indices.
The following expression is used to estimate the tangent
constrained modulus:
M=
v
/ = 2.3(1+e)
v
/C
c
(1)
The general definition of constrained modulus in Equ. (1) is
used in the literature (e.g. Kulhawy and Mayne 1990). There are
several definitions for the constrained modulus depending on
which
v
and which index, C
c
or C
r
, used in Equ. (1). It is
expected that the modulus in the compression range is different
541
than that in the re-compression range. Even in the compression
range, the constrained modulus is dependent on
v
level (Janbu,
1963). Figure (4) introduces the several definitions of the
constrained modulus using consolidation test data from the Idku
site as an example. The Janbu (1963) approach can be used to
define three constrained moduli as defined in Figure (4) and Equs.
(2) to (4); M
i
in the recompression range, M
np
or M
n@p
at
p
and
M
n
in the compression range that is dependent on level of
v
:
M
i
= 2.3(1+e)
p
/C
r
(2)
M
np
= M
n@p
= 2.3(1+e)
p
/C
c
(3)
M
n
= 2.3(1+e)
v
/C
c
(4)
There are investigators (e.g. Sanglerat, 1972, and Abdelrahman
et al., 2005) that are using M
o
at
vo
as in Equ (5)(Fig. 4):
M
o
= 2.3(1+e)
vo
/C
c
(5)
The geotechnical engineer should be cautious as what modulus
is reported or estimated and how it is used in settlement analysis,
because in a lot of literature the reference is given to M without
specifying which modulus is meant such as in Equ. (1). M
o

modulus can be used to estimate both M
i
and M
n
using Equs. (6)
and (7) to be used for settlement analysis in the recompression and
compression ranges, respectively.
M
i
= M
o
OCR(C
c
/C
r
) (6)
M
n
= M
o
(
v
/ p
a
) (7)
where
v
is the average pressure between
p
and the final
pressure due to surface load causing the settlement.
Effective Vertical Stress, kPa
0 100 200 300 400 500
C
o
n
s
t
r
a
i
n
e
d

M
o
d
u
l
u
s
,

k
P
a
0
10000
20000
30000
40000
50000
1 10 100 1000 10000
V
o
id
R
a
tio
0.7
0.8
0.9
1.0
1.1
1.2
1.3
1.4
1.5
0 500 1000 1500 2000 2500 3000
C
o
n
stra
in
e
d
M
o
d
u
lu
s, kP
a
0
10000
20000
30000
40000
50000
M
i
M
o
'
p
M
n-'p
M
n
Idku Site
Figure 4 Definition of tangent constrained modulus concept
Friction Ratio, F
r
= [f
s
/(q
t
-
vo
)] 100, %
1 10
k

=

'
p
/
(
q
t
-
v
o
)
0.0
0.2
0.4
0.6
0.8
1.0
Idku
Metobus
Dammietta 3
Dammietta 4
Port Said 2
El-Gamil
Dammietta 2
Average k = 0.32
(q
t
-
5 6 7 8 9 4 3 2
vo
)/'
vo
1
5
20
10
Robertson (2012)
R
a
n
g
e

F
r
o
m
L
i
t
e
r
a
t
u
r
e
4 PEIZOCONE PENETRATION TESTS
Piezocone Penetration Tests with pore water pressure
measurements (CPTU) were performed at the sites. A l0 cm
2

Piezocone was used to carry out the testing. Records were
made at 2 cm intervals. At each tested depth, cone resistance
(q
c
), pore water pressures behind cone (u
2
) and side friction (f
s
)
were measured. Typical CPTU records at some of the sites
under study are shown in Hight et al. (2000), Hamza et al.
(2003) and Hamza et al. (2005). The corrected tip resistance, q
t
,
can be calculated as q
t
=q
c
+(1-)u
2
, where is a cone
factor. The net cone resistance, q
n
, can be calculated as q
n
= q
t
-
vo
, where
vo
is the total overburden pressure.
5 PEIZOCONE PENETRATION TESTS
5.1. Stress History or Overconsolidation Ratio
Review of the available correlations between
p
or OCR and
Piezocone results was carried out by Lunne et al. (1997), Mayne
(2001), Ladd and DeGroot (2003), Powell and Lunne (2005),
Pant (2007), Mayne (2009), Becker (2010) and Robertson
(2012). The cone parameters used in the correlations include q
c
,
q
t
, q
t
-
vo
, q
t
-u
2
, u. Some of these parameters were used with or
without normalization by
vo
. According to Campanella and
Robertson (1988), there is no unique relationship between OCR
or
p
and measured penetration induced pore water pressures
and if exists, it is poor because the pore pressures measured is
influenced by the location of the u measurement (i.e. u
1
, u
2
or
u
3
), clay sensitivity, over consolidation mechanism, soil type
and local heterogeneity. The most common and widely used
correlation is (e.g. Lunne et al. 1997):

p
= k (q
t
-
vo
) or OCR =
p
/'
vo
= k(q
t
-
vo
)/'
vo
(8)
It should be noted that empirical constant k in both
expressions in Equ. 8 is the same. Table (1) shows a summary
of k values reported in the literature. According to the table, k is
in the range of 0.14 to 0.5. Mayne (2001) showed that k is
slightly dependent on plasticity index, while Becker (2010)
showed that k is slightly dependent on coefficient of horizontal
pressure at rest. Robertson (2012) suggested an expression that
is dependent on (q
t
-
vo
)/'
vo
and sleeve friction ratio, F
r
. The
empirical constant is calculated for the data in this study and is
plotted versus F
r
in Figure (5). The expression suggested by
Robertson (2012) was also plotted on the same plot. Figure (5)
shows that the Robertson (2012) predicts well the range of k.
However, it seems that k is slightly increasing with F
r
. The
calculated k values are in the range of 0.1 to 0.6 (0.18 to 0.4, if
scatter is ignored) with an average of 0.32, which is consistent
with the existing correlations in the literature.
Table 1. Summary of the parameter k from the literature..
Reference k Comment
Lefebvre & Poulin (1979) 0.25- 0.4 Norway & UK sites
Mayne & Holtz (1988) 0.4 World Data
Larson & Mulabdic (1991) 0.29 Scandinavian Soils
Mayne (1991) 0.33
Cavity Expansion & Critical
State Soil Mechanics Analysis
Leroueil et al. (1995) 0.28 Eastern Canada Clays
Chen & Mayne (1996) 0.305 205 Clay sites
Lunne et al. (1997) 0.2 0.5
Mayne (2001) 0.65(I
p
)
-0.23

Mesri (2004) 0.25 0.32 s
u
/
p
=constant interpretation
Abdelrahman et al. (2005) 0.2 0.5 Port Said Site, Egypt
Pant (2007) 0.14 Louisiana Soils 7 Sites
Becker (2010) 0.3 Beaufort Sea Clays K
o
=1.5
0.24 Beaufort Sea Clays K
o
=2.0
Robertson (2012) * SHANSEP & CSSM
* k = [ [(q
t
-
vo
)/
vo
]
0.2
/ (0.25(10.5+7log F
r
)) ]
1.25
where F
r
= f
s
/(q
t
-
vo
)

Figure (5) Empirical constant k for the sites in this study
Ladd and De Groot (2003) proposed the following
SHANSEP type of expression to estimate OCR:
OCR = k
OCR
[(q
t
-
vo
)/'
vo
]
1.25
(9)
Ladd and De Groot reported a value of 0.192 for k
OCR
based
Boston Blue clay experience. Robertson (2009) suggested
general k
OCR
value of 0.25. Robertson (2012) suggested the
expression in Equ. (10) to estimate k
OCR
based on F
r
:
k
OCR
= (2.625+1.75 log F
r
)
1.25
(10)
The data of Delta clay sites was used to back calculate k
OCR

and was plotted versus Fr in Fig. (6). The Robertson (2012)
expression was also plotted on Fig. (6). Figure (6) shows that
Equ. (10) predict well the range of k
OCR
. However, it seems that
k
OCR
is slightly increasing with F
r
. The average k
OCR
of the data
in this study was about 0.23 that is consistent with data in
literature.
542
th
Fr = [f
s
/(q
t
-
vo
)] 100, %
1 10
k
O
C
R

=

O
C
R
/
[
(
q
t
-
v
o
)
/
'
v
o
]
1
.
2
5
0.0
0.2
0.4
0.6
0.8
1.0
Idku
Metobus
Dammietta 3
Dammietta 4
Port Said 2
El-Gamil
Dammietta 2
Average k
OCR
= 0.23
2 3 4 5 6 7 8 9
Robertson (2012)
Equ. (10)
Ladd & DeGroot (2003)
Robertson (2009)

Figure (6) Empirical constant k
OCR
for the sites in the study
(q
t
-
vo
)/p
a
0 2 4 6 8 10 12 14
o

=

M
o
/
(
q
t
-
v
o
)
0
5
10
15
20
25
30
Idku
Metobus
Dammietta 3
Dammietta 4
Port Said 2
El-Gamil
Dammietta 2
Average o = 3.5
R
a
n
g
e

f
r
o
m

l
i
t
e
r
a
t
u
r
e
5.2. Constrained Modulus
Review of the available correlations between M and cone
results for cohesive soil was carried out by Lunne et al. (1997),
Mayne (2001), Pant (2007), and Robertson (2009). Attempts to
correlate M of cohesive soils to cone results have started since
mid sixties of the last century (Sanglerat, 1972). The following
expression shows the general form of the empirical correlation:
M
Subscript
=
Subscript
[q
Parameter
] (11)
The subscript in Equ (11) could be nothing, i, np, n, or o as
in Equs (1 to 5). The empirical constant as well as the cone
parameter, q
Parameter
, used in Equ (11) as reported in literature is
summarized in Table (2). According to the table,
o
is in the
range of 1 to 14. Sanglerat (1972) showed that
o
is inversely
dependent on q
c
. Robertson (2009) suggested that
o
is directly
related to (q
t
-
vo
)/
vo
with an upper limit of 14. The empirical
constant
o
is calculated for the data in this study and is plotted
versus (q
t
-
vo
)/p
a
in Figure (7), where p
a
is a reference pressure
of 100 kPa.. Ignoring some scatter, the calculated
o
values are
in the range of 1 to 8 with an average of 3.5, which is consistent
with the existing correlations in the literature. Sources of scatter
in Figure (7) include but not limited to; sample disturbance with
its influence on the measured compressibility and natural
variation between the location of borehole from which the
samples were extracted and that of the CPTU testing.
Table (2) Summary of components of empirical Equ. (11) in literature
Reference
Subscript
Range q
Parameter
Comment
Bachelier and Parez (1965)
o
2.3-7.7 q
c
Flanders Clay
Sanglerat (1972)
o
1-8 * q
c
France & Spain Clays
Jones & Rust (1995)
o
2.2-3.3 q
c
South African Clays
Pants (2007)
np
3.1 q
t
Louisiana Clay

np
3.27 q
t
-
vo
Louisiana Clay
Kulhawy & Mayne (1990)
8.25 q
t
-
vo

Senneset et al. (1989)
i
5-15 q
t
-
vo

Glava Clay

np
8 q
t
-
vo
Glava Clay
Abdelrahman et al. (2005)
o
1.25 q
t
-
vo
Port Said Clay
Mayne (2009)
5 q
t
-
vo

Vanilla Clays
Robertson (2009)
o
** q
t
-
vo

* Dependent on type of soil and on q
c
values
** For Clays (Ic > 2.2)
o
= (q
t
-
vo
)/
vo

o
14

Figure (7) Empirical constant
o
for the sites in the study
6 SUMMARY AND CONCLUSIONS
1) The results of geotechnical investigations in seven sites in
the Nile Delta clays were used in this paper.
2) The compressibility parameters; OCR, C
c
and C
r
, and M
o
,
were calculated from EOP e-log
v
curves of total 125
consolidation tests carried out on undisturbed samples. The
SQD of the majority of the samples was B to C.
3) The compressibility parameters of each test were paired
with results from neighboring or adjacent piezocone test that
were recorded at the same depths of the samples. Such
pairing allowed for comprehensive review of the existing
empirical correlations to predict compressibility parameters
from in-situ piezocone results.
4) The OCR of the Nile Delta clays can be best predicted
using Equs. (8) and (9) using average k of 0.32 and average
k
OCR
of 0.23. Figs (5) & (6) suggest that k and k
OCR
have the
general tendency to slightly increase with friction ratio, F
r
.
5) The M
o
can be best predicted using Equ. (11) with average
value of
o
of 3.5. Settlement analysis can then be carried out
using M
i
and M
n
that can be calculated using Equs (6) and
(7).
7 REFERENCES
Abdelrahman M., Ezzeldine O.and Salem M. 2005. The Use of Piezocone in
Characterization of Cohesive Soil West of Port Said Egypt, Proc. of 5
th
Int.
Geot. Eng. Conf., Cairo University Egypt, pp. 201-219.
Bachelier M. and Parez L.1965. Cont ribution a letude de la compress ibilite des sols a
laide du penetrometer a cone, Proc. 6
th
Int. Conf. Soil Mech. Found. Eng.,
Montreal, 2, 3-10.
Becker, D. E. 2010. Testing in Geotechnical Design, Geot. Eng. Jour. of the
SEAGS & AGSSEA, Vol. 41, No. 1, pp. 1-12.
Campanella, R.G. and Robertson P. K. 1988. Current status of piezocone test,
Proc. of Int. Symp. on Penetration Testing, Orlando, USA, Vol. 1, pp. 1-24.
Chen B. and Mayne P.W. 1996. Statistical relationships between piezocone
measurements & stress history of clays, Can. Geot. Jour. 33(3), pp. 488498.
Jamiolkowski M., Ladd C.C., Germaine J.T., and Lancelotta R. 1985. New
Development in Field and Laboratory Testing of Soils, Proc. of the 11
th
Int.
Conf. Soil Mech. and Found. Eng., San Francisco, 1, pp. 57-153.
Hamza M., Shahien M. and Ibrahim M. 2003. Ground characterization of Soft
Deposits in Western Nile Delta, Proc. 13
th
Reg. African Conf. Soil Mech.
Geot. Eng., Morocco.
Hamza M., Shahien M. and Ibrahim M. 2005. Characterization and undrained
shear strength of Nile delta soft deposits using piezocone, Proc. 16
th
Int.
Conf. on Soil Mech. and Geot. Eng., Osaka, Japan
Hamza M. and Shahien M. 2009. Effective stress shear strength parameters from
piezocone, Proc.17
th
Int. Conf.Soil Mech. and Geot. Eng., Alexandria, Egypt.
Hight D.W. Hamza M.M. and ElSayed A.S. 2000. Engineering characterization of
the Nile Delta clays, Proc. of IS Yokohama 2000.
Janbu N. 1963. Soil compressibility as determined by oedometer and triaxial tests,
Proc. European Conf. Soil Mech. and Found. Eng. Wiesbaden, 1, 1925.
Jones G.A. and Rust E. 1995. Piezocone settlement prediction parameters for
embankments on alluvium, Proc. Int. Symp. Cone Penetration Testing,
Linkping, Sweden, 2, 5018.
Ladd, C. C. and DeGroot D. J. 2003. Recommended Practice for Soft Ground Site
Characterization, Proc. 12th Panamerican Conf. Soil Mech. and Geot. Eng.,
Cambridge, USA
Larson, R., and Mulabdic, M. 1991. Piezocone tests in clays. Swedish
Geotechnical Institute report no. 42, Linkoping, 240p.
Lefebvre, G. and Poulin C. 1979. A new method of sampling in sensitive clay ,
Canadian Geot. Journal, Vol. 16, pp. 226233.
Leroueil S., Demers D., La Rochelle P., Martel G. and Virely D. 1995. Practical
use of the piezocone in Eastern Canada clays , Proc. Int. Symp. on Cone
Penetration Testing, Linkping, Sweden, 2, 515522.
Lunne T., Robertson P.K., and Powell J.J.M. 1997. Cone Penetration Testing in
Geotechnical Engineering Practice. p. 312.
Mayne,

P.W. 1991. Determination of OCR in clays by piezocone tests using cavity
expansion and critical state concepts. Soils and Foundations 31 (1): 65-76.
Mayne P. W. 2001. Stress-strain-strength-flow parameters from enhanced in-situ
tests, Proc. Int. Conf. on In-Situ Measurement of Soil Properties & Case
Histories, Bali, Indonesia, pp. 27-48.
Mayne P. W., Coop M. R., Springman S. M., Huang A. and Zornberg J. G. 2009.
Geomaterial behavior and testing, State of the Art Lecture, Proc. 17th Int.
Conf. on Soil Mech. and Geot. Eng. Alexandria, Egypt, Vol. 4, pp. 1-96.
Mayne P.W., Holtz R.D. 1988. Profiling stress history from piezocone soundings,
Soils and Foundations, Vol. 28(1), pp. 1628.
Mesri G. 2001. Undrained shear strength of soft clays from push cone penetration
test , Geotechnique 51, No. 2, pp. 167168.
Pant R. R. 2007. Evaluation of Consolidation Parameters of Cohesive Soils Using
PCPT Method. MSc Thesis, Louisiana State University. USA
Powell, J. J. M. and Lunne T. 2005. Use Of Cptu Data In Clays/Fine Grained
Soils, Studia Geotechnica et Mechanica, Vol. XXVII, No. 34, pp. 29-66.
Robertson, P. K. 2009. Interpretation of cone penetration tests a unified
approach, Canadian Geotechnical Journal, Vol. 46, pp. 1337-1355.
Robertson P.K. 2012. Interpretation of in-situ tests some insights, Proc. 4th Int.
Conf. Geot. & Geoph. Site Characterization, ISC4, Brazil, 1, pp 1-22.
Sanglerat G. 1972. The penetrometer and soil exploration, Elsevier, 464 pp.
Senneset K., Sandven R. and Janbu N. 1989. The evaluation of soil parameters
from piezocone tests, Transportation Research Record, No. 1235, 2437.
Terzaghi K., Peck R.B. and Mesri G. 1996. Soil Mechanics in Engineering
Practice, 3rd Ed. John Wiley and Sons, Inc., p. 549.
543
Comportement de la structure de sol amlior par inclusions rigides,
supportant une olienne
Behaviour of soil foundation improved by rigid columns, supporting a wind turbine
Haza-Rozier E., Vinceslas G.
Cete Normandie-Centre/DERDI/CER
Le Kouby A.
Universit Paris Est/IFSTTAR
Crochemore O.
Theolia France
RSUM: Dans le cadre du projet national ASIRi (Amlioration des Sols par Inclusions Rigides, 2006-2011), le CER (Centre
dExprimentation et de Recherche du Cete Normandie-Centre) a instrument la structure de fondation dune olienne. Le principe de
fondation est tel que lolienne est fixe sur une semelle rigide, coule sur une couche granulaire de rpartition de charges, dpos sur
le sol en place, amlior par 84 inclusions rigides. Des dplacements verticaux, des pressions totales transmises en tte dinclusion et
sur le sol sont mesurs. Les capteurs sont installs principalement en priphrie de lolienne, dans les zones qui doivent supporter les
variations de contraintes les plus leves. De plus, des extensomtres cordes vibrantes sont installs dans deux inclusions pour
accder aux descentes de charges. Le comportement de la structure est suivi depuis la phase de terrassement, jusquau montage de la
machine, puis durant le fonctionnement de lolienne, jusqu aujourdhui. La plateforme de travail induit un confinement important
des ttes dinclusions. Les pressions se concentrent en priphrie et les dplacements restent faibles. Leffet de la vitesse du vent sur
la distribution des contraintes est tout fait significatif.
ABSTRACT: Within the French National Project ASIRi (Soil reinforcement with rigid inclusions, 2006-2011), CER
(Experimentations and Researches Centre, Rouen) instrumented foundations of a wind turbine. The foundation principle is such that
wind turbine is fixed on a rigid slab, lying on a granular layer, allowing strength distribution on in-situ subgrade improved by 84 rigid
columns. Vertical displacements and total stress sensors at the head of columns and on soil are measured. Sensors are placed on wind
turbine edge essentially, in areas supporting highest stresses variations. Moreover, vibrating wire extensometers are positioned in two
columns in order to measure load distribution. Behaviour of the structure has been monitored since excavation stage, till machine
construction, and then during service working of the turbine wind until now. Working platform induced an important confinement of
columns heads. Pressures are concentred on edge, displacement are still small. The effect of wind speed on load distribution has been
shown to be significant.
KEYWORDS: Instrumentation, rigid columns, foundation, wind turbine
MOTS-CLS: Instrumentation, inclusions rigides, fondation, olienne
1 INTRODUCTION
Lorsqu'une olienne est construite sur un terrain de
caractristiques mcaniques mdiocres, elle est
traditionnellement fonde sur un rseau de pieux fixs sur sa
semelle de fondation. Cependant, un nouveau type de fondation
fait face cette technique plus traditionnelle : la semelle de
l'olienne repose sur une paisseur de sol, constituant une
couche de rpartition de charges, qui surmonte un rseau
d'inclusions rigides (IR). La qualit mcanique du sol en place
et sa capacit portante sont amliores par la prsence des IR.
Lolienne est alors construite avec une fondation superficielle,
sur un terrain de bonne portance (Figure 1).
Figure 1. Semelle de fondation de lolienne.
Dans le cadre du projet national ASIRi (Amlioration des
Sols par Inclusions Rigides, 2006-2011), du Rseau Gnie Civil
et Urbain, pour mieux comprendre comment une structure
damlioration de sol par IR fonctionne (Brianon 2002), le
CER a instrument un tel systme de fondation. Construite par
Theolia, dans un parc Boralex, dans la rgion de Neuchtel en
Bray (76), lolienne 3.6 a t suivie depuis sa construction en
2009 (Haza-Rozier 2011), sa mise en service en aot 2010,
jusqu ce jour (Haza-Rozier & al. 2012).
Cet article dcrit linstrumentation mise en place, lessentiel
des rsultats de mesures et une approche de leffet du vent sur le
comportement de la structure de fondation.
2 DESCRIPTION DE LOUVRAGE
2.1 Profil gotechnique
La coupe gotechnique locale est constitue de 1,6 m de terre
vgtale et de limon, puis de largile silex jusqu 10,60 m de
profondeur (avec des valeurs de module pressiomtrique entre
2,2 et 18,1 MPa). Apparat alors une frange de craie altre sur
1 m (module pressiomtrique entre 2,8 et 12,6 MPa), puis la
craie de plus en plus saine, jusqu 20 m de profondeur (module
pressiomtrique entre 44 et 200 MPa).
Matelas de rpartition
Bton de propret
2.2 Structure de fondation de lolienne 3.6
Lolienne est constitue dun mt de 78 m, fix sur une semelle
rigide de 18m de diamtre et de 2,5 m dpaisseur en son centre
(en bton, fortement ferraille). Cette dernire est pose sur une
couche granulaire de rpartition de charges, de 80 cm
dpaisseur (sol 0/90 mm sur 70 cm et 0/31,5 mm sur 10 cm en
544
th
partie suprieure), dont les caractristiques mcaniques ont t
dtermines la boite de cisaillement de grande dimension
(500 x 500 mm), du CER : C = 63,7 kPa et = 53,3. Cette
couche granulaire est mise en place sur le sol renforc par 84
inclusions rigides, de 8 m de long et 0,36 m de diamtre
(Figure 2).
Figure 2. Forage et coulage des IR en fond de fouille.
3 INSTRUMENTATION
Linstrumentation est compose de capteurs lectriques de
dplacement et de capteurs de pression totale (nots
respectivement T et CPT dans la suite), positionns en tte
dinclusion et dans le sol. Ils sont installs sur deux niveaux : un
premier, sous la couche de rpartition, au niveau des ttes dIR (
Figure 3, 14 CPT et 8 T) et un second en partie haute de cette
couche, sous la semelle de lolienne (Figure 4, 8 CPT et 11 T).
Figure 3. Plan dinstrumentation au niveau des ttes dIR (
er
niveau).
Les capteurs sont installs principalement en priphrie de la
semelle de lolienne, zone qui doit supporter les contraintes et
les variations de contraintes les plus leves. Ils sont ainsi
disposs sur la ligne des vents dominants, afin de mesurer les
plus forts effets du vent.
De plus, des extensomtres cordes vibrantes (notes CV)
sont installs dans deux inclusions pour y dterminer la
descente de charges. Les capteurs sont relevs automatiquement
toutes les 6 heures depuis le dbut de la construction.
Les mesures sont faites depuis la phase de terrassement,
jusquau montage de la machine, sur une priode de six mois.
Puis, le suivi est men aprs la mise en service de lolienne,
plus de deux ans aprs la pose de linstrumentation.
T20 sur CPT7
direction des vents dominants
semelle comprime le sol
semelle se soulve du sol
Tassement sous socle (CED)
Contrainte sous socle (CPT) 3 bars
Rampe
daccs
Plateforme de
grutage 15 sur CPT4
T19 sur CPT14
22 entre CPT8et G
16 sur CPT9
19 sur CPT6
20 sur A
21sur CPT8
T18 entre Aet D
D
T16 entre CPT6 et C
T15 sur 19
E
T17 entre Aet E
18 sur CPT11
17 sur CPT1
T14 entre CPT2 et B
T13 sur CPT2
T12 sur 15
T10 sur 16
T11 sur CPT5
G
C
B
Figure 4. Plan dinstrumentation en partie suprieure de la couche de
rpartition, sous la semelle (2
me
niveau).
4 RESULTATS
Les mesures sont acquises pendant la construction de louvrage
et aprs sa mise en service. Les donnes propres de lolienne
(vitesse du vent, orientation de la nacelle, puissance produite)
sont acquises depuis le printemps 2012 et permettent une
premire observation de leffet du vent sur le comportement de
la structure de fondation.
CPT1
CPT9
4 bars
semelle comprime le sol
IR ref
(dans buse)
4.1 Transfert des efforts pendant la construction
Depuis la mise en place des CPT, directement sous la semelle
de lolienne, leurs indications ne dpassent pas 65 kPa jusqu
la mise en service. Par contre, les pressions totales mesures sur
9 IR, avant la mise en service de lolienne (Figure 5),
augmentent fortement lors du coulage de la semelle, sauf en son
centre (CPT9), qui reste trs peu sollicit.
Les mesures des CPT1 et CPT2 (comme les CPT7 et CPT8)
diffrent ds ce moment, malgr la proximit des capteurs, mais
leur volutions restent cohrentes.
L'IR situe sous le CPT5 est positionne sous linfluence
dun des appuis de la virole (qui supportera la base du mt de
l'olienne) ; cela explique pourquoi elle est trs fortement
charge. Les diffrences notables entre les valeurs mesures par
les capteurs sinitient cette phase de la construction, pour se
maintenir jusqu la mise en service de lolienne, qui va subir
de fortes sollicitations avant cette date (tests darrt durgence,
survitesse).
Les pressions mesures sur les ttes dIR schelonnent entre
360 kPa (36,6 kN sur lIR) 796 kPa (81 kN), pour une
moyenne de 596 kPa. Les 34 IR situes en priphrie de la
semelle supportent 16 % du poids statique de lolienne.
La pression exerce sur le sol varie entre 100 et 200 kPa,
sans tre maximale en priphrie de la semelle.
En fin de construction, le sol situ dans une frange de 0,5 m
en priphrie de la semelle, reprend 24,4 % du poids de
l'olienne (en considrant une pression moyenne de 120 kPa
exerce).
CPT14
T6
T2
semelle se soulve du sol
Tassement au niveau des ttes dIR (CED)
Effort sur tte IR (CPT) 10 bars et 4 bars
Contrainte sur sol (CPT) 4 bars
Dformation dans les IR (extensomtres corde vibrante)
Rampe
daccs
CPT2
Plateforme de
grutage
CPT4
CPT10
CPT11
CP
CPT5
CPT6
CPT7
CPT8
T3
CPT12
CPT13
T3
T4
T5
T0
T1
T7
T8
AA
545
Figure 5. Pressions totales sur les IR pendant la construction.
Ainsi, cette frange priphrique (soit 22,5 % de la surface
totale de la semelle), reprend 40,4 % de la charge statique totale.
4.2 Tassement
Les tassements sont calculs par rapport un capteur de
rfrence, assurment immobile, afin d'annuler les effets des
variations hydriques et de temprature de la zone exprimentale.
La Figure 6 prsente le tassement du sol de fondation au niveau
dlvation des ttes d'IR.
Figure 6. Tassement du sol de base, au niveau des ttes dIR.
Au cours de la construction, le sol de fondation tasse
lgrement plus que les IR, pour se stabiliser, au moment de la
mise en service, moins de 17 mm (et 10 mm pour les IR). Les
variations releves depuis nexcdent pas 6 mm sur le sol et
en surface de la couche de rpartition de charge.
4.3 Transfert des efforts aprs mise en service
Aprs la mise en service de l'olienne, les mesures montrent
des charges alatoires, du fait des variations de charges
engendres par le vent et la rotation des pales. La variation des
pressions exerces sur les IR, depuis la mise en service de
lolienne (Figure 7), est similaire pour toutes les IR. Les
variations saisonnires sont plus en priphrie de la semelle
quen se rapprochant du centre de la fondation. Ces variations
de pression peuvent avoir une amplitude de 400kPa entre t et
hiver.
Figure 7. Variation des pressions sur IR aprs mise en service.
Aprs mise en service, les pressions exerces sur le sol entre
IR, bien que plus chaotiques que durant la construction, ne
fluctuent quasiment pas. Un suivi de plusieurs annes permettra
de confirmer un tel comportement.
4.4 Dformation dune inclusion rigide
Des extensomtres cordes vibrantes (CV) sont installs dans
deux IR, des profondeurs diffrentes, pour accder aux
dformations de lIR et ainsi au mcanisme de transfert des
efforts dans la colonne. Seuls deux CV ont pu tre mesurs dans
la dure, malgr un fort bruit d aux vibrations de louvrage.
Sur la figure 8 sont traces les dformations mesures en partie
suprieure et 1,36 cm de profondeur de lIR coiffe par le
CPT2. La mesure de la pression exerce sur cette IR est
galement prsente.
Figure 8. Dformation dune IR et pression exerce sur sa tte (CPT2).
Ds le coulage de la semelle, lIR sest dforme, de faon
plus accentue en profondeur. Cela peut sexpliquer car la
plateforme de travail (de 30 cm dpaisseur, mise en place pour
permettre laccs la foreuse en fond de fouille) enserre la tte
de lIR et lempche ainsi de se dformer.
Elle reporte les efforts quelle reoit sa surface, comme
ceux que lui transmet lIR, plus en profondeur, sous sa base.
Une dformation mesure plus importante en profondeur peut
faire apparatre la prsence de frottement ngatif le long de la
colonne. Les fluctuations de pression exerce sur lIR sollicitent
galement lIR plus en profondeur.
546
th
4.5 Effets du vent
La vitesse du vent et lorientation de la nacelle sont enregistres
en continu. En admettant que les ples sont toujours
perpendiculaires la direction du vent, on peut accder la
valeur de la projection de la vitesse du vent sur laxe des vents
dominants. En la superposant celle du vent, on visualise si le
vent souffle dans cet axe ou non.
Lorsque la vitesse du vent dpasse 8 m/s, le pas
dacquisition des mesures des capteurs est automatiquement
rduit moins dune minute, par le biais dune deuxime
centrale dacquisition. Cela permet daccder leur variations,
exclusivement au cours dun vent violent. Notons que seules les
CPT rpondent instantanment en mesure rapide, alors que les
CV et les T ncessitent plusieurs secondes pour se stabiliser.
4.6 Variation des pressions appliques
En sintressant aux capteurs positionns dans laxe des vents
dominants, on peut observer lvolution des pressions
appliques, en fonction de la vitesse et du sens du vent
(Figure 9). Il savre que, dans cet axe, lorsque la nacelle
change de sens, les pressions exerces au niveau des ttes dIR
augmentent dun ct de la semelle pour diminuer de lautre
(entour A sur la
Figure 7). Par vent fort, les variations respectives de contraintes
atteignent 200 kPa (entour B). Dans une moindre ampleur, les
pressions exerces sur le sol prsentent le mme type de
variations.
Figure 9. Pression au niveau des ttes dIR (sol et IR), dans laxe des
vents dominants; vitesse du vent et projection sur laxe.
4.7 volution des tassements en fonction de la vitesse du
vent
Sur la Figure 10 sont superposes la puissance dveloppe par
lolienne, des mesures de CPT et les mesures de tassement
sous la semelle de lolienne et sur le sol au niveau des ttes
dIR, en priphrie de lolienne, sur laxe des vents dominants.
Lorsque lolienne ne produit pas dlectricit (puissance nulle,
entour sur la figure 10), les pressions sous louvrage diminuent
et le sol semble se relaxer, avant de retrouver sa position lorsque
la rotation des ples reprend. La prsence deau en pied de
semelle peut expliquer ce phnomne car elle nest plus chasse
par leffet dynamique des battements dus aux rafales de vent.
5 CONCLUSION
L'instrumentation du chantier d'une olienne du parc olien de
Boralex, au nord de Rouen, fonde sur une structure de sol
amlior par inclusions rigides, a permis de suivre l'volution
des tassements, pressions exerces et dformation dIR et du
sol, au cours de la construction et pendant la vie de louvrage.
En fin de construction, les efforts se concentrent lgrement
sur la priphrie de la semelle de fondation, ce qui montre un
fonctionnement en fondation rigide.
Figure 10. Pression au niveau des ttes dIR (sol et IR), dans laxe des
vents dominant; vitesse du vent et projection sur laxe.
Une fois en service, les pressions varient plus amplement et
voluent avec les variations de temprature saisonnires, qui
impactent la structure de lolienne. La poursuite des mesures
permettra dobserver ventuellement une tendance de
comportement de la structure.
Les tassements du sol et des IR restent faibles. La tte des IR
a tass d1 cm, alors que le sol en place a tass denviron 2 cm
en fin de construction. En service, lamplitude des tassements
ne dpasse pas 5 mm.
Les dformations internes des IR sont difficilement
accessibles par le systme dacquisition install. Cependant, le
mcanisme observ dans la partie suprieure dune IR montre
que la prsence de la plateforme de travail, qui enserre la tte
des IR, a un effet rel sur le comportement mcanique de la
structure, en favorisant le dveloppement de frottement ngatif
le long des colonnes, ce qui ramne plus en profondeur le point
neutre de fonctionnement de la colonne.
Nous poursuivons lanalyse des mesures de tassement et de
dformation dune IR pour accder au comportement dune IR
au cours dun fort coup de vent.
6 REMERCIEMENT
Les auteurs remercient lAgence Nationale pour la Recherche
pour laide financire apporte au montage du projet national
ASIRi, gr par lIREX, dans le cadre du RGCU, ainsi que les
socits Theolia et Boralex qui ont assum une part importante
du cot de linstrumentation. Ils remercient galement la socit
ANTEA, membre du projet national pour le montage du projet,
comme la socit Egis, intervenue lors du montage
exprimental.
7 REFERENCES
Brianon L. 2002. Renforcement des sols par inclusions rigides, tat de
lart en France et ltranger, Irex, Opration du rseau Gnie
Civil et urbain, septembre, 180 p.
Haza-Rozier E. Vinceslas G. Le Kouby A. et Duprez T. 2012.
Instrumentation des fondations dune olienne Cas des inclusions
rigides, Journes Nationales de Gotechniques et de Gologie de
l'ingnieur, JNGG, 4-6 juillet 2012, Bordeaux, France, 561-568.
Haza-Rozier E. 2011. Projet national ASIRi olienne fonde sur
inclusions rigides, rapport de recherche Cete Normandie-
Centre/CER, juillet, 28 p.
547
Seismic Response of Superstructure on Soft Soil Considering Soil-Pile-Structure
Interaction
Influence de l'Interaction sol- pieu- structure sur la rponse sismique de la superstructure
sur sol mou
Hokmabadi A.S., Fatahi B., Samali B.
School of Civil and Environmental Engineering, University of Technology Sydney (UTS), Broadway NSW 2007, Australia
ABSTRACT: This paper presents results of shaking table tests and three dimensional numerical simulations to investigate the
influence of Soil-Pile-Structure Interaction (SPSI) on the seismic response of mid-rise moment resiting buildings supported by end-
bearing pile foundations. Three different cases have been considered, namely: (i) fixed-base structure representing the situation
excluding the soil-structure interaction; (ii) structure supported by shallow foundation on soft soil; and (iii) structure supported by
end-bearing pile foundation in soft soil. Comparison of the numerical predictions and the experimental data shows a good agreement
confirming the reliability of the numerical model. Both experimental and numerical results indicate that soil-structure interaction
induces significant increase in the lateral deflections and inter-storey drifts of the structures on both shallow and end-bearing pile
foundations in comparison to the fixed base structures. This increase in the lateral deformations and in turn inter-storey drifts can
change the performance level of the structure during earthquakes which may be safety threatening.
RSUM : Cet article prsente les rsultats des essais sur table vibrante et trois dimensions simulations numriques pour tudier
l'influence de l'Interaction sol-pieu-structure (ISPS) sur la rponse sismique des btiments pris en charge par les fondations sur pieux.
Trois cas diffrents ont t examins, savoir: (i) la structure de base fixe sans interaction sol-structure; (ii) la structure soutenue par
la fondation superficielle sur sol mou; et (iii) la structure soutenue par la fondation sur pieux dans le sol mou. Les prdictions
numriques et les donnes exprimentales montrent un bon accord. Rsultats exprimentaux et numriques indiquent que l'interaction
sol-structure augmente les dflexions latrales et les drives inter tage des structures en comparaison avec les structures de base
fixes. Cela peut changer le niveau de performance de la structure lors de tremblements de terre qui peuvent tre un problme
d'innocuit.
KEYWORDS: soil-pile-structure interaction, seismic response, shaking table test, FLAC3D, end-bearing pile foundation

1 INTRODUCTION
The problem of soil-pile-structure interaction in the seismic
analysis and design of structures has become increasingly
important, as it may be inevitable to build structures at locations
with less favourable geotechnical conditions in seismically
active regions. Influence of the underlying soil on seismic
response of the structure can be ignored if the ground is stiff
enough, and the structure can be analysed considering fixed-
base conditions. However, the same structure behaves
differently when it is constructed on the soft soil deposit.
Earthquake characteristics, travel path, local soil properties, and
soil-structure interaction are the factors affecting the seismic
excitation experienced by structures. The result of the first three
of these factors can be summarised as free-field ground motion.
However, the foundation is not able to follow the deformation
of the free field motion due to its stiffness, and the dynamic
response of the structure itself would induce deformation of the
supporting soil (Kramer 1996).
Over the past decades, several researchers (e.g. Tajimi 1969,
Gazetas 1991, Shiming and Gang 1998, Hokmabadi et al. 2011,
Carbonari et al. 2011, Tabatabaiefar et al. 2013) have studied
the seismic soil-pile-structure interaction (SSPSI) and the effect
of this phenomena on the response of the structures. The
developed analytical methods for studying the soil-pile-structure
interaction may be categorised into three groups: (i)
Substructure Methods (or Winkler methods), in which series of
springs and dashpots are employed to represent the soil
behaviour (e.g. Hokmabadi 2012); (ii) Elastic Continuum
Methods, which are based on Mindlin (1936) closed form
solution for the application of point loads to a semi-infinite
elastic media; and (iii) Numerical Methods. The substructure
methods are the simplest and most commonly used methods,
however, these methods adopting the substructuring concept
rely on the principle of superposition, and consequently, are
limited to either the linear elastic or the viscoelastic domain
(Pitilakis et al. 2008).
The dynamic equation of motion of the soil and structure
system can be written as:

[M]{}+[C]{}+[K]{u}= -[M]{m}
g
+{F
v
} (1)

where, {u}, {}, and {} are the nodal displacements, velocities
and accelerations with respect to the underlying soil foundation,
respectively. [M], [C] and [K] are the mass, damping, and
stiffness matrices of the structure, respectively. It is more
appropriate to use the incremental form of Equation (1) when
plasticity is included, and then the matrix [K] should be the
tangential matrix and {} is the earthquake induced acceleration
at the level of the bedrock. For example, if only the horizontal
acceleration is considered, then {m}=[1,0,1,0,....1,0]
T
. {Fv} is
the force vector corresponding to the viscous boundaries. This
vector is nonzero only when there is a difference between the
548
th
motion on the near side of the artificial boundary and the
motion in the free field (Wolf 1985).
The present research aims to study the effects of SSPSI on
the seismic response of the superstructure by employing the
fully nonlinear method in which main components of the
interaction including subsoil, pile foundation, and superstructure
are modelled simultaneously. For this purpose, a three-
dimensional explicit finite-difference program, FLAC3D (Itasca
2009), is used to numerically model and examine the influence
of the soil-structure interaction on the seismic response of a 15-
storey moment resiting building. Two types of foundations
including shallow foundations and end-bearing pile foundations
have been considered. The proposed numerical soil-structure
model has been verified and validated against experimental
shaking table test results.

2 SHAKING TABLE EXPERIMENTAL TESTS
2.1 Prototype characteristics and scaling factors
In order to provide a calibration benchmarks for the numerical
simulation and to make quantitative predictions of the prototype
response several of shaking table tests have been conducted.
Previous researchers (e.g. Meymand 1998, Chau et al. 2009)
modeled the superstructure as a simplified single degree of
freedom oscillator in which the behaviour of the soil-structure
system may not be completely conform to reality and the higher
modes would not be captured. In the current model tests, unlike
the previous efforts, a multi-storey frame for the superstructure
is adopted representing most of the dynamic properties of the
prototype structure such as natural frequency of the first and
higher modes, number of stories, and density. The experimental
model tests have been carried out utilising the 33 m shaking
table facilities located at structures laboratory of the University
of Technology Sydney (UTS).
The selected prototype structure is a fifteen-storey concrete
moment resisting building frame with the total height of 45 m
and width of 12 m consisting of three spans, representing the
conventional types of mid-rise moment resisting buildings. The
spacing between the frames into the page is 4 m. Natural
frequency of the prototype building is 0.384 Hz and its total
mass is 953 tonnes. The soil medium beneath the structure is a
clayey soil with the shear wave velocity of 200 m/s and density
of 1470 kg/m3. The horizontal distance of the soil lateral
boundaries and bedrock depth has been selected to be 60 m and
30 m, respectively. The building is resting on a footing which is
4 m wide and 12 m long. For the pile foundations case, a 44
reinforced concrete pile group with equal spacing and pile
diameter of 1.25 m and 30 long are considered. The piles are
embedded into the bedrock representing typical end-bearing pile
foundations.
In order to achieve a reasonable scale model, a dynamic
similarity between the model and the prototype is applied as
described by Meymand (1998). Dynamic similarity governs a
condition where homologous parts of the model and prototype
experience homologous net forces. Although small scale models
could save cost, the precision of the results could be
substantially reduced. Considering the specifications of UTS
shaking table, scaling factor of 1:30 is adopted for experimental
shaking table tests on the scale model which provides the largest
achievable scale model with rational scales, maximum payload,
and overturning moment meeting the facility limitations.
2.2 Shaking table tests model components
The developed soil-structure model for shaking table tests
possesses four main components including the model structure,
the model pile foundations, the laminar soil container, and the
soil mix. Employing geometric scaling factor of 1:30, height,
length, and width of the structural model are determined to be,
1.50 m, 0.40 m, and 0.40 m, respectively. In addition, the
required natural frequency of the structural model is 2.11 Hz.
The model structure has been designed employing SAP2000
(CSI 2010) software to meet the required characteristics, and
finally a 50050010 mm steel plate as baseplate, fifteen
4004005 mm horizontal steel plates as the floors and four
500402 mm vertical steel plates as the columns are adopted.
The completed structural model is shown in Figure 1.

Controlroom
Modelstructure
Displacement
transducers
Shakingtable
LaminarSoilcontainer
Soilmix

Figure 1. Final setup of the shaking table tests for the structure with
end-bearing pile foundation
Similar to the model structure, the model pile is subjected to
the competing scale model criteria. The model piles have a
diameter of 40 mm with L/d ratio of 25. By selecting a
commercial Polyethylene high pressure pipe with Standard
Dimension Ratio (SDR) of 7.4 the model piles fall in the range
of acceptable criteria with 5% deviation from the target value
for EI.
The ideal soil container should simulate the free field soil
response by minimising boundary effects. Since the seismic
behaviour of the soil container affects the interaction between
the soil and structure, the performance of the soil container is of
the key importance for conducting seismic soil-structure
interaction model tests successfully (Pitilakis et al. 2008). A
laminar soil container with final length, width, and depth of
2.10m, 1.30m, and 1.10m, respectively, are designed and
constructed for this study. The employed laminar soil container
consists of a rectangular laminar box made of aluminium
rectangular hollow section frames separated by rubber layers.
The aluminium frames provide lateral confinement of the soil,
while the rubber layers allow the container to deform in a shear
beam manner.
A synthetic clay mixture was designed to provide soil
medium for the shaking table testing considering required
dynamic similarity characteristics. Several mixtures were
examined and finally the desired soil mix (60% Q38 kaolinite
clay, 20% Active-bond 23 Bentonite, 20% class F fly ash and
lime, and water, 120% of the dry mix) produced the required
scaled shear wave velocity of 36 m/s at the second day of its
cure age. Accordingly, the soil density and undrained shear
strength on the second day were determined to be 1450 kg/m
3

and 3.14 kPa, respectively.
The shaking table tests have been carried out in three stages:
fixed-base condition, shallow foundations, and end bearing pile
foundations. Since the properties of the designed soil mix is
time depended, the second and third stages should be carried out
549
in the same age in order to make the results comparable, without
being interrupted by variation of the soil mix dynamic
properties. Two scaled near field shaking events including
Kobe, 1995, Northridge, 1994, and two scaled far field
earthquakes including El Centro, 1940, and Hachinohe, 1968
are adopted. The characteristics of the mentioned benchmark
earthquakes are summarised in Table 1. Displacement
transducers (levels 3, 5, 7, 11, 13, and 15) and accelerometers
(at levels 3, 5, 7, 9, 11, 13, and 15) were installed on the
structure in order to monitor the dynamic response of the
structure and to primarily measure the structural lateral
displacements. The recorded accelerations can be used to check
the consistency and accuracy of obtained displacements through
a double integration in time domain. The final setup of the tests
for the end-bearing pile foundation system on the shaking table
shown in Figure 1.
T d e ake b ns
P M ) Dur S)
is

able 1. Utilise arthqu ase motio
Earthquake Year GA (g) w (R ation (
Northridge 1994 0.843 6.7 30.0
Kobe 1995 0.833 6.8 56.0
El Centro 1940 0.349 6.9 56.5
Hachinohe 1968 0.229 7.5 36.0

3 DEVELOPMENT OF 3D NUMERICAL MODEL
Three-dimensional explicit finite-difference based program
called FLAC3D (Itasca 2009) has been employed to develop the
numerical model for the shaking table tests and to simulate the
response under the seismic loading. Three cases including
fixed-base conditions, the structure supported by shallow
foundations, and the structure supported by end-bearing pile
foundations have been modelled separately and the results are
compared. The dimensions of the numerical models were
chosen similar to the experimental tests. The reason for
choosing the soil deposit thickness of 30 m for the both
experimental and numerical models is that most amplification
occurred within the first 30 m of the soil profile, which is in
agreement with most modern seismic codes calculating local
site effects based on the properties of the top 30 m of the soil
pro
rid and
mo
e-field boundary
con
4.0
records and results in terms of maximum inelastic
eral deflections, determined for the three mentioned cases, are
recorded.
file (Rayhani and El Naggar 2008).
Experience gained from the parametric study helped to
finalise the adopted mesh size and the maximum unbalanced
force at the grid points to optimize the accuracy and the
computation speed simultaneously. The numerical g
del components in FLAC3D are shown in Figure 2.
Adjusting the boundary conditions, in the static analysis in
which the system is under the gravity loads only, the bottom
face of the mesh is fixed in all directions, while the side
boundaries are fixed in the horizontal directions. During the
dynamic time-history analysis, the earthquake acceleration is
applied horizontally at the entire base, while fre
ditions are assigned to the side boundaries.
Solid elements are used to model the soil deposits, and
Mohr-Coulomb failure criterion is adopted. In addition,
Hysteretic damping of the soil is implemented using the built-in
tangent modulus function as developed by Hardin and Drnevich
(1972). The pile elements and superstructure are modelled with
solid elements considering elastic-perfectly plastic behaviour
with yielding criteria for the elements to control the possibly of
inelastic behaviour in both superstructure and piles. As a
calibration, a FLAC3D analysis was first conducted on a
cantilever pile while the pile was fixed at one end into ground
without the surrounding soil and the different lateral loads were
applied on the free end of the cantilever pile. The recorded
deflection from the FLAC3D model shows less than 2%
difference from analytical predictions, confirming the accuracy
of the model. It should be noted that using the structural
elements such as beam and shell elements in FLAC3D (version
) for modelling the superstructure increases the execution
time dramatically and leads to less accurate results.
Because of the different characteristics of the soil and the
superstructure/piles, sliding and separation may occur at the
soilstructure interfaces. Two sets of interface elements are
modelled in this study. For the shallow foundation case, the
interface elements are placed between the foundation and the
soil surface. However, for the pile foundation case, the interface
elements were attached to the outer perimeter of the piles. It
should be noted that in the pile foundation case, there is no
interface or attachment between the foundation and the surface
soil as some gap in the shaking table tests is considered to avoid
any pile-raft behaviour. Therefore, there is not any direct stress
transfer between the foundation slab and the subsoil in the pile
foundation cases. The interfaces were modelled as linear
springslider systems, while the shear strength of the interfaces
was defined by MohrCoulomb failure criterion. The lateral and
axial stiffness of the interface elements are estimated for both
sets separately based on the recommended method given by
Itasca (2009) to ensure that the interface stiffness has minimal
influence on system compliance. Finally, fully nonlinear time-
history analysis is conducted under the influence of the scaled
arthquake e
lat
15storeymodelstructure
Freefieldboundaries
Interfacesbetween
pilesandsoil
Connectionofpiles
tothebaseplate

Figure 2. Numerical grid and model com
structure with end-bearing pile foundation
ponents in FLAC3D for the
with acceptable accuracy and is
4 RESULTS AND DISCUSSION
The average values of the 3D numerical predictions versus
experimental shaking table results for the maximum lateral
displacements of the fixed-base, shallow foundations, and end-
bearing pile foundations were determined and compared in
Figure 3. Evaluation of the predicted and observed values of the
maximum lateral displacements indicates that the trend and the
values of the 3D numerical predictions are in a good agreement
and consistent with the experimental shaking table test results.
Therefore, the 3D numerical model can replicate the behaviour
of the soil-pile-structure system
550
th
rat
comparison with the fix-based case, soil-pile-
structure interaction tends to increase the lateral deformation of
the structure.

ional and appropriate for further studies of the soil-pile-
structure interaction effects.
Accordingly, the maximum lateral deflection of the structure
supported by end-bearing pile foundations is increased by 17%
based on the experimental values and 19% based on the 3D
numerical predictions in comparison to the fixed base structure.
Moreover, the maximum lateral deflection of the structure
supported by shallow foundation is increased by 55% based on
the experimental values and 59% based on the 3D numerical
predictions. Thus, pile foundations reduce the lateral drifts in
comparison to the shallow foundation case. This is due to the
presence of stiff pile elements in the soft soil which increase the
stiffness of the ground and influences the dynamic properties of
the whole system such as the natural frequency and damping.
However, in

0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
0 10 20 30
Maximum Lateral Deflection (mm)
S
t
o
r
e
y

N
u
m
b
e
r
Fixed base Numerical Results
Fixed base Exp. Results
Shallow foundation Numerical Results
Shallow foundation Exp. Results
End_bearing piles Numerical Results
End_bearing piles Exp. Results

Fig
n tends to
increase the inter-storey drifts of the superstructure from life
safe zone toward near collapse or even total collapse.

ure 3. Average values of maximum lateral displacements: Shaking
table experimental values versus 3D numerical predictions
The corresponding inter-storey drifts of the average values of
3D numerical model are plotted in Figure 4. Inter-storey drifts
are the most commonly used damage parameters, and based on
FEMA (BSSC 1997) maximum inter-storey drift of 1.5% is the
defined border between life safe and near collapse levels.
According to Figure 4, seismic soil-structure interactio
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
0 0.5 1 1.5 2 2.5
Inter-storey Drift (%)
S
t
o
r
e
y

N
u
m
b
e
r
Fixed base
Shallow foundation
End_bearing pile foundation

Figure 4. Average experimental inter-storey drifts for: (a) fixed-base
str
tural
period lies in the long period region of the response spectrum
acement response tends to increase.
lling method is rational and is suitable
for
raction are not adequate to
guarantee the structural safety for the moment resisting
oft soils.

Tab

Kra
Piti
olf, J.P. 1985. Dynamic soil-structure interaction. Prentice-Hall,
Englewood Cliffs, New Jersey.

ucture; (b) Structure supported by shallow foundation; (c) structure
supported by end-bearing pile foundation
The natural period of the system increases due to the soil-
structure interaction. Therefore, such increases in the natural
period considerably alter the response of the building frames
under seismic excitation. This is due to the fact that the na
curve. Hence, the displ

5 CONCLUSIONS
In this paper, a three-dimensional finite difference numerical
model on a soil-pile-structure system has been conducted
together with the experimental shaking table tests. By
comparing predicted and observed results, it has been concluded
that the numerical mode
the simulation of the soil-pile-structure interaction under
strong ground motions.
In addition, based on the shaking table results and 3D
numerical investigations it is observed that the lateral
deflections of the structures siting on the end-bearing pile
foundations amplified in comparison to the fixed base model
(approximately 18% in this study). This amplification for the
structure siting on the shallow foundations is more severe
(approximately 57% in this study). Consequently, considering
soil-structure interaction in both cases with and without pile
foundations is vital, and conventional design procedures
excluding soil-structure inte
buildings resting on s

6 REFERENCES
C. 1997. NEHRP Guidelines for the Seismic Rehabilitat BSS ion of
Buildings, 1997 Edition, Part 1: Provisions and Part 2:
Commentary. In: Federal Emergency Management Agency.
bonari, S., Dezi, F., and Leoni, G. 2011. Linear soil-struct Car ure
interaction of coupled wall-frame structures on pile foundations.
Soil Dynamics and Earthquake Engineering 31 (9): 1296-1309.
u, K.T., Shen, C.Y., and Guo, X. 2009. Nonlinear seismi Cha c soil-pile-
structure interactions: Shaking table tests and FEM analyses. Soil
Dynamics and Earthquake Engineering 29 (2): 300-310.
SAP2000 v14 Analysis Reference Manual. CSI (Computers and
Structures Inc.), Berkley, California.
etas, G. 1991. Fo Gaz rmulas and Charts for Impedances of Surface and
Embedded Foundations. Journal of Geotechnical Engineering 117
(9): 1363-1381.
din, B.O., and Drnevich, V.P. 1972. Shear m Har odulus and damping in
soils: desing equations and curves. Journal of the Soil Mechanics
and Foundations Division 98 (7): 667-692.
mabadi, A.S., Fakher, A., and Fatahi, B. 2 Hok 011. Seismic strain wedge
model for analysis of single piles under lateral seismic loading.
Australian Geomechanics 46 (1): 31-41.
mabadi, A.S., Fa Hok kher, A., and Fatahi, B. 2012. Full scale lateral
behaviour of monopiles in granular marine soils. Marine Structures
29(1): 198-210.
atabaiefar, S., Fatahi, B., and Samali, B. Seismic Behaviour of
Building Frames Considering Dynamic Soil-Structure Interaction.
International Journal of Geomechanics (doi: 10.1061/(ASCE)
GM.1943-5622.0000231).
C3D version 4.00 Fast Lagrang FLA ian Analysis of Continua in three
dimentions, User's Manual. Itasca Consulting Group, Inc,
Minneapolis, Minnesota, USA.
mer, S.L. 1996. Geotechnical earthquake engineering. Prentice Hall.
ymand, P.J. 1998. Shaking table scale model tes Me ts of nonlinear soil-
pile-superstructure in soft clay. PhD PhD thesis in Civil
Engineering University of California, Berkley.
Mindlin, R.D. 1936. Force at a Point in the Interior of a Semi-Infinite
Solid. Physics 7 (5): 195-202.
lakis, D., Dietz, M., Wood, D.M., Clouteau, D., and Modaressi, A.
2008. Numerical simulation of dynamic soil-structure interaction in
shaking table testing. Soil Dynamics and Earthquake Engineering
28 (6): 453-467.
hani, M., and El Naggar, M. 20 Ray 08. Numerical Modeling of Seismic
Response of Rigid Foundation on Soft Soil. International Journal of
Geomechanics 8 (6): 336-346.
ming, W., and Gang, G. 1998. Dynamic soil-structure inter Shi action for
high-rise buildings. In Developments in Geotechnical Engineering,
eds. Chuhan Zhang and P. Wolf John: Elsevier. 203-216.
imi, H. 1969. Dynamic Taj Analysis of a Structure Embedded in an
Elastic Stratum. In Proc. 4th World Conf. Earthquake Eng.
Santiago, USA. 53-69.
W
551
Applicability of the RNK-method for geotechnical 3D-modelling in soft rocks
Applicabilit de la RNK-mthode pour la modlisation gotechnique en 3D en roches tendres
Ivi T.
University of Zagreb, Faculty of Civil Engineering, Kaieva 26, 10000 Zagreb, Croatia
Ortolan .
J. J. Strossmayer University of Osijek, Faculty of Civil Engineering, Drinska 16a, 31000 Osijek, Croatia
Kavur B.
Institut IGH d.d., Janka Rakue 1, 10000 Zagreb, Croatia

ABSTRACT: The RNK-method or the Reference Level of Correlation method represents a procedure for spatial engineering-
geological and/or geotechnical modeling, that was tested on many landslides in Croatia. The method gives the opportunity of
differentiation of minimum shear strength zones, zones of different hydraulic conductivities, and zones of various soil densities. The
application and verification of the RNK-method in soft rock formations found on the landslide area in Gorica Svetojanska (Croatia) is
presented. The presentations providing the full set of relevant information needed to develop representative geotechnical profiles are
also shown. The established geotechnical sliding model is verified by measurements of lateral movements in the landslide area and by
results of corresponding stability analyses
RSUM : La RNK-mthode (mthode du niveau de corrlation de rfrence) reprsente une procdure de modlisation spatiale en
gnie gologique et/ou gotechnique, qui a t teste sur plusieurs glissements de terrain en Croatie. La mthode permet la
diffrenciation des zones de la rsistance de cisaillement minimale, des zones des conductivits hydrauliques diffrentes, et des zones
de densit du sol diverse. L'application et la vrification de la mthode RNK aux formations rocheuses tendres, trouvs sur un site de
glissement de terrain Gorica Svetojanska (Croatie) sont prsentes. On prsente aussi un ensemble complet dinformations
pertinentes pour dvelopper les profils gotechniques reprsentatifs. Le modle gotechnique de glissement tabli est vrifi par les
mesures de mouvements latraux dans la zone de glissement, et par les rsultats danalyse de stabilit correspondante.
KEYWORDS: RNK-method, plasticity index, shear strength, slope stability, spatial geotechnical model.

1 INTRODUCTION
1.1 The site description
The village Gorica Svetojanska is located in hills area in north-
western part of Croatia. In last several years the intensive
cracking of the walls of local church has been observed. Also,
the soil movements at the slope with graveyard down the church
have been noticed, as well as damages of the small mortuary
structure.
The church of St. Anastasia (St. Ana, Jana in local
dialect) is situated at the plateau of narrow ridge dominating the
nearby valley (Figure 1). The church at this position is
mentioned in historical parish records from second half of 18
th

century. It was several times reconstructed and strengthened
after damages caused by stronger earthquakes in late 19
th

century.
Figure 1. The St. Ana church with graveyard.

1.2 Local conditions
The site is in seismically active region and in Figure 2 the
frequency of earthquakes (with I > 4) in last 200 years is
shown, supporting the parish records.
The seismic intensities at the church location have been
estimated by common attenuation function compiling the
catalogue records of earthquakes with epicentres in radius R =
75 km from the site (GZ, 2005).
Figure 2. The frequency of moderate and strong earthquakes at the site
In geological profile, generally, the ridges and hills in the
vicinity have the less permeable soft rocks and clayey soils in
upper part, and older, permeable aquifers in lower part of
profile. The aquifers are recharging at higher elevations,
producing artesian or sub artesian groundwater pressures at the
village site. Also, in the vicinity, the mineral water is
commercially extracted and bottled.

552
th
The complex investigations at the location have been made,
including borings and sampling, laboratory tests, water level
measurements and monitoring of church wall movements.
The thorough engineering geology investigations were also
performed, and, in order to obtain reliable geotechnical model
of landslide, the so-called RNK method was used.
2 RNK METHOD - FUNDAMENTAL NOTIONS AND
BASIC DEFINITIONS
The RNK method (RNK-the acronym in Croatian language) or
the Reference Level of Correlation Method (Ortolan 1996) is a
fully developed method for engineering-geological and
geotechnical modelling. It is primarily intended for the landslide
recognition and the analysis of the slope stability of soils and
soft rock formations. However, the sedimentation fingertip
obtained by geotechnical correlation column can be also used
for reliable association of other test results in clayey sloppy
profiles (Ivsic et al., 2005)
The RNK (Reference Level of Correlation) is defined as an
unequivocally recognizable and visually identifiable (or
graphically defined!) bedding plane or any other reference plane
within a structural feature, in relation to which an altitude of all
studied profiles can be unambiguously defined, with individual
point analysis of any material property. Such plane is a part of a
single vertical lithostratigraphical i.e. engineering geological
and/or geotechnical sequence (engineering-geological and/or
geotechnical correlation column).
The importance of correlation for the slip-surface and/or
slip-zone determination is emphasized by Ortolan (1990).
The plasticity index has proven to be a reliable strength
indicator for cohesive materials (Ortolan 1996, Ortolan &
Mihalinec 1998, Ortolan et al., 2009). The highest values of
plasticity index, but also the liquid limit, correspond to the
lowest expected values of friction angle. This fact allows a new
approach to exact geotechnical modelling. Therefore, testing of
Atterberg plasticity limits on point samples can be
recommended for the identification of zones with lowest shear
strengths. The sample length should not exceed 10cm
(sometimes it should be aslittle as several centimetres, and even
several millimetres). The sampling interval of 0.51.0m is
usually considered sufficient.
The correlation between the plasticity index and angle of
internal friction is given in Figure 3, as developed by various
authors, systemized by Ortolan & Mihalinec (1998) and
enriched by new carefully obtained data.
46,0
119
118
117=120
116
111
108
106
104
102
101
100
96
95
92
91
88
68
87
54
115
110
89
90
51
84
19
74
7
82
25
72
26
71
70
14
73
1
75
76
24
3
22
18
31 59
21
30
44
49
9
20
8
36
33
35
69
16
37
77
60
17
80
47
2
32
15
45
48
46
6
83
67
61
42
13
38
50
53
56
52
55
98
58
62
93
39
65
11
12
66
64
40
78
34
4
29
94
41
79
23
5
10
27
28 43
63
86
85
97
99
103
112
113
105
107
109
114
0
5
10
15
20
25
30
35
40
0 10 20 30 40 50 60 70 80 90 100 110 120 1
PLASTICITY INDEX - PI (%)
P
E
A
K

-

P
O
R

R
E
S
I
D
U
A
L

F
R
I
C
T
I
O
N

A
N
G
L
E

-

R

(
30
O
)
Residual friction angle (Ortolan & Mihalinec, 1998)
Peak friction angle (Ortolan & Mihalinec, 1998)
Residual friction angle (1998-2006)
Landslide Hospital Merkur in Zagreb ( 2005/2006)
Landslide Jarpetar near Buje - Istra (2002)
Landslide esmiki west in Zagreb (2002)
Landslide Zalesina: Triassic clays and shales (Ortolan, 1996)
Residual friction angle: Landslide Gorica Svetojanska
Very sensitive clays:
4-8 (OTAWA-KANADA: 8)
Cucaracha
Shale: (15)
Allophane:
JAVA Halloasyte:
JAVA
Halloasyte
(29-32): Clay from Carboniferous
Shales and Mudstones
(25-27): Materials
Containing Hydrous Mica
Soft Clays (47-49)
(20-24): Materials containing
montmorillonite
(100-107):Triassic
clays and shales
P
R

Figure 3. Correlation between index of plasticity and angle of internal
friction both peak and residual.
The following supporting documents are most often used in
the study of landslides: general geological map of the wider area
under study, geotechnical correlation column, and engineering-
geological map with slip-plane contour lines and with clearly
delineated slip areas and hydro-isohypses or hydro-isopiestic
lines at the slip-plane level (Ortolan 1996, 2000).
The geotechnical correlation column is a consistent
engineering-geological and/or geotechnical soil model (design
cross section) in which adequate parameters (defined in
laboratory or in situ either by point method or continuously) can
reasonably be allocated to every defined layer (and portions of
such layers) along the entire height of the vertical sequence of
formations covered by the study. From such geotechnical
correlation column we may in principle differentiate zones of
minimum residual shear resistance, with their thicknesses and
continuities, but also layers with different moisture content,
hydraulic conductivity, natural compaction, compressibility, etc.
The engineering-geological and/or geotechnical correlation
column of an analyzed area is the "key" to the interpretation of
overall engineering-geological and/or geotechnical relationships
in a required number of profiles selected at will for 2D and
spatial analysis, which is especially significant in 3D analysis of
stability.
The consistent use of the RNK-method in the period from
1995 to the present day has resulted in the elaboration of three-
dimensional geotechnical models for some fifty landslides. In
all of these cases the following parameters were successfully
defined: sliding body geometry, pore pressures and shear
strength parameters for materials along zones of minimum shear
resistance. In combination with existing topographical
documents, this enabled accurate stability analyses and
definition of optimum improvement procedures. The Podsused
landslide may be described as one of the most complex urban
landslide projects in the world (Ortolan 1996, 2000). It is
precisely on this project that the RNK-method has been
developed in full detail, and the reliability of the model was
confirmed with photogrammetric measurements (Ortolan et al.
1995) as well as with three-dimensional stability analyses
(Mihalinec & Stani, 1991).
Most of the studied landslides have been stabilized, in all
cases with great success, and the supervisory work conducted
during remedial works provided positive feedback information
about the correctness of adopted engineering-geological and
geotechnical landslide models, (e.g. at the Granice landslide;
Jurak et al., 2004), and about reliability of the engineering-
geological and geotechnical correlation column (design cross
section). On some projects the reliability of the model was
checked and confirmed by appropriate inclinometer, piezometer
and benchmark measurements.
3 DESCRIPTION OF THE LANDSLIDE AND GEO-
TECHNICAL PROPERTIES OF MATERIALS
The topographic presentation of the neighboring terrain in
Gorica Svetojanska with the contour of the landslide is given in
detailed engineering geology map of the area (Figure 4). Results
of laboratory and in situ investigations, presented in form of
geotechnical correlation column are presented in Figure 5.
Plasticity chart with encircled critical geotechnical zone-2 is
presented in Figure 6. Formations found on the landslide
(calcitic clays and clayey marls) date back to the Pontian.
4 ANALYSES
4.1 Wall movements
The investigation program included the measurements of
relative rotation of church walls using several horizontal and
vertical tilt meters, and, also the change of crack widths during
monitoring period (originally found cracks were 15-20mm
wide). The particular results are shown in Figure 7.

553
Figure 4. Detailed engineering geology map of the investigated area
-16
-15
-14
-13
-12
-11
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
5
6
7
8
9
10
11
0 10 20 30 40 50 60 70 80 90 100
PI (ALL BOREHOLES)
WL (ALL BOREHOLES)
SPT (BOREHOLES NEAR
THE CHURCH)
SPT (BOREHOLES AT
THE GRAVEYARD)
PI: SAMPLES FOR RING
SHEAR TEST
LIQUID LIMIT - LL & PLASTICITY INDEX - PI (%); SPT (NUMBER OF BLOWS - TUBE)
RNK
S
A
M
P
L
E

O
R

S
P
T

P
O
S
I
T
I
O
N

W
I
T
H

R
E
S
P
E
C
T

T
O

R
E
F
F
E
R
E
N
C
E

L
E
V
E
L

O
F

C
O
R
R
E
L
A
T
I
O
N

(
R
N
K

m
)

1
2
3
5
4
GEOTECHNICAL
ZONE

0
10
20
30
40
50
60
0 10 20 30 40 50 60 70 80 90 1
P
L
A
S
T
IC
IT
Y
IN
D
E
X
-
P
I (
%
)

Figure 5. Geotechnical correlation column of the landslide.
Figure 6. Plasticity chart of materials from the landslide. The encircled
zone contains samples from preferred slip-zone
The unexpected swaying of eastern part of church was
recorded (i.e. the movements of whole church block had
alterative directions). This has been confirmed by independent
records in horizontal and vertical tilt meters on neighbouring
east and south wall, also accompanied with relative closing or
opening of cracks.
The ground water levels were not measured in the same
frequency, but the collected data indicate possible correlation of
seasonal variations of water levels with the directions of wall
movements.
00
LIQUID LIMIT - LL (%)
SAMPLES FROM BOREHOLES NEAR THE CHURCH
SAMPLES FROM BOREHOLES AT THE GRAVEYARD
SAMPLE FOR RING SHEAR
CL
ML
CI
OL
MI OI
CH
MH OH
SF
SC
GEOTECHNICAL
ZONE - 2
554
th
The whole situation at the site can be described as unstable
(labile) or at the limits of equilibrium. However, even the recent
observations have shown that the net effect of swaying are
slow irreversible displacements in direction of sliding, with
cumulative displacements of 4 - 8 mm in last several years.
Also, some cracks have opened at the western part of church
which was quiet during the intensive monitoring period.
Figure 7. Monitoring of church wall displacements.
4.2 Stability analyses
The presented charts using the RNK method describe the
landslide underground conditions and enable the construction of
geotechnical models for engineering analyses in various cross-
sections.

Figure 8. Model and results of stability analyses
The cross-section A-A (shown in Figure 4) which includes
the church ridge and downhill slope was used for common
stability analyses (Spencer limit equilibrium method). The
layers corresponding to the geotechnical zones in Figure 4, with
several slip surfaces in the layers of lowest strength are shown
in Figure 8. The strength parameters taken in analyses were:
cohesion c=0 kPa (for all layers), and friction angle = 28,
24, 30 (for zones 1, 2-4, 5, respectively). The minimal friction
angle = 24 corresponds to the results of ring shear test and
correlation chart. The ground water levels were varied few
meters from referent level to estimate the influence of possible
variations.
This type of numerical modelling might be understood as
too crude or too approximate for such a complex geologic
situation at the site. However, the results (factors of safety)
reveal that the established geotechnical model (with sequence of
layers, friction angles, water levels) and slip surfaces respecting
the established weakest zones, demonstrate the unstable
conditions - by obtaining the safety factors near Fs = 1. Also
shown is the expected trend of lowering of safety factor with
rise of ground water level.
These rough estimates are demonstrated for static
conditions, implying that even the low or moderate seismic
activity can significantly reduce slope stability.
5 CONCLUSION
The interaction of engineering geology and geotechnics in the
process of designing geotechnical structures is very important.
In the study of landslides or stability levels of natural and
artificially shaped slopes, unequivocal results can be obtained
by the correlation of formations. This can be done by
introducing the reference level of correlation (RNK-method)
and by looking for the zone of minimum shear strength in the
engineering-geological and geotechnical correlation column.
The creation of reliable geotechnical model is a center of
this process, and it is crucial for the quality of the entire project.
The correlation of the friction angle with the liquid limit or
plasticity index is suggested for correct assessment of shear
strength.
6 REFERENCES
GZ. 2005. The catalogue of earthquakes in Croatia, Department of
geophysics, Faculty of Science, University of Zagreb
Ivi, T., Ocvirk, E., Pavlin, . 2005. Geotechnical Aspects of Small
Retention Dam Vir in Croatia. Proc. Ninth International
Symposium on Water Management and Hydraulic Engineering.
Nachtnebel H.P. (ed.). Ottenstein : BOKU-University of Natural
Resources and Applied Life Sciences, 2005. 221-228
Jurak V., Ortolan ., Slovenec D. & Mihalinec Z. 2004. Verification of
Engineering-Geological / Geotechnical Correlation Column and
Reference Level of Correlation (RNK) Method by Observations in
the Slip-Plane Zone. Geologia Croatica 57(2): 191-203
Mihalinec Z. & Stani B. 1991. Three-dimensional slide analysis
procedure (in Croatian). Graevinar, 42(9): 441-447
Ortolan . 1990. Le rle de la methode de correlation dans la
determination des zones de parametres minimaux de resistance au
cisaillement. Proc. of the Sixth Int. Congress IAEG, 6-10 Aug.
1990, Amsterdam. Balkema: 1675-1679
Ortolan . 1996. The creation of a spatial engineering-geological model
of deep multi-layered landslide on an example of the Podsused
landslide in Zagreb (in Croatian). PhD Thesis. University of Zagreb
Ortolan . 2000. A Novel Aproach to the Modeling of Deep Complex
Landslides with Several Sliding Planes. In E. Bromhead, N. Dixon,
M.I. Ibsen (eds) Landslides in Research, Theory and Practice. 3:
1153-1158, Thomas Telford
Ortolan . & Mihalinec, Z. 1998. Plasticity index - Indicator of shear
strength and a major axis of geotechnical modeling. In B. Mari et
al. (eds) Geotechnical hazards, Proc. of the XI-th Danube-
European conference on soil mechanics and geotechnical
engineering, Pore, Croatia, 25-29 May 1998. Balkema: 743-750
Ortolan ., Mihalinec, Z., Stani, B. & Pleko, J. 1995. Application of
Repeated Photogrammetric Measurements at Shaping Geotechnical
Models of Multi-layer Landslides. Proc. 6th Int. Symp. on
Landslides. Balkema: 1685-1691
Ortolan ., Zlatovi S. & Vrkljan I. 2009. Geotechnical 3D modeling in
soft rocks using the RNK method. Rock engineering in difficult
ground conditions soft rocks and karst. Proc. of the reggional
symp. of the ISRM, Eurock 2009, Dubrovnik, Croatia, 29-31
october 2009. CRC Press, Taylor & Francis Group: 489-494.
555
Une nouvelle sonde permettant de mesurer sans extrapoler la pression limite
pressiomtrique des sols
A new probe for measuring the pressuremeter limit pressure of soils without extrapolation
Jacquard C., Rispal M.
Fondasol, Avignon, France
Puech A., Geisler J., Durand F.
Fugro GeoConsulting, Nanterre, France
Cour F.
Calyf, Maisons Lafitte, France
Burlon S., Reiffsteck P.h.
IFSTTAR, Marne-la-Valle, France

RSUM: Une limite actuelle des essais pressiomtriques de type Mnard est lie la difficult datteindre des volumes dexpansion
et des pressions importants sans risque systmatique dclatement. Une nouvelle sonde a t dveloppe qui permet datteindre, mme
sous pressions leves, le doublement du volume du trou et donc la mesure directe de la pression limite conventionnelle du sol. On
dcrit les innovations technologiques qui ont conduit accroitre les performances et la fiabilit des sondes. On prsente ensuite des
essais comparatifs sur diffrents sites montrant les apports techniques et oprationnels du nouveau concept.
ABSTRACT: A present limitation of Menard type pressuremeter tests is due to the difficulty of reaching large expansion volumes and
high pressures without exposing to significant risks of bursting. A new probe has been developed allowing the volume of the hole to
be doubled, even under high pressures: the conventional limit pressure can then be directly measured. Technological innovations
increasing the capabilities and reliability of pressuremeter probes are described. Comparative tests on different sites are presented
demonstrating the technical and operational contribution of the new concept.
MOTS-CLS : essai pressiomtrique Menard, sonde, mesure, membrane
KEYWORDS: Menard pressuremeter test, probe, measurement, membrane
1. INTRODUCTION
La quasi-totalit des essais pressiomtriques raliss partir
de sondes de type Mnard sont arrts avant datteindre la
pression limite du sol, dfinie comme la pression
correspondant au doublement du volume initial du trou
(normes NFP 94-110). Les tentatives pour atteindre cette
pression limite avec les matriels couramment utiliss se
soldent trs frquemment par lclatement de la sonde. Cette
situation nest videmment pas satisfaisante et a conduit
rechercher des amliorations (Cour et al., 2005).
Larticle prsente les caractristiques et les performances
de la sonde pressiomtrique Francis Cour (en abrg sonde
FC) conue de manire atteindre quasi systmatiquement le
doublement du volume de la cavit sans clatement et pour
des niveaux de pression nettement suprieurs ceux tolrs
par les sondes standard.
Des essais comparatifs mens dans diffrentes
formations, avec des matriels standards et avec la sonde FC,
illustrent les capacits de la sonde et ses performances
oprationnelles.
2. SONDE PRESSIOMTRIQUE FRANCIS COUR
La sonde pressiomtrique FC est, selon la norme NF P94-
110-1, une sonde du type G gaine souple.La longueur de la
cellule centrale est de 210mm ; celle des deux cellules de
garde est de 105mm.. Les cellules ont un diamtre extrieur
de 58 mm. La principale originalit de la sonde rside dans
l'adjonction, autour de la gaine extrieure en lastomre,
d'une gaine textile de contention qui a fait lobjet dun brevet
dpos en 2006 par Francis Cour. Cette gaine, de forme
cylindrique au repos (Figure 1b), a la proprit de se dilater
en opposant une trs faible rsistance, jusqu' atteindre un
profil limite partir duquel elle oppose une rsistance trs
leve toute dilatation complmentaire. Dans le cas de la
sonde pressiomtrique, ce profil limite, en forme de fuseau,
correspond un volume d'injection de 1100 cm
3
dans la
cellule centrale (Figure 1b).
Un dispositif similaire de gaine textile de contention a t
mis en application pour la membrane de la cellule centrale
(rsistance propre p
m
= 30 kPa). La cellule, munie de ce
dispositif prsente une section parfaitement cylindrique sur
l'ensemble de sa plage d'injection, jusqu' un volume de
1100cm
3
(Figure 1a) tout en prsentant une faible rsistance
propre.
1a
1
1
Figure 1. Vues de la membrane centrale gonfle 1100 cm
3
(1a), de
la gaine de contention non gonfle (1b) et gonfle 1100 cm
3
(1c)
La nouvelle sonde pressiomtrique est au final compose de
la membrane de la cellule centrale et de la gaine dcrites ci-
dessus, recouvertes d'une sur-gaine en polyurthane et de
lamelles mtalliques (Figure 2).
556
th
Figure 2. Sur-gaine (2a) et lamelles mtalliques (2b) constituant la
sonde FC.
Les proprits particulires de la sonde en termes de
capacit de gonflement en volume et pression, et de
robustesse, sont illustres par le diagramme d'talonnage de la
sonde, l'air libre, pouss 6 MPa (Figure 3).
Figure 3. Courbe dtalonnage de la sonde FC, mene 6 MPa.
3. ESSAIS COMPARATIFS
Des essais comparatifs ont t mens dune part avec une
sonde FC, dautre part avec une sonde classiquement utilise
par la profession sur plusieurs sites. Le mode opratoire a t
identique pour chaque sonde, et a respect les critres de la
norme NF P94-110-1.
Le tableau 1 prsente les caractristiques dtalonnage et
de calibrage raliss sur les quatre chantiers prsents par la
suite. On note que la sonde FC vrifie bien les critres de la
norme NF P94-110-1 dune sonde gaine souple.
Tableau 1. Caractristiques de la sonde FC et du systme CPV-
tubulures.
SN: sonde nue ; TF: Tube fendu ; Vs, pel
selon NF P 94-110-1
Merville Dunkerque Grand Paris Londres
FC SN FC SN FC TF FC
Vs [cm
3
] 522 484 595 475 553 510 483
p
el
[MN/m] 0.27 0.17 0.41 0.34 0.32 0.39 0.35
3.1 Argile des Flandres (Merville)
Largile surconsolide des Flandres est un sol bien connu,
particulirement homogne sur le site exprimental de
Merville. On dispose de nombreux sondages pressiomtriques
et pntromtriques (Puech et al., 2013). On a ralis titre
comparatif deux forages espacs de 5m raliss la tarire
simple 63mm sec, tubs en tte et par passes de forage de
3m. Pour les deux sondes, les conditions opratoires sont
identiques: i) pression diffrentielle (pr+ph)-pk voisine de
0.14MPa, ii) 30m de tubulures coaxiales =3mm, iii) eau
pure. Pour la mesure de grands volumes, la sonde FC a t
couple un CPV pilot PREVO (Jean Lutz S.A.).
Les rsultats sont donns dans le tableau 2. Les valeurs
de Pl* pour la sonde nue (SN) ont t calcules par
extrapolation conformment la norme NF P94-110-1 car le
doublement de la cavit (V
L
=V
S
+2V
1
) na pas t atteint (sauf
essais 7, 10 et 11m). Les valeurs en gras sont obtenues par
interpolation linaire sur les points mesurs. La valeur de Pl*
est toujours mesure directement avec la sonde FC comme
illustr sur le Figure 4 pour lessai 12m.
Tableau 2 : Essais comparatifs dans largile des Flandres
SONDE SN SONDE FC
P
r
o
f
o
n
d
e
u
r

V
L
E
M
Pl* V
L
E
M
Pl*
[m] [cm3] [MPa] [MPa] [cm3] [MPa] [MPa]
6 668 11.5 0.96 680 8.1 0.99
7 686 12.5 1.02 712 9.4 1.01
8 782 15.3 1.15 688 13.1 1.28
9 742 16.5 1.24 716 10.3 1.29
10 656 15.6 1.32 699 14.3 1.62
11 620 14.0 1.33 719 13.0 1.67
12 720 19.7 1.45 763 11.2 1.68
On constate que :
- entre 6 et 9 m les valeurs de pression limite sont en bon
accord ; au-del les valeurs obtenues par la sonde FC sont
lgrement suprieures mais restent dans la limite des
variations locales indiques par les essais au CPT
lectrique raliss proximit immdiate (Figure 5);
- les modules pressiomtriques mesurs la sonde FC sont
un peu plus faibles (E
M
,
FC
/ E
M, SN
~ 0.75) ;
- les rapports E
M
/Pl* sont en consquence plus faibles avec
la sonde FC (E
M
/Pl* ~ 8.5) quavec la sonde SN (E
M
/Pl*
~ 12.5)
0
200
400
600
800
1000
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.
V
o
l
u
m
e

c
o
r
r
i
g

[
c
m
3
]
Pression corrige [MPa]
2a
2b
0
Essai Etalonnage VL=VS+2V1
EM=11.2 MPa
P
1
=0,39
V
L
=763 cm
3
P
2
=0,96
P
L
=1,79
Figure 4 : Essai avec sonde FC dans largile des Flandres (-12m)
557
Figure 5 : Rsultats comparatifs des essais PMT et CPT Merville
3.2 Sables de Dunkerque
Sur ce site correspondant un chantier de production
classique, la sonde FC a t utilise dans les mmes
conditions de forage au bilame sous boue, et avec les mmes
critres darrt que la sonde nue SN, soit trois points aprs le
fluage. Une analyse statistique des valeurs mesures a t
mene : elle a concern 11 sondages au total (5 SN et 6 FC)
Les valeurs ont t regroupes selon 3 classes de sable (1a :
sables moyennement compacts, 1b : sables compacts, 1c :
sables trs compacts). Le tableau 3 montre que pour une mise
en uvre identique les rsultats avec les deux types de sonde
sont quivalents pour un total de 191 essais analyss.
Tableau 3. Valeurs pressiomtriques compares (site de Dunkerque)
E
M
et P
l
* en MPa.
Sonde FC Sonde standard SN
Ratios FC/SN
108
essais EM
Pl*
83
essais E
M
Pl*
E
MFC
/E
MSN
pl*
FC
/pl*
SN
1a 38 10.2 1.5 23 11.4 1.5 0.90 0.97
1b 25 23.8 3.3 32 21.5 2.9 1.11 1.17
1c 45 36.3 4.8 28 37.5 4.7 0.97 1.02
3.3 Rgion parisienne
Deux sondages pressiomtriques ont t raliss 5m de
distance sur un site de la rgion parisienne dans le cadre du
chantier Grand Paris. La sonde classique est de type G avec
tube fendu (TF) et cellule courte de 44 mm. La coupe
lithologique au droit des sondages est :
- 0-7m : alluvions anciennes : limons, sables et graves
- 7 -10m : calcaires de St-Ouen
- 10-20m : sables de Beauchamp.
Figure 6. Essai avec la sonde FC dans les alluvions (-3.5m).
La figure 6 montre lessai 3.5m de profondeur dans les
alluvions avec la sonde FC. Une expansion 700 cm
3
a
permis de mesurer directement la pression limite
Figure 7. Sondages Grand Paris : pressions et volumes atteints pour
chaque essai
Deux essais sur treize raliss avec la sonde TF ont d
tre arrts prmaturment par clatement de la membrane de
sorte que les volumes injects ont d tre limits 440 cm
3
(Figure 7). Avec la sonde FC, aucun clatement nest
dplorer pour des essais mens jusqu 800 cm
3
(limite
correspondant la capacit du CPV utilis), ou arrts 8
MPa, critre darrt fix au cahier des charges (Tableau 4).
Tableau 4. Valeurs pressiomtriques compares (sondages Grand
Paris)-
Sonde FC
Sonde TF
P
r
o
f
o
n
d
e
u
r

V
l

[
c
m
3
]
E
M
F
C

[
M
P
a
]

p
l
F
C
*
[
M
P
a
]

p
l
F
C
3
*
[
M
P
a
]

V
l

[
c
m
3
]
E
M
T
F
[
M
P
a
]

p
l
*
[
M
P
a
]

3.5 697 30 5.5 4.3 750 74 4.5
5.0 667 41 6.3 4.8 970 36 5.9
6.5 931 61 10.7 10.0 1030 36 >3.5
7.9 839 30 2.8 > 1.8 970 19 >1.7
11.0 629 102 8.4 8.0 730 55 5.8
12.5 889 105 >7.8 >5.3 710 128 >7.1
14.0 893 47 5.3 5.1 710 44 4.2
15.5 833 13 2.1 2.0 730 123 >4.1
17.0 - 39 6.1 5.6 610 69 >4.2
18.5 - 152 >7.8 >7.8 710 91 6.8
20.0 883 66 5.8 4.8 610 61 3.7
558
th
4. CONCLUSION ET PERSPECTIVES
La sonde FC constitue une avance technologique
significative dans la mesure des paramtres pressiomtriques
des sols. La conception originale de sa membrane autorise
dans la grande majorit des cas la mesure directe de la
pression limite par doublement effectif du volume de la
cavit.
Sa remarquable rsistance lclatement y compris sous
pression leve permet son emploi dans des terrains
htrognes et rsistants, correspondant au domaine
dutilisation habituel du tube fendu, et autorise ainsi sur les
chantiers des cadences accrues.
La sonde est conforme aux exigences de la norme NFP
94-110-1 pour une sonde gaine souple. Dans des conditions
oprationnelles similaires, la sonde FC et une sonde nue
standard (SN) fournissent des paramtres pressiomtriques
quivalents. Les exemples traits montrent que lextrapolation
de donnes obtenues avec des sondes standard trop faible
capacit dexpansion peut conduire une sous-estimation des
pressions limites.
La gnralisation de ce type de sonde passe par une
amlioration des matriels existants notamment en ce qui
concerne la capacit en volume et pression des contrleurs
pression volume.
On constate pour les deux sondes, une concordance
globale entre modules pressiomtriques (E
M,FC
/E
M,TF
~ 1.1).
En revanche le ratio pl*
FC
/pl*
TF
est de lordre de 1.3.
Si on reprend linterprtation des essais FC (colonne
pl*
FC3
) en limitant lextrapolation trois points aprs P2 (ce
quautorise la norme, et ce qui correspond la faon dont
sont traits la plupart des essais en France), on obtient des
valeurs pl*
FC3
systmatiquement infrieures denviron 10%
aux valeurs pl*
FC
. Lextrapolation limite 3 points conduit
ainsi une sous-estimation des pressions limites.
Corrlativement les valeurs de E
M
/Pl* se trouvent
surestimes.
3.4 Essais haute pression
Des essais haute pression (12 MPa) avec la sonde FC ont t
mis en uvre entre 54 et 66m de profondeur, dans des sables
fins trs compacts de Londres (Thantien). Les deux forages
ont t raliss au taillant en rotation 66 mm avec injection de
boue et mise en place de tubage lavancement. Des essais
haute pression avaient dj t raliss dans ces sables avec
une sonde nue standard et du matriel adapt pour la
circonstance (Massonnet, 2005). Larrt des essais 12 MPa
est d aux limites du contrleur pression-volume et non pas
la sonde qui a permis de mesurer des pressions limites
leves, sans aucun clatement pour les 18 essais. La figure 8
montre un essai 63m de profondeur avec 4 points au-del de
P
2
, ce qui autorise une extrapolation raliste de la pression
limite (Pl*= 13 MPa; E
M
= 112 MPa, pour un volume brut
V
max
= 450 cm
3
; V
max
net= 409 cm
3
).
5. REMERCIEMENTS
Les auteurs remercient la socit Jean Lutz S.A. pour la mise
disposition gracieuse de son matriel et son assistance sur
les chantiers exprimentaux de Fugro et de Fondasol.
6. REFERENCES
AFNOR. 2000. Norme NF P94-110-1. Essai pressiomtrique Mnard.
Cour F., Puech A., Durand F. 2005. Un pressiomtre de nouvelle
gnration. 2005. Proc. ISP5-PRESSIO (1), 63-73
Massonnet R. 2005. Le pressiomtre sous haute pression. Proc.
ISP5-PRESSIO (1), 81-90
Puech A. et Benzaria O. 2013. Effet du mode de mise en place sur le
comportement statique de pieux dans largile fortement
surconsolide des Flandres. Proc. 18 ICSMGE, Paris
Figure 8. Essai avec la sonde FC dans les sables du Thantien (-63m).
559
Long-term Deformation of the Reclaimed Pleistocene Foundation of the Offshore
Twin Airport
Dformations long terme dune fondation de remblai plistocne rcupr sur mer pour un projet
daroport jumel
Jeon B.G.
Samsung C&T Corporation
Mimura M.
Graduate School of Engineering, Kyoto University
ABSTRACT: A series of elasto-viscoplastic finite element analyses is performed to assess the long-term deformation including the
interactive behavior of the reclaimed Pleistocene foundation due to the adjacent construction of the offshore twin airport. Attention is
paid to the modeling of permeability for the Pleistocene sand gravel layers considering the sedimentation environment. The concept of
mass permeability is introduced to model the actual process of dissipation of excess pore water pressure in the field. It is regarded
as the macroscopic capability of permeability for the individual Pleistocene sand gravel layers by evaluating the permeability not of
each element but of the whole layer in one body. The mechanism for the propagation of excess pore water pressure due to
construction of the adjacent reclamation is discussed through the numerical procedure using the concepts of mass permeability. The
concept of mass permeability for the individual Pleistocene sand gravel layers is found to well function to assess the long-term
deformation including the interactive behavior in the reclaimed Pleistocene foundation.
RSUM : Les dformations long terme dun remblai plistocne en mer sont values a partir dune srie danalyses lasto-
viscoplastiques par lments finis. Les interactions dues aux travaux damnagement daroport jumel sont aussi prises en compte.
On vise plus particulirement modliser la permabilit du sable/gravier plistocne en considrant la sdimention du milieu. La
dissipation des surpressions interstitielles in-situ est calcule partir dune permabilit massique de lensemble des couches
sable/gravier. Les mcanismes de propagation de surpressions interstitielles induites par le remblai voisin sont dtermins par
modlisation numrique faisant appel au concept de permabilit massique. Lapplication de ce concept semble tre commode pour
valuer les dformations long terme des couches sable/gravier plistocne en interaction avec dautres ouvrages voisins.
KEYWORDS: elasto-viscoplastic finite element analysis, mass permeability, standard hydraulic gradient
1 INTRODUCTION
The development of coastal areas accomplished in Japan has
been outstanding. Kansai International Airport (KIX) was
constructed in Osaka Bay as two man-made reclaimed islands to
minimize noise and pollution in residential areas as well as to
meet the increasing demand for air transportation. Such a large-
scale offshore reclamation in Osaka Bay is accompanied with
large and rapid settlement of deep Pleistocene clay deposits
(Mimura et al., 2003). Long-term settlement of the Pleistocene
marine foundations due to huge reclamation load has been of
great concern in this project. The seabed deposits of Osaka Bay
have been formed due to the soil supply from the rivers and the
alternating deposits of KIX have been formed due to
sedimentation of clayey soils during transgression and of sandy
to gravelly soils during regression on the sinking base of Osaka
Bay. The Pleistocene clay deposited in Osaka Bay exhibits the
behavior of the quasi-overconsolidated clay without definite
mechanical overconsolidation history. Itoh et al. (2001)
summarized on the basis of the data from elastic wave
exploration and in-situ boring logs that the Pleistocene sand
gravel deposits are not always distributed uniformly in
thickness, consistently and that the amount of fine contents
included in them is significant. The most serious problem
originating from these sand gravel deposits is the permeability
that controls the rate of consolidation of sandwiched Pleistocene
clays. In the sense, the modeling for the quasi-overconsolidated
Pleistocene clay and the evaluation of permeability for the
Pleistocene sand gravel deposits are the significant factors to
assess the long-term behavior of the reclaimed Pleistocene
foundation due to the reclamation of the offshore twin airport.
Mimura and Jang (2004) proposed a concept of compression in
which viscoplastic behavior is assumed to occur even in the
quasi-overconsolidated region less than p
c
for the Pleistocene
clays in Osaka Bay. The procedure has been found to be
versatile and allows for the long-term settlement monitored in
the reclaimed islands in Osaka Port to be described (Mimura
and Jang, 2005a). In the present paper, the numerical procedure
to assess the long-term behavior of the Pleistocene deposits at
KIX in terms of elasto-viscoplasitc FEM is proposed by
introducing the concept of mass permeability and standard
hydraulic gradient for the Pleistocene sand gravel layers. The
validity of the procedure is carefully discussed by comparing
the performed results with in-situ measurements.
2 CONCEPTS OF MASS PERMEABILITY AND
STANDARD HYDRAULIC GRADIENT
Mimura and Jang (2005a) reported when the permeability of
sand gravel layers is considered perfectly drained, one-
dimensional analysis only considering the characteristic of
clayey soil can be adopted for the consolidation problem
without considering the effect of permeability loss in the those
sand gravel layers. However, the sand gravel layers sandwiched
by the Pleistocene clay layers at KIX were recognized not to
function as perfect drainage layers through the in-situ
measurement of excess pore water pressure. Therefore, the two
or three-dimensional analysis that considers the permeability of
the Pleistocene sand gravel layers is required to assess the long-
term behavior of the reclaimed Pleistocene foundation. The
influential factors to evaluate the permeability of sand gravel
layers are the thickness, the horizontal continuity and the fine
contents of them. The permeability of them is different with
places even if they are categorized as the identical ones. But, it
is impossible to evaluate the permeability of sand gravel layers
at every point. It is also very difficult to confirm how the sand
gravel layers under the Pleistocene marine foundation are
560
th
distributed in practice. The concept of mass permeability is
proposed to evaluate the permeability not of each element but of
the whole layer in one body. It is regarded as the macroscopic
capability of permeability for the individual sand gravel layers
by considering the horizontal continuity, the change in thickness
and the degree of fine contents of them. Mimura and Jeon
(2011) evaluated the mass permeability of the Pleistocene sand
gravel layers using the simple foundation model as shown in
Fig.2. The distribution of sand gravel layers not only in the
loading area but also in the area that can rule out the effect of
the hydraulic boundary condition should be considered to assess
the mechanism of the propagation/dissipation of excess pore
water pressure in the coupled stress-flow analysis. In the sense,
on the basis of the assumption that the hydraulic gradient
derived in the representative foundation model having the
horizontally even layer with constant thickness is regarded as
the standard one for the individual Pleistocene sand gravel
layers, the evaluated mass permeability can be the
representative of the capacity of permeability for the individual
Pleistocene sand gravel layers at KIX. The standard hydraulic
gradient is hence applied to the geologically genuine foundation
model that has been developed to consider the actual stress level
not only of the monitoring point but also of the considered area
for the numerical analysis. Due attention should be paid to the
fact that this assumption is only considered in horizontal
position for the individual Pleistocene sand gravel layers.
3 FOUNDATION MODEL AND HYDRAULIC
BOUNDARY
The differential settlement of the individual Pleistocene clay
layers as well as the excess pore water pressure at various
depths, both in the clay and the sand gravel layers, have been
measured at a lot of points of KIX. Figure 1 shows the plan
view of KIX together with the location of representative
monitoring points on the 1
st
phase island. A series of elaso-
viscoplastic finite element analyses is carried out along the
representative section shown by A-A at monitoring point 1 in
Fig.1. Figure.2 shows the representative foundation model
assumed to be horizontally even layer that have a constant
thickness and continuous layer based on the boring data at the
monitoring point 1. Figure.3 shows the geologically genuine
foundation model having the inclined base and layers that is
constructed based on the soil exploration and geological survey
data (Kitada et al, 2011). The clay layers increase in thickness
towards the offing and the sand gravel layers drastically change
in thickness horizontally. The continuity of the individual layers
is still guaranteed even for the geologically genuine foundation
model in the present study. Here, Ma and Ds denote marine clay
and Pleistocene sand gravel layer respectively. Ma13 is the
Holocene marine clay whereas others are the Pleistocene origin.
For the Holocene clay deposit, Ma13, sand drains are driven in
a rectangular configuration with a pitch of 2.0 to 2.5 meters to
promote consolidation. The lateral boundary of the clay layers is
assumed to be undrained while the one of the sand gravel layers
is assumed to be fully drained. Mimura and Jang (2005b)
reported that when the distance to the boundary is set to be
about 10 times of the loading area, the effect of the hydraulic
boundary condition can be ruled out. Based on the findings, the
same condition is satisfied even for the foundation models used
in the present study. The distance to the offshore and onshore
boundary is set to be 10,000m and 5,000m respectively. The
present two foundation models are divided into finite element
mesh consisting of 8,580 nodal points and 8,378 elements.
4 LOADING CONDITION AND SOIL PARAMETERS
The prescribed final overburden due to airport fill construction
amounts to about 430kPa at the 1
st
phase island and about
530kPa at the 2
nd
phase island respectively. The 2
nd
reclamation
is started after about 13years from the 1
st
reclamation. In the
present analysis, the permeable capability evaluated from the
concept of mass permeability for the Pleistocene sand gravel
layers is applied for the present finite element analysis. On the
basis of the findings by Itoh et al. (2001), the relatively high
permeable capability are assumed for Ds1,3 10 because they
have been evaluated as gravelly, horizontally continuous and
Monitoring point 1
N
S
2 phase island
1 phase island
st
nd
S2
S3
(S1)
A (Onshore side)
A' (Offshore side)
(
R
e
p
r
e
s
e
n
t
a
t
i
v
e
s
e
c
t
i
o
n
)
Figure 1. Plan view of Kansai International Airport and the location of
monitoring points on the 1
st
phase island
(
O
n
s
h
o
r
e
s
i
d
e
)
(
O
f
f
s
h
o
r
e
s
i
d
e
)
1 phase island
st
2 phase island
nd
17,880m
10,000m 1440m 5,000m
Ma13
Ds1
Dtc
Ma12
Ds2
Ds3
Ma11U
Ma11L
Ds4
Ds5
Ds6
Ds7
Ds8
Ds9
Ds10
Ma10
Ma9
Doc5&Ma8
Ma7
Ma6
1
4
8
.
0
m
Monitoring point 1
1
3
1
.
0
m
1
7
.
0
m
A' A
1440m
Fi
gure 2. Representative foundation model of KIX for finite element analysis
at representative section
10,000m
1440m
5,000m
A' A
1 phase island
Monitoring point 1 5
8
8
.
7
m
2
0
.
2
m
17,880m
Ma13
Ds1
Dtc
Ma12
Ds2
Ds3
Ma11U
Ma11L
Ds4
Ds5
Ds6
Ds7
Ds8
Ds9
Ds10
Ma10
Ma9
Doc5&Ma8
Ma7
Ma6
st
2 phase island
nd
1440m
(
O
n
s
h
o
r
e
s
i
d
e
)
(
O
f
f
s
h
o
r
e
s
i
d
e
)
Figure 3. Geologically genuine foundation model of KIX for finite
element analysis at representative section
561
having enough thickness. On the other hand, very low
permeable capability is assumed for Ds6 and 7 that have been
evaluated to have insufficient thickness with high degree of fine
contents and poorly continuous. The other layers have been
evaluated as the ordinary permeable capability. The used all soil
parameters for analysis are also exactly the same with that used
by Mimura and Jeon (2011).
5 RESULTS AND DISCUSSIONS
The calculated distribution of excess pore water pressure before
and after the construction of the 2
nd
phase island is shown in
Fig.4 for two foundation models respectively. As shown in
Fig.4, the similar distribution tendency of excess pore water
pressure can be seen for two foundation models. It should be
noted that a large amount of excess pore water pressure still
remains undissipated in the middle Pleistocene clay layers,
Ma10, 9 and Doc5&Ma8 as well as sand gravel layers, Ds6 and
7 before the construction of the 2
nd
phase island because of poor
permeability of sand gravel layers, Ds6 and 7. In contrast, the
excess pore water pressure in the upper and lower Pleistocene
layers such as Dtc, Ma12,11,7,6 and Ds1,3,9,10 is
monotonically dissipated with time because of high
permeability of sand gravel layer, Ds1,3 and 10. At the
completion of the 2
nd
reclamation, a large amount of excess
pore water pressure is concentrated in the upper and middle
Pleistocene layers such as Ma12, 10, 9 and Doc5&Ma8 beneath
the foundation of the 2
nd
phase island. Here, a due attention
should also be paid to the fact that the increased excess pore
water pressure beneath the foundation of the 2
nd
phase island is
propagated to that of the 1
st
phase island. Since the permeability
of the upper and lower Pleistocene sand gravel layers is higher
than the one of the middle layers, a larger amount of excess
pore water pressure in the upper and lower Pleistocene layers is
propagated compared to the one in the middle layers of the
foundation of the 1
st
phase island.
The calculated horizontal distribution of excess pore water
pressure in the representative Pleistocene sand gravel layers
(Ds3, 6, 10) are shown in Fig. 5 at the time before and after the
construction of the 2
nd
phase reclamation for both foundation
models. In the present study, the identical permeable capability
for the individual Pleistocene sand gravel layers in two
foundation models is applied by considering the concepts of
mass permeability and standard hydraulic gradient.
However, in Fig.5, it should be noted that the distribution of
excess pore water pressure near the 1
st
phase island almost
shows a good match for two foundation modes by applying the
concept standard hydraulic gradient whereas the one of the
other region shows the discrepancy distribution with the stress
level. The stress level beneath the foundation of the 1
st
phase
island is almost the same for two foundation models because the
representative model was developed based on the monitoring
point 1 whereas the one beneath the foundation of the 2
nd
phase
island is different each other due to change in thickness of
geologically genuine foundation model. It is noteworthy that
1 phase 2 phase
nd st
0 2000 4000 6000 8000 10000 12000 14000 16000
17880
Distance(m)
-160
-18
-40
-80
-100
-120
-140
-60
E
l
e
v
a
t
i
o
n
(
m
)
-166
Ma13
Ds1
Dtc
Ma12
Ds2
Ds3
Ma11U
Ma11L
Ds4
Ds5
Ds6
Ds7
Ds8
Ds9
Ds10
Ma10
Ma9
Doc5&Ma8
Ma7
Ma6
Completion of 2
nd
phase reclamation
(after 19years from construction)
1 phase
st
-160
-18
-40
-80
-100
-120
-140
-60
E
l
e
v
a
t
i
o
n
(
m
)
-166
0 2000 4000 6000 8000 10000 12000 14000 16000
17880
Distance(m)
Ma13
Ds1
Dtc
Ma12
Ds2
Ds3
Ma11U
Ma11L
Ds4
Ds5
Ds6
Ds7
Ds8
Ds9
Ds10
Ma10
Ma9
Doc5&Ma8
Ma7
Ma6
Before construction of 2
nd
phase reclamation
(Unit: kPa)
Upper
Middle
Lower
Figure 4(a). Contour of excess pore water pressure for representative
foundation model at before and completion of 2
nd
phase reclamation
0 2000 4000 6000 8000 10000 12000 14000 16000
17880
Distance(m)
-30
-619
-80
-130
-180
-230
-280
-330
-380
-430
-480
-530
-580
E
l
e
v
a
t
i
o
n
(
m
)
-6
-26
Ma13
Ds1
Dtc
Ma12
Ds2
Ds3
Ma11U
Ma11L
Ds4
Ds5 Ds6
Ds7
Ds8
Ds9
Ds10
Ma10
Ma9
Doc5&Ma8
Ma7
Ma6
1 phase
st
2 phase
nd
0 2000 4000 6000 8000 10000 12000 14000 16000
17880
Distance(m)
-30
-619
-80
-130
-180
-230
-280
-330
-380
-430
-480
-530
-580
E
l
e
v
a
t
i
o
n
(
m
)
1 phase
st
-6
-26
Ma13
Ds1
Dtc
Ma12
Ds2
Ds3
Ma11U
Ma11L
Ds4
Ds5 Ds6
Ds7
Ds8
Ds9
Ds10
Ma10
Ma9
Doc5&Ma8
Ma7
Ma6
Upper
Middle
Lower
Before construction of 2
nd
phase reclamation
Completion of 2
nd
phase reclamation
(Unit: kPa)
Figure 4 (b). Contour of excess pore water pressure for geologically
genuine foundation model at before and completion of 2
nd
phase
reclamation
500
400
300
200
100
0
0 3000 6000 9000 12000 15000 18000
Ds10
P.W.P
Horizontal distance (m)
500
400
300
200
100
0
Ds6 P.W.P
500
400
300
200
100
0
Ds3 P.W.P
Center of 1
phase island
st
Center of 2
phase island
nd
Representative foundation
Geologically genuine foundation
E
x
c
e
s
s
p
o
r
e
w
a
t
e
r
p
r
e
s
s
u
r
e
(
k
P
a
)
500
400
300
200
100
0
0 3000 6000 9000 12000 15000 18000
Horizontal distance (m)
500
400
300
200
100
Center of 1
phase island
st
Center of 2
phase island
nd
E
x
c
e
s
s
p
o
r
e
w
a
t
e
r
p
r
e
s
s
u
r
e
(
k
P
a
)
Ds3
Ds10
P.W.P
500
400
300
200
100
0
P.W.P
0
Ds6 P.W.P
(a) Before the 2
nd
phase reclamation (b) Completion the 2
nd
phase reclamation
Figure 5. Horizontal distribution of excess pore water pressure for the representative Pleistocene sand gravel layers (Ds3, 6, 10) in a horizontal position
562
th
although the identical permeable capability for the individual
Pleistocene sand gravel layers was applied, the calculated
results of excess pore water pressure could show the difference
with the stress level. The calculated excess pore water pressure
time relations for two foundation models are shown in Fig. 6
together with the measured results for the representative
Pleistocene sand gravel layers at the monitoring point 1. It is
noteworthy that the excess pore water pressure in the upper
(Ds3) and lower (Ds10) Pleistocene sand gravel layers is
increased but the one of the middle layer (Ds6) is not increased
due to the construction of the 2
nd
phase island. The long-term
settlement associated with the phenomenon of propagation of
excess pore water pressure is another serious problem for KIX.
When the excess pore water pressure increases or the
dissipation of excess pore water pressure is hindered due to the
construction of the 2
nd
phase island, the settlement is also
retarded or slight upheaval can happen (see Fig.7). It is also
found that the calculated performance at the monitoring point 1
shows a good match for two foundation models by applying the
concept of standard hydraulic gradient and can also well
describe the whole process of deformation.
6 CONCLUSIONS
The long-term deformation of the reclaimed Pleistocene
foundation of the offshore twin airport was numerically
evaluated through the elasto-viscoplastic finite element analyses
considering the concepts of mass permeability and standard
hydraulic gradient for the Pleistocene sand gravel layers. The
concept of mass permeability was evaluated as the
representative permeable capacity of sand gravel layers of KIX.
The representative permeable capacity of sand gravel layers was
applied to the geologically genuine foundation model by
introducing the concept of standard hydraulic gradient for the
coupled stress-flow analysis. The concept of mass permeability
for the sand gravel layers was found to well function to assess
the process of excess pore water pressure generation/
dissipation/propagation and long-term settlement in the
reclaimed foundations of KIX. The concept of standard
hydraulic gradient was also found to well reproduce the
representative permeable capacity by comparing the calculated
results for two foundation models. The validity and objectivity
of the proposed concepts will be investigated by applying them
to the additional review sections including the monitoring
points S2 or S3 shown in Fig. 2.
7 REFERENCES
Itoh, Y., Takemura, K., Kawabata, D., Tanaka, Y. and Nakaseko, K.
2001. Quaternary Tectonic Warping and Strata Formation in the
Southern Osaka Basin Inferred from Reflection Seismic
Interpretation and Borehole Sequences, Journal of Asian Earth
Science, 20, 45-58.
Kitada, N., Inoue, N., Takemura, K., Fukada, K. and Emura, T. 2011.
Subsurface Structure Model Around Kansai Airport According to
Re- Interpretation of Borehole Data based on Result of KIX18-1
Core. International Symposium on Advances in Ground Technology
and Geo-Information, IS-AGTG, 137-142.
Mimura, M. and Jang, W.Y. 2004. Description of time-dependent
behavior of quasi-overconsolidated Osaka Pleistocene clays using
elasto-viscoplastic finite element analyses, Soils and Foundations,
44(4), 41-52.
Mimura, M. and Jang, W.Y. 2005a. Verification of the Elasto-
viscoplastic Approach Assessing the Long-term Deformation of the
Quasi-overconsolidated Pleistocene Clay Deposits, Soils and
Foundations, 45(1), 37-49.
Mimura, M. and Jang, W.Y. 2005b. Long-term Settlement of the
Pleistocene Deposits due to Construction of KIA, Proceedings of
the Symposium on Geotechnical Aspects of Kansai International
Airport, , 77-85.
Mimura, M. and Jeon, B.G. 2011. Numerical Assessment for the
Behavior of the Pleistocene Marine Foundations Due to
Construction of the 1
st
Phase Island of Kansai International Airport,
Soils and Foundations, 51(6), 1115-1128.
Mimura, M., Takeda, K., Yamamoto, K., Fujiwara, T. and Jang, W.Y.
2003. Long-term settlement of the reclaimed quasi-
overconsolidated Pleistocene clay deposits in Osaka Bay, Soils and
Foundations 43(6), 141-153.
Ds10
0
50
100
150
200
250
300
350
400
Elapsed time(years)
0 10 20 30 40 50 60
Measured
Ds3
0
50
100
150
200
250
300
350
400
E
x
c
e
s
s
p
o
r
e
w
a
t
e
r
p
r
e
s
s
u
r
e
(
k
P
a
)
Completion of 2
phase island
nd
Start of 2
phase island
nd
Ds6
0
50
100
150
200
250
300
350
400
Figure 6. Comparison of measured and calculated excess pore
water pressure with time for the representative Pleistocene sand
gravel layers
Ma6
0
20
40
60
80
100
120
140
Elapsed time(years)
0 10 20 30 40 50 60
160
Measured
S
e
t
t
l
e
m
e
n
t
(
m
)
Completion of 2
phase island
nd
Start of 2
phase island
nd
Ma10
0
20
40
60
80
100
120
140
Ma12
0
20
40
60
80
100
120
140
160
180
200
Figure 7. Comparison of measured and calculated settlement with time
for the representative Pleistocene clay layers
563
Assessment of Scour Potential of a Circular Pier in Silty Sand Using ISEEP
Caractrisation par ISEEP du potentiel d'rosion d'une pile circulaire dans un sable silteux
Kayser M., Gabr M.
Department of Civil, Construction and Environmental Engineering
North Carolina State University, Box 7908, Stinson Drive, Raleigh, NC 27695-7908; PH. (919)515-7904;
ABSTRACT: Work in this paper describes an approach for the assessment of soil scour potential through the use of an In Situ
Erosion Evaluation Probe (ISEEP) that is advanced by water jetting. Soil erosion parameters are assessed for silty sand in terms
of critical stream power, and therefore, critical shear stress, and detachment rate coefficient. Scour depth around a circular bridge
pier was computed using ISEEP data and compared with an empirical approach available in literature for estimating scour depth
in soil similar to the tested in the study. The application of the idea and the utility of this technique to assess scourability profile
are presented and discussed.
RSUM: Le travail prsent dans cet article dcrit une approche pour lvaluation du risque d'affouillement dun sol in situ en
utilisant une sonde drosion quipe de jet deau. Les paramtres drosion sont valus pour les sables limoneux en fonction
dune puissance critique et par consquent en termes dune contrainte de cisaillement critique et dun coefficient exprimant le
taux de dtachement. La profondeur daffouillement autour dune pile de pont a t calcule en utilisant les donnes issues de la
sonde. Elle a t compare celle issue de lapproche empirique pour un sol similaire au sol tudi. Lapplicabilit de lapproche
propose et son utilit pour lvaluation du profil drodabilit sont prsentes et discutes.
KEYWORDS: Bridges, Erosion, Foundation, In Situ, Pier, Probe, Scour, Shear, Soil
1 INTRODUCTION
The assessment of scour and erosion rates of soil profiles
supporting hydraulic structures and critical bridges is vital for
ensuring safe performance under normal flow conditions, as
well as the integrity of their foundation systems during and
after severe storms. Richardson and Davis (2001) highlighted
the importance of assessment of local scour around bridge
piers as it is one of the most common causes of bridge
failure. Several approaches ranging from simple steel
sounding rods to remote sensing have been developed to
assess scour depth after it has occurred. As presented by Lu
et al. (2008) the more sophisticated approaches, including
acoustic doppler and ground penetrations radars, have a high
cost and require frequent maintenance and repair. Even then,
these approaches do not provide an estimate of scour under
future storm events. Current techniques for providing such
information require either the removal of soil samples for
laboratory testing, in a device such as the Erosion Function
Apparatus (EFA) by Briaud et al. (2001), or limiting the
measurements to erodibility of the surface sediments.
Gabr et al. (2012) presented a prototype device, termed
ISEEP (In Situ Erosion Evaluation Probe), for assessment of
scour parameters with depth. ISEEP has been constructed as
simple stainless steel tubes fitted with truncated cone tip. The
cone-tipped vertical probe is attached to a digitally controlled
centrifugal pump that provides controllable and repeatable
water velocity at the tip, with sustained flow rate against any
induced back pressure. As the water jet is induced through
the cone tip, it mobilizes the soil particles. The test data are
analyzed using the stream power (bed shear stress multiplied
by the flow velocity) concept proposed by Annandale (2006)
to account for the nature of the flow conditions induced
during testing. The results from the tests are reduced to
provide critical shear stress (
c
) and a rate of scour per unit
shear stress (k
d
). These two values are used in conjunction
with the applied shear stresses (
applied
) per a given flow type
and as appropriate to the structure being analyzed, to
compute the scour rate (E) using the excess shear model as
follows (Annandale 2006):
) - ( k = E
applied d c
(1)
In this study, experimental work and analyses are
conducted, using ISEEP-estimated data, for evaluating
erosion parameters for a soil with 15% clay and 85% sand.
The soil is classified as silty sand according to the Unified
Soil Classification System. Tests are performed with
different jet velocities and critical stream power value (P
c
)
and the corresponding
c
, and k
d
are evaluated using the data
reduction scheme proposed by Gabr et al. (2012). An
example showing the computation of the scour depth around
a bridge pier using ISEEP-estimated data is presented. The
results are compared with values obtained using empirical
equations reported in literature, and the estimated scour depth
using both approaches is presented and discussed.
2 BRIDGE PIER SCOUR
The magnitude and geometry of local scour at bridge piers in
soil profiles with percent fines content have been
documented in literature (e.g. Hosny 1995, Molinas and
Hosny 1999, Briaud et al. 1999, 2001, and 2004). Hosny
(1995), and Molinas and Hosny (1999) proposed empirical
equations to assess scour depth for saturated and unsaturated
compacted soil with a percent fines that lends a degree of
cohesion to the soil. They reported that the scour depth
decreased as compaction density was increased for the
unsaturated cohesive soil conditions, and scour depth
564
th
decreased with the decrease of the initial water content for
the saturated cohesive soil. Briaud et al. (1999, 2001 and
2004) presented a method termed SRICOS for predicting
scour in cohesive soils, with such an approach being the
most comprehensive to date in literature.
Scouring experiments around cylinders using clay
sand mixtures were carried out by Hosny (1995), Ansari et al.
(2002), and Debnath and Chaudhuri (2010a, 2010b), among
others, with fines fraction in the range of 0.050.4, 0.10.6,
0.21.0 and 0.050.35, respectively. Hosny (1995), and
Debnath and Chaudhuri (2010a, 2010b) concluded that
maximum scour depth decreases with the increase of fines
content whereas Ansari et al. (2002) indicated that the
maximum equilibrium scour depth in sediments with fine
contents could be higher than that of non-cohesive sediments
under similar experimental conditions. Perhaps one reason
for the difference in conclusions is attributed to the nature of
fine being used in the study. In Ansari et al.s (2002) study,
the soil is reported as having zero Plasticity Index (PI). Table
1 shows several empirical equations to estimate scour depth,
with the corresponding fines fraction and Froude number
range for their applicability. The equations proposed by
Hosny (1995), and Debnath and Chaudhuri (2010a, 2010b)
are only applicable for a rather narrow Froude number (i.e.
0.13 0.33) range, in comparison to Ansari et al.s (2002)
range.
Table 1. Empirical Equations for Estimating Scour Around Bridge Piers for Soils with Fine Contents
Reference Equation Conditions Comment
Hosny (1995) d
s
/b =18.9(F
r
/(1+C))
2

C 0.4 and
0.18 F
r
0.33
b = pier diameter, F
r
= Froude number = V/(gd)
0.5
, V =
approach flow velocity, g = gravitational acceleration,
d = depth of flow and C = clay fraction.
Ansari et al.
(2002)
d
smc
/d
sms
= 1.51(C
*
/
*
)
0.2

PI = 0 and
0.16 F
r
0.69
d
smc
= maximum scour depth for cohesive sediments,
d
sms =
maximum scour depth for cohesionless sediments
(estimated using equation proposed by Kothyari et al.
(1992)), C
*
= [%P
c
. C
u
] / [(
s
-
w
).d
a
],
*
= [%P
c
. tan
c

+ (1 %P
c
) * tan
s
] / tan
s
, %P
c
= percentage of clay
content, C
u
= undrained shear strength of soil,
s
= unit
weight of soil, w
= unit weight of water, d
a
= arithmetic
mean diameter,
c
= angle of internal friction for clay
and
s
= angle of internal friction for sand.
Debnath and
Chaudhuri (2010b)
d
s
/b = 8.2F
r
0.79
C
-0.28
(IWC)
0.15
(
s
/V
2
)
-0.38

0.13 F
r
0.20,
W.C. 0.4, C 0.4
and 0.78 1.65
C = clay fraction, IWC = initial water content,
s
= bed
shear strength, = density of water, = V/V
cs
, V
cs
is
critical threshold velocity for sand and V = approach
flow velocity.

3 EXPERIMENTAL PROGRAM
Testing was conducted in a circular chamber with a diameter
of 1.0 m (3.3 ft) and a depth of 1.0 m (3.3 ft). Two 1.5 m long
probe sections, with the bottom section fitted with 19 mm
truncated tip, were used for testing. Figure 1 shows the probe
set up prior to testing.
3.1 Test Soil
The test soil was composed of 15% fine grained particles and
85% sand by dry weight. Percent dispersion of the fine
grained fraction was estimated by performing Double
Hydrometer test. Percent dispersion is the ratio of the dry
mass of particles smaller than 0.005 mm diameter, without a
chemical dispersant, to the same type of data from the
hydrometer test but with a chemical dispersant, expressed as
a percentage. A dispersion value higher than 50% was
obtained for the fine grained soil, and therefore the fine
fraction is classified as dispersive. The sand and the fine soil
components were mixed thoroughly with an electrical mixer,
in a drum, until a uniform mix was obtained. The mixing
process was repeated after the soil was transferred to the test
chamber (shown in Figure 1).
Figure 1. Photograph of the Probe

Set Up in the Chamber Prior to
Testing
The chamber was filled up to 1m mark with the silty sand

soil and approximately 0.45 kN weight was applied on the
top, to induce consolidation, for a week. Several specimens
were then retrieved for physical characterization of the test
soil. Initial water content of the mixture ranged of 18% -
23%. The results from the particle size analysis for three
types of soils are shown in Figure 2, with the test soil
designated as Silty Sand (all soils have been designated
according to Unified Soil Classification System). Table 2
shows the physical and strength properties of the test soil,
with the undrained strength estimated using the Fall Cone
test.
565
0
10
20
30
40
50
60
70
80
90
100
0.001 0.01 0.1 1 10
P
e
r
c
e
n
t

P
a
s
s
i
n
g
Particle Size (mm)
Poorly
Graded Sand
Sandy Clay
Silty Sand

Table 2. Properties of the Test Soil
4 TEST DATA INTERPRETATION
Figure 3 shows the results from the testing using four
different run times. Based on the results from previous
testing in a sand pit, Gabr et al. (2012) evaluated a critical
stream power (P
c
) value = 24 Watts/m
2
for a sand with D
50
=
0.30 mm. Using a similar technique of extrapolation
approach, the data in Figure 3 is extrapolated to zero
penetration rate to yield an average P
c
value of 16 Watts/m
2

for the test soil. Similar to the observation by Gabr et al.
(2012), a minimum of 45 sec run time is needed to provide
a reliable measurement of the penetration rate. To calculate
critical shear velocity from the P
c
, the following equation is
used (Annandale, 2006):
qh P
c
(2)
where, is unit weight of water, q is the discharge per unit
area, and h is the hydraulic head including the jet velocity
head.
0
0.5
1
1.5
2
2.5
3
3.5
10 100
P
e
n
e
t
r
a
t
i
o
n

r
a
t
e

(
c
m
/
s
)
Stream Power (Watts/m
2
)
Run time = 15s
Run time = 30s
Run time = 45s
Run time = 60s
Figure 3. Extrapolation of the Stream Power to Assess

Critical Stream Power Value (P
c
)
Using Equation 2, a critical velocity is back calculated as
0.32 m/s corresponding to a critical stream power value of 16
Watts/m
2
. As flow field changes around the pier, and
therefore the flow-related shear stress, the shear stress below
which no scour is assumed to take place is estimated using
the equation proposed by Briaud et al. (1999):
)
10
1
log
1
( 094 . 0
2
max

e
R
V

where, is water density, R
e
= VD/ is Reynolds number, V
is depth average flow velocity at the location of the pier if the
bridge were not there, D is pier diameter and is kinematic
viscosity of water. In this case, assuming a pier diameter of 1
m and a depth of flow= 2 m, the critical shear stress value is
estimated equal to 1.75 Pa.
Figure 4 shows the equivalent penetration rate per shear
stress function for the averaged test data. During testing, the
probe tip is in close proximity of the soil mass, and erosion
occurs within the jet potential core. The applied shear stress
in this case is estimated using the relationship presented by
Annandale (2006) as:
2
U C
f
(4)
where: = applied shear stress to bed in N/m
2
, U = average
velocity of water at the tip (m/s), = density (kg/m
3
), and C
f

is a friction coefficient = 0.016 according to Annandale
(2006).

Figure 4. Computed Scour Around Circular Bridge Pier on Silty
Sand Bed; Range between Arrows is Values Estimated using
Empirical Equation .
The slope of the data for the 45 and 60 secs in Figure 3
provides a parameter equivalent to the detachment rate
coefficient (k
d
) proposed by Mehta (1991). The k
d
values of
0.017 cm/sec per N/m
2
and 0.015 cm/sec per N/m
2
are
estimated, respectively, for the test soil at run times of 45 and
60 seconds. In comparison, k
d
values of 0.017 cm/sec per
N/m
2
and 0.013 cm/sec per N/m
2
were observed respectively
for sand at run times of 45 and 60 seconds by Gabr et al
(2012). The K
d
value obtained at 60 sec for the test soil is
approximately 13% higher than the value obtained for sand.
This observation agrees with the conclusion made by Ansari
et al. (2002), where the authors indicated that in a lower fines
content (<20%) type of soil, non-plastic fine particles are
carried away as the resistance due to cohesion becomes
insignificant.
5 SCOUR AROUND BRIDGE PIERS
Local scour around bridge piers occurs due to induced shear
stresses associated with flow field changes. Ettema et al.
Dry unit
weight
(kN/m
3
) =
17.7
Mean Particle
Diameter, D
50
(mm) = 0.26
Undrained Shear
Strength, C
u
(kPa)
=5-8
PI=
Non-
Plastic

Figure 2. Grain Size Distribution of Test Soil: Silty
Sand
566
th
(2011) indicated that the estimation of time-dependent clear
water scour magnitude at bridge piers remains a challenge
during the limited duration of excessive flow as, for example,
in the case of a storm surge. Equation 3 proposed by Briaud
et al. (1999) provides an estimate of the
max
as a function of
the flow velocity at a round pier. However, in order to assess
the scour depth with time using ISEEP data, a reduction in
max
with the progression of scour depth is needed. While a
significant number of studies have been performed for the
assessment of maximum scour at piers, these approaches
were not specifically concerned with evolution of shear
stresses with time and flow field. Data presented by Briaud
et al. (1999) indicated a nearly linear relationship between
the
t
max
and scour depth) /d (pier diameter), where
t
is
the shear stress with the progression of scour depth with a
minimum value of
critical
. An iterative approach is used to
estimate
t
since the maximum depth of scour is not known
apriori.
The scour depth around a bridge pier is estimated for a
flow velocity range of 1.0 m/s to 2.0 m/s (Froude number
0.23 to 0.45) with a pier diameter of 1 m and a depth of flow
= 2 m. The computations are performed based on the ISEEP
data and compared with the values from the empirical
equation by Ansari et al. (2002), as the conditions for Ansari
et al. (2002) empirical equation are in agreement with the
percent fines in the test soil. Figure 4 shows scour depth,
normalized with respect to the pier diameter ratio, versus
time for different flow velocities. The scour depth from the
equation by Ansari et al. (2002) is within 4.6-8 D (D = pier
diameter) for a flow velocity range of 1.0 m/s to 2.0 m/s,
which is higher than values estimated using the ISEEP data.
A reason for the deviation can be attributed to the fact that
the maximum shear stress equation developed by Briaud et
al. (1999) was for clay, while the soil in this study is 85%
sand. Furthermore, the application of Ansari et al. (2002)
approach required the definition of the scour level in sand
first which can widely vary depending on the parameters
assumed.
6 SUMMARY AND CONCLUSIONS
The ISEEP approach developed by Gabr et al. (2012) was
used to provide parameters for evaluating scour potential
with time in a 15%-85% silty sand mixed bed. Soil erosion
parameters included critical shear stress and detachment rate
coefficient. Higher detachment rate was obtained for the silty
sand than the sand soil. Application of the ISEEP data to
assess magnitude of scour with time for a circular bridge pier
indicated a scour depth on the order of 1-6 m versus 4.6-8 m
estimated by an empirical equation in literature. The
difference in results is attributed to the difference in the
approach for scour computation and the limitations of
estimating the evolution of shear stresses with time and flow
field. In this case, the applicability of the empirical equation
was somewhat limited since the testing conditions deviated in
terms of soil type and moisture conditions.

7 ACKNOWLEDGEMENTS
This work is supported by the US Department of Homeland
Security under Award Number: 2008-ST-061-ND 0001. The
views and conclusions contained in this document are those
of the authors and should not be interpreted as necessarily
representing the official policies, either expressed or implied,
of the U.S. Department of Homeland Security.
8 REFERENCES
Annandale, G. W. and Parkhill, D. L. 1995. Stream Bank Erosion:
Application of the Erodibility Index Method. Proceedings of
International Water Resources Engineering Conference, 2,
1570-1574.
Annandale, G.W. 2006. Scour Technology: Mechanics and Practice.
McGraw Hill, New York.
Ansari, S.A., Kothyari, U.C. and Ranga Raju, K.G. 2002. Influence
of Cohesion on Scour around Bridge Piers. Journal of
Hydraulic Research, 40 (6), 717729.
Briaud, J.-L, Ting, F. C. K., Chen, H. C., Gudavalli, R., Perugu, S.
and Wei, G. 1999. SRICOS: Prediction of Scour Rate in
Cohesive Soils at Bridge Piers. Journal of Geotechnical and
Geoenvironmental Engineering, 125 (4), 237246.
Briaud, J.-L., Ting, F.C.K., Chen, H.C., Cao, Y., Han, S.-W. and
Kawk, K. 2001. Erosion Function Apparatus for Scour Rate
Predictions. Journal of Geotechnical and Geoenvironmental
Engineering, 127 (2), 105-113.
Briaud J.-L., Chen, H.-C., Li, Y., Nurtjahyo, P., Wang, J. 2004. The
SRICOS-EFA method for complex piers in fine grained
soils, Journal of Geotechnical and Geoenvironmental
Engineering, ASCE, Vol 130, No. 11, p1180-1191.
Debnath, K. and Chaudhuri, S. 2010a. Laboratory Experiments on
Local Scour Around Cylinder for Clay and ClaySand Mixed
Beds. Engineering Geology, 111, 5161.
Debnath, K. and Chaudhuri, S. 2010b. Bridge Pier Scour in Clay-
Sand Mixed Sediments at Near-Threshold for Sand. Journal of
Hydraulic Engineering, 136(9), 597-609.
Ettema, R., Constantinescu, G. and Melville, B. 2011. Evaluation of
Bridge Scour Research: Pier Scour Processes and Predictions ,
Contractors Final Report for NCHRP Project 24-27(01),
National Cooperative Highway Research Program, March
2011, 181.
Gabr, M., Caruso, C., Key, A. and Kayser, M. 2012. Assessment of
In Situ Scour Profile in Sand Using a Jet Probe. Journal article
accepted in ASTM Geotechnical Testing Journal, In Press.
Hanson, G. J. and Cook, K. R. 2004. Apparatus, Test Procedures,
and Analytical Methods to Measure Soil Erodibility In Situ.
Applied Engineering in Agriculture, Vol. 20, No. 4, 455-462.
Hosny 1995. Experimental Study of Scour around Circular Piers in
Cohesive Soils. Ph.D. Dissertation, Colorado State University.
Kothyari, U.C., Garde, R.J. and Ranga Raju, K.G. 1992. Temporal
variation of scour around circular bridge piers. Journal of
Hydraulic Engineering, 118(8), 1091-1105.
Lu, J-Y., Hong, J-H., Su, C-C., Wang, C-Y. and Lai, J-S. 2008. Field
Measurements and Simulation of Bridge Scour Depth
Variations during Floods. Journal of Hydraulic Engineering,
134(6), 810-821.
Mehta, A. J. 1991. Review Notes on Cohesive Sediment Erosion. In:
N.C. Kraus, K.J. Gingerich, and D.L. Kriebel, (eds.), Coastal
sediment 91, Proceedings of Specialty Conference on
Quantitative Approaches to Coastal Sediment Processes,
ASCE; pp.40 53.
Molinas, A. and Hosny, M. 1999. Experimental Study on Scour
around Circular Piers in Cohesive Soil. Publication No.
FHWA-RD- 99-186, Federal Highway Administration, U.S.
Department of Transportation, McLean, VA.
Richardson, E.V., and Davis, S.R., 2001, Evaluating Scour at
Bridges (4th ed.), Federal Highway Administration Hydraulic
Engineering Circular No. 18, FHWA NHI 01-001.

567
Practical Reviews on CO
2
Sequestration in Korean Sedimentary Basins and
Geophysical Responses of CO
2
-injected Sediments
Le comportement pratiques sur la squestration du CO
2
dans les bassins sdimentaires corens et
responses gophysiques de CO
2
injectes sdiments
Kim A.R., Cho G.C.
Department of Civil and Environmental Engineering, KAIST, Korea
Kwon T.H.
Department of Civil and Environmental Engineering, Washington State University, USA
Chang I.H.
Geotechnical Engineering Research Division, Korea Institute of Construction Technology, KICT, Korea, Co.
ABSTRACT: Geological CO
2
sequestration is an effective means of reducing the emission of carbon dioxide. The Korean
government aims to reduce CO
2
emissions by 30% comparing to the usual amounts of emissions by 2020. It is expected that
geological CO
2
storage technology will account for more than 10% of the reduction of CO
2
emissions. The forward strategies and
technologies of CO
2
sequestration in Korea need to be determined depending on the geological conditions of potential sites in Korea;
moreover, the geophysical characteristics of CO
2
and the reservoirs depend on the geological conditions. However, previous domestic
studies related to geological conditions and the geophysical behavior of Korean sedimentary basins are rare thus far, with only a few
studies focusing on numerical modeling. This study aims to review the geological characteristics of CO
2
storage projects around the
world and in Korea while also discussing the suitability for CO
2
sequestration. Moreover, a laboratory approach simulating an in-situ
high effective stress condition with silty sand from the Bukpyeong basin is attempted in an effort to determine the geophysical
behaviors. This study offers an improved understanding of the possibility and potential of CO
2
sequestration in Korea.
RSUM : La squestration gologique du CO
2
est un moyen efficace de rduire les missions de dioxyde de carbone. Le
gouvernement coren a pour objectif de rduire les missions de CO
2
de 30 % lchance de 2020. Il est prvu que la technologie
gologique de stockage de CO
2
reprsentera plus de 10% de la rduction des missions de CO
2
. Les stratgies futures et les
technologies de squestration du CO
2
en Core doivent tre dtermines en fonction des conditions gologiques des sites potentiels en
Core, d'ailleurs, les caractristiques gophysiques de CO
2
et les rservoirs dpendent des conditions gologiques. Toutefois, les
prcdentes tudes nationales relatives aux conditions gologiques et gophysiques sur le comportement des bassins sdimentaires de
la Core sont rares ce jour, avec seulement quelques tudes mettant l'accent sur la modlisation numrique. Cette tude vise
examiner les caractristiques gologiques des projets de stockage de CO
2
dans le monde et en Core tout en discutant de leur
pertinence pour la squestration du CO
2
. En outre, une approche de laboratoire simulant un tat in situ haute contrainte effective
avec du sable limoneux du bassin Bukpyeong est tente dans le but de dterminer les comportements gophysiques. Cette tude
amliore la comprhension et la possibilit ainsi que le potentiel de squestration du CO
2
en Core.
KEYWORDS: CO
2
sequestration, korean marine sediment, geological condition, geophysical behavior
1 INTRODUCTION
Recently, several methods have been proposed to mitigate
carbon dioxide (CO
2
) emissions and to decrease the atmospheric
concentration of CO
2
, including material recycling, the usage of
renewable energy, and nuclear fusion. Among these, carbon
capture and storage (CCS) strategies are considered as effective
methods of reducing the atmospheric concentration of CO
2
in a
relatively short time at a low cost compared to other
technologies (Espinoza et al. 2011; Pires et al. 2011). In light of
this, approximately 40 CCS projects (including pilot-to-
commercial scale applications) are in the planning or
operational stages around the world (Hosa et al. 2010).
It has been reported that the increasing rate of CO
2
emissions
in South Korea is the highest among OECD member countries,
making Korea the seventh largest CO
2
emitting nation in the
world (BP 2011). Since 1999, the atmospheric concentration of
CO
2
in Korea has always been higher than the global average
(Figure 1). The Korean government plans to reduce its CO
2
emissions by 30% compared to the current business as usual
(BAU) value by 2020 (i.e., about 244 Mton/yr; Presidential
Committee on Green Growth, 2011). It is prospected that
geological CO
2
storage technology will account for more than
10% of global CO
2
emissions (approximately 25 Mton/yr). As a
part of this effort, pilot (10000 tons of CO
2
) and demonstration
(100000 tons of CO
2
) scale CO
2
sequestration projects, capable
of storing more than 1 Mton of CO
2
in total, are currently
planned and being conducted with commercial considerations in
Korea (Presidential Committee on Green Growth 2011). This
effort mostly relies on existing geologic data and resources that
were gathered during oil explorations and drilling projects.
However, the current geologic information on onshore and
offshore deep subsurface areas of Korea is insufficient.
Comprehensive geological exploration and database
construction activities are critical for characterizing, selecting,
or at least screening potential storage sites for CO
2
sequestration.
Figure 1. Atmospheric concentration of CO
2
of the world and in Korea.
Data were gathered from Climate Change Information Center
(www.climate.go.kr) and National Oceanic and Atmospheric
Administration (www.noaa.gov).
568
th
Proper strategies and technologies for storing CO
2
in
geologic formations must be determined with deep
consideration of the geological conditions of the potential sites.
Geological conditions (e.g., the pressure, geothermal gradient,
geology, geochemical characteristics, and mineralogy) govern
the interpretation of the geophysical responses of CO
2
-storage
reservoirs. Therefore, a fundamental understanding of the
geophysical responses of CO
2
-containing sediments in Korean
sedimentary basins is required. However, while a few studies
have been conducted on numerical reservoir modeling, there
have been few efforts to examine the geological suitability for
CO
2
storage and on geophysical characterizations of Korean
sedimentary basins to date.
This study provides a review and discussion of the
geological conditions and suitability of the potential CO
2
storage sites in Korea as well as CO
2
storage sites around the
world. Moreover, laboratory experiment results on the P-wave
velocity and the electrical responses of CO
2
-injected sediments
are presented, in which high in-situ effective stress conditions
were simulated on natural samples cored from the eastern
Bukpyeong basin, which is one of the candidate sites for
geologic CO
2
storage.
2 SITE CHARACTERIZATION FOR CO
2
SEQUESTRATION
2.1 Site-dependent strategies
CO
2
can be stored in various geological formations, such as
(1) deep saline formations, (2) coal beds, (3) depleted oil and
gas fields, and (4) oil and gas reservoirs during enhanced oil and
gas recovery efforts (EOR and EGR). Coal beds absorb and
contain gaseous CO
2
in micro-pores in coal. However, the
temperature and pressure effects on the CO
2
trapping process in
coals are not well understood (Larsen, 2003). CO
2
storage in
depleted oil and gas fields or the use of CO
2
for EOR (or EGR)
have been proven as effective GCS methods due to the
geological suitability for CO
2
storage, the existence of
geophysical and geological data, and the ready-made
infrastructure used for oil and gas production. However, several
problems remain poorly identified, including well plugging,
leakage induced by the overpressure of pore fluids, and the
injection depth. Deep saline formation is the most promising
method for safe and effective CO
2
storage due to its vast
capacity. The potential storage capacity of deep saline
formation is expected to be at least 1000 GtCO
2
, which is
approximately 200 to 300 times higher than the potential
storage offered by oil or gas fields and coal seams (IPCC,
2005). In particular, sedimentary basins that have permeable
formations (e.g., sandstone) with overlying low-permeable seals
(caprocks) are effective for both CO
2
injection and CO
2
leakage
prevention. The target depth is deeper than 1000 meters below
the ground surface.
2.2 Selected sites for CO
2
sequestration in America, Europe,
and Asia
There are more than 800 sedimentary basins around the
world (St John et al., 1984). Approximately 40 CO
2
storage
projects are under operation or in planning in North America,
Europe, Australia, and Asia. Among them, ten onshore projects
in the USA and two onshore projects in Canada are being
conducted. Two onshore projects and three offshore projects in
Europe are being undertaken, and two demo onshore projects in
Japan are being tested (Hosa et al., 2010). Additionally, other
potential sedimentary basins have been investigated and
proposed for pilot-scale testing. For instance, the Alberta basin
in Canada, where natural hydrocarbon resources had been
found, was evaluated to be the most suitable basin in Canada
owing to the existence of adjacent infrastructure (Bachu, 2003).
The offshore Gippsland basin in Australia is considered as an
effective target for CO
2
storage due to its complex stratigraphy,
high injectivity, low-permeable marginal reservoir, the
existence of several depleted oil fields, and its long migration
pathways (Gibson-Poole et al., 2008). In China, an ECBM
(enhanced coalbed methane recovery) pilot test and a single-
well micro-pilot test were successfully performed at the South
Qinshui basin (Wong et al., 2010).
One of the most well-known CO
2
storage attempts is the
Sleipner project, targeting the Utsira Sand formation, which was
launched in 1996. It was the first commercial-scale project to
store CO
2
in a saline formation. The geologic condition of this
site is a brine-saturated sandstone layer (250 m thick) with an
overlying thin shale cap layer. Its storage capacity is expected to
be 25 MtCO
2
(Hosa et al., 2010). The Nagaoka project at
Nagaoka City, Japan, was the first pilot-scale attempt in Asia. In
this pilot-scale test, CO
2
was injected into a Haizume-formation
sandstone layer. The injection efficiency differs between the
two formations (e.g., the Utsira Sand formation and the
Haizume formation). In detail, the Utsira Sand formation (2800
tons/day) has a storage capacity of approximately 70 times that
of the Haizume formation (a maximum 40 tons/day), as the
permeability of the Haizume formation (i.e., 6 mD) is much
lower than that of the Utsira Sand formation (i.e., 5 D) (Hosa et
al., 2010) though both formations have relatively high porosities
(37% for the Utsira formation and 22.5% for the Haizume
formation) and similar injection depths (about 10001100 m).
Therefore, it can be tentatively concluded that the permeability
is a major controlling parameter for CO
2
injectivity rather than
the porosity or injection depth.
2.3 Geologic characteristics of sedimentary basins in Korea
There are several potential geologic formations for GCS on
the Korean Peninsula. The porosity, storage capacity, and
geologic characteristics of those proposed sites are listed in
Table 1. However, the geological information of offshore basins
is still poorly identified due to insufficient exploration.
Table 1. Porosity and storage capacity of potential Korean geological
storage sites
1
The values are from Kim et al. (2011) and
2
MEST (2008).
Korean offshore sedimentary basins have thickness ranges
from 3 km to 10 km (MEST, 2008). The vast coverage area and
high porosity of Korean offshore sedimentary basins are
expected to show that these have larger storage capacities than
onshore basins. Moreover, Korean offshore sedimentary basins
show geologically structural similarity with natural hydrocarbon
reservoirs, which indicates good suitability for GCS. For an
example, the Gunsan basin and the Jeju basin show high
potential for GCS because the geologic structures are similar to
natural hydrocarbon reservoirs in analogous Chinese basins
(Hong et al., 2005). The Ulleung basin contains natural gas
deposits and is located more than 1000 m below the sea level.
Thus, structural trapping may be feasible (Hong et al., 2005).
To evaluate the storage and economic efficiencies of
sedimentary basins in Korea, a systematic and quantitative
evaluation method (Bachu 2003) was employed in this study.
Fifteen criteria (e.g., geological characteristics, basin resources,
maturity, and infrastructure, among others) are considered with
weight factors to assess the suitability (Table 2). Bachus (2003)
method classifies the proposed sites with dimensionless values
between 0 and 1. The value can be used as a decision criterion
Basin
Porosity
(%)
Storage Capacity
(Mton)
Geologic
Characterstics
Bukpyeong N.A. 877
1
Saline aquifer
Ulleung 10.6
2
3,018
2
EGR
Jeju 15.7
2
95,101
2
Saline aquifer
Gunsan 10
2
254
2
Saline aquifer
Heuksan N.A. N.A. Saline aquifer
Pohang N.A. 38
2
Saline aquifer
569
for assessing the suitability of the proposed sedimentary basins
for geological sequestration and for comparing it with basins in
other countries in which pilot- and commercial-scale projects
are already underway.
Table 2. Scores and weight assigned to the criteria and classes for
assessing sedimentary basin in terms of their suitability for CO
2
sequestration in geological media (Bachu, 2003)
Scores Weight
Criterion
j=1 j=2 j=3 j=4 j=5
Tectonic setting 1 3 7 15 15 0.07
Size 1 3 5 9 0.06
Depth 1 3 5 0.07
Geology 1 3 7 0.08
Hydrogeology 1 3 7 0.08
Geothermal 1 3 7 0.10
Hydrocarbon potential 1 3 7 13 21 0.06
Maturity 1 2 4 10 0.08
Coals and CBM 1 2 5 0.04
Salts 1 2 3 0.01
On/offshore 1 4 10 0.10
Climate 1 2 4 7 11 0.08
Accessibility 1 3 6 10 0.03
Infrastructure 1 3 7 10 0.05
CO
2
Sources 1 3 7 15 0.09
The evaluation results are summarized in Figure 2. Among
other sites, the Ulleung basin was evaluated to be the most
suitable site for geologic CO
2
storage in Korea due to the
presence of nearby infrastructure as constructed for natural gas
recovery.
Figure 2. Scores for suitability for of Korean sedimentary basins.
However, the proposed sites in Korea are less feasible for
geologic CO
2
storage compared to the basins in Canada and
Australia. Specifically, the Korean basins show low scores on
the following criteria: size, hydrocarbon potential, maturity, and
infrastructure. Most Korean sedimentary basins, except the
Ulleung basin, are estimated to be of a small-to-medium size,
whereas the Otway and Gippsland basins in Australia are
categorized as large and the Alberta and Williston basins in
Canada are known to be large to giant in size. Moreover, a lack
of boring studies and geophysical exploration exacerbate the
problems of low maturity and insufficient infrastructure. Thus,
more data acquisition and exploration are required to enhance
the reliability of numerical modeling and simulations for the
first Korean pilot project. Additionally, a new alternative
approach involving the use of deep saline formations should be
considered for safe and economic CO
2
sequestration in Korea
with consideration of the geological characteristics of Korean
basins and their limitations as regards CO
2
injection.
3 GEOPHYSICAL RESPONSES OF CO
2
-CONTAINING
SEDIMENTS
While CO
2
leakage from CO
2
stored sites could cause
serious environmental problems, geophysical survey techniques
are viable methods to detect CO
2
leakage and to identify CO
2
movement. Therefore, understanding the geophysical responses
of Korean sediments during CO
2
injection and storage is
important to ensure safety.
The most widely used geophysical methods are seismic
surveys using P-waves and electrical resistivity surveys
(Nakatsuka et al., 2010). In particular, feasibility of P-wave
surveys for the detection of CO
2
has been examined in
laboratory tests (Shi et al., 2007; Siggins et al., 2010; Xue and
Lei, 2006). Also, it is well known that CO
2
-containing
formations have less stiffness than brine-saturated formations
do (Daley et al., 2008; Lazaratos and Marion, 1997; Mito and
Xue, 2011). Because the physical properties of unconsolidated
sediments are significantly affected by the effective stress as
well as the formation characteristics, achieving an in-situ
effective stress condition is critical to obtain reliable physical
properties of CO
2
-containing sediments, though this may not be
the case for cemented porous media. This section presents the
geophysical responses of sediments in which an in-situ effective
stress condition of the potential CO
2
storage sites was achieved.
3.1 Experimental study
The sediment sampled from the Bukpyeong basin (located
on the east coast of Korea; see Table 1) was used in this study.
The silty sand sample was compacted into a rigid-walled vessel
and was saturated with water. Vertical in-situ effective stress of
15 MPa was then applied. The final porosity resulting from the
applied stress condition was estimated to be 49%.
0
0.2
0.4
0.6
0.8
1
S
c
o
r
e
Table 3. Properties of test specimen
property Soil type
(USCS)
Specific
gravity
Permeability
(cm/s)
D50
(mm)
specimen Silty sand 2.73 4.2*10
-5
0.15
The test was performed on saturated silty sand under a
supercritical temperature condition (35C) under the effective
stress (15 MPa). The pore water pressure was maintained at 8
MPa using a back pressure regulator. The specimen was then
flooded with supercritical CO
2
. After a pre-determined amount
of CO
2
was injected into a water-saturated sample, the
geophysical responses (P-wave velocity V
P
and electrical
resistivity) were measured and the amount of injected CO
2
was
estimated by measuring the amount of water expelled from the
vessel.
3.2 Results and analysis
The total amount of water expelled by weight was 29.86 g;
thus, approximately 30% of the pore water was displaced with
supercritical CO
2
, indicating that the CO
2
saturation rate was
~30%. The density and solubility of CO
2
at the target
temperature and pressure are 591.85 kg/m
3
(NIST) and 1.35
mol/kg of water (Duan and Sun 2003), respectively.
Accordingly, the total calculated amount of injected CO
2
was
17.29 g.
Figure 3 shows the decrease in V
P
during the CO
2
injection
process. The decrement range is relatively low compared to pre-
existing studies (Shi et al. 2007; Siggins et al. 2010). This can
be explained by the fact that the CO
2
injectivity in sandy soil
specimens is lower than that in sandstone. The electrical
resistivity increased rapidly as CO
2
was introduced. The
convergence of the electrical resistivity indicates a fully
saturated condition in which no more CO
2
can be injected (see
Figure 4). The overall results show that CO
2
can be detected by
measuring the geophysical properties; however, the injected
570
th
amounts and readings of the movement of CO
2
are not highly
precise.
Daley, T., Myer, L., Peterson, J., Majer, E., and Hoversten, G. 2008.
Time-lapse crosswell seismic and VSP monitoring of injected CO
2
in a brine aquifer. Environmental Geology, 54(8), 1657-1665.
Duan, Z., and Sun, R. 2003. An improved model calculating CO
2
solubility in pure water and aqueous NaCl solutions from 273 to
533K and from 0 to 2000 bar. Chem. Geol. 193, 257-271.
1720
1722
1724
1726
1728
1730
0 50 100
P
-
w
a
v
e

v
e
l
o
c
i
t
y

(
m
/
s
)
Time (min)
Espinoza, D., Kim, S., and Santamarina, J. C. 2011. CO
2
geological
storage Geotechnical implications. KSCE Journal of Civil
Engineering, 15(4), 707-719.
Gibson-Poole, C. M., Svendsen, L., Underschultz, J., Watson, M. N.,
Ennis-King, J., van Ruth, P. J., Nelson, E. J., Daniel, R. F., and
Cinar, Y. 2008. Site characterization of a basin-scale CO
2
geological storage system: Gippsland Basin, southeast Australia.
Environmental Geology, 54(8), 1583-1606.
Figure 3. The effect on the P-wave velocity of the sediment during CO
2
injection.
Hong, S. K., Lee, H., Egawa, K., Choi, T., Lee, M. K., Yoo, K. C., Kim,
J. H., Lee, Y. I., and Kim, J. M. 2009. Preliminary evaluation for
carbon dioxide storage capacity of the Chungnam, Taebacksan,
Mungyeong and Honam basins. Journal of the Geological Society
of Korea, 45(5), 449-462 (in Korean).
150
Hosa, A., Esentia, M., Stewart, J., and Haszeldine, S. 2010. Injection of
CO
2
into saline formations: benchmarking worldwide projects.
Chemical Engineering Research and Design,
doi:10.1016/j.cherd.2011.04.003.
80
82
84
86
0 100 200 300 400 500 600 700
E
l
e
c
t
r
i
c
a
l

R
e
s
i
s
t
i
v
i
t
y

(
)
Time (sec)
IPCC, 2005, Special Report on Carbon Dioxide Capture and Storage.
In: Metz., B., Davidson, O., de Coninck, H. C., Loos, M., Meyer, L.
A., (eds.), Prepared by Working Group III of the Intergovernmental
Panel on Climate Change, Cambridge University Press, Cambridge,
UK and New York, NY, USA
Kim, J. M., Kim, J. H., and Park, S. W. 2011. Evaluation of CO
2
storage
capacity of Bukpyeong Basin using three-dimensional modelling
and thermal-hydrological nemerical modelling. 1st Korea CCS
Conference (Abstract), Jeju, April 13-15 (in Korean).
Larsen, J.W. 2003. The effects of dissolved CO
2
on coal structure and
properties. International Journal of Coal Geology, 57, 63-70.
Ministry of Educational Science and Technology, 2008,
Characterization and evaluation of geologic formation for
geological sequestration of carbon dioxide. 21
st
Century Frontier
R&D Program (in Korean).
Figure 4. The electrical resistivity of the sediment during CO
2
injection.
4 CONCLUSION
The presented study explores practical reviews on Korean
marine sediments for CO
2
sequestration in relation to geological
and geotechnical considerations. The geological conditions of
off-shore sedimentary basins in Korea were investigated, and
the suitability of the basins for CO
2
storage were evaluated. The
Ulleung basin were found to be the most suitable site for GCS,
although their scores were lower than the scores of some basins
where CO
2
storage is currently undergoing or pilot-tested in
Canada and Australia. Geophysical behavior of CO
2
-storing
sediments is available for field application to monitor CO
2
movements and leakages. A laboratory scale experiment
simulating the in-situ condition for measuring geophysical
properties, and the results showed that CO
2
can be detected by
measuring geophysical properties but further study is required
to exact understanding geophysical behavior of CO
2
-storing
Korean marine sediments.
Mito, S., and Xue, Z. 2011. Post-Injection monitoring of stored CO
2
at
the Nagaoka pilot site: 5 years time-lapse well logging results.
Energy Procedia, 4, 3284-3289.
Nakatsuka, Y., Xue, Z., Garcia, H., and Matsuoka, T. 2010.
Experimental study on CO
2
monitoring and quantification of stored
CO
2
in saline formations using resistivity measurements.
International Journal of Greenhouse Gas Control, 4(2), 209-216.
NIST Thermophysical Properties of Fluid System, 2007. National
Institute of Standards and Technology, Available on line:
(http://webbook.nist.gov/chemistry/fluid). Accessed: October 2007.
Pires, J. C. M., Martins, F. G., Alvim-Ferraz, M. C. M., and Simoes, M.
2011. Recent developments on carbon capture and storage: An
overview. Chemical Engineering Research and Design, In Press,
Corrected Proof.
Hosa, A., Esentia, M., Stewart, J., and Haszeldine, S. 2010. Injection of
CO
2
into saline formations: benchmarking worldwide projects.
Chemical Engineering Research and Design,
doi:10.1016/j.cherd.2011.04.003.
Shi, J.-Q., Xue, Z., and Durucan, S. 2007. Seismic monitoring and
modelling of supercritical CO
2
injection into a water-saturated
sandstone: Interpretation of P-wave velocity data. International
Journal of Greenhouse Gas Control, 1(4), 473-480.
5 ACKNOWLEDGEMENTS
Siggins, A. F., Lwin, M., and Wisman, P. 2010. Laboratory calibration
of the seismo-acoustic response of CO
2
saturated sandstones.
International Journal of Greenhouse Gas Control, 4(6), 920-927.
This research was supported by a grant from the Korea Electric
Power Corporation (KEPCO) and by the Block Funding Project
(GP2012-030) of Korea Institute of Geoscience and Mineral
Resources (KIGAM) transferred from the Energy Efficiency &
Resources Program of the Korea Institute of Energy Technology
Evaluation and Planning (KETEP) grant, funded by the
Ministry of Knowledge Economy, Republic of Korea.
St John, B., Bally, A. W., and Klemme, H. D. 1984. Sedimentary
provinces of the world hydrocarbon productive and nonproductive.
American Association of Petroleum Geologists, Tulsa.
Wong, S., Macdonald, D., Andrei, S., Gunter, W. D., Deng, X., Law, D.,
Ye, J., Feng, Z., and Ho, P., 2010, Conceptual economics of full
scale enhanced coalbed methane production and CO
2
storage in
anthracitic coals at South Qinshui basin, Shanxi, China.
International Journal of Coal Geology, 82(3-4), 280-286.
6 REFERENCES
Bachu, S. 2003. Screening and ranking of sedimentary basins for
sequestration of CO
2
in geological media in response to climate
change. Environmental Geology, 44(3), 277-289.
Xue, Z., and Lei, X. 2006. Laboratory study of CO
2
migration in water-
saturated anisotropic sandstone, based on P-wave velocity imaging.
Exploration Geophysics, 37(1), 10-18.
BP. 2011. BP Statistical Review of World Energy June 2011. The report
can be downloaded at the website
http://www.bp.com/statisticalreview.
571
Using Multi-scale Sediment Monitoring Techniques to Evaluate Remediation
Effectiveness of the Tsengwen Reservoir Watershed after Sediment Disasters
Induced by Typhoon Morakot
Lin B.-S., Ho H.-C., Hsiao C.-Y., Keck J., Chen C.-Y., Chi S.-Y.
Disaster Prevention Technology Research Center, Sinotech Engineering Consultants, Taiwan
Chien Y.-D., Tsai M.-F.
Soil and Water Conservation Bureau, Council of Agriculture, Taiwan
ABSTRACT: The 2009 typhoon Morakot dumped more than 3,005 mm of rain in mountain areas of the Tsengwen reservoir
watershed and caused unprecedented landslide and sediment-related disasters. Subsequently, the storage capacity of the Tsengwen
reservoir was drastically reduced. In order to increase the longevity of the reservoir and also protect ecosystems and the peoples living
in the upper portions of the watershed, the Taiwan Executive Yuan implemented the "Tsengwen, Nanhua, Wushantou Reservoir
Remediation and Water Resources Protection Act". This study aims to use multi-scale sediment monitoring techniques including field
investigations and multi-stage remote sensing data to identify sediment migration patterns associated with remediated areas of the
Tsengwen reservoir watershed after typhoon Morakot and to guarantee the effectiveness of remediation efforts. A case study of the
Longjiao creek in Tsengwen Reservoir watershed shows that remediation works can not only reduced sediment production due to
erosion and landslides, but future sediment production will also be suppressed. The reduction of sediments carried by the Tsengwen
river will also lead to an increase in the service life of the Tsengwen reservoir.
RSUM : En 2009, le typhon Morakot a dvers plus de 3 005 mm de pluie dans les rgions montagneuses du bassin versant du
rservoir et provoqu des glissements de terrain Tsengwen sans prcdent et les catastrophes sdiments. Par la suite, la capacit de
stockage du rservoir Tsengwen a t considrablement rduit. Afin d'augmenter la longvit du rservoir et aussi de protger les
cosystmes et les populations vivant dans les parties suprieures du bassin versant, le Taiwan Yuan excutif mis en place le
"Tsengwen, Nanhua, Wushantou rservoir d'assainissement et de protection des eaux Loi sur les ressources". Cette tude vise
utiliser multi-chelle des techniques de surveillance des sdiments, y compris les enqutes sur le terrain et sur plusieurs priodes
donnes de tldtection pour identifier les schmas de migration des sdiments associs des zones assainies du bassin versant du
rservoir Tsengwen aprs le typhon Morokot et de garantir l'efficacit des efforts d'assainissement. Une tude de cas du flux Longjiao
dans le bassin versant du rservoir Tsengwen montre que les travaux de rhabilitation peuvent non seulement rduit la production de
sdiments cause de l'rosion et des glissements de terrain, mais la production de sdiments avenir seront galement supprims. La
rduction des sdiments charris par le fleuve Tsengwen conduira galement une augmentation de la dure de vie du rservoir
Tsengwen.
KEYWORDS: Tsengwen reservoir watershed, typhoon Morakot, sediment disasters, remediation effectiveness.
1 INTRODUCTION
In 2009, typhoon Morakot brought heavy rainfall up to 3,005
mm, which was recorded at Alishan rainfall guage station of the
Tsengwen reservoir watershed over a five day period. Also, the
consecutive 72 hour accumulated rainfall exceeded historical
records in Taiwan (SWCB, 2011). This typhoon event induced
massive sediment-related disasters within the watershed, which
caused about 91,080,000 m
3
of sediment in the reservoir and
exceeded the original design level (5,610,000 m
3
/yr).
Afterward, Taiwan government passed Tsengwen, Nanhua,
Wushantou Reservoir Remediation and Water Resources
Protection Act and planned a project for managing and
remediating sediment problems. The primary goals of the
proposed project are to reduce reservoir turbidity levels, extend
the service life of the dam and protect security of the upstream
residents.
Sediment transport and deposit within the watershed is an
unavoidable natural process. It is very important to do field
survey and monitor periodically especially in major sediment
source areas including old debris flow, large-scale landslide and
massive alluvial soil or river terrace deposits. Many researches
has pointed out that the sedimentation of Tsengwen Reservoir
has been serious in flood season due to intense geological
activity. Recently, under the effect of global climate change, the
probability of extreme weather occurrence has increased. In the
mountain area, it can be observed that the magnitude of
disasters caused by water-sediment flows, induced by high
intensity and long duration rainfall events, has increased (Lo et
al., 2012; Lin et al., 2012). The mode of sediment transport can
be classified in different ways, according to the mechanics of
sediment transport process, from suspended load to debris flow.
Therefore, the sediment deposited in the Tsengwen reservoir
watershed comes from long-term deposits of the trunk river and
soil erosion as well as slope landslides due to 2009 typhoon
Morakot. It significantly affects water supply to residents and
rapidly reduced storage capacity of reservoir. For validating and
proving the effectiveness of remediation efforts after typhoon
Morakot event, the study integrates multi-scale sediment
monitoring techniques to collect time-dependent monitoring
data and spatiotemporal remote sensing information including
watershed scale, high-resolution airborne LiDAR DTMs. Then,
using the data obtained from the remediated environmental area,
remediation effectiveness of the Tsengwen reservoir watershed
with regard to suppression of soil erosion, vegetative recovery
rate, variation in amount of landslide and sediment trapping
efficiency are quantified. Finally, the proposed procedure of this
study will assist us to track remediation effectiveness, and
reduce sediment yield entering a reservoir or trap eroded
sediment for effective watershed management.
572
th
2 OVERVIEW OF ENVIRONMENT
2.1 Study area
As illustrated in Figure 1, the Tsengwen reservoir is located in
the southwestern portion of Taiwan. It is the most important
water resource conservation hydraulic structure and the largest
dam in Taiwan. The dam is 400 m in width and 133 m in height,
and mainly serves irrigation, municipal water and power supply
functions for the southern plains and downstream area of Chiayi
county, Tainan and Kaohsiung city shown in Figure 2. The
watershed area of Tsengwen Reservoir is approximately 481
km
2
, where Tsengwen river is the trunk river originating from
Alishan mountain at elevation of 2,609 m a.s.l. The watershed
shape is similar to a quadrilateral with elevations gradually
increasing from southwest to northeast, and ranging from 100 m
a.s.l and 2,700 m a.s.l. In general, most hill slopes are steeper
than 28.8 and approximately represents over 60% of the study
area. As for the aspect, slopes are mostly west-facing and
southwest-facing in the watershed. There are many fault line
and geologic structures and the geological condition of the
watershed consists mostly of sandy shale, siltstone, and isolated
areas of muddy sandstone, which are prone to more severe
weathering and become weak layers in the rock strata. These
conditions make the slope unstable during heavy rainfall or
strong earthquake shaking. Hence, during the typhoon and flood
season, the combination of huge rainfalls and local weak
geological conditions easily permit the occurrence of sediment
landslides (Lo et al., 2012).
Due to high topographic relief, annual average temperature
ranges from 24 in the plains and 11 in the mountainous
parts of the watershed. According to Alishan rainfall gauge
station, average annual accumulated ranges from 1,950 to 4,980
. Recent extreme rainfalls have caused annual accumulated
rainfall of Taiwan to increase, especially for Alishan, where,
since 2005, annual rainfalls have exceeded 5,000 (see Figure
3). This rate is double the annual average precipitation (2,500
) for Taiwan and over four times of world annual average
precipitation. Rainfall distribution increases from the plains to
the mountains and is mostly concentrated between May and
September when the watershed receives approximately 80% of
the overall annual rainfall.
Figure 1. Graphical location of Tsengwen reservoir watershed in
Taiwan (local coordinate system: TWD97).
Figure 2. Topographic map of Tsengwen reservoir watershed (local
coordinate system: TWD97).
1
9
5
9
1
9
6
0
1
9
6
1
1
9
6
2
1
9
6
3
1
9
6
4
1
9
6
5
1
9
6
6
1
9
6
7
1
9
6
8
1
9
6
9
1
9
7
0
1
9
7
1
1
9
7
2
1
9
7
3
1
9
7
4
1
9
7
5
1
9
7
6
1
9
7
7
1
9
7
8
1
9
7
9
1
9
8
0
1
9
8
1
1
9
8
2
1
9
8
3
1
9
8
4
1
9
8
5
1
9
8
6
1
9
8
7
1
9
8
8
1
9
8
9
1
9
9
0
1
9
9
1
1
9
9
2
1
9
9
3
1
9
9
4
1
9
9
5
1
9
9
6
1
9
9
7
1
9
9
8
1
9
9
9
2
0
0
0
2
0
0
1
2
0
0
2
2
0
0
3
2
0
0
4
2
0
0
5
2
0
0
6
2
0
0
7
2
0
0
8
2
0
0
9
0
1000
2000
3000
4000
5000
6000
7000
8000
Gr
Figure 3. Time series of annual rainfalls at Alishan gauge station.
2.2 Relation between reservoir sedimentation and major
typhoon events
Presently, the greatest challenge of Tsengwen reservoir is
sedimentation. Whether caused by anthropogenic or natural
factors, both of them directly triggers problems such as
increased turbidity and reduced reservoir storage volume.
Figure 4 illustrates the historical trends of sedimentation in
Tsengwen reservoir concerning major typhoon events.
According to the figure, since completion of reservoir
construction, typhoons repeatedly hit the Tsengwen reservoir. It
can be found that the peaks in the historical sedimentation curve
of Tsengwen reservoir correspond to major typhoon events.
Before 2008, the annual average sedimentation volume is
4,760,000 m
3
and still lower than the yearly designed value of
5610,000 m
3
. However, 2009 typhoon Morakot brought around
91,080,000 m
3
of sediment into Tsengwen Reservoir, which
occupies about 12% of the reservoir capacity. After the 2009
typhoon Morakot, the annual average sedimentation rapidly
increases to 7,060,000 m
3
, exceeding the yearly designed value
by 12.6 times. It is truly believed that massive amounts of
sediment washed downstream. Also, this event seriously
resulted in debris flows and large large-scale landsides along
river flanks and close to human inhabitation in upstream areas,
and threatens longevity of reservoir and significantly affects
water supply to the south area in Taiwan.
Figure 4. Historical sedimentation curve of Tsengwen reservoir with the
major typhoon events
573
3 MULTI-SCALE SEDIMENT MONITORING AND
EVALUTION METHOD TO REMEDIATION
EFFECTIVENESS
Multi-scale sediment monitoring techniques is used in
Tsengwen reservoir watershed to study remediation
effectiveness and topographical changes. This section describes
the method about how to systematically study and analyze soil
erosion, landslide areas, and sediment trapping in the check
dams from easily measured physical quantities such as depth,
area, and volume by collecting time-dependent monitoring data
and multi-stage remote sensing information in a watershed
scale. A case study of the Longjiao subwatershed was chosen to
be validated with remediation effort. The above proposed
methods are detailed separately below.
3.1 Depth-based evaluation method
Soil erosion estimates were often based on empirical equations,
such as USLE, MUSLE, and RUSLE, etc. These empirical
equations are limited regionally and by spatial distribution of
rainfalls. Therefore, this study focused on the different
vegetated slopes to design erosion pins by some research reports
(Schumn, 1956). Site surveys were conducted to measure
surface erosion depth to investigate the state of slope soil after
erosion from rainfalls. The result was used to assess the
inhibition rate of soil erosion from both remediated and non-
remediated hillslopes in order to understand the efficiency of
remediation. To quantify the soil erosion suppression ratio (SSR)
from the measured soil erosion depth of several erosion pins
(see Figure 5) embedded in remediated and non-remediated
hillslopes, we used an index value to depict efficiency of soil
erosion retention after completing remediation. Higher index
values indicate higher soil erosion suppression. Therefore, this
study uses this index (SSR) to understand the remediation
effectiveness of the hillslopes. SSR is defined as follows
% 100 (%)
DR
DN DR
E
E E
SSR
(1)
Where SSR is soil erosion suppression ratio (%)E
DR
is surface
eroded soil depth of remediated hillslope (mm)E
DN
is surface
eroded soil depth of non-remediated hillslope (mm).
Figure 5. Schematic layout and photos of erosion pins embedded in
remediated/non-remediated hillslope.
3.2 Area-based evaluation method
To understand the evolution of vegetation coverage of the
Tsengwen reservoir watershed resulting from remediation
efforts, multi-spectral high-resolution satellite images from
different periods are adopted to analyze the ratio of green cover
to assess the vegetation restoration after remediation.
Normalized Difference Vegetation Index (NDVI) is currently a
popular method to assess vegetation coverage (Kriegler et al.,
1969). The NDVI is calculated from these individual
measurements as follows:
VIS NIR
VIS NIR
NDVI
(2)
where VIS and NIR stand for the spectral reflectance
measurements acquired in the visible (red) and near-infrared
regions, respectively. The NDVI value is normalized between -
1.0 to 1.0. Values of NDVI above a certain threshold
correspond to vegetation coverage area and values below the
threshold correspond to non-vegetation coverage areas, as
shown in Figure 6.
Figure 6. Orginal satellite image and the classified image from
results of NDVI.
Once the NDVI has been used to classify the images into
vegetated and non-vegetated zones, the ratio of vegetation
coverage (VR) involved with the total area is estimated, as
shown in Eq. (3).
% 100 (%)
C
V
A
A
VR
(3)
Where A
c
is a given watershed area, and A
v
is vegetated area
within a given watershed.
3.3 Volume-based evaluation method
This study compiled satellite images to identify the landslide
distribution. Number of landslides, existing landslide area,
incremental landslide area, and spatial distribution in key
regions were obtained through digital interpretation to
understand its evolution. Further, this was complemented with
multi-period terrain data, established by airborne LiDAR, to
quickly obtain information on terrain changes in each sub-
catchment area and assess the effectiveness of the remediation
projects. To evaluate the effectiveness of check dams, this study
utilized airborne LiDAR (Light Detection And Ranging)
technology to survey and produce high resolution DEMs of the
Tsengwen reservoir watershed. The pre-event DTM is
subtracted from the post-event DTM. A negative value in the
grid represents failure or erosion, and positive value indicates
deposits. Variation in volume of a grid can be obtained by
multiplying this value by the area of the unit grid (see Figure. 7).
The total volume of landside material and sediment trapped by
the check dam can also be obtained from multiple LiDAR
generated DTMs. Then, sediment discharge and trapping
efficiency of dams can be precisely calculated. It can also be
applied to monitor the accumulated volume of sediment on the
confluence between tributaries and river, growth of alluvial fan,
and large scale wedge like slope failures. Comparison of
LiDAR DEMs from different periods can also indicate terrain
migration and be used to trace sediment transport from
tributaries, especially in extreme typhoon disasters. Sediment
trapping ratio (STR) can be assessed by measuring the volume
of deposited sediment in front of the check dams (Sophie et al.,
2008). If STR after remediation is higher than before
remediation, it means that check dams are effectively
controlling sediment transport and have adequate remediation
efficiency levels. The sediment trapping ratio can be expressed
as
% 100 (%)
y
d
V
V
STR
(4)
Where V
d
is the trapped volume in the check dam(m
3
)V
y
is
the sediment yield from upstream (m
3
).
574
th
19 19 16
17 15
15 10 3
19 19 16
17 15
15 10 3
16 14 16
13 11
14 11 6
16 14 16
13 11
14 11 6
-3 -5 0
-4 -4
-1 1 3
-3 -5 0
-4 -4
-1 1 3
=
Results 1st DEM Grid Map 2nd DEM Grid Map
-
No data Erosion Deposition
Variation of Volume
= Result x Unit grid area
Grid Subtraction
Figure 7. Grid subtraction in post-event and pre-event digital terrain
elevations
4 CASE STUDY
In this paper, case study chooses Longjiao creek located at
downstream area of Tsengwen Reservoir (see Figure 2) for
proving the effectiveness of remediation efforts after typhoon
Morakot, based on real data from multi-scale sediment
monitoring techniques. Then, the proposed depth-area-volume
based methods are all applied to evaluate the ratios of soil
erosion suppression, vegetation coverage and sediment trapping
in the following sections.
4.1 Soil erosion suppression
To effectively measure surface soil loss on remediated and non-
remediated hillslopes, ten erosion pins were installed on each of
the hillslopes types to monitor eroded soil depth for each
rainfall events. The monitoring period is from May 14., 2011 to
Oct. 04, 2011. Figure 8 is the diagram of the average
accumulated eroded soil depth on remediated and non-
remediated hillslopes. According to the figure, soil erosion of
remediated hillslopes is obviously lower than the remediated.
Compared with others, remediated hillslopes can reduce soil
erosion by about 1.6 mm. This data is input into Eq. (1) and
then the SSR of Longjiao subwatersehd is obtained as below
% 23 . 61 % 100
7 . 18
25 . 7 7 . 18
(%)
SSR
(5)
The calculated result shows that remediation of hillslops could
reduce erosion amounts by 61.23% of soil loss per unit area and
time. It is evident that remediation can accelerate environmental
vegetation recovery and under good practical sediment control.
7.25
18.7
0.00
4.00
8.00
12.00
16.00
20.00
Figure 8. Diagram of average accumualted eroded soil depth on
remediated and non-remediated hillslopes.
4.2 Vegetation coverage
Utilizing NDVI, multi-stage vegetation recovery of the overall
Longjiao subwatershed after remediation was assessed for the
five events. According to Eq. (2) and (3), the ratios of
vegetation coverage were calculated. Figure 9 shows that after
remediation, typhoon Fanapi and typhoon Namodol repeatedly
affected Longjiao subwatershed but the vegetation coverage
ratio (VR) still remained over 80%. This value was estimated by
satellite images and is better than the ratio after typhoon
Morakot. Again, these results show that remediation including
check dams, river bed foundation, and revetment as well as
excavation of deposited sediments, can effectively reduce
sediment yield.
Figure 9. Evolution of vegetation coverage in pre-remediation and post-
remediation for Longjiao subwatshed.
4.3 Sediment trapping
Soil and Water Conservation Bureau (2011) has collected three
high-precision digital elevation models from aerial orthoimages
and airborne LiDAR. These measurements can be divided into
pre-remediation and post-remediation. Further, sediment yield
is the total volume of terrain changes such as slope failures and
river erosions by grid subtraction of DTMS. Sediment trapping
ratio (STR) can be assessed by measuring the amount of
sediment trapped in front of the check dams, which has been
listed in Table 1. Compared with the results listed in table 1,
post-remediation STR of Longjiao subwatershed is significantly
higher than pre-remediation by 17.18 times. In the meantime,
the sediment yield after remediation is lower than before
remediation. Through the above results, it was found that
sediment yields were effectively controlled.
Table1 List of Sediment trapping ratio of Longjiao subwatershed
Stages Duration
Sediment yield
(m
3
)
Sediment trapping
(m
3
)
STR (%)
pre-remediation 2008~2010 1,548,300 34,540 2.2
post-remediation 2010~2011 149,143 56,373 37.8
5 CONCLUSIONS
This study systematically integrates multi-scale sediment
monitoring techniques to analyze soil erosion, vegetation
coverage, and sediment trapping from easily measured physical
quantities such as depth, area, and volume in a watershed scale.
Thorough the case study, it suggests that remediation in
Tsengwen Reservoir Watershed are certainly effective and are
able to reduce sediment production and soil loss entering a
reservoir.
6 REFERENCES
Soil and Water Conservation Bureau (SWCB), 2010. Multi-scale
environmental measurements and evaluation of conservation and
management efficiency for Tsengwen reservoir watershed (in
Chinese).
Lo, W.-C., Lin, Bor-Shiun., Ho, H.-C., Keck Jeff k, Yin, H.-Y. and H.-
Y., Shan 2012. A simple and feasible process for using multi-stage
high-precision DTMs, field surveys and rainfall data to study
debris-flow occurrence factors of Shenmu area, Taiwan, Natural
Hazards and Earth System Sciences, 12, 3407-3419.
Lin, Bor-Shiun, Cheng-Yang Hsiao,Wai-Yi Leung and Shu-Yeong Chi
2012. Using Airborne LiDAR Technology to Analyze Landslide
Hazards in Shih-men Watershed, European Geosciences Union,
General Assembly 2012, 14, EGU2012-2884, Vienna, Austria,
22~27, April.
Schumm, S. 1956. Evolution of Drainage Systems and Slopes in
Badland at Perth Amboy, Bulletin of Geological Society of
America, 67, 597-646.
Kriegler, F.J., Malila, W.A., Nalepka, R.F., and Richardson, W., 1969.
Preprocessing Transformations and Their Effects on Multispectral
Recognition, Proceedings of the Sixth International Symposium on
Remote Sensing of Environment, 97-131.
Sophie L., Tim W. E., Peter B. H., and David J. T. 2008. Sediment
trapping by a tree belt: processes and consequences for sediment
delivery, Hydrological Processes, 22(17), 3523-3534.
575
Practice and development of the piezocone penetration test (CPTu) in geotechnical
engineering of China
La pratique et le dveloppement de l'essai de pntration au pizocne (CPTu) en Chine
Liu S., Cai G., Du Y.
Institute of Geotechnical Engineering,Southeast University, Nanjing 210096, China
Puppala A.J.
Department of Civil Engineering, The University of Texas at Arlington, Arlington, Texas
ABSTRACT: The cone penetration test (CPT) technique is widely used in field site investigation due to its fast, repeatable, and cost-
effective advantages. It can provide near-continuous information of soil properties and has a strong theoretical background. In this
paper, the history and current development status of CPT, particular the cone penetration test with pore pressure measurement (CPTu)
in China practice is systematically introduced. The relationship between international standardized CPTu and China CPT is proposed
based on a great number of soils. The paper then presents the review and comparison of the soil characterization methods based on
CPTu tests results in China, including stress history, deformation, consolidation and permeability characteristics.
RSUM: Lessai de pntration au cne (CPT) est largement utilis dans les enqutes de terrain pour ses avantages rapides,
reproductibles et rentables. Il peut fournir des informations quasi continues des proprits du sol et il a une solide thorie. Dans cet
article, lhistoire et ltat et dveloppement du CPT, notamment lessai de pntration au pizocne (CPTu) qui peut mesurer la
pression deau interstitielle est systmatiquement introduit en Chine. La relation entre les rsultats normalise interationale CPTu et
des CPTs en Chine est propose selon beaucoup des donnes. Par la suite, le document prsente lexamen et la comparaison des
mthodes de caractrisation des sols par les rsultats des CPTUs en Chine, y compris lhistorique des contraintes, des dformations,
des caractristiques de permabilit et de consolidation.
KEYWORDS: Site investigation, CPT, CPTu, Engineering Characterization

1INTRODUCTION
The cone penetration test (CPT) has been used for decades to
investigate the properties of soil in situ. Essentially, the test
consists of pushing a penetrometer with a standard geometry
(cylindrical with a diameter of 35.7 mm and a conical point
with an apex angle of 60
o
) into the soil at a rate of 20 mm/s,
while measuring a number of parameters. The cone penetration
test (CPT) is widely used in-situ testing method, especially in
soft soil exploration. As a new kind of in-situ test technique, the
piezocone penetration test (CPTu) has been attracting wide
attention and widely used in the western developed country. It
has been increasingly used because of its important advantages,
such as simplicity, speed and continuous profiling. The
piezocone, which provides near-continuous measurements of tip
resistance, sleeve friction, and pore water pressure induced
during the penetration, appears to be a powerful tool for
determining the stress history of soft clay deposits.
The mechanical CPT like Dutch cone was developed by
foreign engineers in Shanghai in the early 1930s (Liu and Wu,
2004). In 1954, the Holland mechanical CPT was first
introduced into China. In 1964, the first electric single bridge
CPT was independently produced in China with only one
measurable parameter (e.g., total specific penetration
resistance). Later, the double bridge CPT was developed to
measure the tip resistance and sleeve friction independently in
the 1970s, which is currently used in Chinese standards. The
Holland CPTu was introduced into China in the early 1990s by
Nanjing Hydraulic Research Institute, but its follow-up
development is very slow. The multifunctional CPTu was
introduced into China in 2005 by Southeast University
researchers. In the following years, the related theory analysis
and application practice of CPTu in China have been developed
rapidly. In this paper, the relationship between international
standardized CPTu and China CPT is proposed based on a great
number of soils. The paper then presents the review and
comparison of the soil characterization methods based on CPTu
tests results in China, including determination of stress history,
deformation, consolidation and permeability characteristics.
2 COMPARISON BETWEEN INTERNATIONAL
STANDARDIZED CPTU AND CHINA CPT
Due to the inner geometry of a cone penetrometer, the ambient
pore pressure will act on the shoulder area behind the cone.
Therefore, the measured cone resistance should be corrected by
the unequal area effect for the data presentation and
interpretation. In literature works, most of the correlations were
developed based on CPT with cone cross area of 10 cm
2
as per
international standardized CPT and CPTu tests. However, in
China, both 15 cm
2
and 20 cm
2
CPT devices are frequently
used. Therefore, the internationalization of Chinese CPT is
inevitable (Liu and Wu, 2004).
When different sizes of CPT and CPTU are employed, the
question of scale effects inevitably arises. For piezocones
ranging in area from 5 to 15 cm
2
, the usual assumption, based
on experience summarized by Lunne et al. (1997), is that scale
effects are negligible in soil layers of sufficient thickness
relative to the cone diameter: that is, quantities such as the cone
resistance and excess pore pressure do not depend on the size of
the piezocone. Powell and Lunne (2005) compared the results
using the 10 cm
2
and 15 cm
2
piezocones in UK clays. The
comparison of various cone sizes and configurations between
China CPT and international standard CPTu device at 28 field
testing sites is presented. To avoid the variability, all the tests
were performed by the same operators. The elevations of the
576
th
ground surface at different sites were measured and the
difference of elevation may be considered.
To quantify the differences between CPT and CPTU
measurements, the ratios of the CPTU to CPT cone resistance
and sleeve friction measurements were calculated for each site.
The reference line positioned at an average CPTU to CPT ratio
equal to one in the plots of average ratios represents the
theoretical value if soil variability was eliminated and if there
was no effect of cone size. In general, the ratios of cone
resistance and sleeve friction measurements fluctuate near one,
and the measured values increase with depth. For the soft clay
sites (Figure 1), the average fs ratio of the friction sleeve is
always significantly greater than the average qc ratio. For the
topsoil such as fill and silty clay, the ratios CPTU to CPT
fluctuate drastically. The relationships of derived key
parameters are developed between China CPT and international
CPTu (Table 1). From the perspective of engineering practice, it
is concluded that q
t
= 1.03 q
c
f
s
-
CPTu
= 1.05 f
s
-
CPT
. The
empirical equation can be used as useful engineering tool to
guide use of 10 cm
2
international CPTu in China.
16
14
12
10
8
6
4
2
0
0 1 2 3 4
D
e
p
t
h
/
m
Coneresistance
Sleevefriction
Averageof CPTU/CPTratios

Figure 1. Statistical analysis of q
c
and f
s
ratios

Table 1. Conversion relationships between CPTu and CPT parameters.

Soil types Regression equation
Cone resistance Sleeve friction
Soft clay q
t
= 1.04q
c
f
s-CPTU
= 1.01 f
s-CPT

Clay q
t
= 1.02q
c
f
s-CPTU
= 1.07 f
s-CPT

Stiff clay q
t
= 1.01q
c
f
s-CPTU
= 1.06 f
s-CPT

Silt q
t
= 1.03q
c
f
s-CPTU
= 1.06 f
s-CPT

Silty sand q
t
= 1.03q
c
f
s-CPTU
= 1.03 f
s-CPT

3 EVALUATION OF ENGINEERING
CHARACTERISTICS BASED ON CPTU TESTS IN CHINA
3.1 Stress history
Since the advent of CPTU in geotechnics, nearly 20 different
methods have been suggested for interpreting the
preconsolidation pressure and the overconsolidation ratio of
clays (Mayne 1991). In this study, the three sites with all
sensitive clay deposits in Jiangsu province of eastern China are
selected (Liu et al., 2007). These Quaternary clay deposits are
located at Lianyungang, Changzhou and Nanjing respectively.
Whenever possible, the OCR values interpreted from various in
situ tests were compared with the oedometer values for
Lianyungang marine clay. At the other test sites, in addition to
oedometer results, some field OCR values were deduced from
field performance observation of trial embankments. These field
values provide a reliable basis for evaluating the validity of the
various interpretation methods in Jiangsu clays. Table 2
presents a summary of the typical property values of the soil
layers.

Table 2. Typical property values of the soils.

Site Soil type
Water
content/%
Liquid
limit
Plasticity
index
Lianyungang muck 79.6 75.6 35.8
Changzhou clay 23.5 41.8 19.0
Nanjing
mucky
silty clay
41.4 36.4 13.3

Figure 2 shows the relationship between net tip resistance
(q
t

vo
) and the preconsolidation pressure (
`
p
) measured in the
laboratory oedometer test on high-quality samples. Here, n as
shown in Fig. 1 is the number of data available. In Lianyungang
marine clay deposits, the correlation is excellent (r
2
= 0.99) for
all the data, and the preconsolidation pressure varies between 29
and 139 kPa. In Changzhou clay site, the correlation is good (r
2

= 0.95) for the data, and the preconsolidation pressure varies
between 812 and 1789 kPa. It can be seen from Figure 5 that the
relationship between net tip resistance and the preconsolidation
pressure of Nanjing clay site is also pretty good (r
2
= 0.98).
Consequently, we can obtain the value of Nt factor, which is
equal to 2.7 according to the correlation relationship for
Lianyungang lightly overconsolidated clay. Similarly, for
Changzhou lightly to moderately overconsolidated clay, the N
t

factor is 2.2. For Nanjing backswamp clay deposit, the N
t

factor is 2.5. Consequently, the equation defining the correlation
of Lianyungang marine clay site can be expressed as follows:

7 . 2
0 ' v t
p
q
(1)

100 1000
10
100
1000
'
p
-
o
e
d
/
k
P
a
q
t
-
v 0
/ k Pa
Li any ungang n=24, r
2
=0. 99
Changz hou n=28, r
2
=0. 95
Nanj i ng n=20, r
2
=0. 98

Figure 2. The relationship between preconsolidation pressure measured
in oedometer test and net tip resistance

3.2 Deformation modulus
The one-dimensional constrained modulus, M, as measured in
an oedometer test, has been expressed in terms of a coefficient,
m, and cone resistance:
c m
q M (2)
where
m
is a correlation factor. In practice, it has been
usual to correlate the modulus M to a penetration resistance. To
estimate one-dimensional constrained modulus M, the
correlation with net cone resistance (q
t
-
v0
) is used in the form
(Kulhawy and Mayne, 1990):

8.25 8.25
n t
M q q
vo
(3)

Comparisons between M from CPTU with laboratory
oedometric modulus for various types of soil proposed by
Kulhawy and Mayne (1990) showed that the ratio M -
CPTU
/ M -
lab
could equal to 2.21 for high-plasticity clays and silty soils. In
Figure 3 the constrained modulus estimated with relationship
(3) is plotted against that determined by laboratory oedometric
tests, carried out on all the types of Jiangsu lagoonal soils. In
our case, the ratio M-
CPTU
/M-
lab
is always greater than the unity
and is not influenced by the type of soil or by its cone resistance
value (Cai et al. 2010).
577

0 4 8 12 16 20 24 2
0
4
8
12
16
20
24
28
8
Kulhawy and Mayne(1990)
M
-CPTU
=2.21M
-Lab
R
2
=0. 98
M
-
C
P
T
U
/
M
P
a
M
- l ab
/MPa
CH
CL
ML

Figure 3. Measured versus predicted constrained modulus values

To examine the possibility for better correlations to estimate
the constrained modulus from CPTu data, the corrected cone tip
resistance (q
t
) and the net cone resistance (q
t
-
vo
) were plotted
against the laboratory measured constrained modulus as shown
in Figures 4(a) and 4(b). A linear correlation was obtained
between M and qt as follows:
M =3.36q
t
R
2
=0.78 (4)
And the following linear correlation was also obtained
between M and (q
t
-
v0
) given as follows:
M =3.73(q
t
-
v0
) R
2
=0.92 (5)
The arithmetic mean and standard deviation of (M
-CPTU
/M
-
lab
) are 0.85 and 0.23 for the first correlation (M = 3.36q
t
),
whereas 1.02 and 0.29 for the second correlation [M=3.73(q
t
-
vo
)].

0 1 2 3
0
2
4
6
8
10
M
e
a
s
u
r
e
d

M
/
M
P
a
q
t
/ MPa
M=3. 36q
t
R
2
=0. 78

(a)

0 2 4 6 8
0
5
10
15
20
25
30
M
e
a
s
u
r
e
d

M
/
M
P
a
( q
t
-
v 0
) / MPa
M=3. 73( q
t
-
v 0
)
R
2
=0. 92

(b)
Figure 4. (a) q
t
~Measured M ; (b) (q
t
-
v0
) ~Measured M

3.3 Consolidation and permeability properties
Many theoretical and semi-empirical methods have been
proposed for deriving the coefficient of consolidation from
CPTu dissipation data. Teh and Houlsby (1991) proposed a
relationship between a dimensionless time factor and c
h
value
based on numerical analysis of dissipation pore pressure with
the consideration of soil rigidity index parameter. Robertson et
al. (1992)

reviewed some dissipation data from piezocone tests,
and concluded that the predicted coefficient of consolidation by
Teh and Houlsby (1991) solution compared well with reference
values from laboratory tests and field observations. Schnaid et
al. (1996) and Danziger et al. (1997) showed that, when Teh and
Houlsbys approach was employed to interpret various CPTU
results, the calculated values of c
h
were of the same order of
magnitude as those measured independently in oedometer tests
in the laboratory. Abu-Farsakh and Nazzal (2005) compared
seven CPTU methods and showed that Teh and Houlsby (1991)
and Teh (1987) methods can estimate c
h
value better than the
other prediction methods.
Teh and Houlsby (1991) proposed a relationship between a
dimensionless time factor and c
h
value. The dimensionless time
factor, T
*
, is defined as:

5 . 0 2
*
r
h
I r
t c
T
(6)

where c
h
= coefficient of consolidation in horizontal
direction; r = radius of cone, typically 17.85 mm; I
r
= rigidity
index, G/S
u
. Among the methods available for evaluating c
h

from piezocone dissipation tests, the one proposed by Teh and
Houlsby (1991) is probably most widely used (Robertson et al.
1992). Teh and Houlsbys (1991) equation is as follows:
2
50
50
r
t
T
c
h

(7)

where the time factor T
50
is related to the location of the
filter element and cone size. For a cone with a cross-sectional
area of 10 cm
2
and with a shoulder filter element, T
50
= 0.245
(Teh and Houlsby 1991). The t
50
is the measured time for 50%
dissipation. The method proposed by Teh and Houlsby (1991)
was used here to interpret the coefficient of consolidation for
the pore pressure dissipation curves in this study.
The coefficient of permeability in the horizontal direction
can be estimated from a CPTU dissipation test and by means of
the correlation factor (k
h
/k
v
) proposed by Jamiolkowski et al.
(1985). Baligh and Levadoux (1980) recommended that the
horizontal coefficient of permeability could be estimated from
the expression:
h
v
w
h
RRc k
'
0
3 . 2
(8)
where RR = compression ratio in the overconsolidated
range, and can be obtained from the consolidation tests at the
corresponding stress level.
A comparison of the consolidation coefficient values
measured by CPTu dissipation test and laboratory oedometer
test is presented in Figure 5 in which the CPTU test measures c
h

values, whereas the conventional oedometer test measures c
v
. It
can be seen that the c
v
values measured by oedometer test are
lower than the c
h
values measured by CPTu tests. The c
h
values
of the lacustrine clay measured by the CPTu tests are generally
4-6 times larger than the c
v
values measured by the
conventional oedometer test, indicating anisotropic
characteristic of the soil.

25
20
15
10
5
0
0 1 2 3 4 5
D
e
p
t
h

(
m
)
Coefficent of consolidation (10
-3
cm
2
/s)
CPTU
Oedometer

Figure 5. Comparison of cv and ch profiles measured from CPTu and
laboratory tests

578
th
The k
v
values measured by oedometer test and k
h
values
deduced from CPTu test are compared with each other in Figure
6. The values of coefficient of permeability from back-analysis
and falling-head permeability tests are also presented in Figure
6. The comparison shows that the k
h
values measured by
oedometer are lower than those obtained from CPTu test within
1-2 order of magnitude. The coefficient of permeability of
Taihu lacustrine clay is in the order of 10
-5
-10
-7
cm/s. The k
h

value measured by falling-head permeability tests agrees well
with that determined by CPTu tests.
6 REFERENCES
Abu-Farsakh, M.Y. & Nazzal, M.D. 2005. Reliability of piezocone
penetration test methods for estimating the coefficient of
consolidation of cohesive soils. Journal of the Transportation
Research Board. No. 1913, 62-76.
Baligh, M. M. & Levedoax, J. N. 1980. Pore pressure dissipation after
cone penetration. PhD Thesis, Massachusetts Institute of
Technology. Cambridge, Mass, 80-111.
Cai, G., Liu, S., Tong, L., 2010. Field evaluation of deformation
characteristics of a lacustrine clay deposit using seismic piezocone
tests. Engineering Geology, 116(3-4), 251-260.

25
20
15
10
5
0
1E-8 1E-7 1E-6 1E-5
k
h
(cm/s)
D
e
p
t
h

(
m
)
CPTU
Falling-head tests
Asaoka back-analysis
Oedometer

Danziger, F.A.B., Almeida, M.S.S. & Sills, G.C. 1997. The significance
of the strain path analysis in the interpretation of piezocone
dissipation data. Geotechnique, 47(5): 901-914.
Jamiolkowski, M., Ladd, C.C., Germaine, J.T., and Lancellotta, R.
1985. New developments in the field and laboratory testing of soils:
State of the art. Proceedings of the 11th International Conference
on Soil Mechanics and Foundation Engineering, San Francisco, 12-
16 August 1985. A.A. Balkema, Rotterdam. Vol. 1, pp. 57-153.
Kulhawy, F.H., and Mayne, P.W. 1990. Manual on estimating soil
properties for foundation design. Report No. EL-68000, Electric
Power Research Institute, EPRI, August 1990. Palo Alto. 306 pp.
Liu, S. Y. and Wu, Y. K. 2004. On the state-of-art and development of
CPT in China. Chinese Journal of Geotechnical Engineering, 26(4),
553-556.
Liu, S. Y. Cai, G. J. Tong, L. Y. Du, G. Y. 2007. On preconsolidation
pressure of clays from piezocone tests. Chinese Journal of
Geotechnical Engineering, 29(4): 490-495.
Figure 6. Comparison of k
h
profiles measured by different methods

4 CONCLUSIONS
The comparison of various cone sizes and configurations
between China CPT and international standard CPTu device at
28 field testing sites is presented. The relationships of derived
key parameters are developed between China CPT and
international CPTu. From the perspective of engineering
practice, it is concluded that q
t
= 1.03 q
c
f
s-CPTU
= 1.05 f
s-CPT
.
The empirical equation can be used as useful engineering tool to
guide use of 10 cm
2
international CPTu in China. The field
CPTu tests were carried out in Jiangsu sites to evaluate the
stress history. Consequently, we obtained the value of N
t

factor, which is equal to 2.7 according to the correlation
relationship for Lianyungang lightly overconsolidated clay. The
results show that ratio of M derived from Kulhawy and Mayne
1990s method to that determined from laboratory oedometric
tests, M
-CPTU
/M
-lab
, practically equals to 2.21 for high plasticity
clays. A quick estimation of the magnitude of coefficient of
consolidation c
h
is proposed by pore pressure dissipation (type
u
2
) tests from the CPTu database. Comparisons of the results
obtained by different methods indicate that the values of
horizontal coefficient of consolidation determined by CPTu are
typically 4 to 6 times those of laboratory tests. The coefficient
of permeability values measured by laboratory tests are less
than by almost 1-2 orders of magnitude with that determined by
CPTu tests in Jiangsu soft clays.
Lunne, T., Robertson, P.K., Powell, J.J.M., 1997. Cone penetration
testing in geotechnical practice, Blackie Academic & Professional,
Chapman & Hall, London.
Mayne, P.W. 1991. Determination of OCR in clays by piezocone tests
using cavity expansion and critical state concepts. Soils and
Foundations, 31(2): 65-76.
Powell, J. J. M. and Lunne, T. 2005. A comparison of different sized
piezocones in UK clay. Proceedings of the 16th International
Osaka, 729-734.
Robertson, P.K., Sully, J.P., Woeller, D.J. et. al. 1992. Estimating
coefficient of consolidation from piezocone tests. Canadian
Geotechnical Journal, 29(4):539-550.
Schnaid, F., Sills, G.C., Soares, J.M. & Nyirenda, Z. 1996. Predictions
of the coefficient of consolidation from piezocone tests. Canadian
Geotechnical Journal, 34, No. 2, 315-327.
Teh, C.I. 1987. An analytical study of the cone penetration test. PhD
Thesis, Dept. of Civil Engineering, Oxford university.
Teh, C.I. & Houlsby, G.T. 1991. An analytical study of the cone
penetration test in clay. Geotechnique, 41(1): 17-34.

5 ACKNOWLEDGEMENTS
The work in this paper was funded by the National Natural Scie
nce Foundation (Grant No. 41202203) of China,Twelfth five-y
ear National Science and Technology Support Plan (Project No
. 2012BAJ01 B02) and the Key Project of Natural Science Foun
dation (Grant No. BK2010060) of Jiangsu Province of China.

579
The use of hydro test results for design of steel tanks on stone column improved
ground - a case history
Lemploi des rsultats des essais hydrauliques dans ltude des rservoirs en acier sur le sol
amlior par colonnes de pierre histoire de cas
Matei L.
Geokon-Zagreb, Zagreb, Croatia; University of Rijeka, Faculty of Civil Engineering, Rijeka, Croatia
Mihaljevi I., Grget G.
Geokon-Zagreb, Zagreb, Croatia
Kvasnika P.
University of Zagreb, Faculty of Mining geology and petroleum engineering, Zagreb, Croatia
ABSTRACT: This paper describes hydro tests performed on five large storage tanks (80.000 m
3
each) located at the Sisak Oil
Terminal, Croatia. Because of its small stiffness and low water permeability, foundation soil for each tank was improved with 660
stone columns. In order to reduce the risk of accidents, such as fire, breach or leak, a crude oil storage tank requires stringent security
measures. In the case of the Sisak tanks, the hydro tests were conducted as part of technical monitoring to determine a set of
documented and interconnected activities which would provide proof of proper functioning of all elements of a tank structure. In case
of critical deviations from the operation expected, such activities ensure that such deviations are removed or corrected on time by
taking necessary measures approved by experts. Under a procedure for the hydro tests, the phases of tank and bund filling were
defined and each phase was followed by visual inspection and measurements of settlements and deformations of the steel structure.
Design directions for future foundation and hydro tests of tanks were made accordingly.
RSUM : Les essais hydrauliques conduits sur cinq rservoirs de grande taille (chacun de 80,000 m3 ) situs dans le Terminal
ptrolier de Sisak en Croatie, sont dcrits dans louvrage. Compte tenu de petite rigidit et permabilit leau peu importante, le sol
de fondation pour chaque rservoir a t amlior avec 660 colonnes de pierre. Les mesures trs rigoureuses doivent tre prises pour
les rservoirs ptrole brut afin de rduire le risque daccidents tels que feu, rupture ou fuite de ptrole. Dans le cas des rservoirs de
Sisak, les essais hydrauliques ont t conduits dans le cadre de la surveillance technique dont le but tait de dfinir une srie des
activits bien documentes et interconnectes vises prouver le fonctionnement impeccable de tous les lments structurels du
rservoir. Dans le cas dune dviation critique par rapport au fonctionnement normal, ces activits permettent llimination ou la
correction prompte de ces dviations en prenant les mesures appropries approuves par les experts. Dans la procdure pour les essais
hydrauliques, les phases de remplissage du rservoir et de la cuvette de rtention ont t dfinies, et chaque phase a t suivie par une
inspection visuelle et par mesurage du tassement et des dformations de la construction en acier. Les instructions dtudes sont
fournies pour les essais hydrauliques et les essais des fondations futurs.
KEYWORDS: steel tank, stone columns, hydro test, monitoring, settlement
1 INTRODUCTION
During the years 2010 and 2011, five new crude oil storage
tanks were built at the Sisak Oil Terminal.
80.000 m
A-2510
N
50 100 150 200 m 0
3
80.000 m
A-2509
3
80.000 m
A-2508
3
80.000 m
A-2507
3
80.000 m
A-2511
3
A A
cross section A-A
bund
floating roof
tank
Figure 1. The layout plan of the tanks
All the tanks are of the same size and have an identical steel
structure with a floating roof and steel bund wall designed in
accordance with API 650. The diameter of the tanks and bunds
is 73,2 m and 78,2 m respectively. The tanks have 80.000 m
3
in
volume and their total and overflow height is 20,6 m and 19,5 m
respectively (Figure 1.).
Foundation soil is horizontally stratified and, therefore, the
soil under all the tanks is of almost the same properties. Because
of its small stiffness and low water permeability, the foundation
soil for each tank was improved with hundreds of stone
columns, which is a technology applied in similar cases of soil
improvement (Raju et al 2004, Ambily and Gandhi 2004).
In order to prevent industrial accidents, viz. fire, breach or
leak of a tank, etc. to happen, a crude oil storage tank requires
special safety measures. For this reason, all tank development
stages such as ground investigations, design, construction,
hydro tests and exploitation, were strictly controlled according
to a highly elaborated plan as laid down in API 653 and EN
14015.
On the basis of in situ and laboratory tests, a numerical
model was created in Plaxis, and all phases of hydrostatic tests
were checked before testing.
2 DESIGN REQUIREMENTS
In the near vicinity of the new tanks, three 80.000 m
3
tanks with
floating roofs were installed 30 years ago. During the hydro test
performed on one of them, the yielding of foundation soil
580
th
occurred, which caused tank shell deformations. Such
deformations affected normal operation of the floating roof and
consequently made the use of the tank impossible. The case
described above has not been fully documented, but this is the
reason why the investor imposed strict requirements for tank
behaviour.
Based on the documented cases of soil yielding during tank
foundation works (Bell & Iwakiri, 1980) criteria for design,
construction and use of tanks have been established.
Criteria for maximum total and differential settlement of
tanks were determined according to Marr, et al. (1982), API-650
and API-653. Allowable differential settlements are the
maximum allowable design limits for deformation of the tank
after allowance has been made for construction tolerances.
These comprise combinations of: (a) tilt of the tank; (b) tank
floor settlement along a radial line from the perimeter to the
tank centre; and (c) settlement around the perimeter of the tank.
Foundation and foundation soil are subjected to the highest
load during hydro tests when a tank is filled with water having a
density of 1 t/m
3
. Later, during tank use, loads on foundations
and foundation soil are lower by about 15 % because the tanks
are filled with crude oil having a density of 0,85 t/m
3
.
3 OVERVIEW OF GROUND INVESTIGATION WORKS
At the site, 45 geotechnical boreholes were drilled of which
three were 70 m deep. In addition to the boreholes, 18 CPTU
tests were also carried out. From the boreholes, undisturbed soil
samples were continually taken or SPTs performed. A
piezometer was installed in one borehole and a level of ground
water monitored over a number of years. Soil classification tests
as well as strength, stiffness and water permeability tests were
carried out in a laboratory. The investigations showed that the
soil is horizontally stratified.
4 DESCRIPTION OF FOUNDATION
The foundation soil was improved with hundreds of stone
columns. After the soil had been prepared in this way, the tank
shell and bund wall were installed on rigid reinforced concrete
ring while the tank bottom was placed directly on the bedding
prepared.
4.1 Soil improvement
As the foundation soil is horizontally stratified, the soil under
all the tanks has almost the same properties. Because of its
small stiffness and low water permeability, the foundation soil
for each tank was improved with about 660 stone columns.
The depth of the improved soil was approximately 18 m. The
spacing between stone columns varied depending on their
location on the layout plan. Considering that tank structure is
susceptible to planar tilt settlement and non-planar settlement,
stone columns were spaced more closely on the perimeter below
the foundation ring and centre to achieve stronger effect of
improvement.
The quality of improvement was checked by CPTU and
SASW tests and geodetic surveys carried out in control fields
before and after soil improvement. In addition, the data relating
to the installation of stone columns were analyzed. Among other
things, the volume of the gravel pressed into foundation soil
was determined. For each tank, it was found to be about 3% of
the volume of the foundation soil improved. As geodetic
surveys showed negligible soil upheave (a few millimetres), it
can be considered that all the stone pressed into the soil
increased directly its density, i.e. soil compaction.
4.2 Concrete ring foundation
The shells of both the tanks and bund walls were mounted
directly on a rigid reinforced-concrete foundation ring of
rectangular cross-section b/h=350/(260-370) cm, with a central
drainage gutter having a width b=60 cm and variable height
h=4080 cm. Outside and inside ring diameters are
D
out
=79,00 m and D
in
=72,00 m respectively.
4.3 Bedding of steel tank bottom
The bottoms of the steel tanks were mounted directly on the
multi-layer bedding prepared as described below.
The foundation soil was levelled and a layer of gravel of
grain size 0-64mm was placed. The layer had 70 cm in
thickness. To prevent soil pollution in case of tank leak, a
HDPE geomembrane was installed in the bedding. The
geomembrane was placed between a geosynthetic clay liner and
clean sand to protect it from damage. Above the geomembrane,
cathodic protection was installed. Additional reinforcement of
the soil below the tanks was achieved by placing sand in
geocells of 20 cm in height.
5 GEOTECHNICAL DESIGN ANALYSIS
On the basis of in situ and laboratory tests, an axisymmetric
numerical model was created in Plaxis 2D-V8. The material
behaviour is represented by the Hardening Soil model.
In a numerical analysis, soil materials of five types were
used and their description and some properties are shown in
Table 1. The analysis included tank installation stages, hydro
tests and tank exploitation.
Table 1. A description of stratified foundation soil
Layer depth [m] description
(1)
approx
0 6
Surface layer of stiff clay
k=0,0002[m/day]; E
oed
ref
=9,4[MPa];
E
ur
ref
=30[MPa]; p
ref
=100[kPa];
m=0,409
(2)
approx
6 13
Layer of soft clay
k=0,0002[m/day]; E
oed
ref
=7,8[MPa];
E
ur
ref
=23,4[MPa]; p
ref
=100[kPa];
m=0,376
(3)
approx
13 20
Sand with silt and clay
k=0,02[m/day]; E
oed
ref
=16[MPa];
E
ur
ref
=50[MPa]; p
ref
=100[kPa]; m=0,5
Soil below the tanks was improved
with stone columns as designed
k=1[m/day]; E
oed
ref
=35[MPa];
E
ur
ref
=90[MPa]; p
ref
=100[kPa]; m=0
(4)
approx
20 70
Alternating layers of clay and sand
with silt
k=0,0002[m/day]; E
oed
ref
=10[MPa];
E
ur
ref
=40[MPa]; p
ref
=100[kPa];
m=0,376
k - permeability; E
oed
ref
reference edometric modulus at
reference stress p
ref
; E
oed
=E
oed
ref
(/p
ref
)
m
edometric modulus;
E
ur
ref
= unload/reload modulus
6 HYDRO TEST
6.1 Introduction
In the case of the Sisak tanks, hydro tests were conducted as
part of technical monitoring to determine a set of documented
and interconnected activities which would provide proof of
proper behaviour of all elements of a tank structure.
In case of malfunction or critical deviations from the
expected behaviour, such activities would ensure that these
deviations are removed or corrected on time by taking necessary
measures approved by experts.
Under a procedure for a hydro test, the phases of tank and
bund filling and emptying were defined; after each phase had
581
been completed, visual inspection and measurements of
settlements and deformations of the steel structure were made.
The next phase of the hydro test could begin only after the
analysis of the results of measurements obtained from a
previous one had been made.
Monitoring of each tank settlement involved geodetic
surveys and measurement of settlements by a horizontal
inclinometer. The geodetic surveys included 24 points on the
outside perimeter of the foundation and one point on each of six
manholes allowing access to horizontal inclinometer pipes (HI).
Settlement measurements were made in three 100-m long pipes
running below the tank centre and horizontally declined by 60.
The manholes were located eight meters outside of the
foundation perimeter.
A monitoring programme was planned and carried out in a
similar manner as described in the paper by Berardi and
Lancellotta (2002).
6.2 Hydro test results
By way of illustration, the results of settlements obtained from
the hydro test carried out on Tank A-2507 are given. Figure 2
shows time history of tank and bund filling and emptying
together with the graphs showing averaged settlements of HI
pipes at manholes HI, points on the foundation perimeter and
centre.
-8
-7
-6
-5
-4
-3
-2
-1
0
s
e
t
t
l
e
m
e
n
t

[
c
m
]
manhole R=47,5 m
tank periphery - R=39,5 m
tank center R= 0 m
2
7
/0
2
/1
0
0
6
/0
3
/1
0
1
3
/0
3
/1
0
2
0
/0
3
/1
0
2
7
/0
3
/1
0
0
3
/0
4
/1
0
1
0
/0
4
/1
0
1
7
/0
4
/1
0
2
4
/0
4
/1
0
0
1
/0
5
/1
0
0
8
/0
5
/1
0
1
5
/0
5
/1
0
2
2
/0
5
/1
0
2
9
/0
5
/1
0
0
5
/0
6
/1
0
date [dd/mm/yy]
0
2
4
6
8
10
12
14
16
18
20
w
a
t
e
r

l
e
v
e
l

[
m
]
hydro test for
tank A-2507
tank
bund wall
Figure 2 Time history of the hydro test performed on Tank A-2507
According to settlement criteria (Marr et al 1982), the design
defined allowable total and differential settlements for different
settlement patterns. Thus, during the hydro test, the allowable
total settlement of the tank perimeter and tank centre were 15
cm and 31 cm respectively. In a calculation, they were
estimated to be 11 cm and 19,5 cm respectively. However, the
results of measurement obtained for such settlements as shown
on the graph were 3,6 cm and 7,2 cm respectively.
Figure 3 shows total settlements of the tank perimeter for the
two representative phases of the hydro tests in which the largest
settlements occurred at the highest loads. In the third phase, i.e.
when the tank was emptied, permanent (plastic) deformations
occurred.
0 50 100 150 200 250
tank perimeter [m]
-4.5
-4.0
-3.5
-3.0
-2.5
-2.0
-1.5
-1.0
t
o
t
a
l

s
e
t
t
l
e
m
e
n
t

[
c
m
]
hydro test for tank A-2507
date / water level
26.3.2010 tank=17 m; bund=17m
07.04.2010 tank=19 m; bund=0m
07.06.2010 tank=0 m; bund=0 m
plane of rigid tilt
Figure 3 Total settlements of the tank perimeter occurred during the
hydro test on Tank A-2507
Figure 4 illustrates the two phases in which the largest
differential settlements of tank perimeter at the highest load
occurred as well as the phase following tank emptying in which
permanent (plastic) deformations occurred.
Design allowable differential settlements for the cases of
planar tilt settlement and non-planar differential settlement was
44 cm and 0,8 cm respectively. The planar tilt settlement and
non-planar differential settlement obtained by calculation were
1,3 cm and 0,34 cm respectively. The results of measurement
obtained for such settlements were 1,0 cm and 0,32 cm
respectively.
0 50 100 150 200 250
tank perimeter [m]
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
n
o
n
-
p
l
a
n
a
r

d
i
f
f
e
r
e
n
t
i
a
l

s
e
t
t
l
e
m
e
n
t

[
c
m
]
-1.0
-0.5
0.0
0.5
1.0
p
l
a
n
a
r

t
i
l
t

d
i
f
f
e
r
e
n
t
i
a
l

s
e
t
t
l
e
m
e
n
t

[
c
m
]
26.3.2010 tank=17 m; bund=17m
07.6.2010 tank=0 m; bund=0m
planar diff. settl. 26.3.2010
planar diff. settl. 07.06.2010
non-planar diff. settl. 26.3.2010
non-planar diff. settl. 07.06.2010
Figure 4 Differential settlements at the perimeter of Tank A-2507
Figure 5 shows total settlements for the three representative
phases of the hydro test measured by horizontal inclinometer.
-50 -40 -30 -20 -10 0 10 20 30 40 50
tank diameter [m]
-8
-7
-6
-5
-4
-3
-2
-1
0
t
o
t
a
l

s
e
t
t
l
e
m
e
n
t

[
c
m
]
horizontal inclinometers along tank diameter
date / water level
26.3.2010 tank=17 m; bund=17m
07.04.2010 tank=19 m; bund=0m
07.06.2010 tank=0 m; bund=0 m
manhole GR-28
tank periphery GR-13
tank center
manhole GR-25
tank periphery GR-1
Figure 5 Total settlements of Tank A-2507 (cross-section)
582
th
7 CONCLUSIONS
Soil improvement described in this paper involved stone
columns installed below each tank. The gravel material pressed
into the soft soil is equivalent to a layer of about 30 cm in
thickness. Since the geodetic surveys showed negligible soil
upheaval (a few millimetres), it can be concluded that the soil
improvement prevented equivalent settlement of 30 cm.
The hydro test showed, as demonstrated in the case of Tank
A-2507, that total settlements are relatively small, i.e. smaller
than estimated by calculation (Figure 6).
For the purpose of comparing actual settlements with those
given in behaviour criteria, the settlements of the tank bottoms
are shown so that displacements corresponding to a rigid body
rotation are given separately from displacements resulting from
non-planar differential settlement. Figure 3 illustrates that the
bedding was mostly displaced as a rigid body, while non-planar
differential settlement was slight. For this reason, it is sure to
say that maximum values of the settlements and their shapes are
within the values required by the relevant standard (Figure 4).
02
0
4
0
6
0
8
0
1
0
0
1
2
0
1
4
0
1
6
0
1
8
0
2
0
0
load [kPa]
-22.0
-20.0
-18.0
-16.0
-14.0
-12.0
-10.0
-8.0
-6.0
-4.0
-2.0
0.0
s
e
t
t
l
e
m
e
n
t

[
c
m
]
tank center - measured
tank periphery - measured
tank center - Plaxis - calculated
tank periphery - Plaxis - calculated
Figure 6. A comparison of measured and calculated results of
settlements obtained from the hydro test on Tank A-2507.
It was found that the settlements, after the tank had been
emptied, were smaller although they had the same shape. This
proves that the deformations after tank emptying are mostly
elastic (Figure 4). As tank loads by crude oil are less than those
by water, it is expected that subsequent displacements at
operating load will be less than those recorded in hydro tests,
and that no further non-planar differential settlement of the tank
bedding will occur. The same goes for the other four tanks
(Figure 7). The diagram of the settlements of all five tanks
shows that such settlements are about the same when the tanks
are subjected to the same load. As this is normally expected in
the case of horizontally stratified soil, this is proof of proper and
correct measurement of displacement.
As seen in Figure 6, the settlements obtained by calculation
were significantly greater than those measured. An explanation
for different values of settlements should be thoroughly
investigated in further numerical analysis which will take into
consideration the fact that columns and soil act together as
recommended in Ambily and Gandhi (2004).
In the case of the Sisak tanks, the hydro tests proved correct
functioning of the floating roofs, watertightness of the shells
and bottoms, and rigidity of the foundation structure for all
tanks.
The strictly applied procedures regarding soil investigation,
design, hydro test and exploitation ensured safety in execution
and further use. Considering the safety risks and loss of
investment in case of non-allowable differential settlements of
the soil, it is clearly understandable, yet in some cases
disregarded, why such procedures must be applied.
The data collected about the behaviour of the tanks during
the hydro tests were well documented and could be used to
improve design of tanks.
02
0
4
0
6
0
8
0
1
0
0
1
2
0
1
4
0
1
6
0
1
8
0
2
0
0
load [kPa]
-4.0
-3.0
-2.0
-1.0
0.0
s
e
t
t
l
e
m
e
n
t

[
c
m
]
hydro test -
tank periphery
- measured
A-2507
A-2508
A-2509
A-2510
A-2511
Figure 7. Total settlements of tank perimeters during hydro tests
performed on all five tanks
8 ACKNOWLEDGEMENTS
The writers particularly wish to thank Z. Korica and M. Bago
for their support and assistance.
9 REFERENCES
Ambily, A.P.& Gandhi, S.R. 2004. Analysis of hydro test results for
steel tank on stone column improved ground. Proceedings of the
Indian Geotechnical Conference held at NIT Warangal, 420-423.
API-650: American Petroleum Institute (API) standard 650: Welded
tanks for oil storage.
API-653: American Petroleum Institute (API) standard 653: Tank
Inspection, Repair, Alteration, and Reconstruction
Bell, R.A. & Iwakiri, J. 1980. Settlement comparison used in tank-
failure study. Journal of the Geotechnical Engineering Division
106, GT2, 153-169.
Berardi, R. & Lancellotta, R. 2002. Yielding from field behavior and its
influence on oil tank settlements. Journal of geotechnical and
geoenvironmental engineering 5, 404-415.
EN 14015:2004, Specification for the design and manufacture of site
built, vertical, cylindrical, flat-bottomed, above ground, welded,
steel tanks for the storage of liquids at ambient temperature and
above
Marr W.A., Ramos J.A. & Lambe T.W. 1982. Criteria for settlement of
tanks. Journal of the Geotechnical Engineering Division 108, GT8,
1017-1039.
Raju, V.R., Hari Krishna, R. and Wegner, R. 2004. Ground
improvement using Vibro Replacement in Asia 1994 to 2004 a 10
year review, Proceedings of 5th Int. Conf. on Ground Improvement
Techniques, Kuala Lumpur, Malaysia.
583
Interrelationship between deformation moduli from CPTU and SDMT tests
for overconsolidated soils
La corrlation entre le module de dformation de CPTU et de tests SDMT
pour les sols surconsolids
Mynarek Z., Gogolik S.
University of Life Sciences, Poznan, Poland
Sanglerat G.
Ingnieur civil des Ponts et Chausses, Lyon, France

ABSTRACT: At the area of Poland glaciations caused overconsolidation in deep layers of the subsoil. It is imperative to take into
account this fact to calculate the differential settlements of structures subjected to great moments, such as wind turbines. Paper
presents the results obtained from the deformation characteristics estimated from CPTU and SDMT tests in clays Vistula and Riss
glaciations with interbedded layers of fluvioglacial sands.
RSUM : En Pologne les glaciers ont provoqu la surconsolidation des couches profondes. Il est impratif de tenir compte de ce fait
pour calculer les tassements diffrentiels des structures soumises des moments importants telles que les eoliennes. On prsente les
rsultats obtenus partir des caractristiques de compressibilit values partir des essais classiques CPTU et SDMT dans des
argiles de Vistula et Riss avec intercalations de couches de sables fluvioglacieres.
KEYWORDS: deformation modulus of overconsolidated soils, CPTU, SDMT

1 INTRODUCTION
Determination of representative values of constrained moduli
and deformation moduli of soils found in the subsoil is a topical
research problem. It is generally known that deformation and
strength parameters may be determined using laboratory and in
situ tests. In the laboratory method the key element in the
evaluation of quality in case of e.g. an oedometric test is
connected with the quality of samples collected for analyses
(Mynarek 2003, Tanaka 2007). This problem is particularly
evident in overconsolidated deposits. This fact indicated that
soil deformation parameters need to be determined in situ using
DMT, CPTU or SDMT method. Static penetration plays a
particularly important role in forecasting values of deformation
modulus of soils, as with the use of this method we may obtain
a continuous picture of changes in moduli in the subsoil in a 1-
D or 3-D system (Mynarek et al. 2007). The other testing
techniques determine values of moduli pointwise. In CPTU the
constrained deformation modulus is determined from
correlation relationships. For this reason calibration or
assessment of quality of the identification of this modulus using
SDMT is of considerable practical importance (Marchetti 1999).
This paper discusses this problem together with an assessment
of interrelationships between modulus G
0
from CPTU and
SDMT tests.
2 INVESTIGATIONS OBJECTS
Subsoil structure in Poland is highly complicated in terms of
their stratigraphy and lithology. The contact zone of the
building structure with the subsoil is comprised primarily of
deposits from the two last glaciations, as well as different forms
of glacilacustrine deposits.

Figure. 1 Location of investigated plots in the region of Poland

The subsoil of the investigated area (Fig. 1) comprises
glacial tills of the Riss and Vistula glaciations, Quaternary and
Pleistocene soils and also Holocene fluvial deposits. The effect
of diversification in terms of the genesis and lithology of soils
in the discussed locations is shown in CPTU classification
systems (Lunne et al. 1997) (Fig.2). In turn, Figure 3 presents
examples of geotechnical profiles and results of CPTU and
DMT. CPTU tests were performed using a Hyson 200 kN static
probe by ap van den Berg, while dilatometer tests were
performed with an original seismic dilatometer by Marchetti.
584
th

Figure 2. CPTU Soil classification chart (Lunne et al. 1997).

Figure 3. Geotechnical profile with CPTU and SDMT characteristics.
3 CONCEPT FOR THE IDENTIFICATION OF THE SOIL
PRECONSOLIDATION EFFECT IN SUBSOIL
Identification of the relationship between the genesis of subsoil
and a measure determining the overconsolidation rate, eg.
overconsolidation ratio OCR, is a complex problem. Of the two
discussed methods, CPTU and SDMT, the chance to determine
reliable OCR values is greater for SDMT, since the effect of
preconsolidation is strongly related with the geostatic stress
ho
.
For this reason in order to obtain a continuous picture of
changes in OCR of the subsoil in the examined locations
calibration was performed for OCR values determined using
cone resistance Q
t
, applying OCR values determined by SDMT.
In the approach three groups were identified for the discussed
locations: with complete drainage sands (group I),
intermediate soils (group II) and clays (group III) (Fig. 2). The
groups of intermediate soils and clays were identified based on
the content of the clay fraction and the plasticity index I
p
.
The values of OCR from dilatometer testing for soils of
groups II and III were calculated from the relationship
(Marchetti (1999):

OCR = 0.5 (K
D
)
1,56
(1)
where: K
D
horizontal stress index
The value of OCR for soils in CPTU testing was determined
using the nomogram proposed by Wierzbicki (2010), in which
OCR values are established on the basis of cone resistance Q
t

and the plasticity index of soil I
p
is considered. In the case of
non-cohesive soils OCR values were also assessed applying a
diagram proposed by Wierzbicki (2010). This diagram uses
both tests, i.e. CPTU and DMT, as it is constructed on the basis
of the formula proposed by Mayne (2000).

OCR = 5.04 K
0
1,54
(2)

Figure 4. Relaationship between coefficient OCR
DMT
/OCR
CPTU
and
coefficient OCR
CPTU
(group II and III).

Figure 5. Relationship between OCR
DMT
and OCR
CPTU
after calibration
(group II and III)

It results from Figs. 4 and 5 that the relationship between
OCR values from both tests has a high statistical evaluation.
This fact makes it possible to construct a direct dependence
between cone resistance Q
t
and OCR from SDMT (Fig. 6).
Values of OCR determined from this dependence were used to
supplement data for statistical analysis and next in the profiles
at different levels
v0
, where SDMT testing was not performed.

Figure 6. Relationship between cone resistance q
t
and OCR
DMT

coefficient.
585
4 THE RELATIONSHIP BETWEEN DEFORMATION
MODULI AND SHEAR MODULI FROM CPTU AND SDMT
4.1 Constrained deformation modulus M

Concepts for the determination of modulus M from CPTU and
SDMT differ fundamentally. This results from the different
techniques of parameter measurements, which are introduced to
the relationship between the measured parameter and modulus
M. Another factor is connected with the different location of
CPTU and SDMT on the characteristics shear modulus
0

shear strain (Mayne 2000). In the case of CPTU modulus M is
determined from the relationship (Sanglerat 1972, Mayne 2000)

M
CPTU
=
1
(q
t
-
v0
) (3)

In DMT the dilatometer modulus E
D
is the starting point for
the determination of modulus M
DMT
(Marchetti 1999)

M
DMT
= f (E
D
, K
D
, I
D
) (4)
where K
D
horizontal stress index according to Marchetti
(1999)
A modified original formula according to Marchetti for the
determination of R
M
for preconsolidated glacial tills was given
by Lechowicz et al. (2011).
4.2 Shear modulus G
0

A function which describes the relationship between modulus
G
0
or G from SDMT or SCPTU and variables which describe
parameters of the soil medium was given by Lee and Stokes
(1986), Jamiolkowski et al. (1995)

G
0
= f (
v0
, e
0
, OCR, S
r
, C, K, T) (5)

where: (
v0
geostatic effective vertical stress, e
0
initial void
ratio, OCR overconsolidation ratio, S
r
degree of saturation,
C- grain characteristics, K soil structure, T- temperature.

This relationship may be used to forecast values of modulus
G
0
directly on the basis of cone resistance Q
t
.
4.3 Analysis of results
4.3.1 Constrained moduli M
CPTU
, M
DMT

To calibrate the relationship between moduli M
CPTU
and M
DMT
individual moduli were determined from the following
formulas. For the CPTU test according to Mayne (2000) (eq. no.
3).
This formula was verified by oedometric tests. The analysis
showed that for the tested loams and clays the mean value of
coefficient
i
was close to 8.25.
Modulus M
CPTU
for non-cohesive soils was calculated from
dependencies supplied by Lunne et al. (1997) depending on
values q
c
and including the degree of preconsolidation in these
deposits.
Moduli M
DMT
were calculated prior to calibration from
original formulas proposed by Marchetti et al. (1999).

Figure 7. Relationship between constrained modulus M
0
CPTU
and M
0
DMT
.

Figure 8. Relationship between constrained modules M
0
CPTU
and M
DMT

after calibration.

For preconsolidated deposits the moduli determined by
SDMT are higher than those from CPTU (Fig. 7). Obtained
relationships fully confirm the opinion by Marchetti et al.
(1999) on this subject.
Calibration of both moduli in order to describe their changes
in the subsoil with changes in
v0
is presented in Figs. 8. It was
assumed in the calibration process that modulus M
CPTU
is the
reference point.
4.3.2 Shear moduli G
0
DMT
, G
0
CPTU

The determination of shear modulus from CPTU G
0
CPTU
was
based on empirical dependencies
for non-cohesive soils (group 1) (after Hegazy, Mayne
1995)

V
s
= 12,02 q
t
0,319
f
s
-0,0466
(6)

G
0
= V
s
2
(7)

for cohesive soils (groups 2 and 3) relationship
determined using multi linear regression

Group 2
G
0
= 41,44 q
t
+ 0,31 G
v0
+ OCR 1,71 (8)
586
th
Group 3
G
0
= 41,20 q
t
+ 0,37 G
v0
+ 0,88 OCR 28,53 (9)

Figure. 9. Relationship between shear modulus G
0
DMT
and G
0
CPTU

Figure 10. Relationship between measured G
0
DMT
and G
0
CPTU
after the
calibration.

Figure 11. The model of subsoil stiffness calculated on the G
0
values
from CPTU, calibrated by SDMT results.

In order to obtain a continuous picture of changes in the shear
modulus G
0
the dependence of G
0
CPTU
was calibrated using
measured values of G
0
DMT
(Fig. 10). Moduli G
0
CPTU
determined
from this relationship may be used in the construction of a
model of rigidity for the subsoil composed of soils of varied
genesis. An example of such a model for the foundation of a
wind turbine is presented in Fig. 11. The model was constructed
using Inverse Distance Weighting Method (IDW) according to
Mynarek et al. (2007).
5 CONCLUDING REMARKS
On the basis of the conducted investigations it may be
concluded that the simultaneous use of CPTU and SDMT
provides a continuous picture of changes in rigidity of subsoil
composed of soils with diverse genesis. The effectiveness of
these methods is emphasized by the high statistical evaluation
for the dependence between deformation and shear strength
moduli from both tests. However, to determine this dependence
it is necessary to apply a calibration function. The calibration
function needs to be specified for the soils, which should be
grouped depending on their grain size, since this variable also
influences relationships between parameters measured in CPTU
and SDMT.
After calibration this relationship may be a useful tool in the
construction of a model for rigidity of subsoil based on shear
strength moduli G
0
or M
0
moduli.
6 REFERENCES
Hegazy Y.A. & Mayne, P.W. 1995. Statistical correlations between Vs
and CPT data for different soil types. Proc. of Symposium on Cone
Penetration Testing, Vol. 2:173-178.Swedish Geotechnical Society:
Linkping.
Jamiolkowski M., Lancellotta R., Lo Presti D.C.F. 1995. Remarks on
the stiffness at small strain of six Italian clays. International
Symposium on Pre-failure Deformation Characteristics of
Geomaterials. Hokkaido vol.2: 817-836.
Lechowicz Z., Rabarijoely S., Galas P., Kiziewicz D. 2011. Settlement
evaluation of spread foundations on heavily preconsolidated
cohesive soils. Annals of Warsaw University of Life Sciences
SGGW, Land Reclam.Nr. 43 (2,) pp.113-120.
Lee S. H. H., Stoke K. H. 1986. Investigation of low amplitude shear
wave velocity in anisotropics materials. Geotechnical Report No.
GR 86-6, Civil Engineering Department, University of Texas,
Auatin.
Lunne T., Robertson P. K., Powell J.J.M. 1997. Cone Penetration
Testing in Geotechnical Practice, Blackie Academic &
Professional.
Marchetti S., Monaco P., Calabrese M., Totani G. 1999. The flat
dilatometer test. A report to the ISSMGE Committee TC-16.
Mayne P.W. 2000. Stress-Strain-Strength-Flow Parameters from
Enhanced In-Situ Tests. Proc. International Conference on In-Situ
Measurement of Soil Properties and Case Histories, Bali,
Indonesia: 27-48.
Mynarek Z. 2003. Influence of quality of in-situ tests on evaluation of
geotechnical parameters of subsoil. Proc. of 13
th
European
Conference on Soil Mechanics and Geotechnical Engineering.
Prague. vol.3.
Mynarek Z., Wierzbicki J., Woyski W. 2007. An approach to 3-
subsoil model based on CPTU results. Proc. of 14th European
Madrid.
Sanglerat G. 1972. The penetrometer and soil exploration. Elsevier,
Amsterdam.
Tanaka H., Nishida K. 2007. Suction and shear wave velocity
measurements for assessment of sample quality. Proc. of the 3
rd
International Workshop on in-situ tests and sample disturbance of
clays. Studia Geotechica et Mechaica No. 1. Technical University ,
Wrocaw.
Wierzbicki J. 2010. Evaluation of subsoil overconsolidation by means
of in situ tests at the aspect of its origin. (in Polish). Rozprawy
Naukowe nr 410. Wydawnictwo Uniwersytetu Przyrodniczego w
Poznaniu. ISSN 1896-1894, 182 pp.
587
Le Gomcamtre, un nouvel essai in situ adapt la mesure des caractristiques
hydro-mcaniques du sol
The Geomechameter test, a new in-situ apparatus adapted to the measurement of the hydro-
mechanical characteristics of the soil
Monnet J.
UJF-Grenoble 1 CNRS UMR 5521, Laboratoire 3SR, Grenoble
RSUM : Un nouvel appareil dessai in situ, le Gomcamtre a t conu et construit. Cette volution du pressiomtre utilise les
forces produites par un coulement d'eau vertical descendant autour de la sonde de mesure pour gnrer un champ de gravit artificiel.
Cet coulement hydraulique permet de rgler la contrainte effective verticale au niveau de la sonde de mesure. L'influence de cette
contrainte est prise en considration dans l'interprtation des rsultats de lessai. l'origine conu pour la caractrisation mcanique
des sols et notamment pour la mesure indpendante des caractristiques de rsistance (cohsion c, angle de frottement des
caractristiques de dformation (module dlasticit E), l'appareil a t modifi pour la mesure de la permabilit (coefficient k) et de
la sensibilit lrosion du sol. Cette nouvelle version de lappareil de 2004 est quipe de la saisie numrique pour la pression, pour
le dplacement radial de la membrane et dune camra vido pour la mesure de turbidit. Un des avantages de cet appareil est la
mesure simultane des principales caractristiques hydro-mcaniques du sol, notamment dans les digues et des remblais. Le sable fin
dHostun a t choisi comme matriau de rfrence pour l'tude exprimentale du modle rduit test en laboratoire. Les limons de
lIsre ont t choisis pour exprimenter le prototype de ce nouvel appareil dessai in situ. Les rsultats des mesures au gomcamtre
sont valids par la comparaison aux rsultats de lessai triaxial pour les caractristiques mcaniques et lessai Lefranc pour la
permabilit. Cette exprimentation est dveloppe au sein du Projet national Erinoh.
ABSTRACT: A new in situ testing apparatus, the Geomechameter, has been designed and built. It is an evolution of the
pressuremeter, using the forces generated by water flow around the measurement probe. The hydraulic flow allows to control the level
of the vertical stress at the test level. The influence of this stress is taken into account in the interpretation of the test results.
Originally designed for the soil mechanical shearing resistance (cohesion c and friction angle ), deformation resistance (Young
modulus E), the apparatus was modified for the measurement of the permeability (coefficient k) and the sensibility to erosion. This
new version of the 2004 apparatus is equipped with numerical gauge for pressure and radial displacement and video camera for
turbidity measurement. One of its interests is the simultaneous measurement of the main hydro-mechanical characteristics of the soil
inside the soil mass, for the dams and embankments. Hostun thin sand was chosen as a material to undergo the experimental study in
laboratory. Isre loam was chosen to experiment the model of this in-situ apparatus. Results of the geomechameter are validated by
comparison of mechanical characteristics obtained by the triaxial test and by the Lefranc injection test for the permeability. This
experiment is developed with the help of the Erinoh project.
MOTS CLEFS : Erosion interne, essais in situ, modlisation numrique, pressiomtre
KEYWORDS: Internal Erosion, In-Situ Test, numerical modelling, Pressuremeter
1 INTRODUCTION
L'essai pressiomtrique peut tre considr comme un essai de
cisaillement entre les contraintes radiales et circonfrentielles
dans la condition de dformation plane, avec la contrainte
verticale qui est la contrainte normale applique sur le plan de
cisaillement (Baguelin et al., 1978). La limite thorique de l'essai
pressiomtrique est lie au fait que la contrainte verticale est
donne par le poids des terres au repos. Ce test peut tre
considr comme un essai de cisaillement unique et il peut tre
utilis pour dterminer soit l'angle de frottement interne (Hughes
et al., 1977 ; Monnet, 1990 ; Monnet & Khlif, 1994 ; Monnet,
2012) ou la cohsion du sol ; il peut tre utilis aussi dans la
conception des travaux de gnie civil (Monnet & Allagnat,
2002). Lorsque la cohsion et l'angle de frottement interne sont
dtermins conjointement, ils sont relis entre eux dans
l'interprtation du pressiomtre, si bien que la valeur de la
cohsion dpend de l'angle de frottement interne.
Lessai au Gomcamtre a plusieurs avantages.
- Il contrle le niveau de contrainte 3D autour de la sonde. Le
principe de l'essai au Gomcamtre est de crer un gradient
hydraulique (Figure 1) pour contrler la contrainte verticale en la
rglant une valeur approprie. Ce nouvel appareil rgle la
contrainte radiale et circonfrentielle par lexpansion de la
sonde. Il permet le contrle de l'tat tridimensionnel des
contraintes autour de la sonde, la paroi du forage.
- Il permet la mesure de la cohsion et de frottement. Une srie
de trois essais raliss avec trois valeurs diffrentes du gradient
hydraulique externe i (Eq.3) permet dimposer trois contraintes
verticales diffrentes. Ces trois contraintes verticales
correspondent trois courbes diffrentes dexpansion radiales
qui sont autant de courbe de cisaillement diffrentes. Ceci
permet de dterminer cohsion et l'angle de frottement interne.
Cette possibilit de mesure locale du couple cohsion-frottement
est d'un grand intrt en Gnie Civil.
- Il permet la mesure du coefficient de permabilit et du
coefficient de consolidation. Le pressiomtre peut galement
mesurer la dissipation de la pression interstitielle autour du
forage lorsque la consolidation est atteinte dans un dlai d'une
heure ou plus (Clarke et al., 1979). Le Gomcamtre amliore
cette mesure par une dtermination simultane du module de
cisaillement et de la permabilit et n'a pas besoin d'attendre
jusqu' la consolidation finale.
- Il permet la mesure du risque de l'rosion interne du sol dans
les barrages et les digues. Ce risque aussi appel suffusion
rsulte des exfiltrations. La suffusion semble tre la principale
588
th
cause des incidents hydrauliques observs sur les barrages et les
digues. Sur lensemble des barrages qui prsentent des
dommages ou des ruptures, dans 46% des cas on trouve un
phnomne d'rosion interne (Foster et al. 2000). En France, 70
cas critiques ont dj t dtect.
Dans cet article, nous prsentons la nouvelle version du
Gomcamtre qui a t adapt pour la mesure de la suffusion du
sol dans les digues. Pour valider les principes thoriques de
fonctionnement du Gomcamtre, il est ncessaire d'utiliser un
sol standard qui est connu, homogne et dont l'tat initial est
parfaitement dfini. Tous ces paramtres peuvent tre mesurs
dans une chambre d'talonnage. L'essai de ce nouveau dispositif
est ralis avec une dmarche qualit o chaque tape est
soigneusement contrle. L'essai en chambre d'talonnage est la
premire opration de contrle quil est ncessaire deffectuer.
Nous prsentons une premire srie de test de validation de
lappareil dans des conditions relles in-situ.
2 DESCRIPTION DU GEOMECAMETRE
2.1 La sonde du gomcamtre
Le prototype de la sonde Gomcamtre est constitu de six
modules diffrents.
- Le module 0 relie la tige de forage la sonde.
- Le module 1 : le packer suprieur isole la circulation
hydraulique autour de la sonde de la partie suprieure du forage.
- Le module 2 : il a deux fonctions diffrentes, l'injection de
l'eau dans le sol et la mesure de la turbidit de l'eau extraite la
base de la sonde.
- Le module 3 : c'est la partie centrale de la sonde. Ses
fonctions consistent imposer une pression contrle au forage
par lexpansion dune membrane et mesurer le dplacement
radial au niveau de la sonde.
- Le module 4 : c'est la cellule de pompage qui sert extraire
l'eau du sol.
- Le module 5 : le packer infrieur qui isole la circulation
hydraulique autour de la sonde du bas du forage.
La sonde doit rpondre plusieurs spcifications :
- Il doit tre possible de dmonter les diffrentes parties pour
les besoins d'entretien de lappareil (changement de
membrane,).
- L'indpendance des quatre circulations liquides doit tre
assure, notamment au niveau des liaisons entre les modules
(pompage, injection, pression de l'air, la pression de l'eau).
- La mesure des pressions, des dformations, des dbits, de la
turbidit, doit tre ralise numriquement et stocke sur micro-
ordinateur.
- Lexpansion de la sonde, ainsi que la mesure de la
dformation doit tre possible jusqu la pression limite
(doublement du rayon du forage).
2.2 Evolution de la sonde du gomcamtre
Cette nouvelle version du gomcamtre (Figure 2) permet
de dterminer le risque de suffusion du sol, par la mesure de la
turbidit de leau extraite, en utilisant une vido camra
embarque dans le module 2.
3 TUDE THORIQUE
3.1 Etude analytique : circulation hydraulique autour de la
sonde gomcamtre
Le Gomcamtre (Figure1) est un appareil qui utilise le flux
hydraulique autour d'une sonde gonflable pour augmenter
localement la contrainte effective verticale dans le plan moyen
de la sonde. La charge hydraulique diminue le long de la
trajectoire lorsque l'eau se dplace dans le sol de la tte jusqu'au
pied de la sonde. Une force est applique aux particules de sol
dans la direction de l'coulement. Dans lessai au
Gomcamtre, ces forces sont semblables l'action des forces
de gravit. Cette action peut tre considre comme l'action
inverse de l'coulement de l'eau de boulance obtenu par la
relation de Terzaghi (Eq.1) :

w w sat crit
i (1)
Figure 1: Le principe de lessai gomcamtrique
Figure 2: Le Gomcamtre, version 3, juin 2012
Lorsque le coefficient de permabilit est suprieur 10
-8
m/s, l'expansion de la sonde n'a aucune influence sur la pression
d'eau interstitielle (Cambou et Bahar, 1993 ; Frank et Nahra,
1986) et la pression effective peut tre utilise, mais si le
coefficient de permabilit est infrieur 10
-10
m/s, le test ne
peut pas tre effectu car le sol devient non drain. Une
modlisation numrique de l'essai (Senouci et Monnet 1999)
montre que dans une unit de volume du sol, la force applique
par le dbit hydraulique peut tre estime au moyen de la
relation (Eq.2). Dans lessai au Gomcamtre, l'eau est injecte
dans le sol de la cellule d'injection une pression d'injection (p
i
).
Aprs circulation dans le sol, l'eau est pompe par la cellule de
589
pompage avec une pression (p
p
). Le gradient hydraulique (Eq.3)
est obtenu par la diffrence entre la charge hydraulique
l'injection et au pompage. L'augmentation de la contrainte
effective verticale (Eq.4) est associe un coulement
hydraulique de gradient i dans le sol, au niveau de la sonde de
longueur l
e
. La contrainte verticale impose au niveau de la
sonde et la profondeur simule sont ainsi trouves (Eq.5-6):
V i F
w
A = . .

(2)
(3)
(4)
(5)
(6)
Figure 3: Comparaison entre les rsultats numriques du Gomcamtre
pour une profondeur simule Z
sim
=2.74m et la courbe exprimentale
correspondante avec i = 5.25 ( = 14,2 kN/m
3
, o' = 50 kPa)
Figure 4 : Variation de la charge hydraulique autour de la sonde du
Gomcamtre rsultat Plaxis
3.2 Etude numrique : Simulation de lessai par Plaxis
La modlisation numrique (Senouci, Monnet 1999) par
programme d'lments finis Plaxis montre que la variation du
gradient hydraulique calcul donne une variation de la contrainte
verticale (Eq.5) qui permet de dfinir une profondeur simule de
la sonde (Eq.6). Les diffrences entre la courbe numrique
trouv par Plaxis la profondeur simule de 2,74 m (Eq.6) et la
courbe exprimentale pour le gradient hydraulique
correspondant (i = 5,25) sont trs faibles (Figure 3).
La variation de charge hydraulique calcule par Plaxis dans
la masse du sol (Figure 4) montre l'augmentation de la charge
hydraulique impose par la cellule d'injection (sur la gauche,
hauteur moyenne) et la dcharge hydraulique impose par la
cellule de pompage ( gauche en bas). On voit dans le plan
mdian que la charge hydraulique reste gale constante -2,6m.
Le gradient i (Eq.3) est donc constant dans le plan moyen,
laugmentation de contrainte (Eq.4) est galement constante dans
le plan moyen.
Figure 5 : Influence du gradient hydraulique sur lexpansion de la sonde
du gomcamtre
Figure 6 : Un exemple dessai au Gomcamtre 2,2m de profondeur
4 TUDE EXPERIMENTALE
4.1 Essai au laboratoire Vrification de linfluence du
gradient hydraulique sur lexpansion de la sonde
Des essais ont t raliss en chambre de calibration avec le
sable dHostun au poids volumique de 16kN/m
3
(Figure 5). Ils
montrent que la courbe dexpansion pour un gradient de 3 est au-
dessus de la courbe sans coulement. Le gradient hydraulique
produit une augmentation de la rigidit apparente du sol qui peut
tre interprte comme une augmentation de la contrainte
verticale.
4.2 Essai in situ - Dtermination des caractristiques
mcaniques
( )
e p i
l H i = H
( )
Des essais in situ ont t raliss sur les digues de lIsre (Figure
6). Pour lessai prsent, le dbit est denviron 40 l/min avec 5 h
de temps d'essai. Ce temps est ncessaire pour atteindre un dbit
stabilis satur, afin que la contrainte verticale soit modifie par
la sonde du gomcamtre. La simulation de la contrainte
verticale est 119kPa pour une contrainte verticale au repos de
55kPa, ce qui correspond une augmentation de 64kPa. Lessai
au gomcamtre permet de mesurer, le module lastique 5MPa,
sur le cycle dchargement rechargement. La comparaison avec
les courbes dexpansion thoriques permet la dtermination de la
cohsion 5kPa et de langle de frottement 30,5.
Des essais triaxiaux de contrle ont t effectus sur des
chantillons de limon recompacts en conditions draines la
mme densit. Les rsultats sont indiqus (Tableau 1). On peut
noter pour niveau de contrainte (100kPa) proche de celui
' 2 . .
vi e m
l i o o + = . ' Z
' '
+ =
w si v
i w
Z l i Z + + 2 ' .
2
e
l i ) ( '
w v
oo =
e sim
= ). 1 (
590
th
impose par le gomcamtre, un module de Young qui est trs
proche de la valeur mesure in situ avec une diffrence de 10 %,
une cohsion nulle et un angle de frottement 33,5 qui est
lgrement plus grand de 3 que celui trouv in situ. Cette
diffrence peut tre explique par le fait que lchantillon triaxial
est remani, reconsolid, et que cette procdure exprimentale a
supprim la cohsion.
4.3 Essais in situ - Dtermination de la permabilit
Le Gomcamtre permet de mesurer la permabilit par la
mesure du dbit inject, avec la relation de Mnard (Eq.7) :
5 CONCLUSION
Un nouvel appareil de mesure in situ a t construit pour
tester le sol dans des conditions de contrainte tridimensionnelle.
Le modle rduit du gomcamtre a montr sa capacit
augmenter la contrainte verticale autour de la sonde pendant
l'expansion de la sonde. Ces conditions ont permis de dterminer
les caractristiques de cisaillement des limons de lIsre en
lasticit, cohsion, frottement avec une bonne prcision.
Linjection de leau produite par le gomcamtre a permis de
mesurer la permabilit du limon.
La sonde du Gomcamtre sera quipe d'une micro camra
vido pour la mesure de la turbidit de l'eau extraite et ainsi
dtecter le seuil de suffusion lintrieur des massifs de sols
comme les digues.
( )
(
=
2
1
ln
2
0 0
r
l
p p l
Q
k
w
t

(7)
Lessai au gomcamtre a t ralis le long des berges de
lIsre, dans les limons sablonneux. Le coefficient de
permabilit obtenu est de 1,4 10
-6
m/s. Ces rsultats peuvent
tre compars avec les essais Lefranc (Tableau 2), qui ont t
raliss sur le mme site pour le Symbhi. A lexception du
forage P250, la permabilit moyenne mesure par l'essai
Lefranc est 1,1 10
-5
m/s et la permabilit minimale est de
1,1.10
-6
m/s. La permabilit mesure par lessai au
gomcamtre se trouve dans l'chelle des valeurs de l'essai
Lefranc, mais proche de la valeur minimale mesure. D'autres
expriences doivent tre ralises pour confirmer et expliquer
une telle diffrence.
Figure 7 : Mesure de la suffusion entre un point dinjection
et un point de pompage
Tableau 1 : Rsultats de lessai triaxial sur un chantillon
remani reconsolid des Limons de lIsre
o
3
kPa
E
MPa
v c
kPa
|
degr
60 2,7 0,47
100 4,5 0,42
200 10 0,3
300 18,9 0,28
0 33,5
Tableau 2 : Rsultats de lessai Lefranc test raliss sur les
digues de lIsre
Profils Profondeur k
m m/s
P 252 2 2,8 10
-5
P 252 4 6.6 10
-6
P 250 3 1,1 10
-6
P 248 2 > 10
-3
P 248 5,8 6,6 10
-6
P 248 7,5 1,1 10
-5
4.4 Dtermination de la suffusion
La micro camra utilise sur le Gomcamtre a t calibre
avec diffrentes concentrations de sol rod. La densit en
couleur rouge a t utilise pour la turbidit de leau. Le systme
a t calibr dans un rservoir rempli d'un mlange de sable et de
gravier, et la turbidit de l'eau des effluents a t mesure. Il
semble qu'une suffusion apparaisse pour un gradient hydraulique
environ 3. Aprs une valeur d'un gradient hydraulique de 6,
l'analyse de l'image montre que l'eau est claire, mais que parfois
certaines particules passent devant la camra et modifient la
valeur de la densit de couleur rouge (Figure 7). Les essais de
juin 2012 nont pas permis de raliser correctement la mesure in-
situ. De nouveaux essais sont programms pour juin 2013.
6 RFRENCES
Baguelin F, Jzquel J.F., Shield D.H., 1978, The pressuremeter and
foundation engineering, Series on Rock and Soil Mechanics, Vol. 2,
Trans. Tech. Publication, 335-406.
Cambou B., Bahar R., 1993, Lutilisation de lessai pressiomtrique
pour lidentification de paramtres intrinsques du comportement du
sol, Revue Franaise Gotechnique (N63): 39-53.
Clarke B.G., Carter J.P., Wroth C.P., 1979, In-situ determination of the
consolidation characteristics of saturated clays, Proc. 5
th
Europ.
Conf. SMFE, Brighton, Vol.2, 207-213.
Foster M., Fell R., Spannagle M., 2000, The statistics of embankment
dam failure and accidents, Canadian Geotechnical Journal, Vol. 37,
pp. 10001024
Frank R., Nahra R., 1986, Contribution numrique et analytique ltude
de la consolidation autour du pressiomtre, Rapport recherche
LCPC (N137)
Hughes J.M.O., Wroth C.P., Windles D., 1977, Pressuremeter tests in
sand, Geotechnique, Jnl 27 (N 4): 455-477.
Monnet J, 1990, Theoretical study of elasto-plastic equilibrium around
pressuremeter in sands, Proc. 3rd Int. Symp. Pressuremeter,
Oxford, 137-148.
Monnet J., Allagnat D. 2002, Design of a large soil retaining structure
with pressuremeter analysis, Geotechnical Engineering 155, Issue 1,
71-78.
Monnet J., Khlif J, 1994, Etude thorique de lquilibre lastoplastique
dun sol pulvrulent autour du pressiomtre, Revue Franaise
Gotechnique (N67): 71-80
Monnet J., 2012, An Elasto-Plastic analysis of the Pressuremeter
Test in Granular Soil part 1: theory , European J. of Environmental
and Civil Engineering, Vol.16, N6, June 2012, 699-71
Senouci S.M., Monnet J, 1999, Modlisation numrique du
Gomcamtre, Revue Franaise de Gotechnique (N88): 21-35.
591
Analytical approach for determining soil shear strength parameters from CPT
and CPTu data
Approche analytique pour dterminer la rsistance au cisaillement d'un sol sur la base d'essais
CPT et CPTu
Motaghedi H., Eslami A., Shakeran M.
Dept. of Civil Engineering, Amirkabir University of Technology, Tehran, Iran

ABSTRACT: The common approaches for soil strength parameters determination from CPT data are on the basis of bearing capacity
and cavity expansion theories. A new method is proposed for C, prediction using all quantities, q
c
, u and f
s
from CPTu considering
bearing capacity mechanism of failure at cone tip and direct shear failure along penetrometer sleeve. One advantage of this method is
improvement the accuracy in the case of erroneous data by using all three output of CPTu. Laboratory test results, the two sets of
nonlinear equations by the proposed approach and existing correlations of C and angle parameters have been compared applying on
a data base compiled from four sources. It has been considered that the internal friction angle which is obtained by current methods is
almost relatively higher than the measured values. Also, the comparison indicates good consistency and low scatter for the proposed
method.

RSUM: Les approches communes pour les paramtres de rsistance des sols, dtermins par CPT, sont bases sur la capacit de
cisaillement et les thories dexpansion des cavits. Une nouvelle mthode a t propose pour C, et utilise toutes les quantits de
prvision, q
c
, u et f
s
de CPTu, en considrant la capacit de cisaillement et le mcanisme de failure dans type paramide et failure
cisaillement direct, le long du pntromtre manchon (sleeve). Une des avantages de cette mthode est damliorer exactitude dans le
cas des donns fausse utilisation, tout les trois sortie de CPTu. Les rsultats des essais du laboratoire, les deux combinaisons des
quations non linaires, lapproche propose et les corrlations existantes de c et langle de sont compares sur quatre bases de
donnes. On considre que la friction interne obtenue par la mthode courante est toujours relativement plus grande que la valeur
mesure, aussi la comparaison montre la bonne consistance et le bas scatter pour la mthode propose.

KEYWORDS: Soil shear strength, Cohesion and friction parameters, CPT and CPTu data, Bearing capacity theory

1 INTRODUCTION
Geotechnical investigation by CPTu provide continuous vertical
profile of cone tip resistance (qc), sleeve friction (f
s
) and pore
water pressure (u
2
) in every inch of the subsoil depth (Lunne et.
al, 1997). The CPTu test is used in soft to medium deposits, and
not applicable in cemented sand, hard clay and gravelly strata.
The penetrometer is a useful tool to identify of thin layers where
the traditional sampling procedures cannot be employed. Also,
using the CPTu test may distinct the liquefiable or collapsible
soil layers around 50 mm thickness in depth (Tavenas and
Leroueil, 1987), (Eslami and Fellenius, 2004).
In alluvial soils containing gas, determining undrained shear
strength by traditional sampling procedures and using UU
triaxial tests may lead to conservative results. In granular soils,
determining the friction angle () as one of the major soil
strength parameters by using direct shear or triaxial tests
involves uncertainties due to sampling difficulties, confining
pressure simulation and limitations of size effects (Mitchell and
Durgunoglu, 1983). The main advantage of CPTu versus other
in situ test procedures is the relatively elimination of
undisturbed sampling, performance in real condition regarding
stress level and geological aspects. Furthermore, by using the
continuous data in one inch interval of depth, shear strength
parameters (C,), can be obtained which have significant role
in geotechnical designs.

2 SHEAR STRENGTH PARAMETERS BY CPTU DATA

Two main theories have been implemented for the estimation of
shear strength parameters by using CPT and CPTu results;
bearing capacity (Janbu and Senneset, 1974), (Durgunoglu,
1975) and cavity expansion (Vesic, 1972) approaches. The
methods which are based on bearing capacity theories; for
penetrometer penetration mechanism, it is assumed that cone tip
resistance (q
c
) is equivalent with ultimate load of a deep circular
foundation in subsoil and leads the soil mass to be failed.
Whereas, failure assumption in cavity expansion theory is based
on required pressure for forming of deep hole in an elastic-
plastic environment which is fitted with the pressure needed for
creation and cavity expansion in the same volume under
identical conditions. So far, Muromachi, 1972, Schmertmann,
1978, Mitchell and durgunoglu, 1983, Robertson and
Campanella, 1988, Kulhawy and Mayne, 1990 have studied on
determination of shear strength parameters from CPT and CPTu
data which solely have presented S
u
in fine grained or in
granular soils.

3 ANALYTICAL MODEL FOR C AND BY CPTu DATA

By applying two basic equations on determination of the deep
foundation bearing capacity, one for tip and other for
penetrometer sleeve, using the effective bearing capacity instead
of total stress approach and extension of the relationships, a
dual equation system with two unknowns, can be achieved as
below under static loading conditions.

(1)

Considering deep bearing capacity factors proposed by
(Junbu, 1974 base failure model) and applying the analytical
Eslami and Fellenius, (1997) model based on CPTu results, the
rel ions can be summarized as follows: at

592
th

(2)

Eq. 3 is expressed according to empirical results for at
. Also, N
q
can be achieved from Eq. 4 which is shown as
below:

(3)

(4)

Jamiolkowski and Robertson, 1988 presented a correlation
for as function of and mean in situ stresses:

(5)
Where and are the vertical total and
effective stresses, respectively.
The lateral stress increases by increasing the relative
density. Usually, in calculation, it is assumed that the lateral
stress value is equal to resistant horizontal stress by acceptable
accuracy as follows:

(6)

(7)

(8)

(9)

By substitution Eqs. 2 to 9 in two basic Eq.1 can be achieved
two sets of equation.10 as follow:

(10)

4 EXPERIMENTAL RECORDS FOR EVALUATION

Geotechnical properties and information including experimental
results from the data base of four sites have been compiled.
These records are containing 25 series of CPT and CPTu data
and shear strength parameters measured by laboratory tests
which are used for evaluating developed model. The site
specifications are briefly reviewed as follows:
Site No. 1, Narenjestan tourism complex, (Mandro Co.,
2012); site is located in southern bank of Caspian Sea in
Mazandaran Province, Iran. According to borehole operations
results, observation and field tests from ground level silty sand
with medium dense deposits is located to the depth of 7.5 m.
Following the depth of 7.5 m the firm silt layer with high
plasticity exist with thickness of 2 m. From depth of 9.5 m
down to end of boring poorly graded, silty sand and sand are
located with of dense condition and classified as an SM, SP.
Site No. 2, Narges Hotel complex, (Sham-e Co., 2012); is
located in southern Caspian Sea Shore in the suburb of Sari city
in Iran. The observation of three boreholes by rotary drilling
indicate that the superficial soil layer consists of alluvial gray
sea sand with some silt which exists to the depth of 11m.
According to USCS this layer is classified as SP, SM or SP-SM.
Between depth of 10m to 14m fine clay and silt layer are
located in dirty green color with the thickness of 1m to 4m
which is classified as CL. The bottom layer is containing fine
sea alluvial sand which is observed in depth of 11m to 30m and
is classified as SM. Also, the ground water level is located
below 3m of ground surface. For determining soil shear strength
parameters of filed soil stratification, direct shear, triaxial and
uniaxial tests are accomplished on samples. Also, according to
SPT records in subsurface depths around 10m, the N values are
ranged from 22 to 35, which represent medium to dense relative
density for upper layer. The N

values in depth of 10m to 14m
and 14m to 30m vary from 12 to 25 and 22 to 45, respectively,
and classified as dense to high dense coarse grained deposit.
The CPTu profile in Sari Narges Hotel site is shown in Fig. 1.
Site No. 3, East Changi, (Choa et al. 2004); site is a recovery
site which is located in eastern costal of Changi Airport in
Singapore. From geotechnical investigations, it is observed that
the geomaterial is a kind of soft to medium clay.
Site No. 4, University of Texas which is known as A&M
Site, (Briaud and Gibbens, 1994). It is one of the international
site of study in geotechnical basis and is located in Texas
Province, USA. Soil deposits are formed of silty sand.

Figure 1. CPTu profiles in Narges Hotel Complex, (Sham-e Co., 2012)

The accumulated results of analytical procedure in 25 cases
and also laboratory test results are presented in Table 1.

Table 1. Shear strength parameters according to proposed method and
laboratory test results for 25 measure cases
C (kPa)
Site
No.
Soil
type
q
E
(MPa)
f
s
(kPa)
Lab
test
proposed Lab
test
Proposed
SM 30 22 4 3.7 30 32
SP 13 65 4.5 5 30 31
SM 11 50 4 3.8 31 32
MH 13 40 50 49 4 6
SP 5 55 4 5 33 31
SM 22 110 4 3.5 33 31
SM 40 150 4 4.5 35 36
SM 28 140 4 5 35 37
593
SM 30 135 6 5.2 34 36
SM 18 60 6 8.4 38 37
SM 6 75 3 2.5 32 31
SM 5 30 0 1 32 32
SM 7 80 6 6.6 31 32
CL 4 70 29 30 2 2.5
SM 4 90 58 57 2 2.7
CL 2 14 30 29 12 12
ML 2 55 28 29 14 15
SM 2 27 5 6.1 16 16
CH 2 56 57 56 5 6
CL 2 78 35 36 8 9
IV SM 1 6 0 1 33 34
IV SM 7 30 0 1 36 38
IV SM 8 60 0 1.5 32 33
IV SM 6 38 1 1.5 - 8
IV SM 9 75 9 8.4 - 10

5 VALIDATION OF RESULTS AND DISCUSIONS

The accomplished geotechnical study in each site has been upon
borehole excavations. The samples dependent on soil
stratification and from different depths are taken as distributed
and undistributed specimens. For determining the shear strength
parameters, direct shear, uniaxial and triaxial tests are done on
samples in laboratory. Meanwhile, because of high quality of
sampling in triaxial test and logicality of the test results in
laboratory, it can be more adequate. Four practical cases include
CPT and CPTu test results associated with laboratory test
results and SPT records are used for evaluating the proposed
analytical relations.
The measurement results by laboratory tests and also,
prediction by using analytical procedure, are presented in
Table 1. Evaluation of results expressed the fact that the
suggested procedures not only can spontaneously predict and
determine both shear strength parameters but also it contain
acceptable and reasonable results. Fig. 2 is associated to
evaluation and comparison between laboratory results and
suggested analytical model for determining the cohesion
parameter. The measured and predicted C values show good
agreement which denotes the capability of analytical approach.
Also, Fig. 3 shows the comparison between measured values
and analytical procedure results for internal friction angle within
the range of study in four sites. As for the laboratory results
which are achieved from drained triaxial test and suggested
analytical model, it is observed that the proposed analytical
procedures based on CPT and CPTu in cases with cohesion and
internal friction angle, almost has identical to laboratory results.
The laboratory results are compared with different presented
procedures by researchers are shown in Fig. 4a to 4f. According
to graphs, the achieved friction angle values by other procedures
are always greater than the suggested analytical procedure
values and laboratory results. Meanwhile, it is observed that the
friction angle values from Meyerhof, (1974) results are closer to
bisector line indicating close agreement between the predicted
and measured values. Moreover, the presented analytical
procedure and laboratory results have more coincidence and are
closer to actual values. While, the values obtained from current
methods, are more than the experimental results and analytical
method.
The current procedures do not contain any recommendation
for soil cohesion and it is one of the advantages for the proposed
procedure. Also, it is not depending only one of the test outputs
rather, the entire CPT and CPTu outputs such as q
c
, f
s
and u are
used in equations, hence the error creation reaches to minimum
value in inaccurate records, because of the simultaneous
employment of each three output quantities, the other
advantages in the presented analytical procedure contrary to
traditional procedures. Furthermore, the shear strength
parameters derived from actual subsurface failure mechanisms
condition in cone tip and sleeve has been realized reasonably in
proposed relations.

Fig. 2. Comparison between the measured cohesion in laboratory and
estimated cohesion by proposed method

Fig. 3. Comparison between estimated and measured values for friction
angle

Fig. 4. Comparison between estimated and measured values for friction
angle
594
th
6 CONCLUSION

Geotechnical study by CPT or CPTu can determine continuous
vertical profile of cone tip resistance (q
c
), sleeve friction (f
s
) and
pore water pressure (u) in every inch of the subsoil depth.
Hence, the shear strength parameters can be well determined
which have major role in geotechnical design.
In this study two main theories have been implemented for
the estimation of shear strength parameters by using CPT i. e.,
bearing capacity in cone tip and direct mode of shear failure in
along penetrometer jacket. So far, different researchers have
studied on determination of shear strength parameters from CPT
and CPTu data which solely have presented S
u
in fine grained or
angle in granular soils. The entire of CPTu data, q
c
, f
s
and u
are used to calculate C and , via bearing capacity theory and
shear stress relation at failure condition. By combining these
relations and applying the proposed analytical Eslami and
Fellenius, (1997) model based on CPTu results and direct shear
failure along cone sleeve, the drained shear strength parameters
values include cohesion and internal friction angle can be
derived simultaneously.
In proposed procedure the error creation reaches to minimum
value through inaccurate records, because of the simultaneous
use of each three output quantities. The existence methods for
determining the internal friction angle are rely on only one of
the test outputs (depending only to q
c
) while the inaccurate
records creates more error in shear strength parameters. But,
three parameters q
c
, f
s
and u are dependent on friction angle in
presented procedure and lead to prorate the error cases. The
current procedures do not contain any recommendation for soil
cohesion and it is one of the advantages in the proposed
procedure. The presented procedure differs from common
procedure results by increasing fine grains in soil. Comparison
with 25 data sets of C and from laboratory tests and predicted
by the proposed method indicate good agreement and
consistency.
7 REFERENCES
Briaud J.L. and Gibbens R.M. 1994. Test and Prediction results for Five
large Footing on Sand, FHWA prediction Symp, ASCE Spec, Publ.
41, 255-262.
Campanella R.G. Robertson P.K. and Gillespie D. 1983. Cone
Penetration testing in deltaic soils. Canadian Geotechnical Journal
20(1) ,23-35.
Durgunglu H.T. 1975. Penetration tests of cohesion soils. Proceedings,
ASCE, Speciality Conference on In-Situ Measurements of Soil
Parameters
Eslami A. and Fellenius B.H. 1997. Pile capacity By Direct CPT and
CPT
u
Methods Applied to 102 Case Histories. Canadian
Geotechnical Journal 34( 6), 880-898.
Eslami, A. and Fellenius, B.H. 2004. CPT and CPTu Data for Soil Profile
Interpretation: Review of Methods and a Proposed New Approach.
Iranian Journal of Science and Technology, Transaction B 28(1), 69-
86.
Gottardi G. and Tonni, L. 2009. Analysis and interpretation of piezocone
data from the Treporti test site for the evaluation of compressibility
characteristics of silty soils. DISTART Technical Reort NO. 226,
University of Bologna.
Jamiolkowski M. and Robertson P.K. 1988. Closing Adress: Future
Trends for Penetration Testing. Geotechnology Conference
Penetration Testing in UK, Birmingham.321-342.
Janbu N. and Senneset K. 1974, Effective stress interpretation of in situ
static penetration tests. Proceedings of the European Symposium on
Penetration Testing, ESOPT, Stockholm. 22, 81-93.
Kulhawy F.H. and Mayne. P.H. 1990. Manual on estimating soil
properties for foundation design. Electric Power Research Institute,
EPRI.
Lunne T. Robertson. P.K. and Powell J.J.M. 1997. Cone penetration
testing in geotechnical practice, Blackie Acad. Chapman and
Hall/Routledge Press, London,
Mandro. Consulting Engineers Final Report. 2012. Site Investigation
and Geotechnical Survey for Narenjestan Hotel Babolsar located in
Southern Caspian Sea, North of Iran.
Maple. Maplesoft, A Division of Waterloo Maple Inc, 1981-2010.
Mayne P.W. 2007. Cone penetration testing. NCHRP Synthesis 368,
Transportation Research Board, National Academies Press,
Washington, D.C
Mayne P.W. Peuchen J. and Bouwmeester D. 2010. Estimation of soil
unit weight from CPT. Proc., 2
nd
International Symp. On Cone
Penetration Testing, CPT'10, Huntington Beach, CA.
Meyerhof G.G. 1983. Scale effects of pile capacity. Journal of the
Geotechnical Engineering Division, ASCE. 108(GT3). 195-228.
Mitchell J.K. and Durgunoglu. H.T. 1983. Cone resistance as measure of
sand strength. Journal of the Geotechnical Engineering Division,
ASCE. 104(GT7),. 995-1012.
Na Y.M. Choa V. The C.I. and Chang M.F. 2004. Geotechnical
parameters of reclaimed sandfill from the cone penetration test.
Canadian Geotechnical Journal. 42(1), 91-109.
Robertson P.K. 2009. CPT interpretation a unified approach, Canadian
Geotechnical Journal. 49 (11), 1337-1355.
Robertson P.K. and Campanella R.G. 1988. Guidelines for geotechnical
design using CPT and CPTu. University of British Columbia,
Vancouver, Department of Civil Engineering, Soil Mechanics
Series 120.
Robertson P.K. Woeller D.J. and Finno W.D.L. 1992. Seismic Cone
penetration test for evaluating liquefaction Potential under cyclic
loading. Canadian Geotechnical Journal. 29(4), 685-95.
Senneset K. and Janbu N. 1985. Shear strength parameters obtained
from static cone penetration tests. Strength Testing of Marine
Sediments; Laboratory and In Situ Measurement. Symposium, San
Diego, 1984, ASTM Special technical publication, STP 883, 41-54.
Sham-e Consulting Engineers, Final Report. 2012. Site Investigation and
Geotechnical Survey for Narges Hotel sari located in Southern
Caspian Sea, North of Iran.
Tavenas F. and Leroueil S. 1987. State of the art on laboratory and insitu
stress- strain-time behavior of soft clay. Proc. Intl. Symp. on
Geotechnical Engineering of Soft Soils, Mexico City, 1-146.
Vesic A.S. 1972. Expansion of cavities in infinite soil mass. Journal of
the soil Mechanics and Foundations Division, ASCE, .265-290.

595
th
Use of penetration testing for determination of soil properties in earth dam
ploi des essais de pntration pour dterminer les proprits de sol pour barrages en
terre
Mulabdic M.
nned, due to bad
order to assure
mpacted dam and
in situ test on the
on the crest.
sed, since standard
acting clay. It has
, but also that one
nkments.
e mauvaise
avoir des donnes
t les proprits de
s essais in situ sur
CPT) et les essais au dilatomtre (DMT) in
otentiel de ces essais dans la description des
de ces essais sont bases sur les
ontr que les essais CPT et
DMT sont utiles pour lanalyse dtat du sol compact, mais quil faut tre trs attentif dans lemploi des procds standard
e ces essais quand il sagit des essais pour les ouvrages en remblai.
cone test, flat dilatometer test, interpretation
e, essai au pizocne, essai au dilatomtre plat, interprtation
irrigation
e deepest
ay from
it was
ly follow
zoned
d degree
ng of the
covered
e were
d. It w
against
echni
ld ha
should have been checked in detail. The site test
program consisted of drilling boreholes for gett
samples for laboratory testing of clay, of penetrat
testing CPT and a flat dilatometer test (Marchetti
dilatometer DMT). This paper presents the results of
analysis of the properties of clay in the dam based on in
situ (CPT and DMT) and laboratory testing. Only
boreholes in the crest were used for the analysis, see
Fig. 1.

Em
University in Osijek, Croatia
ABSTRACT: The Paper describes a case of a small earth dam for which remediation work was pla
construction and possible damage to the dam that could have occurred during filling of the retention. In
relevant data for the remediation design solution it was necessary to determine the current state of the co
properties of the clay fill in the dam. Investigation work consisted of drilling boreholes and performing
dam, and of laboratory testing of soil samples. CPT and DMT in situ tests were carried out nearby the boreholes
Potential of these in situ tests in describing physical and mechanical properties of the clay was analy
methods of interpretation of these tests are based on natural soils, while the dam was constructed by comp
been shown that CPT and DMT tests are useful in describing properties of a compacted clay embankment
should be cautious in using common methods of interpretation of their test results in case of earth fill emba
RSUM : Larticle dcrit le cas d'un petit barrage en terre, pour lequel une remise en tat est planifie cause d
ralisation et du danger potentiel dendommagement au cours de remplissage de la retenue. Afin d
pertinentes pour les techniques de confortement il a t ncessaire de dterminer ltat actuel du barrage e
largile utilise dans la construction du barrage. Les travaux de reconnaissance ont compris les forages et le
le barrage, ainsi que les essais en laboratoire. Les essais de pntration au cne (
situ ont t faits auprs des trous de forage dans la crte du barrage. Le p
proprits physiques et mcaniques dargile est analys, tant donn que les interprtations
sols naturels tandis que largile a t mise en uvre dans le barrage par compactage. It est dm
dinterprtation des rsultats d
KEYWORDS: earth dam, compacted clay, piezo
MOTS-CLS : barrage en terre, argile compact

1. INTRODUCTION

A small earth dam was built as a part of a future
system. The dam was about 10-meter high at th
point in depression, and was constructed of the cl
its vicinity. During the construction of the dam
noticed that the construction company didn't ful
the design requirements and criteria related to
construction, replacement of foundation soil an
of compaction of the lifts of clay. During the filli
lake, when only few meters of dam slopes were
with water, problems with bottom discharg
observed and filling of water had to be stoppe
decided that the dam should be checked for safety
sliding and deformability, for which geot
properties of compacted clay in the dam shou
been
as
cal
ve
ing
ing
ion

Figure 1. Position of in situ tests and boreholes on the dam; most
important work was done on the crest (line B2-C4)

From the samples taken during drilling boreholes
specimens were formed for the laboratory testing
program, which comprised the testing of physical and
mechanical properties of clay from the dam. Fig. 2 shows
the plasticity of clay from the dam, determined on
samples from the B2 and B5 boreholes.
596
th
th

city; there
n deg e
that water content w
according to Proctor test and t
ticity was used than t
latome
est of
dam. General interpretation of test results of these t
his c
erefore
stand
th tests
relev
was used
ne in CPT cone. It should be no
that there is not much experience presented in literat
d DMT testing in compacted clay.
for the
are based
Lunne et
r, exc t
for the part deeper than 9 m as measured from the crest.
That required careful cone filter saturation with glycerine.
CPT soil-type identification was done according to a
widely used chart (Robertson, 1990), and in doing so clay
of low plasticity was identified in most cases, with some
thin layers of silty clay (see plasticity chart in Fig. 2).
Pore pressures measured behind the cone (u
2
) were
almost zero, or slightly negative, in all depths.
Figure 2. The clay from the dam was of low plasti
were zones of silty clay at some depths

It was determined that generally clay compactio
was under 95 % of Proctor value,
re
as
hat
hat
a bit higher than w
opt
clay of somewhat lower plas
which was defined by the design solution.

2. CPT AND DMT TESTS

Four CPT test-boreholes and three flat di
(DMT) test-boreholes were realised, along the cr
ter
the
wo
ase
it
ard
.
ant
test types is established for natural soils, and in t
there is compacted clay human made soil. Th
was necessary to check the applicability of
interpretation methods to compacted clay, for bo

Both tests were conducted according to
standards (EN 1997 Part 2:2006). Glycerine
as fluid in porous sto

d te
ure
All covering CPT an
empirical and theoretical expressions
interpretation of test results of these two tests
on natural soils (Larsson and Mulabdic, 1991,
al, 1996, Marchetti, 1980).

2.1. Soil identification

Clay in the embankment was never under wate ep

Figure 3. DMT (M1, M2, M3) and CPT tes
dam height (c
ts (C1, C2), over the
ross-section along the crest). Both, CPT and DMT
bankment it seems
ver the dam height;
ed Mv, and CPT test
d a sandy-silty to
clayey-silty thin
ata for undrained
ding to Marchetti
tion is related to I
d

pe it should satisfy
0.1 < I
d
< 0.6. Since the value of I
d
in compacted clay of
to be about or higher than 2
(suggesting a sandy or sandy-silty soil type), and there
ould be concluded
compacted soil
and .

2.2. Undrained shear strength by CPT
test is calculated
et al, 1997)
tests revealed inhomogeinity in the clay em
that almost every lift of clay can be spotted o
DMT test illustrate interpreted versus requir
interpreted versus required cu

On the other hand, DMT test detecte
silty-sandy soil type, with very rare
layers. Therefore there were almost no d
strength in DMT interpretation. Accor
(1980), soil type in DMT test interpreta
= (p
1
-p
0
) / (p
0
-u
0
), and for clay soil-ty
the dam was found
was no in situ pore pressure in soil, it c
that p
0
was too small, due to structure of
absence of in situ pore pressures

Undrained shear strength from CPT
according to common expression (Lunne
k
v c
u
N
q
s
0

(1)

Value for N
k
=15 was used in this case, which is the mean
value of proposed values for natural soils (suggested
values are N
k
=11-19), and it was confirmed to be
applicable for compacted clay as well (Fig. 4).

597
th
0
50
100
150
200
250
0 1000 2000 3000
q
T
-
vo
(kPa)
4000
S
u
(
k
P
a
)

pacted clays was found to be
applicable, based on comparable UU and CPT test results
n of modulus of vertical deformation was
made for relevant results for oedometer and CPT test.
Lunne et al (199 ession for
modulus of vertical deformation from CPT test when
Figure 4. Nk=15 Nk = 15 for com

2.3. Modulus of vertical deformation

Compariso
7) critically analyse expr
determined as

) ( 25 , 8
0 v T
q M

In this case it seems that this value should be de
factor of two (Fig. 5). This might be due to the
this gener
(2
vided y
fact t
al expression has limitations, and becou
oedometer tests were performed on submerged specime
while CPT and DMT tests were performed on clay fill
the embankment that was not submerged. Values of Mv
from DMT test were the highest of these three (Fig. 6)
(Fig. 6).

)
b
hat
se
ns
in

would be respected, then big
Figure 5. Relationship between laboratory determined modulus
of vertical deformation and corrected tip resistance for CPT test,
around B5 borehole

Based on a limited number of available test results, the
expression Mv=4, 3 (q
t
-
vo
) seems to better fit test results
than the equation (2). Modulus seems to be half of the
value suggested by that commonly used equation. If
relationship between DMT-M
v
and LAB M
v
from Fig 6
ger portion of the
then required
compared to situation illustrated in Fig 3 (M proj) .

embankment would show lower values
0
1
2
3
4
5
6
7
8
0 40 80 120 160 200
Mv DMT M3
Mv CPT C2
9
10
Mv lab
(m)
Mv (MPa)

from oedometer (on
r than from CPT
if compared to DMT
(performed on clay layers that
were not submerged)

Fig. 7 presents the sets of CPT and DMT tests with a
view to illustrate soil resistance in relation to depth. It
seems that the results of tests from different locations are
very similar throughout the depth of testing.

d
Figure 6. Modulus of vertical deformation
submerged specimens) was much smalle
interpretation (equation (2)) or even lower
standard interpretation values
0
2
4
6
8
10
12
14
0 100 200
M3
M2
M1
Mv (MPa)
d (m)
0
2
4
6
8
10
12
14
0 10 2
q
c
[MPa]
0
C1
C2
d (m)

Figure 7. CPT and DMT tests in cumulative presentation

598
th
th
st resu
ussed
ls. Th
tion h
e we
bmerg
bmerg
v
for
situ t
ll. O
guide
in situ evaluated parameters to laborato
const
ts) we
mic te
ed abo
s wa
ls. Sin
rigid
arded
echani
unced
standard
diagrams. Although velocities generally increase w
depth, there are weaker and stronger intervals at certain
depths in M2 and M1 boreholes. The M3 location sho
constant increase in shear wave velocity by depth.

The influence of presence of water on DMT te
interpretation in terms of modulus M
v
was disc
Mulabdic and Bruncic (2000) for natural soi
concluded that error in water depth assump
limited influence on interpreted M
v
values. Her
dealing with compacted clay, never being su
and obviously soil would be softer if it were su
That is, it is difficult to predict soil modulus M
state of a submerged embankment fill from an in
performed on a non-submerged embankment fi
comparison as shown in Fig. 6 can be used as a
correcting
lts
in
ey
ad
are
ed,
ed.
the
est
nly
for
ry
ant
re
sts
ve
ve
ce
ity,
as
cal
in
values, but even then correction would not be
with depth.

Tests marked as M1, M2 and M3 (DMT-tes
performed in one run as standard tests and seis
(SDMT), using a special seismic probe install
blade (Cavallaro et al, 2006). Fig. 8 show
velocities measured in 0,5-meter depth interva
velocity is a measure of soil structure and its
variability of those two parameters should be reg
a basic indication of the variability of soil m
properties. These variabilities are more prono
wave velocity diagrams than in CPT and DMT
ith
ws
0,0
0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
5,0
5,5
6,0
0 100 200 300 400
6,5
7,0
7,5
8,0
M1
M2
M3
d (m)
v (m/s)

nt SDMT
locations, depth intervals 0,5 m

3. CONCLUSIONS

The paper presented the case of an earth dam of a poor
construction quality. In order to characterize clay fill in
the embankment in terms of its physical and mechanical
properties, CPT and DMT tests were performed in
addition to borings and laboratory testing. The purpose of
e compacted clay
ties in a
erent positions on
ir results
standards. Based
itu and laboratory
e drawn from this
omogeneous clay
th, both in static
T), (2) common
results should be
priate corrections
d clays, since they
we deal with the
of importance for
erties whether the
at the time of
MT tests showed
be valuable aid in
oth in terms of
anical properties;
cularly useful; (5)
in situ test
order to properly
account for effects of the presence of water (submerged
ed), specific structure of compacted soil,
specific stress distribution and limited experience in
compacted
KNOWLEDGEMENTS

y Elektroprojekt,
analysis
.
dam, Elektroprojekt,
Zagreb, 2006.
ial design - Part 2-
ocone tests in clay.
rt No. 42, Linkping,
J.M. 1997. Cone
Practice, E & FN
Spon, pp 312.
ilatometer", ASCE
pp 299-321.
Mulabdi, M. and Bruni, A. (2000.). Prilog analizi primjene
dilatometra Marchetti, Graevinar, Vol 52, No. 1, pp 9-17
(in Croatian).
Robertson,P.K. 1990. Soil classification using the cone
penetration test. Canadian Geotechnical Journal, 27 (1),
151-8.
Cavallaro, A., Grasso, S. and Maugeri, M. 2006. Clay Soil
Characterization by the New Seismic Dilatometer
Marchetti Test (SDMT), Proc. 2
nd
international flat
dilatometer conference.
Figure 8. Measured shear wave velocity at differe
these two in situ tests was to determin
condition and its physical and mechanical proper
continuous profile by depth and at diff
dam crest. The tests and the interpretation of the
were performed according to accepted
on analyses of all test results from in s
tests the following conclusions can b
case: (1) CPT and DMT detected inh
conditions very clearly along the dep
testing and in seismic testing (SDM
interpretation of CPT and DMT test
used with caution, allowing for appro
when tests are performed in compacte
are developed for natural clays, and here
compacted man made soil; (3) it is
the analysis and perception of clay prop
embankment is dry or submerged
performing in situ tests; (4) CPT and D
remarkable repeatability and proved to
characterizing embankment quality, b
inhomogeinity and physical and mech
SDMT results also proved to be parti
local correlations between laboratory and
results should always be used, in
or non-submerg
using in situ tests for the characterization of
soils.

4. AC
Cooperation with designing compan
Zagreb, Croatia, in the planning, execution and
of investigation work is highly appreciated

5. REFERENCES

Design solution documentation for Opatovac
EN 1997-2:2006. Eurocode 7: Geotechnc
Ground investigation and testing.
Larsson, R. and Mulabdi, M. 1991. Piez
Swedish Geotechnical Institute, Repo
pp 240.
Lunne, T., Robertson, P.K., Powell, J.
Penetration Testing in Geotechnical
Marchetti S. 1980. "In Situ Tests by Flat D
Journal GE, Vol. 106, No. 3, March 1980,
599
Diagnosis of earth-fills and reliability-based design
Diagnostic de remblais de terre et conception base sur la fiabilit
Nishimura S., Shuku T.
Graduate School of Environmental Science, Okayama University
Suzuki M.
Institute of Technology, Shimizu Co.
ABSTRACT: This research deals with the maintenance strategy of geotechnical structures such as earth-fill dams and river
embankments. To determine the soil parameters, the standard penetration test (SPT) N-values are frequently used. Firstly, a
statistical model for the N-values is determined from sounding test results. In this research, Swedish Weight Sounding (SWS) tests,
simpler than SPT, are employed. Secondly, an indicator simulation is conducted to interpolate the spatial distribution of the N-values,
and the results are utilized to find degraded areas inside the embankments and to maintain the embankments. Based on the statistical
model for the N-values, the shear strength parameter is derived through the empirical relationships, and then a reliability analysis of
the embankments is conducted considering the variability of the internal friction angle. Finally, the effect of improving the
embankments is discussed, comparing the calculated risks of the original state with the improved and restored state.
RSUM : Cette recherche porte sur la stratgie de maintenance des structures gotechniques comme les barrages en remblais de
terre et les digues fluviales. Les valeurs N du test de pntration standard (SPT) sont frquemment utilises pour dterminer les
paramtres du sol. Premirement, le modle statistique de N valeurs est dtermin partir des rsultats des essais de sondage. Dans
cette recherche, on utilise le sondage par poids sudois (SWS), plus simple que le SPT. Deuximement, une simulation indicatrice est
effectue pour interpoler la distribution spatiale de N valeurs, et les rsultats sont utiliss pour trouver les zones dgrades l'intrieur
des remblais, pour l'entretien des remblais. Bass sur le modle statistique de N valeurs, les paramtres de rsistance au cisaillement
sont dduits des relations empiriques, ensuite, l'analyse de fiabilit des remblais est effectue en tenant compte de la variabilit de
l'angle de frottement interne. Finalement, l'effet de l'amlioration du remblai est discut, en comparant lanalyse de risque calcule
partir de ltat initial et de ltat des remblais amliors et restaurs.
KEYWORDS: earth-fill dam reliability-based design, indicator simulation, statistical model of N-value

1 INTRODUCTION
There are many earth-fill dams for farm ponds in Japan. Some
of them are getting old and decrepit, and therefore, have
weakened. Making a diagnosis of the earth-fills is important for
increasing their lifetime, and an investigation of the strength
inside the embankments is required for this task. In the present
research, firstly, the spatial distribution of the strength
parameters of decrepit earth-fills is discussed, and an
identification method for the distribution is proposed. Although
the strength of earth-fills is generally predicted from the
standard penetration test (SPT) N-values, Swedish Weight
Sounding (SWS) tests are employed in this research as a simpler
method of obtaining the spatial distribution of the N-values.
SWS tests are advantageous in that they make short interval
exams possible, because of their simplicity.
To mitigate disasters, improvement works are conducted on
the most decrepit earth-fill dams. Since there is a recent
demand for low-cost improvements, the development of a
design method for optimum improvement works at a low cost is
the final objective of this research. A reliability-based design
method is introduced here in response to this demand.
Generally, the identification of the spatial correlation of soil
parameters is difficult, since the usual sampling intervals are
greater than the spatial correlation. Therefore, sounding tests
are convenient for determining the correlation lengths. Tang
(1979) determined the spatial correlation of a ground by cone
penetration tests (CPT). Cafaro and Cherubini (1990) also
evaluated the spatial correlation with CPT results. Uzielli, et al.
(2005) considered several types of correlation functions for CPT
results. Firstly, statistical models for the N-values are
determined from the SWS test results. Secondly, the
relationship between the SPT and the SWS N-values is
modeled, including the transformation error term. The N-value
distributions derived from SWS are spatially interpolated with
the indicator simulation (Journel and Huijbregts 1978), which is
one of the geostatistical methods. The simulated spatial
distribution of the N-values can be used for the health
monitoring of the inside of an embankment. To evaluate the
risk to earth-fill dams, due to the earthquakes, the circular slip
surface (CSS) method is used as the stability analysis method
along with the soil-water coupling finite element method. The
finite element method is used to estimate the normal and the
shear stress values on the slip surfaces. In this study, the Monte
Carlo method (MCM) is combined with the CSS method to
obtain the probability of failure. The procedure for the CSS
method, combined with the MCM, has also been conducted by
Shinoda, et al. (2006) and Yoshida, et al. (2005).
The strength parameter, namely, internal friction angle ,
derived from SWS tests, is considered to be the probabilistic
variable in this research. Additionally, two transformation error
terms, namely, the error terms from the SWS N-value to the
SPT N-value, and from the N-value to the internal friction angle,
are introduced to the MCM. Finally, the risk to an earth-fill
dam is calculated from the costs that would be incurred due to
embankment failure and probability failure. In this study, the
effect of improving an embankment is evaluated as a reduction
in risk between the original and the improved states.
Diagnosis of earth-fills and reliability-based design
Diagnostic de remblais de terre et conception base sur la fiabilit
Nishimura S., Shuku T.
Graduate School of Environmental Science, Okayama University
Suzuki M.
Institute of Technology, Shimizu Co.
600
th
2 INSITU TEST RESULTS
Although high-density sampling is required in order to evaluate
the spatial distribution of soil parameters, the amount of data is
not sufficient in the general sampling plans. In such cases,
sounding is a convenient way to identify the spatial distribution
structure of soil parameters. In this research, an embankment at
Site H is analyzed, for which SWS tests were conducted at 9
points, at 5-m intervals, along the embankment axis, as shown
in Figure 1. The soil profile of the embankment is categorized
as intermediate soil.
Generally, the strength parameters are assumed based on
standard penetration tests (SPT) with the use of empirical
relationships. In this research, Swedish weight sounding tests,
which are simpler than SPT, are employed instead of SPT.
Inada (1960) derived the relationship between the results of SPT
and SWS. Equation (1) shows the relationship for sandy
grounds, and the relationship is shown in Figure 2.
N
SWS
= 0.67N
SW
+0.002W
SW
(1)
in which NSWS is the N-value derived from SWS, NSW is the
number of half rations and WSW is the total weight of the
loads. Based on this data, the variability of the relationship is
evaluated in this study, and the coefficient of variation is
determined as 0.354. The determined -limits are also shown in
Figure 2 with broken lines. Considering the variability of the
relationship, the SPT N-values are derived by
5m
Plan view of test points
Top of embankment
x
No.1 No.2 No.3 No.4 No.5

Figure 1. Plan view of embankment and testing interval.

J
J
J
J
J
J
J
J
J
J J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
J
JJ
J
J
J JJ
J
JJ
J J
J
J JJ
JJ
JJJ
JJJ
J
J
J
JJJ
J
JJJJJ
J
J
JJJ
JJ
J
J
J
J
J
J J
J
JJJ
JJ
JJ
50 100 150 200 250
N
sw
J
0
5
10
15
20
25
30
0
2
5
0
5
0
0
7
5
0
1
0
0
0
N
W
sw
(N)

Figure 2. Relationship between SWS results and SPT N-values.
N
SPT
= 1+0.354
r
( )
N
SWS
(2)
in which c
r
is an N(0,1) random variable.
3 STATISTICAL MODEL OF N-VALUES
3.1 Determination method
A representative variable for the soil properties, s is defined by
Equation (3) equation as a function of the location X=(x, y, z).
Variable s is assumed to be expressed as the sum of the mean
value m and the random variable U, which is a N(0,1) type
normal random variable in this study.
s X ( )= m X ( )+U X ( )
(3)
The random variable function, s(X), is discretized spatially into
a random vector s=(s
1
,s
2
,...,s
M
), in which s
k
is a point estimation
value at the location X=(x
k
, y
k
, z
k
). The soil parameters, which
are obtained from the tests, are defined here as S=(S
1
,S
2
,..., S
M
).
Symbol M signifies the number of test points. Vector S is
considered as a realization of the random vector s=(s
1
,s
2
,...,s
M
).
If the variables s
1
, s
2
,...,s
M
constitute the M - variate normal
distribution, the probability density function of can then be
given by the following equation.
f
S
s ( )= 2t ( )
M 2
C
1 2
exp
1
2
s m ( )
t
C
-1
s m ( )

`
)
(4)
in which m=(m
1
,m
2
,...,m
M
) is the mean vector of random
function s=(s
1
,s
2
,...,s
M
) and is assumed to be the following
regression function. In this research, a 2-D statistical model is
considered, namely, the horizontal coordinate x, which is
parallel to the embankment axis, and the vertical coordinate z
are introduced here, while the other horizontal coordinate y,
which is perpendicular to the embankment axis, is disregarded.
k
= a
0
+ a
1
x
k
+ a
2
z
k
+ a
3
x
k
2
+ a
4
z
k
2
+ a
5
x
k
z
k (5)
in which (x
k
, z
k
) means the coordinate corresponding to the
position of the parameter s
k
, and a
0
, a
1
, a
2
, a
3
, a
4
, and a
5
are the
regression coefficients.
C is the MM covariance matrix, which is selected from the
following four types in this study.
C = C
ij | |
=

o exp x
i
x
j
2
l
x
z
i
z
j
l
z ( )
(a)
o
2
exp x
i
x
j ( )
2
l
x
2
z
i
z
j ( )
2
l
z
2
{ }
(b)

o
2
exp x
i
x
j ( )
2
l
x
2
+ z
i
z
j ( )
2
l
z
2

`
)
(c)
N
e
o
2
exp x
i
x
j
l
x
z
i
z
j
l
z ( )
(d)
(6)

i, j =1,2, , M

in which the symbol [C
ij
] signifies a i-j component of the
covariance matrix, o is the standard deviation, and l
x
and l
z
are
the correlation lengths for x and z directions, respectively.
Parameter N
e
is the nugget effect. The Akaikes Information
Criterion, AIC (Akaike 1974) is defined by Equation (7),
considering the logarithmic likelihood.
AIC = 2 max ln f
S
S ( ) { }+ 2L
= Mln2t + min lnC + S m ( )
t
C
1
S m ( ) { }
+ 2L
(7)
in which L is the number of unknown parameters included in
Equation (4). By minimizing AIC (MAIC), the regression
coefficients of the mean function, the number of regression
coefficients, the standard deviation, o, a type of the covariance
function, the nugget effect parameter, and the correlation
lengths are determined.
3.2 Determination of statistical model of SWS N-values
The mean function and the covariance function of the SWS N-
values, N
SWS
, are determined with MAIC, and the mean and the
-limits are exhibited in Figure 3. Although the covariance
functions given by Equation (6) were examined, the available
correlation lengths were not identified. Therefore, additional
mean functions are examined. Since the periodic tendency,
601
whose period is about 10 m along the horizontal axis, is found,
the term sin{(x/5-1/2)} was added to Equation (5). The
determined mean function is
m =1.98 + 0.816sin
x
5

1
2
|
\

|
.
| t

(

(
+ 0.157z
(8)
The covariance function is determined by
C
ij
= 0.75 ( )
2
exp x
i
x
j
6.14 z
i
z
j
0.63
( )
i = j (
C
ij
= 1.24 ( )
2
i = j ( )
)
(9)
The horizontal correlation length is identified to be
approximately ten times of the vertical one. Since this rate is
similar to the values published previously (e.g. Soulie et. al.
1990), the correlation lengths identified here are judged to be
appropriate. The boundary between the base ground and the
embankment is determined based on the SWS results.

The N-distribution predicted based on the determined
statistical models with aid of the indicator simulation method
(Deutsch and Journel 1990), which is one of the geo-statistical
methods, and interpolates the point-estimated N-values, is
exhibited in Figure 4. The horizontal periodicity of the N-
values is presented according to the figure.
4 RELIAIBILITY-BASED DESIGN OF A FILL-
EMBANKMENT
4.1 Statistical model of an embankment
A stability analysis is conducted and the risk is evaluated for an
earth-fill dam at Site H to analyze the transversal section, the
mean of the equation. As a mean function, Equation (12) is
proposed by averaging Equation (8) along the x axis, while the
covariance function is defined as Equation (13), in which
coordinate x is replaced by y of Equation (9), and depth z is
replaced by elevation h. This assumption is based on the reason
why the embankments are compacted horizontally in the
construction, and the correlation structure at the same elevation
is homogeneous.
m =1.89 + 0.157z
(12)
C
ij
= 0.75 ( ) exp y
i
y
j
2
6.14 h
i
h
j
0.63
( )
i = j ( )
C
ij
= 1.24 ( )
2
i = j ( )
(13)
The analytical sections of the original embankment, and the
improved and restored embankment are exhibited in Figure 5.
The embankment is improved by constructing an inclined core,
and by covering the original embankment with the additional
soil for reinforcement. The material properties are given in
Table 1. The soil parameters are determined from the SPT N-
values and the laboratory soil tests. The Bs means the
embankment material; it is determined from the N-values based
on the SWS results to consider the spatial distribution. The
effective internal friction angle |'=|
d
, is obtained from the
conversion, namely, Equation (14) (Hatanaka and Uchida
1996). In the equation, 5.3c
f
is the conversion error, in which c
f

is an N(0,1) type normal random variable, and the ratio of 5.3 is
the standard deviation.

Figure 3. Spatial distribution of NSWS and statistical model.

15
13
11
9
7
5
3
1
0 5 10 15 20 25 30 35 40
D
e
p
t
h

(
m
)
Horizontal Coordinate x (m)
0
5
10
15
20
25

Figure 4. Predicted spatial distribution of N-value.

Bs
Ac
As
Gr

(a) Original embankment.

Bs
Ac
As
Gr
Core
Block
Rigid soil

(b) Restored embankment.
Figure 5. Cross sections and critical slip surfaces of embankments.
|' = 20N
1
( )
0.5
+ 20+ 5.3c
f
(14)
N
1
= N
SPT
o
v
' /98 ( )
0.5
(15)
in which o
v
' is the effective vertical stress.
4.2 Reliability analysis
In the stability analysis, the pore water pressure is required; it is
calculated with a saturated-unsaturated seepage finite element
analysis (e.g., Nishigaki 2000). In the restored embankment, the
water table level is dramatically reduced by the existence of the
impermeable zone. Consequently, this reduction can make the
embankment stable.
The circular slip surface method is employed as the stability
analysis in this study. For uncertain factors, the random
numbers are assigned, and the stability of the embankments is
evaluated as the probability of failure with the use of the Monte
Carlo method. For the reliability analysis, Equation (16) is
defined as a performance function, in which the internal friction
angle is a probabilistic parameter. As the load of the
earthquake, the design earthquake intensity of 0.15 is
considered.
g = t
fi
t
si ( )
l
i
i=1
n
(16)
602
th
where t
f
and t
s
are the shear strength and the shear force on the
slip surface exhibited in Figure 6, which shows a slip surface
across a finite element. In the figure, l
i
is the length of the slip
surface of element i, and n is the number of elements, which a
slip circle crosses. The strength, t
f
is defined by the Mohr-
Coulomb law of Equation (17). Normal stress o
n
and shear
force t
s
are defined in Figure 6, and calculated with the soil and
water coupling finite element method in this study. In the finite
element analysis, the pore pressure is estimated in the saturated
zone identified with the saturated - unsaturated seepage
analysis, and the negative pore water pressure in the unsaturated
zone is disregarded. This assumption can simplify the analysis
and make an evaluation for the stability that is on the safe side.
t
f
= c' +o
n
' tan|'
(17)
o
n
' =
o
z
' +o
y
'
( )
2
+
o
z
' o
y
'
( )
2
cos2u t
yz
sin2u
(18)
t
s
=
o
z
' o
y
'
( )
2
sin2u + t
yz
cos2u
(19)
in which c' is the effective cohesion, |' is the effective internal
friction angle, o
z
' and o
y
' are the vertical and the horizontal
stresses, t
yz
is the shear stress, and u is the angle between a
horizontal plane and a slip surface. The probability failure is
evaluated with Equation (20) through the use of the Monte
Carlo method.
P
f
= Pr obability g < 0 ( )
(20)
For the internal friction angle |' of the embankment material Bs,
is dealt with as a random variable. Firstly, the random numbers
considering the spatial distribution derived from Equations (12)
and (13) are assigned to the N
SWS
. Secondly, the random
variable N
SPT
is evaluated by Equation (2) by considering the
conversion error c
r
, and then the |' is obtained with Equation
(14), including the conversion error term 5.3c
f
. The Monte
Carlo method is iterated 1000 times.
4.3 Risk evaluation
Two cases of the original embankment and the restored one are
compared, whose cross sections are shown in Figures 5(a) and
(b). In the figures, the representative slip surfaces, which give
the minimum safety factors, are exhibited. In Table 2, the
results of the reliability analysis are shown, in which F
s
is the
average factor of safety, P
f
is the probability of failure, C
f
is the
failure cost, including the damage to houses, agricultural
facilities, and farm lands, and C
F
is the value of the expected
failure cost. The average factor of safety is almost 1.0 for the
original embankment, and the probability of failure is nearly
20%, which seems very high. For the restored embankment, on
the other hand, the probability of failure is nearly zero and the
evaluated reduction in risk is drastic, at a value of 39,400,000
JPY. The reduction value means the effect of the improvement
work for the embankment.
Table 1. Parameters of embankment materials.

f
s

Figure 6. Slip surface across an element
Table 2. Result of reliability analysis.
Fs P
f
C
f
C
F

Original 1.04 0.197 3,940
Restored 1.62 0 0
Unit (10000 JPY)
2,000
2,000
C
F
= C
f
P
f

5 CONCLUSIONS
(1) A method to determine the statistical models of the soil
strength was presented. The indicator simulation, which is one
of the geostatistical methods, was employed. With the proposed
procedure, a detailed spatial distribution of the N-values was
exhibited.
(2) Based on the determined statistical model of the internal
friction angle, including the spatial distribution of the N-values,
the two conversion errors, from the SWS N-value to the SPT N-
value, and the SPT N-value to the internal friction angle, the
reliability analysis was conducted for an earth-fill embankment,
and the probability of failure was evaluated for the original state
of the embankment and the restored state of the embankment.
By comparing the risks between the original state and the
restored one, the effect of the improvement work of the
embankment was evaluated
6 ACKNOWLEDGEMENTS
This work was partly supported by JSPS KAKENHI
(23248040).
7 REFERENCES
Akaike H. 1974. A new look at the statistical model identification. IEEE
Trans. on Automatic Control, AC-19 (6), 716-723.
Cafaro F. and Cherubini C. 2002. Large sample spacing in evaluation of
vertical strength variability of clayey soil. Journal of Geotechnical
and Geoenvironmental Engineering 128 (7), 558-568.
Deutsch C.V. and Journel A.G. 1992. Geostatistical Software Library
and Users Guide, Oxford University Press.
Inada M. 1960. Usage of Swedish weight sounding results. Tsuchi-to-
Kiso, J. of JSSMGE 8 (1), 13-18 (in Japanese).
Journel A.G. and Huijbregts Ch.J. 1978. Mining geostatistics, Academic
Press.
Hatanaka M. and Uchida A. 1996. Empirical correlation between
penetration resistance and internal friction angle of sandy soils.
Soils and Foundations 36(4), 1-9.
Nishigaki M. 2001. AC-UNSAF3D User's Manual. (in Japanese).
Shinoda M., Horii K., Yonezawa T., Tateyama M. and Koseki J. 2006.
Reliability-based seismic deformation analysis of reinforced soil
slopes. Soils and Foundations 46 (4), 477-490.
Soulie P., Montes P. and Silvestri V. 1990. Modelling spatial variability
of soil parameters. Canadian Geotechnical Journal 27. 617-630.
Tang W.H. 1979. Probabilistic evaluation penetration resistances.
Journal of the geotechnical engineering, ASCE, 105(GT10). 1173-
1191.
Uzielli M., Vannucchi and Phoon, K. K. 2005. Random field
characterization of stress-normalized cone penetration testing
parameters. Geotechnique 55(1), 3-20.
Yoshida, I., Arakawa, T., Kitazume, T. and OOtsu H. 2005. Study on
seismic probabilistic safety assessment of a slope, Journal of
geotechnical engineering, JSCE, No.785, 27-37.
603
th
1
Correlation between cone penetration rate and measured cone penetration
parameters in silty soils
Corrlation entre le taux de pntration dun cne et des mesures de paramtres de pntration au
cne dans les sols limoneux.
Poulsen R., Nielsen B.N., Ibsen L.B.
Aalborg University, Department of Civil Engineering, Aalborg, Denmark
ABSTRACT: This paper shows, how a change in cone penetration rate affects the cone penetration measurements, hence the cone
resistance, pore pressure, and sleeve friction in silty soil. The standard rate of penetration is 20 mm/s, and it is generally accepted that
undrained penetration occurs in clay while drained penetration occurs in sand. When lowering the penetration rate, the soil pore water
starts to dissipate and a change in the drainage condition is seen. In intermediate soils such as silty soils, the standard cone penetration
rate may result in a drainage condition that could be undrained, partially or fully drained. However, lowering the penetration rate in
silty soils has a great significance because of the soil permeability, and only a small change in penetration rate will result in changed
cone penetration measurements. In this paper, analyses will be done on data from 15 field cone penetration tests with varying
penetration rates conducted at a test site where the subsoil primary consists of sandy silt. It is shown how a reduced penetration rate
influences the cone penetration measurements e.g. the cone resistance, pore pressure, and sleeve friction.
RSUM: Dans cet article, on montre comment un changement dans le taux de pntration dun cne affecte les mesures de
pntration de cne, d'o la rsistance du cne, la pression interstitielle et la friction manche en sol limoneux. Le taux normal de
pntration est de 20 mm/s, et il est gnralement admis que la pntration se produit dans de l'argile non draine alors que la
pntration se produit dans le sable drain. Lors de l'abaissement du taux de pntration, l'eau interstitielle du sol commence se
dissiper et un changement de l'tat de drainage est vu. Dans les sols intermdiaires, tels que les sols limoneux, le taux de pntration
de cne standard peut conduire un drainage des conditions qui pourraient tre non draines, partiellement ou totalement dcharge.
Cependant, l'abaissement du taux de pntration dans les sols limoneux a une grande importance en raison de la permabilit du sol et
seulement un petit changement dans le taux de pntration se traduira par des mesures de pntration au cne chang. Dans ce
document, les donnes de 15 essais sur le terrain de pntration au cne, avec diffrents taux de pntration menes sur un site d'essai
o le premier sous-sol se compose de limon sableux, sont analyss. Linfluence dune rduction du taux de pntration sur les
mesures de pntration dun cne, par exemple la rsistance du cne, la pression de pore, et la friction manchon, est dmontre.
KEYWORDS: Silt, CPT, penetration rate, cone resistance, pore pressure, sleeve friction, drainage, in situ testing.
1 INTRODUCTION
The Cone Penetration Test (CPT) is an in situ testing method
that todays geotechnical engineers often make use of when
determining soil parameters, and classifying soil type. The
standard rate of penetration is 20 5 mm/s, (ASTM 2007), and
while the cone is pushed into the ground the cone resistance,
(q
c
), pore pressure (u
2
s
), and depth (d) are
measured.
During the penetration, the pore water starts to dissipate, and
the dissipation for sands occurs so quickly that the penetration
appears as fully drained, whereas the dissipation happens over
time for clays, for which reason the penetration is undrained in
clays. For intermediate soil, such as silty soils, the penetration is
somewhat in between; that is partially drained.
According to several researchers (Silva and Bolton 2005,
Lehane et al. 2009, Kim et al. 2008, Schneider et al. 2008,
Chung et al. 2006, House et al. 2001), the drainage is dependent
on the soil permeability, compressibility and penetration rate.
The soil permeability and compressibility are both connected to
the soil type. However, the penetration rate is regardless of soil
type 20 mm/s.
When the penetration rate is lowered, the pore water
dissipates (change in drainage condition) which results in an
increased cone resistance (Lehane et al. 2009, Kim et al. 2008,
Chung et al 2006, House et al. 2001). For this reason, the largest
cone resistance that could be obtained corresponds to a fully
drained penetration. This effect has been shown by several
researchers (Chung et al. 2006, House et al. 2001 and Randolph
and Hope 2004) from laboratory tests in clay. Kim et al. (2008)
also conducted laboratory as well as field cone penetration tests
in cohesive soil and found that the soil behaves undrained for a
penetration rate of 20 mm/s and partially drained for a
penetration rate of 0.05 mm/s.
According to Poulsen et al. (2011a), the change in
penetration rate and hence drainage condition has a greater
impact in silty soils where the standard rate of penetration often
induces a partially drained penetration.
This paper analyses data from 15 field cone penetration tests
conducted with a penetration rate varying from 60 to 0.5 mm/s.
Only a short description of the method for the cone penetrations
tests will be given. The results and the interpretation of how a
change in the penetration rate affect the measured parameters,
hence the cone resistance, pore pressure, and sleeve friction will
be given.
2 DESCRIOTION OF EXEPERIMENTAL PROGRAM
The aim of the research is to examine how a change in the cone
penetration rate affects the measured cone penetration
parameters when conducting cone penetration tests (CPT). The
research was carried out at a test site located in the northern
Jutland in Denmark, more specifically at a field near the town
Dronninglund.
604
th
2.1 Test site soil stratigraphy
The soil stratigraphy was at the test site identified by means of
two geotechnical boring results. In addition, soil samples were
collected and laboratory tests were executed in order to classify
the soil. Both test results show that the subsoil consists of sandy
silt with clay stripes from approx. 4.5 to 11.4 m below ground
level. Above 4 m, the soil consists of silty sand, and below 11 m
the soil consists of clay with sandy silt stripes. In addition, the
groundwater was encountered at approximately 0.2-0.6 m below
the ground level. Generally, the soil is much layered and
inhomogeneous which makes the soil difficult to classify.
However, in Poulsen et al. (2012a), the soil was overall
classified as sandy silt with clay stripes.
2.2 Cone Penetration Tests
In order to examine how a change in the cone penetration rate
affects the measurements, various cone penetration tests have
been conducted. A total of 15 CPTs with five different
penetration rates were conducted; these were 60, 20, 5, 1, and
0.5 mm/s. All CPTs were conducted with a distance of
approximately 3 m. This was done to make sure that the
drainage of each CPT would not affect the drainage of the other
CPTs. The location of the 15 CPTs can be seen in Figure 1.
60mm/s
20mm/s
5mm/s
1mm/s
0.5mm/s
577670 577660 577650 577640
6336550
6336540
6336530
East coordinate(m)
N
o
r
t
h

c
o
o
r
d
i
n
a
t
e

(
m
)
Borings
Figure 1 Location of the 15 CPTs with penetration rates of 60, 20, 5, 1
and 0.5 mm/s. The coordinates are given UTM coordinates.
During the execution of the CPTs, the cone resistance, (q
c
),
pore pressure (u
2
s
), depth (d), and the
penetration rate (v) were measured. A more detailed description
of the test site, experimental programme and the validity of the
tests can be found in Poulsen et al. (2012b).
Because of the layered and inhomogeneous soil, the
measured CPT parameters are very fluctuating and hence
difficult to interpret. As a result, the data has been smoothed for
every 50 cm, which was concluded acceptable in Poulsen et al.
(2012b).
3 EFFCT OF PENETRATION RATE IN SILT LAYER
The soil layer classified as sandy silt with clay stripes was
located between 4.5 to 11.4 m below ground level. Only this
layer has been analysed since it is considered to be the silt layer
where the effect of the penetration rate is clearest. As a result,
the following graphs only contain results from 4.5 to 11.4 m.
In the following, it is analysed how a change in cone
penetration rate affects the measured cone resistance, pore
pressure and sleeve friction respectively. As described in
Poulsen et al. (2012b), the soil layer consists of many stripes,
which gives a very fluctuating result for the measured cone
penetration parameters. In order to clearly visualise the effect of
a change in the penetration rate, only the penetration rates of 60
and 0.5 mm/s have been included. This is done as it is the
extreme points corresponding to undrained and fully drained
that are of interest, and the penetration rates of 60 and 0.5mm/s
are closest to these conditions. Consequently, the data from the
CPTs conducted with a penetration rate of 20, 5 and 1 mm/s
have been excluded in the figures.
3.1 Pore pressure
In Figure 2, the smoothed pore pressure from 4.5 to 11.4 m
for the CPTs conducted with a penetration rate of 60 and 0.5
mm/s can be seen.
-200 0 200 400 600 800 1000
5
6
7
8
9
10
11
u
2
(kPa)
D
e
p
t
h

(
m
)

60 mm/s
0.5 mm/s
u
0
Figure 2. Comparison of the smoothed pore pressure conducted with a
penetration rate of 60 and 0.5 mm/s. The figure contains results from 3
CPTs test for each penetration rate. The plotted u
0
is an average value.
Figure 2 shows that changing the penetration rate from 60 to
0.5 mm/s results in a decreased pore pressure. This is because
the drainage conditions change when the penetration rate is
decreased. From Figure 2, it seems as though the CPT
conducted with a penetration rate of 0.5 mm/s corresponding to
fully drained penetration, since the measured pore pressure is
close to or equivalent to u
0
.
However, it is not possible to conclude if the CPT conducted
with a penetration rate of 60 mm/s corresponds to undrained or
partially drained penetration. Nevertheless, by lowering the
penetration rate, the penetration changes from undrained or
partially drained to fully drained, which results in a lower pore
pressure.
3.2 Cone resistance
In Figure 3, the smoothed cone resistance from 4.5 to 11.4 m
for the CPTs conducted with a penetration rate of 60 and 0.5
mm/s can be seen.
Figure 3 likewise shows that changing the cone penetration
rate from 60 to 0.5 mm/s results in a change in the cone
resistance. However, a decreased penetration rate results in an
increased cone resistance. The changes observed in the cone
resistance are like the pore pressure caused by changes in
drainage conditions where the penetration changes from
undrained or partially drained to fully drained. This results in a
higher cone resistance.
3.3 Sleeve friction
In Figure 4, the sleeve friction from 4.5 to 11.4 m for the
CPTs conducted with a penetration rate of 60 and 0.5 mm/s can
be seen.
Contrary to the pore pressure and cone resistance, Figure 4
does not show any correlation between the sleeve friction and
cone penetration rate.
605
0 5 10 15 20
5
6
7
8
9
10
11
q
t
(MPa)
D
e
p
t
h

(
m
)

60 mm/s
0.5 mm/s
Figure 3 Comparison of the smoothed cone resistance conducted with a
CPTs test for each penetration rate.
0 20 40 60 80 100 120 140 160 180
5
6
7
8
9
10
11
f
s
(kPa)
D
e
p
t
h

(
m
)

60 mm/s
0.5 mm/s
Figure 4. Comparison of the smoothed sleeve friction conducted with a
CPTs test for each penetration rate.
4 CORRELATION BETWEEN PENETRATION RATE
AND MEASURED PARAMETERS
The order of the change in the cone penetrations parameters that
can be anticipated is however difficult to read from Figure 2,
Figure 3 and Figure 4. As a result, the mean value of the entire
silt layer from all CPTs (CPTs with penetration rate of 60, 20, 5,
1 and 0.5 mm/s) can be plotted in a semi logarithmic plot. This
has been done for the sleeve friction in Figure 5.
Just as Figure 4, Figure 5 does not show any correlation
between the mean sleeve friction and the mean penetration rate.
According to Lunne et al. (1997) the sleeve friction does not
give consistent results during cone penetration. The results
shown in Figure 4 and Figure 5 substantiate this, for which
reason caution must be taken when using the sleeve friction to
analyse CPT data.
The mean value for the pore pressure and cone resistance
plotted against the mean penetration rate in a semi logarithmic
plot is seen in Figure 6 and Figure 7. It can be seen that a
correlation between the pore pressure and the penetration rate
(Figure 6) and cone resistance and the penetration rate (Figure
7) exist.
0.1 1 10 100
70
80
90
100
110
120
130
140
Penetration rate (mm/s)
f
s

(
k
P
a
)

Mean values
Figure 5. The mean sleeve friction plotted against the mean penetration
rate from 4.5 to 11.4 m below ground level. No correlation seems to
exist. The standard rate of penetration of 20 mm/s has been marked with
a dotted line.
0.1 1 10 100
0
50
100
150
200
250
300
350
u
2

(
k
P
a
)

Fitting curve
Mean values
Figure 6. Correlation between the mean pore pressure and the mean
penetration rate from 4.5 to 11.4 m below ground level. The standard
rate of penetration of 20 mm/s has been marked with a dotted line.
0.1 1 10 100
0
2
4
6
8
10
12
q
t

(
M
P
a
)

Fitting curve
Mean values
Figure 7. Correlation between the mean cone resistance and the mean
penetration rate from 4.5 to 11.4 m below ground level. The standard
rate of penetration of 20 mm/s has been marked with a dotted line.
606
th
The correlations appear to be linear, however this cannot be
true since there must exist an upper and lower boundary for the
cone resistance and pore pressure corresponding to fully drained
and fully undrained. The mean values can instead be fitted by an
expression from Chung et al. (2006), which has been modified
by Poulsen et al. (2012b). The expression is given in (1) and (2)

(2)
where is the pore pressure (kPa), is the cone resistance
(MPa), is the penetration rate (mm/s), is the reference
penetration rate equal to 20 mm/s and a
u
, b
u
, c
u
, m
u
, a
q
, b
q
, c
q
,
and m
q
are fitting constants.
The value of corresponds to undrained penetration,
whereas corresponds to a fully drained penetration.
From Figure 6 and Figure 7, it is not possible to see when the
penetration is undrained or fully drained from the mean values.
For this reason, the constants a and b must be assumed.
However, in Figure 2 the penetration is close to fully drained for
a penetration rate of approximately 0.5 mm/s. This gives an
estimate of the constants a
u
+ b
u
and a
q
+ b
q
. The value of the
other constants can be seen in Table 1, and the fitting curves for
the pore pressure and cone resistance can be seen in Figure 6
and Figure 7.
Table 1. Derived value for fitting constants.
a b c m
Pore pressure, u 350 -290 1.2 1.1
Cone resistance, q
5.3 3.8
3.1 0.9
By lowering the penetration rate so that the penetration
occurs as drained, the cone resistance increases. This can be
expressed as (3) (Poulsen et al. 2012b):
(3)
Where is the cone resistance corresponding to
drained penetration (MPa), is the measured cone
resistance determined with a penetration rate of 20 mm/s (MPa),
and is a coefficient of drainage. The coefficient of drainage,
can for the Dronninglund silt be set to 1.0-1.7 depended on
whether is undrained ( =1.7), fully drained
( =1.0), or how close to fully drained it is.
5 CONCLUSIONS
This paper has shown how a change in the penetration rate
affects the measured cone penetration parameters in silty soil.
When using cone penetration tests (CPT) with the standard rate
of penetration of 20 mm/s, the penetration will appear as fully
drained in sandy soils and undrained in clayed soils. However,
for silty soils the standard rate of penetration of 20 mm/s results
in a partially drained penetration. In order to examine which
affect a changed penetration rate has in silty soils on the
measured cone penetration parameters (cone resistance, pore
pressure, and sleeve friction), 15 CPTs with varying penetration
from 60 to 0.5 mm/s have been conducted.
Results from the cone penetration tests conducted with a
penetration rate of 60 and 0.5 mm/s were compared for the cone
resistance, pore pressure and sleeve friction. It was shown that
both the pore pressure and cone resistance gave different results
for a penetration rate of 60 and 0.5 mm/s. The pore pressure
measured with a penetration rate of 0.5 mm/s corresponded to
drained penetration, which resulted in the highest cone
resistance. For the sleeve friction, no correlation was seen.
In addition, a correlation between the mean pore pressure
and mean penetration rate, and mean cone resistance and mean
penetration rate was however seen when plotting the mean
penetration rate in a semi logarithmic plot.
Compared to the normal penetration rate of 20 mm/s, a
decrease in the penetration rate leads to an increase in the cone
resistance due to drainage. The increase can be expressed by a
coefficient of drainage, , that is equal to 1.0 for fully drained
penetration and 1.7 for undrained penetration. The increase
depends on whether the normal penetration rate of 20 mm/s has
been conducted under undrained, partially drained or fully
drained conditions.
For that reason, it can be concluded that a correlation
between the cone penetration rate and the cone resistance and
pore pressure exists. It is an important factor that the cone
resistance is dependent on drainage condition and consequently
the penetration rate. Particularly if a project requires knowledge
of both the undrained soil parameters and the drained soil
parameters. In this case, it can be useful to know when the
penetration is partially drained and how to convert it to a fully
drained penetration or undrained penetration.
6 ACKNOWLEDGEMENTS
The project is funded by DONG Energy and associated with the
EUDP programme Monopile cost reduction and demonstration
by joint applied research funded by the Danish energy sector.
The funding is sincerely acknowledged.
7 REFERENCES
ASTM. 2007. Standard Test Method for Electronic Friction Cone and
Piezocone Penetration Testing of Soils. D5778-07, ASTM,
Pennsylvania.
Chung, S. F., Randolph, M.F., and Schneider, J.A. 2006. Effect of
Penetration Rate on Penetrometer Resistance in Clay. J. Geotech.
Geoenviron Eng.. 132(9), 1188-1196.
House, A. R., Oliveira, J. R. M. S., and Randolph, M.F. 2001.
Evaluating the Coefficient of Consolidation using Penetration
Tests. Int. J. of Physical Modelling in Geotechnics. 3, 17-26.
Kim, K., Prezzi, M., Salgado, R., and Lee, W. 2008. Effect of
Penetration Rate on Cone Penetration Resistance in Satuated
Clayey Soils. J. Geotech. Geoenviron Eng.. 134(8), 1142-1153.
Lehane, B. M., OLoughlin, C. D., Gaudin, C., and Randolph, M. F.
2009. Rate effects on penetrometer resistance in kaolin.
Gotechnique. 41(1), 41-52.
Lunne, T., Robertson, P. K., and Powell, J. J. M. 1997. Cone
Penetration Testing in Geotechnical Practice. 1
st
ed., Spon Press,
New York, NY, USA
Poulsen, R., Nielsen, B. N., Ibsen, L. B. 2011. Effect of Drainage
Conditions on Cone Penetration Testing in Silty Soils. Proc. 64th
Canadian Geotechnical Conference and 14th Pan-American
Conference on Soil Mechanics and Engineering. Toronto, ON,
Canada
Poulsen, R., Ibsen, L. B., Nielsen, B. N. 2012a. Difficulties Regarding
Determination of Plasticity Index of Silty Soils by use of
Casagrande and Fall Cone Methods. Proc. of Nordic Geotechnical
Meeting. Copenhagen, Denmark
Poulsen, R., Nielsen, B. N., Ibsen, L. B. 2012b. Field Test Evaluation of
Effect on Cone Resistance Caused by Change in Penetration Rate.
Proc. of Nordic Geotechnical Meeting. Copenhagen, Denmark
Schneider, J. A., Randolph, M. F., Mayne, P. W., and Ramsey, N. R.
2008. Analysis of Factors Influencing Soil Classification Using
Normalized Piezocone Tip Resistance and Pore Pressure
Parameters. J. Geotech. Geoenviron Eng.. 134(11), 1569-1576.
Silva, M. F., and Bolton, M. D. 2005. Interpretation of centrifuge
piezocone tests in dilatants, low plasticity silts. Proc., Int. Conf. on
Problematic Soils. Eastern Mediterranean University, Famagusta,
N. Cyprus.
607
Sampling method and pore water pressure measurement in the great depth (-400m)
Mthode de mesure de pression interstitielle de l'eau d'chantillonnage en grande profondeur
( 400m)
Rito F.
OYO Corporation
Emura T.
Kansai International Airport CO.,LTD
ABSTRACT: Pleistocene clay and sand layers are deposited in the great depth under Holocene soft clay layer at Kansai international
airport area. Since the weight of the reclamation soil is heavy because of its depth of sea water which is reached -20m, it has become
the very important issue that the characteristics of Pleistocene clays are investigated correctly. For this reason, the new type sampling
method which has been called Koken wire line system was developed and the undisturbed samples were obtained by this sampling
system. Sample quality which obtained from great depth was estimated using the range of the strain which was re-consolidated to in-
situ effective stress by constant strain rate consolidation test. As a result, it was confirmed that the sample quality of these samples had
good quality. On the construction phase of reclamation, cone type measuring equipment of pore pressure for Pleistocene clay and new
type measuring equipment of pore pressure for sand were developed and the excess pore water pressure was measured. As a result of
the examination of these data, the measured value has been had high accuracy. Therefore, the consolidation characteristic of
Pleistocene deposit of Kansai international airport area has been estimated more correct by these useful data.
RSUM : Largile Plistocne et couches du sable sont dposes dans la grande profondeur sous Holocene couche de l'argile
douce Kansai rgion aroportuaire internationale. Depuis le poids du sol de la rclamation est lourd cause de sa profondeur d'eau
de mer qui en est atteinte -20m, il est devenu la question trs importante que les caractristiques d'argiles Plistocne sont enqutes
sur correctement. Pour cette raison, la nouvelle mthode de l'chantillonnage du type qui a t appele 'Koken installent le systme de
la ligne' a t dvelopp et les chantillons non drangs ont t obtenus par ce systme de l'chantillonnage. Gotez la qualit qui a
obtenu de grande profondeur a t estime utiliser la gamme de la tension qui a t rconsolide in-situ stress efficace par preuve
de la consolidation du taux de la tension constante. En consquence, il a t confirm que la qualit de l'chantillon de ces
chantillons avait la bonne qualit. Sur la phase de la construction de rclamation, matriel de la mesure du type du cne de pression
du pore pour argile Plistocne et nouveau matriel de la mesure du type de pression du pore pour le sable a t dvelopp et la
pression de l'eau du pore en excs a t mesure. Par suite de l'examen de ces donnes, la valeur mesure a t eue la haute exactitude.
Par consquent, la caractristique de la consolidation de dpt Plistocne de Kansai que la rgion aroportuaire internationale a t
estime plus correct par ces donnes utiles.

KEYWORDS: Pleistocene clay, Koken wire-line system, Pore water pressure measurement.

1 INTRODUCTION
Kansai international airport has constructed in the Osaka bay
area. In this area, Pleistocene clay and sand layers are
deposited into the great depth under a Holocene soft clay
layer. Since the weight of the reclamation soil is heavy
because of its depth of sea water which is reached 20m in
depth, it has become the very important issue that the
characteristics of Pleistocene clays are investigated correctly.
Therefore, it has been required high quality sampling and
high-precision consolidation test for the samples deposited
such great depth of 400m in depth.
On the construction phase of reclamation, the measurement of
pore water pressure for Pleistocene clay and sand layers has become a
important issue to improve settlement analysis in addition to the
measurement of settlement. As the target depth of the measurement of
pore water pressure reaches 300m in depth, the piezometer and the
permeability test equipment which are usually used in shallow depth
can not use such great depth. The cone type measuring equipment of
pore water pressure for Pleistocene clay named GD-CONE and the
new type measuring equipment of pore water pressure for sand
named H-MHT have been developed. We have been able to measure
the pore water pressure in great depth using these new equipment.
2 SAMPLING FROM GREAT DEPTH USING KOKEN
WIRE LINE METHOD
Port and Airport Research Institute has tried to improve wire
line boring method for the investigation method at the port
and harbor area (Matsumoto K., et al.1981). As a result, new
wire line method called Koken wire line method has
developed. The characteristic of this method is to be able to
obtain undisturbed samples which are stiff clay and sand in
great depth. Koken wire line method has applied for boring
and sampling method of Kansai international reclamation
project.
The system of Koken wire line method is shown in Figure
1 (Okumura T., et al.1982). Three types of specific samplers
have made for Koken wire line method. The structure of these
samplers is shown in Figure 2 and Table 1. Thin-walled tube
sampler with fixed piston is used for soft and stiff clay whose
unconfined compressive strength is under 2MN/m
2
. Denison
sampler which is rotary double-tube sampler is used for more stiff
clay whose unconfined compressive strength is over 2MN/m
2
. Rigid
sampler which is double-tube sampler fixed outer tube and inner tube
is used for stiff sand and gravel.

608
th
Pump
Mud screen
Mixer for mud water
Boring machine
Tank for clean water
Tank for sludge
Suction tank
Tank for mud water
Winch
Casing pipe
(=10,8,6)
Sea level
Sea bottom
Boring rod
(=135mm)
Outer tube
Rigid sampler or
Denison sampler
Wire rope
(=9mm)
Water pressure
Thin-walled tube sampler
with fixed piston
Sample
Sampling tube

Figure 1. System of Koken wire line method.

Bit
Guide ring
Outer cube
Inner cube
Sampling tube
Drain hole
Ball cone clamp
Shoulder ring
Wire
Piston Piston rod Valve for air extraction
Denison sampler
Cutting wedge
Outer tube
Sampling tube
Ball check
Shoulder ring
Latch
Wing
Bit
Inner tube
Spring
Thrust bearing
Thin-walled tube sampler with fixed piston
Rigid sampler
Metal crown
Outer tube
Inner tube
Sampling tube
Shoulder ring
Latch assembly
Spring
Bit
Guide ring
Ball check

Figure 2. Samplers for Koken wire line method.

3 ESTIMATION OF SAMPLESQUALITY OBTAINED
FROM GREAT DEPTH
The samples obtained from great depth which is up to 400m
in depth are influenced not only mechanical disturbance but
also stress release. As mechanical disturbance can avoid
improving sampling technic, stress release cannot avoid even
what kind of method. Therefore, the estimation of sample
quality has to be examined quantitatively to interpret the
results of constant strain rate consolidation test (CRS test).
The volumetric strain which is consolidated to
the in-situ effective stress,
v0
, can be used as
an indication of sample quality. The relationship between
change of
v0
and quality of samples having various degree of
sample disturbance is shown in Table 2 (Andersen A. and
Kolstad P. 1979). The relationship between e/e
0
and sample
quality, where e is change in void ratio in recompressing a
sample to in-situ effective stress and e
0
is initial void ratio, is
shown in Table 3 (Lunne T., et al. 1997). According to this
figure, when the range of e/e
0
is within 0.07, the sample can
regard good quality.
The change of
v0
ande/e
0
profile of Pleistocene clay
samples with recompression to the corresponding in-situ
effective vertical stress is shown in Figure 3. With few
exceptions, the range of
v0
varies within 2% to 4%. It is
apparent that the majority of samples of Pleistocene clay are
good quality. The values of e/e
0
vary in a narrow band of
0.04 to 0.07, and are not sensitive with depth. It may be
recalled that if e/e
0
is within 0.07, the soil samples can be
considered to be good quality.
The relation between
v0
and OCR, e/e
0
and OCR of
Pleistocene clay samples is shown in Figure 4. It is obvious
that OCR is almost constant with depth, and is independent of
v0
and e/e
0
. As the consequence, it is clear that the samples
obtained from great depth in the Osaka bay have good and
uniform quality.

Table 2. Relationship between volumetric strain (
v0
) and sample
uality. q
v0
(%) Sample quality
1 Very good
12 Good
24 Fair
410 Poor

10 Very poor

Table 3. Relationship between changing rate of void ratio (e/e
0
) and
ample quality. s
e/e
0

Overconsolidation
Ratio
Very good
to excellent
Good
to fair
Poor
Very
poor
1-2 0.04 0.04-0.07 0.07-0.14 0.14
2-4 0.03 0.03-0.05 0.05-0.10 0.10
For particular clay multiply e/e
0
by e
0
/(l+e
0
) to get the criteria in terms of
V0

Table 1. Specification of samplers.
Sampler Sampling tube
Applicated
soil property Sampler Name
Outer
diameter
(mm)
Length
(mm)
Remarks Material
Inner
diameter
(mm)
Thickness
(mm)
Angle of
edge
()
Ratio of
inside
diameter
(%)
Length
(mm)
Remarks
Soft clay
Thin-walled tube sampler
with fixed piston
108 4358
Hydraulic type
sampler

Stainless steel
(SUS-304)
90 2 6 0 1250 -
Stiff clay
Thick-walled tube sampler
with fixed piston
108 4358
Hydraulic type
sampler

Same as the
above
81.1 4 6 0 1170
Exchange of
edge blade is
possible
Stiff clay Denison sampler 108 2850
Projection length
of edge blade
(20-50mm)

Same as the
above
81.1 4 6 0.5 1000
Exchange of
edge blade is
possible
Stiff sand and
gravel
Rigid sampler 108 2875 -
Same as the
above
90 2 - - 1000 -
609

- 350
- 300
- 250
- 200
- 150
- 100
- 50
0
0 1 2 3 4 5 6
vo

(

m
)
vo=4%
- 350
- 300
- 250
- 200
- 150
- 100
- 50
0
0. 00 0. 02 0. 04 0. 06 0. 08 0. 10
ee
o
(

m
)
e/e
0
=0. 07
D
e
p
t
h

(
m
)

D
e
p
t
h

(
m
)

Figure 3. Result of
v0
and e/e
0
obtained by CRS test.

0. 0
0. 5
1. 0
1. 5
2. 0
2. 5
0. 00 0. 02 0. 04 0. 06 0. 08 0. 10
(
O
C
R
)
0. 0
0. 5
1. 0
1. 5
2. 0
2. 5
0. 0 1. 0 2. 0 3. 0 4. 0 5. 0 6. 0
vo
(%)
(
O
C
R
)
O
v
e
r
c
o
n
s
o
l
i
d
a
t
i
o
n

r
a
t
i
o

(
O
C
R

)

O
v
e
r
c
o
n
s
o
l
i
d
a
t
i
o
n

r
a
t
i
o

(
O
C
R

)

Figure 4. Relationship between OCR and
v0
, e/e
0
.
4 MEASURMENT OF PORE WATER PRESSURE IN
GREAT DEPTH
4.1 Cone type measuring equipment of pore water pressure for
Pleistocene clay (GD-CONE)
The measurement of pore water pressure for clay layer is used a
push-in type piezometer (JGS 1313-2003) .This type of
piezometer has a merit which can seal the measuring section
completely and can measure a correct pore water pressure.
Therefore, this is usually used in shallow depth and cannot use
such great depth which is up to 350m in depth because of the
capacity of sensor and the penetrating power of cone.
The cone type measuring equipment of pore water pressure
for Pleistocene clay in great depth called GD-CONE has been
developed. The structure of this equipment is shown in Figure 5.
The characteristics of this cone are as follows:
The tip part which is the penetrating part is very thin in order to
decrease the penetrating resistance and promote the dissipation
of pore water pressure. Its diameter is only 15mm to 20mm.
The upper part of the tip becomes thicker gradually. Its diameter
is 41mm to 56mm. This part is penetrated into the small
borehole, which is drilled in advance, to seal the testing section
completely.
GD-CONE is connected with AQ rod whose outer diameter is
44.5mm and installed into the borehole. During installation of
GD-CONE, the center riser fixed to AQ rod is used in order to
install into the pre-drilled small borehole correctly.
The pressure gauge of GD-CONE has used a crystal oscillation sensor
which has wide pressure range and high sensibility. The maximum
pressure range is 5MPa and the sensibility has 0.01%FS. The
compensation of atmospheric pressure has been done by using another
pressure gauge on the ground.
4.2 New type measuring equipment of pore water pressure for
sand (H-MHT)
The measurement of pore water pressure for sand layer in great
depth is used a new type measuring equipment of pore water
pressure called H-MHT. The structure and test procedure of H-
MHT is shown in Figure 6. The characteristic of H-MHT are as
follows:
As the principle of measurement is simple, the reliable
measurement is possible easily.
As the pressure gauge of H-MHT has used a crystal oscillation
sensor too, the high pressure caused in the great depth can
be measured highly precise.
H-MHT can obtain equilibrium water table in a short time because the
specific air valve which is joined to the AQ rod can shut the test
section in order to promote dissipation of pore water pressure
rapidly.

Metal filter
Tip part
Crystal oscillation sensor
Connection to AQ rod
(Outer diameter: 44.5mm,
Inner diameter: 34.9mm)
Length 2,540mm
(300mm)
Penetrating part Sealing part Pore water pressure measuring part
Figure 5. Structure of cone type measuring equipment of pore water pressure for Pleistocene clay (GD-CONE).

610
th

Wire line rod
T
e
s
t
i
n
g

s
e
c
t
i
o
n

T
e
s
t
i
n
g

s
e
c
t
i
o
n

Water table in measuring pipe is
fallen by being given gas pressure
Packer
Three-way
Pressure regulation
Distributor
Nylon tube for packer
pressurization
Pressure sensor code
Converter Converter
Computer Computer
Gas tube
Air valve
AQ rod
Wire line rod
Underwater connector
Packer
(Expanded state)
Underwater connector
Drainage pipe
Crystal oscillation sensor
Guide for water flow

Figure 6. Structure of new type measuring equipment of pore water pressure for sand (H-MHT).
4.3 Results and discussion
4.3.1 Estimation of pore water pressure using hyperbolic
method
The diameter of the tip part of GD-CONE is designed of thin
size because pore water pressure is needed to dissipate rapidly.
The standard method to GD-CONE continue to measure for
three hours after penetrating. The adoption value of pore water
pressure is calculated by hyperbolic method using the data
measured after three hours from penetration. In order to
accurate the adoption value, the long hours measurement, 38
hours, has carried out. The results of these data are shown in
Table 4. Since the results of the long hours measurement and
the adoption value by hyperbolic method using three hours
measurement are almost same, the accuracy of hyperbolic
method has been confirmed.
Table 4. Application of hyperbola method for measuring result of pore
ater pressure. w
No.
Altitude
(CDL-m)
Investigation
case
Measuring
hours(sec)
Pore water pressure
of last measuring
time (kPa)
Pore water pressure
by hyperbola
method (kPa)
1 263.47 11260 2760 2700
2 263.47
Standard
measurement
12110 2800 2702
3 264.02
Long
measurement
138917 2700 2702
4.3.2 Reliability of sealing of measurement section of GD-
CONE
Three patterns of penetration of GD-CONE, which is varied
from 30cm, 60cm to 90cm length, have carried out. The results
are shown in Figure 7. These data are almost same despite the
penetrating length. In addition, since the result of pore water
pressure is not same to the mud water pressure of the bore hole,
the seal of measuring section is regarded completely.

k
P
a

6000

5000

4000

3000

2000

1000

0

0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000
sec
1
2
3

2960kPa
Note
Pushing length: 30cm
Mud water pressure=2,960kPa
P
o
r
e

w
a
t
e
r

p
r
e
s
s
u
r
e

(
k
P
a
)

Time (sec)

Figure 7. Difference of test result of pore water pressure by penetrating
length.

4.3.3 Comparison examination of the results by GD-CONE
and H-MHT for Pleistocene clay
The comparative experiments used both GD-CONE and H-
MHT has been carried out in same depth, which is 170m in
depth. The results are shown in Table 5. Since these data are
almost same, the appropriate of GD-CONE and H-MHT method
for measuring equipment for great depth can be confirmed.

Table 5. Comparison between measuring result of pore water pressure
sing GD-CONE and H-MHT. u
Measuring
method
Altitude
(CDL-m)
Pore water pressure
p (kPa)
Excess pore water
pressure u (kPa)
GD-CONE 172.77 2006 256
H-MHT 169.12172.77 1984 257
5 CONCLUSIONS
Kansai international airport which has been constructed in the
Osaka bay far from 5km from the land area in order to solve
noise pollution has been able to take off and landing of airplane
using two runways whose length is about 4,000m. The
consolidation settlement of 2nd runway, which is related to this
paper, is almost the same like the consolidation analysis. For the
future, the management of consolidation settlement shall be
important for the operation of the airport while the
consolidation settlement will continue for long times. It is
important that the continuing study for the settlement of
Pleistocene clays using the in-situ observation data.
6 REFFERENCES
Matsumoto K., et al. 1981. Undisturbed sampling method using wire
line sampler (in Japanese). Sampling symposium.
Okumura T., et al. 1982. Soil investigation at Kansai international
airport The investigation for great depth- (in Japanese).
Mechanization for construction.
Andresen A. and Kolstad P. 1979. The NGI 54-mm Samplers for
Undisturbed Sampling of Clays and Representative Sampling of
Coarser Materials. State of the Art on Current Practice of Soil
Sampling, Progressing of the International Symposium of Soil
Sampling. Singapore. 13-21.
Lunne T., et al. 1997. Sample disturbance effect in soft low plastic
Norwegian clay. Symposium on Recent developments in Soil and
Pavement Mechanics. Rio de Janeiro. 81-102.
Japanese geotechnical standard (JGS 1313-2003). 2004. Method for
Measuring Pore Water Pressure using Electric Transducer (in
Japanese). Japanese Standards for Geotechnical and
Geoenvironmental Investigation Methods Standards and
Explanations- . 368-376.
611
Une mthode de classification de la sensibilit des sols au moyen du pizocne
Soils sensibility classification method from piezocone data
Serratrice J.-F.
CETE Mditerrane, Aix en Provence, France
RSUM : Une mthode de classification des sols naturels partir des mesures au pizocne est propose. La mthode procde en
deux tapes et en rfrence aux rsistances draines et non draines mesures l'appareil triaxial sur les sols carotts au pralable dans
le mme site. La mthode est prsente puis deux exemples d'interprtation sont proposs et comments partir de mesures recueillies
en sondages. Les tendances d'volution des mesures au pizocne avec la profondeur dans les massifs argileux homognes rcents
sont examines ensuite.
ABSTRACT: A method of classification of natural soils from piezocone measurements is proposed. The method proceeds in two
steps with reference to drained and undrained strength provided by triaxial testing on soil previously sampled into the same site. The
method is presented and two examples of interpretation are proposed and discussed from data collected in-situ. Evolution trends of
piezocone responses with depth in recent homogeneous clayey deposits are then examined.
KEYWORDS: Soil, penetrometer test, piezocone, triaxial testing, shear strength, soil classification
1 INTRODUCTION
Une mthode de classification des sols est propose ici sur la
base des mesures au pizocne et par analogie avec les
rsistances dtermines l'appareil triaxial. Cette mthode
s'inscrit dans la ligne des mthodes de classification proposes
dans le pass. Senneset et al. (1982) sont les premiers avoir
introduit la variable B
q
dfinie comme le rapport entre la
pression d'eau nette u
2
u
0
et la rsistance nette q
t

v0
, o
v0
et la contrainte verticale totale la profondeur considre dans
le massif, u
0
la pression hydrostatique, q
t
la rsistance de pointe
et u
2
la pression d'eau mesure en position "u
2
". Parez et Fauriel
(1988) ont propos un guide de classification (et non pas un
abaque) bas sur la reprsentation de q
t
en fonction de B
q
, qui
s'inspire de celui propos par Senneset et Jambu (1984). Parez
et Fauriel (1988) rappellent ce titre que le guide qu'ils
proposent " ne dispense pas de raliser, sur chaque chantier,
un forage carott.".
Par la suite, Robertson et al. (1986), puis Robertson (1990),
ont propos une classification qui fait intervenir les trois
composantes mesures par l'intermdiaire de variables
normalises de la rsistance de pointe q
t
et du frottement latral
unitaire f
s
en accompagnement de la variable B
q
.
Fellenius et Eslami (2000) ont propos un abaque donnant la
rsistance effective q
E
(q
E
= q
t
u
2
) en fonction du frottement
latral unitaire f
s
. Cette classification prsente l'avantage
d'utiliser directement les mesures pntromtriques ou une
combinaison linaire de celles-ci. Cette rsistance "effective"
avait t introduite par Senneset et al. (1982).
Schneider et al. (2008) proposent un cadre de classification
des sols d'aprs les donnes du pizocne, qui associe la fois
la rsistance de pointe q
t
et la pression d'eau u
2
sous formes
normalises notes Q et U respectivement. Les mesures sont
reprsentes dans trois diagrammes qui combinent les deux
variables Q et U avec B
q
, chacun de ces diagrammes tant plus
pertinent qu'un autre, selon la nature du sol, pour tablir la
classification.
La mthode par analogie l'essai triaxial, propose ici,
prconise une utilisation directe de la mesure u
2
. Elle procde
en deux tapes : 1) classification des sols en rfrence leurs
rsistances triaxiales draines et non draines ; 2) valuation de
la sensibilit des sols. Ainsi, les trois mesures du pizocne sont
utilises directement, sans recours des variables normalises.
Cette mthode d'interprtation par analogie l'essai triaxial est
prsente tout d'abord, puis deux exemples d'illustration sont
comments ensuite.
2 PRSENTATION DE LA MTHODE
2.1 Prsentation
La mthode de classification des sols d'aprs les mesures au
pizocne se fonde sur l'analogie qui peut tre tablie entre les
comportements des matriaux observs en laboratoire
l'appareil triaxial et le fonage d'un pizocne dans ces mmes
matriaux. La mthode procde en deux tapes, en partant des
mesures brutes q
t
, f
s
et u
2
.
A l'tape 1, la rsistance de pointe est dcompose en une
partie isotrope et une partie dviatoire en tenant compte de la
pression d'eau u
2
. Pour cela, il est fait rfrence aux rsistances
draines et non draines mesures pralablement en laboratoire
l'appareil triaxial. Cette dcomposition permet de classer les
sols, en distinguant les sols argileux dans lesquels se
dveloppent de fortes pressions d'eau, des sols sableux dans
lesquels ces pressions sont gales la pression hydrostatique ou
sont ngatives. L'interprtation se fonde sur les caractristiques
de rsistance mesures partir d'chantillons carotts dans le
site.
L'tape 2 consiste identifier les sols sableux sensibles, de
faible compacit et peu rsistants, exposs au risque de
liqufaction notamment. La mthode se fonde sur les variations
relatives de f
s
et q
t
induites par la densification d'un sol. Le
principe de la classification l'tape 2 s'appuie sur des donnes
de la littrature et sur des donnes pntromtriques recueillies
dans diffrents sites en France.
L'intrt de la mthode rside dans l'utilisation simultane
des trois mesures fournies par le pizocne et qui portent en
elles l'effet de la profondeur sur la rsistance (effet du poids des
terres en tant que pression de confinement), pour des sols qui
612
th
peuvent tre considrs comme normalement consolids dans
leur tat naturel.
2.2 Classification l'tape 1
Des variables quivalentes sont dfinies pour le pizocne par
analogie aux variables p, p
T
et q de l'essai triaxial. Pendant
l'essai triaxial, p est la pression effective moyenne, p
T
la
pression totale moyenne et q le dviateur. La rsistance du sol
est dfinie dans le plan des contraintes effectives (p, q) par une
droite de Coulomb de caractristiques C
qc
et M
c
(ordonne
l'origine et pente) dont dcoulent les proprits effectives c' et '
(cohsion et angle de frottement). En contraintes totales et dans
le plan (p
T
, q), les caractristiques de rsistance sont C
q cu
, M
c cu
,
c
cu
,
cu
. (l'indice
cu
signifie consolid non drain).
Figure 1. Principe de la classification des sols l'tape 1 d'aprs les
donnes du pizocne et les rsistances triaxiales.
Dans la mthode par analogie l'essai triaxial, les variables
quivalentes dfinies partir des mesures du pizocne sont p'
c
,
p
c
et q
t
, une pression moyenne effective quivalente p'
c
, une
pression moyenne totale quivalente p
c
et la rsistance de pointe
q
t
. Les rsistances des sols mesures au pizocne sont
supposes s'exprimer au moyen des mme droites de rsistance
dans le plan des contraintes effectives quivalentes (p'
c
, q
t
) et
des contraintes totales quivalentes (p
c
, q
t
), comme indiqu sur
le graphique de la figure 1.
Tableau 1. Caractristiques des rsistances non draines des sols.
Sols c
cu
(kPa)
cu
()
Argiles, limons argileux 24 19,5
Limons, limons sableux 33 25,4
Sables, sables limoneux 50 36,9
Dans cet exemple et d'aprs les essais triaxiaux, la droite
dfinie en "contraintes effectives" possde les caractristiques
M
c
= 1,5 et C
qc
= 0, soit ' = 36,9 et c' = 0. Les trois autres
droites reprsentent les rsistances "consolides non draines"
dont les proprits
cu
et c
cu
sont donnes dans le tableau 1 et
qui reprsentent les rsistances des "argiles et limons argileux",
des "limons et limons sableux" et des "sables et sables
limoneux".
Ainsi, dans le graphique (p'
c
, q
t
), la mesure de q
t
permet de
dterminer la contrainte moyenne effective quivalente p'
c
gnre par la pointe sur la droite effective (M
c
, C
qc
). La
contrainte moyenne totale quivalente p
c
s'obtient alors en
ajoutant la mesure de u
2
dans la direction isotrope, ce qui donne
les coordonnes (p
c
, q
t
) du point reprsentatif de l'tat de
contraintes totales quivalentes autour du cne. Parmi les trois
droites dfinies en contraintes totales pour reprsenter les
rsistances non draines des sols, la droite la plus proche de ce
point permet de classer le sol.
Un exemple est donn sur la figure 1. Trois mesures
effectues au pizocne dans trois facis diffrents du mme
sondage y sont reportes (figurs pleins). Les pressions p'
c
sont
d'abord recherches sur la droites des rsistance effectives (en
trait pais). Puis, aprs adjonction de la pression u
2
, la classe de
sol est dtermine par la droite de rsistance totale la plus
proche (en traits fins, figurs creux). Aprs classification, les
figurs reprsentant les classes de sols sont reports sur les
diagrammes donnant q
t
, f
s
et u
2
en fonction de la profondeur,
comme sur les figures 3 et 4, puis sur tous les autres
diagrammes drivs. Il apparat clairement sur ces exemples que
les plus fortes pressions u
2
sont attribues aux limons argileux,
alors que les plus faibles pressions sont attribues aux sables et
sables limoneux.
2.3 Recherche de la sensibilit des sols l'tape 2
Aprs cette premire tape de classification des sols, l'tape 2
consiste dtecter les sols sensibles. Les sols sont considrs
comme sensibles s'ils appartiennent au quart infrieur gauche du
diagramme (lgf
s
, lgq
E
) dlimit par les deux courbes d'quations
q
E
= 2000 f
s
2
et q
E
= 2 f
s
0,16
(q
E
et f
s
en MPa).
Figure 2. Identification des sols sensibles dans le diagramme (f
s
, q
E
).
Ces courbes sont indiques sur le diagramme (f
s
, q
E
) de la
figure 2, en chelles arithmtiques. Les sols sensibles sont
symboliss par des figurs en rapport avec les croix reprsentant
les classes de sol identifies l'tape 1.
3 EXEMPLES
Ce paragraphe prsente deux exemples d'interprtation des
donnes recueillies au pizocne.
3.1 Exemple 1
Le premier exemple concerne des donnes recueillies pendant
une campagne de sondages au pizocne. Trois familles de sols
apparaissent, qui sont indiques dans le tableau 1. Leurs
caractristiques de rsistance ont t mesures au triaxial
partir des chantillons carotts dans le site. Les trois familles de
sols sont frottantes, avec des rsistances effectives communes
(c' = 0 et ' = 36,9 ). Le tableau 1 indique les rsistances non
draines correspondantes.
La classification des sols l'issue des tapes 1 et 2 de la
mthode par analogie est reprsente sur la figure 3 pour l'un
des six sondages. Des sols sableux sensibles apparaissent entre
4 et 6,5 m de profondeur et entre 7,5 et 9,5 m, puis en des points
isols plus en profondeur.
613
Figure 3. Classification des sols d'aprs les donnes d'un sondage au
pizocne et l'issue des tapes 1 et 2 de la mthode de classification
par analogie l'ess
iaxiales mesures au laboratoire sur les
chniques parmi lesquelles figurent des
sondages au pizocne.
dans le site de Soccer Field Glck
(D'apr 4).
ai triaxial.
D'aprs la description des sols tablie l'issue des carottages,
les sols argileux reprsentent 16,8 % de l'ensemble des
chantillons, les sols limoneux 8,5 % et les sols sableux 74,7 %,
pour un linaire effectif de 26,5 m de carottage. Ces
pourcentages sont comparer aux proportions des trois familles
de sols fournies par les six pizocnes, qui sont : 9,4 % pour les
sols argileux, 12,1 % pour les sols limoneux et 78,5 % pour les
sols sableux. Une bonne correspondance apparat ainsi entre les
carottages et les sondages au pizocne interprts en rfrence
aux caractristiques tr
chantillons carotts.
3.2 Exemple 2
Cetin et al. (2004) proposent une revue de diffrents cas de
rupture de pentes qui se sont produites sur le rivage de la baie
d'Izmit lors du sisme de Kocaeli du 17 aot 1999 en Turquie.
Plusieurs sites de cette rgion ont t explors en 2000 au
moyen de diffrentes te
ableau 2. Nature des sols T
s Cetin et al., 200
profondeur (m) sol
0 0,5 ou 1,5 remblai d'argile silteuse brune
0,5 ou 1,5 2,5 sable silteux et silts
2,5 16,5 argile silteuse molle trs plastique
Les donnes numriques de quelques sondages au pizocne
ont t importes directement d'une base de donnes
(http://peer.berkeley.edu). Les mesures du sondage SF5 ralis
dans le site de Soccer Field Glck sont utilises ici titre
d'illustration. Le tableau 2 indique le profil des terrains donn
par Cetin et al. (2004) jusqu' 16,5 m de profondeur.
Les sols rencontrs sont des sdiments fins rcents de nature
varie. Des sols sableux apparaissent dans tous les sites
diffrentes profondeurs. Mais des sols silteux et argileux
apparaissent aussi et souvent prs de la surface. La frange
superficielle comprend un deux mtres de remblais, qui sont
ouvent dtects comme sensibles.
et 2 de la mthode de classification par analogie l'essai
iaxial.
re 4 pour le sondage au pizocne
SF e 0 et 2,6 m de profondeur.
sensibles entre 1 et 2 m de
profondeur.
s
Figure 4. Classification des sols d'aprs les donnes du sondage au
pizocne SF5, site de Soccer Field, Glck (Turquie), l'issue des
tapes 1
tr
Figure 5. Extrait du profil de la figu
5, entr
La classification des sols du sondage SF5 l'issue des tapes
1 et 2 de la mthode par analogie l'essai triaxial est
reprsente sur la figure 4. Les rsistances triaxiales draines et
non draines du tableau 1 ont t adoptes pour effectuer cette
interprtation des donnes. Il apparat des sols sableux sensibles
entre 0,8 et 2,5 m de profondeur, des sols limoneux et argileux
jusqu' 16,5 m avec des pressions u
2
positives, puis, au-del, des
sols sableux rsistants et non sensibles. La figure 5 montre un
extrait du profil pntromtrique du sondage SF5 en surface et
marqu par des sols sableux
4 VOLUTIONS AVEC LA PROFONDEUR
Bon nombre d'enregistrements pntromtriques obtenus dans
des argiles molles et publis dans la littrature font apparatre
une augmentation quasi-linaire des mesures avec la
profondeur. Des droites reprsentent l'volution moyenne de la
rsistance q
t
et de la pression u
2
en fonction de la profondeur (et
614
th
parfois du frottement latral unitaire f
s
). De tels profils
apparaissent plus rarement dans les dpts sableux, dont la
lith
t l'volution de la
cohsion non draine avec z et la relation donnant l'volution de
la icale effective ' avec z :
= 15 c
u
(
= ( -
w
) z
so
avec le poids volumique du sol et le poids volumique de
l'e = k
3
l v n
e
ort des coefficients de
co pressibilit C
c
et de gonflement C
s
, 1 C
s
/C
c
), les
rel apparaissent :
= 12,2 c + p (4)
coefficient des terres au repos K = '
h0
/'
v0
= 0,5, p
0
s'crit
p
0
+
0
. m
3
et
w
= 10 kN/m
3
:
= 47 z (q en kPa et z en m ) (5)
T en ntr les
Exe le q
t
/z (kPa/ f
s
/z (kP u
2
/z (kPa/
ologie et l'tat initial sont caractriss gnralement par une
forte variabilit spatiale.
L'exprience acquise dans diffrents sites en France
confirme ces tendances. Les gradients d'augmentation de q
t
avec
la profondeur z varient autour 30 50 dans les argiles, ceux de
u
2
, autour de 25 40. Pour la rsistance de pointe, ce rsultat
s'explique par la relation empirique donnan
contrainte vert
v0
c
u
= '
v0
/ 3
q
t
'
v0
1)
it :
q
t
= 5 ( -
w
) z (2)
w
au. Avec 18 N/m
3
et
w
= 10 kN/m , i ie t :
q
t
= 40 z (q
t
en kPa et z en m) (3)
La formulation base sur le modle Cam-Clay de Chang et
al. (2001) aboutit des rsultats semblables. En admettant par
exemple que M
c
= 1,2 pour le critre de Coulomb (' = 30 ), un
degr de surconsolidation OCR = 1 (argile normalement
consolide), un indice de rigidit I
r
= 100 pour un sol mou
(rapport entre l module lastique de cisaillement G et la
cohsion non draine c
u
, I
r
= G / c
u
) et le rapport de
compressibilit = 0,9 (rapp
m
ations suivantes
c
u
= 0,322 '
v0
q
t u 0
u
2
= 5,68 c
u
+ p
0
o p
0
est la pression moyenne effective (p
0
= ('
v0
+ 2'
h0
)/3,
'
h0
contrainte effective horizontale). Puis, en admettant un
0
= 2 '
v0
/ 3 u Enfin, avec = 18 kN/
c
u
= 2,5 z (c
u
en kPa et z en m)
q
t t
u
2
= 30 z (u
2
en kPa et z en m)
ableau 3. P tes des profils p omtriques dans sols argileux.
mp m) a/m) m)
1 31 0,35 37
2 38 0,55 31
Ces relations donnent des ordres de grandeur des gradients
compatibles avec l'observation. Le tableau 3 indique les pentes
values dans les niveaux argileux des profils pntromtriques
nce aux rsistances draines et non draines mesures
pr
ndages
ca
argileux homognes et rcents, l'exprience fait
apparatre des gradients d'volution des mesures q
t
et u
2
avec la
nt tre encadrs par des ordres de grandeur
rptitifs.
Cet
Rob
erratrice J.F. Identification des sols argileux, limoneux et sableux du
plateau deltaque du Var partir de sondages au pizocne, Soumis
au Bulletin des Laboratoires des Ponts et Chausses.
des figures 3 et 4.
5 CONCLUSION
Une mthode a t propose d'identification des sols partir des
donnes mesures au pizocne. Cette mthode procde en
deux tapes : tape 1, classification des sols ; tape 2,
identification des sols sensibles. L'exploitation des donnes fait
rfre
alablement l'appareil triaxial sur les sols carotts dans le
site.
La mthode propose s'accorde aux mthodes en usage en
matire de reconnaissances pntromtriques, en comprenant
une tape d'identification des sols (profiling), puis une tape de
recherche des sols sensibles. Cette mthode cherche tirer parti
des essais triaxiaux pour interprter les donnes
pntromtriques, ce qui suppose que les reconnaissances
gotechniques prvoient la fois la ralisation de so
rotts et de sondages pntromtriques, pour aboutir une
analyse ddie du site. Elle ne vise pas revtir un caractre
universel, en utilisant une classification unique des sols.
Concernant les enregistrements au pizocne recueillis dans
des dpts
profondeur qui peuve
6 RFRENCES
in O.K., Youd T.L., Seed R.B., Bray J.D., Stewart J.P., Durgunoglu
T., Lettis W., Yilmaz M.T. (2004) Liquefaction-induced lateral
spreading at Izmit Bay during the Kocaeli (Izmit)-Turkey
earthquake. J. of Geotech. and Geoenv. Engng., vol. 130, n 12,
1300-1313.
Chang M.F., Teh C.I., Cao L.F. (2001) Undrained cavity expansion in
modified Cam clay: II Application to the interpretation of the
pizocne test. Gotechnique, vol. 51, n 4, pp. 335-350.
Fellenius B.H., Elsami A. (2000) Soil profile interpreted from CPTu
data. "Year 2000 Geotechnics", Geotech. Engng. Conf., Asian
Institute of Technology, Bangkok, Thailand, 27-30 novembre 2000,
18 p.
Parez L., Fauriel R. (1988) Le pizocne. Amliorations apportes la
reconnaissance des sols. Revue Franaise de Gotechnique, n 44,
pp. 13-27.
ertson P.K., Campanella R.G Rob ., Gillepsie D., Greig J. (1986) Use of
piezometer cone data. Use on in situ tests in geotechnical
engineering; Proc ASCE Speciality Conference In Situ '86,
Blacksburg, pp. 1263-1280.
ertson P.K. (1990) Soil classification using the cone penetration
test. Canadian Geotech. J., vol. 27, n 1, pp. 151-158.
neider J.A., Randolph M.F., Mayne P.W., Ramsay N. Sch R. (2008)
Analysis of factors influencing soil classification using normalized
piezocone tip resistance and pore pressure parameters. J. Geotech.
and Geoenv. Engng., vol. 134, n 11, pp. 1569-1586.
Senneset K., Jambu N., Svano G. (1982) Strength and deformation from
cone penetration tests. Proc. 2nd Euro. Symp. on Penetration
Testing, ESOPT-2, Amsterdam, vol. 2, pp. 863-870.
neset K., Jambu N. (1984). Shear strength parameters obta Sen ined from
static cone penetration tests. Proc. on Strength Testing on Marine
Sediments. Laboratory and In-situ Measurements. ASTM Special
Technical Publication 883, Symp. San Diego, pp. 41-54.
S
615
th
1
Correction of soil design parameters for the calculation of the foundation based
on the results of barrettes static load test
Correction des paramtres de conception du sol pour le calcul sur la base des rsultats de test
de barrettes de charge statique
Shulyatiev O., Dzagov A., Bokov I., Shuliatev S.
Gersevanov institute for soil bases and underground structures, e-mail:niiosp35@yandex.ru.
ABSTRACT: Geotechnical investigations and design works were being performed in 2008-2010 for the construction of Okhta-center
high-rise tower in St. Petersburg. Unique static load tests of 80 m deep barrettes were performed in 2010 as part of design process. 3
barrettes were tested simultaneously as a group and two were tested individually. The tests were planned in such a way as to get the
standard values of bearing capacity of barrettes, and to clarify the parameters of soil needed for the calculation of the barrette
foundation.The comparison of the bearing capacity values with the values calculated using Russian and foreign building codes is
provided. Limitations of currently used codes are shown.
RSUM : Les tudes gotechniques et les travaux de conception ont t ralises en 2008-2010 pour la construction du centre Okhta
haute tour Saint-Ptersbourg. Uniques essais de charge statique de 80 m de profondeur barrettes ont t ralises en 2010 dans le
cadre du processus de conception. 3 barrettes ont t tests simultanment en tant que groupe et deux ont t tests individuellement.
Les tests ont t planifis de manire obtenir les valeurs standard de la capacit portante des barrettes, et de prciser les paramtres
du sol ncessaires pour le calcul de la fondation barrette. La comparaison des valeurs de capacit portante avec les valeurs calcules
l'aide des codes de construction russes et trangres est fournie. Limites des codes actuellement utiliss sont affichs.
KEYWORDS:piles, barrettes, static load test, shaft friction, FEM, bearing capacity, high-rise building.

1 INTRODUCTION
In recent decades, in Russia there is a steady increase in the
number of tall buildings being built, of which a substantial part
is the building higher than 150 m.
Building heigher then 150m need a special approach to
design. Existing bulding codes in Russia and other countries as
well, can not fulfill the requirements of modern day high-rise
construction. For foundation constructions existing codes are
limited by relatively small depth of ground investigation and
testing loads.
In the current RF building codes plate loading test is
considered as the reference method for soil Young modulus
estimation. According to codes soil modulus determined by
other methods should be adjusted to plate loading test modulus.
It is not always possible, given the great depth of the soil used
as the bearing layer of high-rise building foundation.
This paper discusses the engineering properties of Vendian
clay as a bearing layer of Okhta tower high-rise building in St.
Petersburg. According to building design its pile foundation
will be embedded in Vendian clay layer lying deeper then 45 m
from ground surface.
Building design The project has a device for high-rise
building with pile foundation bearing on Vendian clay layer,
lying with a mark of -45 m B.S.V.
Laboratory tests on odometer and triaxial schemes were
made during ground investigations to study the properties of
Vendian clays. Given the depth of bearing layer pressuremeter
test were selected as in-situ test method.
Laboratory testing of soil extracted from great depth usually
complicated by disturbanceof soil samples, caused by stress
relief and preparation of samples for testing, and by the
complexity of high-precision measurements of deformation of
the sample (especially true for high stiffness soils).
Pressuremeter test, in turn, has no alternatives for soil testing
in-situ at greater depths. Design value of pressuremeter Young
modulus needs to be adjusted to plate loading test modulus, and
if that is not possible soil anisotropy factor needs to be
determined for conversion of soil modulus in the horizontal
direction to the modulus in the vertical direction.
Trial Barrette static test was scheduled as part of the pile
foundation design process. Given the high testing load
"Osterberg" method were considered economically effective.
Given specifics of the method, in addition to pile bearing
capacity assesment, one can provide design engineer with the
possibility of making deep plate loading test.
Figure1 .Location of test barretes.
Trial Barrettes were made from the surface of the soil. The
working part of a 65-m barrette was made of B40 grade
reinforced concrete. Barrettes were constructed by the
conventional technology - in the trenches, excavated under the
protection of bentonite slurry.
616
th
th
2
2 GEOLOGICAL CONDITIONS
Geological conditions of the construction site can be
described as unfavorable for civil engineering and especially for
high-rise buildings. Weak water-saturated soils lies to the depth
of 30 ... 35 m. Underneath these soils is a layer of moraine
deposits of small thickness. From a depth of 45 m liesVendian
clay. Rock, commonly used as a bearing stratum for high-rise
buildings are located at depths of over 200 m. Considering the
aforementioned facts Vendian clay was selected as the bearing
layer for the Okhta tower pile foundation. Vendian clay is
relatively strong soil and classified as hard clay and weak rock
at the same time. Despite the relatively high strength properties
Vendian clay exhibits long-term development of deformations
in time under load. It should be noted that engineering
properties of these soils in Saint-Petersburg is mostly unstudied.
3 TESTING SETUP
The test program was design in such a way as to achieve the
following goals:Determination of the bearing capacity of
barrette and its individual fragments;Determination of load-
settlement characteristics for "top-down" loading
scheme;evaluation of the Young modulus for the underlying the
barrette base;Evaluation of interface strength on the shaft of the
barrette.
Three of the five tested group barrettes were equipped with
loading device installed in two levels, two single piles in one
level. Single-level and two-level testing scheme and barrette
part nameingare shown in Fig. 2
Figure2. Scheme of barrete parts namings
Barrettes with one level jacks were tested in one phase, the
pile with two levels - in two stages. In the first stage the lower
part of the pile is loaded with the lower level of the jack. In the
second stage the upper jack level creates load on the middle part
of barrette. During a first stage of testing upper level jacks are
closed and load transfer through them is not different from a
solid barrette section. During loading of the upper level, the
hydraulic system of lower jacks is open into atmosphere,
making them closing or opening freely. During the testing of the
upper level when the lower level is open and jacks are retracted,
the entire load of the upper level of loading is transmitted to the
shaft of the barrette middle fragment. As the criteria for test
advancement standard RF deformation stabilization criteria 0.1
mm/h was used. It is 2.5 times more rigorous then the standard
European 0.25 mm/h. Results comparison with two different
stabilization criteria showed, that application of the criterion of
0.25 mm/h underestimates the magnitude of barrete base
displacement by 30%. Choice of stabilization criteria is
especially important when the testing jack located near the
barrette base in clay soils, as in this case, due to soil
consolidation deformation process is much slower.
4 TEST RESULTS
4.1 Test of the lower parts of barrettes B1...B5.
The value of the load reached during first stage of testing
was 40MN and 48 MN for second stage.
Bearing capacity of the lower parts of the test group barrette
was 90% of the bearing capacity for a single barretes test, due to
the group effect.
Load-settlement characteristics for barrettes B1B5 shown
on Figure 3 and shows that settlement of Barrette B2, located
between Barrette B1 and B3 is 15% more than that of Barrette
B1. This effect is referred to well-known concepts of group
effect in pile groups. Pile settlement in the group always
exceeds the settlement of single pile, and the settlement of
central pile is highest. With the growth of the number of piles in
the group this effect expected to increase. By means of
mathematical modeling of group testing and achieving the same
group effect, which was observed in the trial, one can confirm
the accuracy of the model input parameters, and to validate its
use for the calculation of the entire foundation.
The elastic component of the Barrette B1 ... B5 base
settlement is 13 ... 20%, and the residual inelastic component
reaches 79 ... 87.6% (Fig. 3), i.e. much of the ground under the
base of Barrette undergone plastic deformation.
In the analysis of Fig. 3 it may be seen that load-settlement
characteristics can be divided into several stages. In the first
phase, with a load values up to 5MN, load held by the shaft
friction on the surface of Barrette part, and movement up to 1
mm recorded. At the 2nd stage of loading barrete part is moved
and load being transferred to barrete base. Soil underneath the
barrete disturbed by drilling began to compact under load.
Settlement of barrete base increases linearly with load until
2040 MN load value is reached. As the barrette part
movement increases, shaft frictions on its side reaches a
maximum value and remain constant to the rest of stage 1. Due
to this effect further increment of load transferred directly to the
barrette base. The final stage is characterized by an increase of
settlement increment per unit increment of load, indicating that
the transition of the ground under the base of barrette to the
plastic state.
Figure 3. Stage 1 test results.
In order to clarify the shaft bearing capacity for bottom part
of Barrette B4, the loading increments in the first stages of the
load testing has been reduced from 5MN to 2.5 MN, which led
to an increase in the number of stages in the load range of up to
20 MN from 4 to 8 . An interesting finding was the fact that,
regardless of the number of stages loading time spent on testing
barrette B4 and B5, was similar and was 277 and 259 hours,
respectively.
Concluding the analysis of bottom level testing one can
mention high repeatability of results, which indicates the
617
th
3
homogeneity of the subsoil under the Barrette base and the good
quality of their production.
4.2 Test of the middle parts of barrettes B1...B5.
At the end of the first stage of testing bottom level jacks are
retracted, and their hydraulic system is open. In this
configuration, the lower level of jacks do not transfer the load
from a upper jacks level on the base. In this case, loading of the
upper jack level resisted only by shaft resistance of the middle
barrette part, allowing to accurately determine shaft friction
value.
Figure 4. Stage 2 test results.
Compared with the test of the first stage (lower jack level)
results show significantly greater variation in the ultimate shaft
friction values. Load-settlement curves for the second-stage
tests are shown in Fig. 4.
Load-settlement curves shown on Fig. 3 characterized by
initial almost flat part, with a slightly longer length for Barrette
B2. The angle of the load-settlement curves for B1 and B3
began to increase after the load value of 20 ... 25 MN, and for
B3 35 MN.
The presence of a longer horizontal part on load-settlement
curve for B2 may be due to heterogeneity of soil conditions
along the tested barrette part, or, more likely, due to its central
position in the group.
4.3 Back-analysis of test results performed by FEM.
One of the most effective tools for the analysis of load test
made by the Osterberg scheme is the reverse calculation method
with regard to elastic-plastic soil properties by means of FEM.
The reverse calculation has several objectives: 1) Calibration
of design parameters of adopted soil model 2) evaluation of the
bearing capacity of single pile in the top down loading
conditions 3) assessment of the applicability and adequacy of
the chosen soil model.
The starting point for the reverse calculations is the soil
properties obtained by laboratory testing. By varying individual
soil model parameters one can identify the most important of
them, and then achieve convergence between experimental and
calculated results.
The first iteration of calculation based on laboratory
determined soil properties showed that the calculated values of
barrete upward movement is 6 times larger than the
experimental values, and downward movement is overestimated
by 2 times. This suggests that the characteristics of soils,
provided through laboratory testing are very different from the
characteristics of the soil in-situ.
Taking into account observed discrepancy the objective was
to find such soil characteristics, which would have shown the
best convergence of calculation with the experiment. Barrette
movement and stress along its body were chosen as
convergence criteria between the experimental and calculated
values.
Figure 5. Comparative charts of data obtained from the
experiment and the results of the calculation.
Comparison of modeling results with the experimental data
shown in Figure 4. The graph shows the results of calculations
with adjusted characteristics of soil is almost equal to the results
of the test. It should be noted that non-compliance of Barrette
behavior during unloading caused by incorrect rapid unloading
procedure.

Table 1.Adjusted by FEM calculations soil properties.
GE
number
Soil
classification
Density,
g/cm
3

,
MPa
Poisson ratio

,

,
kPa
11 Hard clay 21.3 50 0.25 17 150
12 Hard clay 22.2 200 0.2 25 330
12 Hard clay 21.1 105 0.22 18 200
13 Hard clay 22.3 252 0.18 27 491

5 COMPARISON OF OBTAINED SHAFT UNIT
FRICTION WITH BASIC CALCULATION METHOD
RESULTS.
The main purpose of the second phase of the test was to
determine the specific shaft resistance values for middle parts of
Barrette B1 ... B3. Resistance value is determined by dividing
the applied load on the shaft surface area of the middle part of
barrette.
Figure 6. Shaft friction-movement curves.
As can be seen from Fig. 5, for barrette movements of up to
20 mm shaft friction increasing drastically to 190 ... 290 kPa. A
further increase in displacement to 60 mm results in a small
(about 60 kPa) monotonic increment of resistance. The peak
(maximum) value of the shaft resistance was not clearly
618
th
th
4
observed as the resistance monotonically increases up to 100
mm movement. By analyzing the obtained movements-
resistance curves it can be noted that values of shaft resistance
for Vendian clays can be taken as corresponding barrette shaft
resistance for 40 mm (according to RF building codes)
movement.
Design shaft resistance values provided for assessment of
pile bearing capacity by SP24.13330.2011 at depths greater than
40 m are assumed constant and equal to 50 kPa, which is 4.6 ...
6.4 times less than the values obtained in static Barrette test. For
comparison, ultimate values of shat friction were also calculated
for the most common approaches worldwide: the -method
(Skempton 1959) using undrained strength parameters and
Mohr-Coulomb law. In calculation with the alpha method lower
(=0.3) and upper bound (=0.6) values of alpha were
examined. Resulting specific values of shaft friction values
were 250 and 500 kPa respectively. For calculation by Mohr-
Columb earth pressure at rest coefficient K
0
were taken for non-
consolidated soil by well-known Jaky equation and as for
overconsolidated soil with OCR=2,5. Resulting K
0
values were
0.66 and 0.99 respectively. Factor of 0.5 for interface strength
also applied according to SP24.13330.2011. Specific shaft
friction values obtained by this calculation method were 270
kPa for K
0
=0.66 and 460 kPa for K
0
=0.99.
Thus, the lower limit of the specific shaft friction, calculated
using the mechanical properties of soils were within the values
obtained by the results of static Barrette tests, and the upper
limit value was higher on average of 1.7 times. One of the
reasons for this discrepancy may be that the central parts of
barrette during the first stage of test (lower part testing) has the
5...7 mm upward movement, during which partial mobilization
of shaft friction forces in opposite direction were observed.

6 CONCLUSIONS
Trial works confirmed the technical feasibility of barrettes
construction with cross-section of 1.5 x 3.0 m length of 85 m by
the standard "slurry wall" technology in difficult sub-soil
conditions of St. Petersburg.
It is advised to implement Osterberg testing technique (by
cast in pile submerged jack) for the deep foundation of high-rise
buildings. The method allows to use pile parts as an anchor
system and to clearly determine the values of unit shaft friction
and base resistance. It is recommended to install two levels of
jacks in a pile: one near the base of the pile, and the second in
the middle of the main bearing layer. It is critical to install
several levels of strain gauges in the pile along its length.
Pile testing at construction site should be seen not only as
method to determine pile bearing capacity but as an effective
method to calibrate design parameters of adopted soil model,
and to assess its applicability and adequacy. Soil parameters
provided by ground investigation can by checked and adjusted
if necessary. During design process of tower foundation,
obtained results of unit shaft friction and base resistance should
be used as the control values, against which the calculation
results are checked.
The calculations made on the basis of experimental data
showed what the values of the mechanical properties of soils
determined by the laboratory testing has severely
underestimated soil strength and deformability parameters due
to sample disturbance, the influence of the scale factor & etc.
As a result of the tests it was found that the Vendian clays
can provide high values of shaft friction and base resistance.
The experimental values of shaft friction and base resistance
exceed the ultimate values provided by codes by 4 ... 6 and 1.6
timesrespectively.

7 REFERENCES
Aspects of pile testing for high-rise buildings on the example of ODTS
Okhta tower.High-rise buildings journal 6 (2012).96-
99.PetrukhinV.P., ShuliatevO.A., BokovI.A., ShuliatevS.O.
Cast-in-situ Bored Piles in London Clay, Geotech,Vol.9.Skempton,
A.W., and Northey, R.D. (1952)
619
Characterization and Settlement Modeling of Deep Inert Debris Fills
Caractrisation et modlisation du tassement de dpts pais de gravats inertes
Somasundaram S., Khilnani K., Shenthan T.
Advanced Earth Sciences, Irvine, California, USA
Irvine J.
Irvine Geotechnical, Pasadena, California, USA
ABSTRACT: Inert debris fills are difficult to characterize and model by normal geotechnical methods, due to their inherent
heterogeneity, very large particle size, and nested and voided structure. The approach taken to characterize a 54 m deep inert debris
fill, model its settlement behavior under seismic loading and groundwater level rise, and develop remedial measures to render it
suitable for development is presented. Fines migration into open cavities and collapse of nested structure were determined to be the
primary settlement mechanisms for this material. An upper bound estimate of cavity volume vulnerable to fines migration and
collapse was made based on the results of large scale in-situ density and gradation tests. Settlement was estimated for various
percentages of cavities becoming filled, and compared to case histories of dry fill settlement from the San Fernando and Northridge
earthquakes. The proposed remedy involved partial removal of the debris fill and replacement as a compacted fill cap to attenuate the
surface expression of differential settlement occurring in the underlying debris fill. Surface manifestation of settlement was simulated
using FLAC. Charts were developed relating cap thickness to surficial manifestation of differential settlement.
RSUM : Les dpts de gravats inertes sont difficiles caractriser et modliser par les approches gotechniques usuelles, en
raison de leur htrognit intrinsque, de la grande taille des particules qui les constituent, et de leur structure lacunaire et embote.
On prsente une approche utilise pour caractriser un dpt de gravats inertes de 54 m d'paisseur, modliser son comportement de
tassement sous chargement sismique et sous l'effet d'une monte du niveau de la nappe phratique, et dvelopper des mesures de
remdiation en vue de le rendre propre l'utilisation. On a pu montrer que la migration des fines dans les cavits ouvertes, et
l'crasement des structures embotes, constituent les mcanismes principaux responsables du tassement pour ce matriau. Une
estimation par excs du volume des cavits vulnrables par la migration des fines et crasement a t tablie sur la base d'essais
grande chelle de densit in-situ et de granulomtrie. Le tassement a t estim pour divers proportions de remplissage de cavits, et
compar des observations historiques de tassement de remblais secs suite aux sismes de San Fernando et de Nothridge. Le remde
propos implique un retrait partiel du dpt de gravats et son remplacement par une couche de remblai compact, en vue de minimiser
l'expression en surface des tassements diffrentiels survenant dans le dpt de gravats sous-jacent. Le dplacement en surface a t
simul en utilisant le logiciel FLAC. La relation entre l'paisseur de la couche de protection et l'incidence en surface du tassement
diffrentiel a t exprime sous forme d'abaques.
KEYWORDS: inert debris landfills; debris fills; seismic settlemen

1 INTRODUCTION
Inert debris landfills in urban areas are increasingly becoming
potential sites for industrial / commercial redevelopment due to
scarcity of vacant land and a desire by local communities to turn
blighted areas into revenue sources. These fills, generally placed
in abandoned mine pits, could be over 50 m deep and typically
consist of uncontrolled fills of construction and demolition
(C&D) debris. Due to their inherent heterogeneity and very
large particle size they are difficult to characterize and model by
normal geotechnical methods. This case study presents the
approach taken to characterize a deep inert debris fill, model its
settlement behavior under seismic loading and groundwater
level fluctuations, and develop remedial measures to render it
suitable for development.
The inert debris fill, located in the City of Irwindale in
southern California, consists of over 8 million cubic meters of
C&D waste placed over a period of 15 years within a 54 m deep
abandoned open pit gravel mine covering a footprint of 22
hectares. The lower 2 to 12 m of the pit was filled with
hydraulically placed silt, a by-product of aggregate mining
operations. Review of placement records indicated that the inert
debris fill above the silt layer consists of a succession of 1 to 3
m thick lifts of rubble consisting mostly of broken concrete,
brick, tile and asphalt capped with 15 to 30 cm thick lifts of
sandy and silty soils. The soil layers were generally placed and
compacted above each rubble lift to provide a suitable surface
for rubber tired traffic. The entire inert debris fill is capped with
a 3 m thick layer of compacted soil to allow for utility
excavation and structure foundation at the finished surface.
Placement records indicate that initially the rubble fills were
placed with some degree of material processing (crushing of
oversize concrete clasts) and compaction. However, much of
the inert debris fill was loosely end dumped with little or no
control of lift thickness, particle size or compaction. The
groundwater level was approximately 36 m below the ground
surface during filling, but could rise by about 12 m based on
historic records. An idealized profile of the fill stratigraphy is
shown in Figure 1.

1 to 3 m
Potential highest
groundwater level
Previous highest
groundwater level in fill
150 to 300 mm
(24 m below ground surface)
(36 m below ground surface)
Compacted Fill Cap
(3 m)
Rubble Fill
(Thickness varies)
Silt Deposit
(Thickness varies)

Figure 1. Debris Fill Stratigraphy
620
th
The site is vulnerable to relatively high levels of seismic
loading, with a design peak ground acceleration of
approximately 0.53g per the building code. Deaggregation
analysis indicated the corresponding moment magnitude to be
6.7. The area is zoned for industrial or commercial
development. The owners are evaluating remedial measures to
make the site suitable for building development.
There are no industry-accepted standards or case histories to
predict settlements of inert debris fill containing significant
oversize fragments and significant open cavities. Case histories
of seismic settlements of unsaturated fills are generally limited
to earthfill/rockfill dams and compacted soil fills. Laboratory
cyclic simple shear test data relating cyclic shear strain to
volumetric strain, that may be used to estimate the settlement of
unsaturated fills under seismic shaking, are limited to sands
(Silver and Seed 1971, Pyke et al 1975), and finer grained
compacted fills (Stewart et al 2002). Charles (2008) documents
case histories of long-term settlement and collapse potential of
uncontrolled opencast mining backfills in Britain. The City of
Irwindale is currently conducting a laboratory study to evaluate
the potential for wetting induced settlements (hydrocollapse) in
inert debris fills.
2 FIELD INVESTIGATIONS
Field investigations for this site included Becker hammer
borings, surface and downhole geophysical surveys, downhole
video logging, test excavations and large scale in-situ density
and grain size distribution tests. Neither the Becker penetration
tests (BPTs) nor the surface and downhole seismic surveys,
proved to be suitable to characterize the heavily nested and
voided nature of the fills. The presence of very large size
fragments appear to significantly skew the measured Becker
blow counts and shear wave velocities, making these methods
incapable of adequately differentiating between well
compacted, grading code - compliant fills (derived from the
same debris materials), and the loose debris fills with
voids/cavities. This conclusion has been confirmed by studies
performed by the City of Irwindale at other debris fill sites
(Geomatrix, 2007).
Mapping of two deep test excavations to 21 m depth in the
poorly controlled debris fill, confirmed the layered filling
pattern consisting of thick rubble fill lifts capped by thin soil
layers. The layered filling pattern was also apparent in the BPT
logs. The rubble fill consisted of concrete clasts and blocks up
to 2 m in size (with abundant rebar), mixed with brick, tiles,
asphalt concrete, crushed glass and variable amounts of soil
infill. Large voids, cavities and nesting were very common.
Eight large diameter ring density tests (1.8 m diameter x 1.5
m deep) performed as per ASTM D5030 in the inert debris fill
at various depths (ranging from 5 to 15 m below ground
surface) in the test excavations, and eight sand cone tests
performed on soil layers or soil rich fills gave the following
results.

Table 1. Results of In-situ Density Tests

In-situ Dry Density (gm/cc) Material
Range Average
Average Void
Ratio
Inert Debris Fill 1.22 2.03 1.77 0.43 (e
t
)
Soil Layers 1.45 1.86 1.64 0.62 (e
s
)

The in-situ densities of the inert debris fill were compared
to field maximum achievable density (MAD) tests performed on
inert debris materials placed in 30-cm thick lifts and compacted
by 50 passes of heavy earthmoving equipment (combination of
Caterpillar 820 front end loader and 825 compactor). The
corresponding MAD dry densities ranged from 2.03 to 2.13
gm/cc.
A qualitative evaluation of the voided / nested structure of
the inert debris fill was performed by measuring the rate of
water percolation in large diameter test holes. After completing
the large diameter in-situ density tests, the plastic sheeting used
to line the test hole was pulled out and the water level drop was
monitored. The water levels dropped very rapidly (emptied in a
matter of minutes) in test holes in debris fills, while the water
levels stayed full for several days in the MAD tests holes,
confirming the presence of significant voids / cavities in the
debris fill.
Field bulk gradation tests performed on the bulk samples
excavated from the density test pits showed the following
distribution:

Table 2. Summary of Field Gradation Test Results

Material Size Range (%) Average (%)
Boulders (>300 mm) 3 to 23 11
Cobbles (>75 mm) 10 to 25 18
Gravels (>19 mm) 6 to 20 14
Finer than (19 mm) 44 to 66 57

Visual observations of the materials removed from the test
excavations suggest that the oversize fraction is greater than the
amounts listed above, since representative amounts of very
large concrete clasts could not be included in the material from
1.5 m diameter test holes. The actual boulder size fraction (>
300 mm) was estimated to be in excess of 20 percent by weight.

3 SETTLEMENT MODEL
The settlement model used in the analysis considered the
layered nature of the debris fill consisting of a succession of 1 to
3 m thick voided and nested rubble lifts capped by 15 to 30 cm
thick loose to medium dense soil lifts. The total debris fill may
be considered to consist of nested oversize clasts (defined as
materials lager than 19 mm for purposes of this analysis), infill
soils (minus 19 mm fraction that partially fills the cavities
between clasts and also caps individual layers of rubble), and
cavities (Figure 2).

Figure 2. Debris Fill Structure
When subjected to seismic loading and/or saturation due to
groundwater rise, the predominant mechanisms of settlement in
the debris fill are considered to be partial filling of the cavities
by fines migration (cap soils migrating into the underlying
nested rubble), and collapse of the nested structure. Volumetric
compression of the infill soils and soil lifts will also take place,
but they are considered to be significantly smaller than the two
dominant settlement mechanisms. The volume of cavities
between the nested clasts, as a percentage of the total volume of
fill, will, therefore, form an upper bound of the potential
volumetric strain / settlement of the fill. The volume of cavities
in the fill (Figure 2) as a ratio of the total fill volume, was
estimated as shown below, based on the void ratio of the entire
debris fill, e
t
(calculated from large diameter ring density tests),
void ratio of the infill soils, e
s
(calculated from the sand cone
density tests), the ratio of weight of clasts to weight of infill
621
soils, R (from particle size distribution tests), and the specific
gravity of the clasts (Gc) and infill soils (Gs).
The ratio of volume of cavities to volume of solids, e
c
, and
the relative volume of cavities with respect to the total volume
for debris fill (Pc) may be expressed as:
e
c
= e
t
e
s
/ (1+R.Gs/Gc) (1)
P
c
= e
c
/ (1+ e
t
) (2)
Based on the above equations, and using the average values
of e
t
(=0.43), e
s
(=0.62), R (43%/57% = 0.75) and specific
gravity (Gs = 2.65; Gc = 2.4), the average volume of cavities
within the poorly controlled debris fills was calculated at 6.6%
of fill volume. The calculated volume of cavities agrees well
with field experiment estimates of cavity volume made at other
inert debris fill sites in Irwindale with similar materials and
filling practices. Those evaluations included a controlled in-situ
pilot grouting test which resulted in a grout take of 4.4 to 7.2%
of total volume, and an in-situ dynamic compaction test which
resulted in a volume reduction of 5 to 7% of total fill volume
(AMEC, 2008).
However, not all of the calculated cavity volume is available
for fines migration / collapse. Actual volumetric strain and the
resulting settlement is proportional to the volume of cavities
that are closed or filled with fines in the event of an earthquake
or hydrocollapse caused by rise in groundwater level. This is a
function of many factors including the grain size distribution of
the oversize clasts, accessibility of cavities to overlying infill
soils, cohesion of infill soil and intensity and duration of
seismic shaking, and cannot be reliably estimated in the absence
of material-specific physical modeling. Therefore, a parametric
settlement evaluation considering various percentages (p) of
total cavity volume becoming filled was performed. The results
are summarized as average settlement versus depth plots (Figure
3). The settlements shown in Figure 3 for each value of p,
represent the average of the calculated settlements at six BPT
locations across the site. Although the total thickness of debris
fill was similar at each location (approximately 33 m), the
thickness of the poorly controlled, layered rubble fill vulnerable
to fines migration/collapse was variable (ranging from 15.6 to
25.0 m).
The average settlement corresponding to 20% of cavities
filled (p = 20%), was computed at 28 cm (approximately 1.32%
of poorly controlled debris fill thickness or 0.85% of total debris
fill thickness). The latter value compared favorably with some
case histories of dry compacted fills in southern California
which settled by 0.6 to 0.9 percent of fill thickness during the
M6.6, 1971 San Fernando, and the M6.7, 1994 Northridge
earthquakes, under ground accelerations comparable to the
design ground motions for the site. Considering the significant
heterogeneity of the debris fills, the seismic settlements could
be higher or lower than that predicted for p = 20%. To bracket
this uncertainty, seismic settlements under the design
earthquake were calculated for p ranging from 10% to 30%.
The resulting settlements ranged from 0.4 to 1.1 percent of total
debris fill thickness.
A 12 m thick zone of debris fill immediately above the
current groundwater level could become saturated if the
groundwater level was to rise to the historic high groundwater
level. This zone has not been saturated since the time of
placement. Settlement due to groundwater saturation was
considered to result from the same mechanisms of fines
migration and collapse, and was assumed to be of the same
order of magnitude as the seismic settlements. These
settlements, estimated to range from 75 mm to 150 mm, occur
approximately 24 m below ground surface (the depth of the high
groundwater level below ground surface). Because the same
mechanisms (migration of sands into open voids and collapse)
apply to both seismic settlement and settlement due to
groundwater rise, the two components of settlement (seismic
and hydrocollapse) are not considered to be cumulative.
0
5
10
15
20
25
30
0 100 200 300 400 500
D
e
p
t
h

(
m
)
Average Total Seismic Settlement (mm)
% of Cavities
Filled (p)
0%
10%
20%
30%
Figure 3 Distribution of Seismic Settlement with Depth
4 REMEDIAL MEASURES
The remedial measures recommended for limiting settlement at
the site to within agency-defined guidelines or structural
tolerances, consisted of partial removal of the existing debris fill
and replacement with a properly processed and compacted fill
cap. The required cap thickness could also be achieved by a
shallower removal and replacement combined with in-situ
ground improvement of the lower part of the debris fill by
dynamic compaction. With increasing thickness of cap, the fill
thickness left in place that is vulnerable to settlements would
decrease. The cap will also help attenuate the differential
settlement taking place at depth as it manifests at the surface of
the fill cap.
The surface manifestation of settlement was simulated by
numerical modeling using FLAC. A representative two-
dimensional cross section across the entire site was considered.
The fill cap was modeled as a non-linear elastic perfectly
plastic Mohr-Coulomb material. The initial shear modulus for
the cap was based on the average shear wave velocity of 268
m/sec measured in the compacted fill. The modulus
degradation curve was based on the Seed-Idriss relationship for
sand. The calculated seismic / hydrocollapse settlement of the
debris fill underlying the fill cap, was applied as nodal
displacement boundary conditions at the base of the cap. Since
the thickness of poorly controlled rubble fill and the
corresponding settlements are variable across the site, the nodal
displacements were specified as randomly varying over the
range of settlements calculated at the 6 BPT locations.
The nodal displacements (
n
) were generated as follows:
n
=
min
+ r. (
max
-
min
)
where, r is a random number between 0.0 and 1.0 (determined
by a random number generator for the numerical analyses) and
min
and
max
are the minimum and maximum values,
respectively, of calculated seismic/hydrocollapse settlements,
for a given value of p. The specified random nodal
displacements were applied at 1.5 m horizontal intervals along
the base of the cap. The modeling was performed for p = 10%,
20% and 30%.
Typical FLAC analysis results as illustrated in Figure 4,
show the original and deformed shape (grid) of a segment of the
fill cap as a result of the random differential settlement applied
at the base of the cap, for cap thicknesses of 12, 18 and 24 m,
respectively. As the fill cap thickness increases, the magnitude
of the total and differential settlement of the material left in
622
th
place decreases, and the attenuation of the surface manifestation
of differential settlement increases. For the case illustrated in
Figure 4, the differential settlement at the base of the cap
decreases from 122 mm to 43 mm as the cap thickness increases
from 12 m to 24 m. The corresponding maximum differential
settlement at the surface (over a 9-m horizontal distance)
decreases from 56 mm to 8 mm.

Figure 4. Sample Results from FLAC Analysis

The results of the surficial manifestation analyses, presented
as plots of surficial total and differential settlements versus
thickness of fill cap (for a range of assumed values of cavities
filled by migration of fines and collapse, p), are plotted in
Figure 5. This chart was used to select a suitable thickness of
removal and replacement based on the differential and total
settlement tolerance of the proposed structures.

5 CONCLUSIONS
The seismic and hydrocollapse settlement potential of
uncontrolled inert debris fills containing significant oversize
clasts could not be evaluated by conventional means.
Laboratory testing of representative material was not feasible
because of particle size limitations. BPTs and seismic shear
wave velocity surveys were ineffective in differentiating well
compacted fills from uncontrolled fills. An alternative approach
consisted of the following steps:
Based on the results of large scale in-situ density
and grain size distribution tests, an upper bound estimate of
cavity volume was made (approximately 6.6% of total
debris fill volume).
Figure 5. Surface Man station of Settlements
percent of total fill thickness with an average of
ake ground motions similar to
tabase of
observed settlements under moderate seismic events.
AM ce II
Cha
Geo tration tests, Reliance
Pyk
tional shaking. J .Geotech. Engrg. Div., ASCE, 101(4),
Silv
loading. J. Soil Mech. and Found. Div., ASCE, 97(9), 1171-
Stew
eotechnical
and Geoenvironmental Engineering, 127(11) 905-919
ife

A settlement model based on partial filling of cavities by
fines migration and collapse of nested structure was
developed. Parametric analyses of various degrees of
cavities filling were performed to account for heterogeneity
of the debris fill and to obtain a range of likely settlements.
Estimated settlements due to seismic shaking ranged from
0.4 to 1.1
0.85%.
The predicted settlements from this model were compared
to published case histories of seismic settlement of
unsaturated fills under earthqu
the design ground motions.
The proposed remedy for rendering the site suitable for
building development was partial removal of the uncontrolled
debris fill and replacement as a properly compacted fill cap.
Based on numerical modeling, charts were developed relating
thickness of fill cap to estimated surficial differential settlement.
To meet local building code requirement of maximum 25 mm
differential settlement over a 9-m length, 22 m of removal and
replacement will be necessary. The depth of removal and
replacement may be reduced, provided the differential
settlement tolerance of the structure is increased by structural
improvements such as stiffened foundation systems including
mat foundations, post tensioned slabs and grade beams. The
reliability of predictions by this approach may be increased by
physical modeling of debris fill settlement under the effects of
seismic shaking and saturation, and developing a da
9 m
(a) 12 m Compacted Soil Cap
75 m 90 m 120 m 105 m
Max.
diff
= 55 mm
Max.
diff
= 120 mm
Max.
diff
= 8 mm
Max.
diff
= 43 mm
Max.
diff
= 105 mm
Max.
diff
= 33 mm
(b) 18 m Compacted Soil Cap
(c) 24 m Compacted Soil Cap
diff
- Diffential settlement over horizontal distance of 9 m
- Original grid
- Deformed grid (distortion magnified by a factor of 20)

6 REFERENCES
EC Geomatrix, Inc., 2008. Closure geotechnical report, Relian
landfill improvements,Vulcan Materials Company, Irwindale.
rles J.A. 2008. The engineering behavior of fill materials: the use,
misuse and disuse of case histories. Gotechnique 58 (7), 541-570.
matrix, 2007. Documentation of Becker pene
landfill improvement, Azusa and Irwindale..
e R, Chan C.K. and Seed H.B. 1975. Settlement of sands under
multidirec
379-398
er M.L. and Seed H.B., 1971. Volume changes in sands during
cyclic

1182
art J.P., Bray J.D., McMahon D.J,. Smith P.M., and Kropp A.L.
2001. Seismic performance of hillside fills. Journal of G

623
Site Characterization for the HZM Immersed Tunnel
Caractrisation du site pour le tunnel immerg HZM
Steenfelt J.S., Yding S., Rosborg A.
COWI, Copenhagen, Denmark
Hansen J.G.
Ben C. Gerwick, COWI Group Company, Oakland, USA
Yu R.
COWI China, Beijing, Peoples Republic of China
ABSTRACT: The 36 km long HZM Link, crossing the Pearl River estuary between Hong Kong in the east and Macao and Zhuhai in
the west is rated one of the most important current infrastructure projects in China. It is slated for completion in 2016 and consists of a
world record length of 6 km immersed tunnel, two artificial transition islands and some 30 km bridges with a dual three lane motor-
way. In order to provide the structural designers with the requisite input for proper soil structure interaction analysis a very extensive
site characterisation was carried out comprising geotechnical boreholes, CPTUs and seismic testing with associated advanced labora-
tory testing. This paper describes the results and calibration of geotechnical boreholes, CPTUs and advanced laboratory tests to pro-
vide the requisite tool for inference of ground stratification and stiffness variation to be used in the structural modelling of the im-
mersed tunnel, the design of piles and dredging slopes.
RSUM : La liaison HZM de 36 km de long qui traverse lestuaire du fleuve Pearl entre Hong-Kong lest, Macao et Zhuhai
louest, est considr comme tant lun des plus importants projets dinfrastructure en Chine. Le projet qui doit tre achev en 2016
est compos dun tunnel immerg dune longueur record de 6 km, de deux les artificielles de transition et denviron 30 km de pont
autoroutier deux fois trois voies. Afin dobtenir les lments essentiels pour lanalyses de linteraction entre les fondations et les
structures, une campagne de sondages gotechniques trs dtaille a t mene comprenant des forages, des tests de pntration au
cne (CPTU) et des sondages sismiques ainsi que les tudes en laboratoire correspondantes. Cet article dcrit les rsultats obtenus et
mthodes de calibration des forages, CPTU et des essais en laboratoire mis en uvre afin dobtenir les lments de base ncessaire
pour la dtermination des caractristiques mcaniques des sols utiliser pour la modlisation des lments du tunnel immerg, la d-
finition des pieux de fondation et ltude des pentes de dragage.
KEYWORDS: Site characterization, immersed tunnel, CPTU, triaxial testing, undrained shear strength, settlements, spring stiffness.

1 INTRODUCTION
The Hong Kong-Zhuhai-Macao (HZM) Link crosses the Pearl
River Estuary in south-eastern China in the Guangdong prov-
ince connecting Hong Kong at Shek Wan, Lantau Island to the
Pearl at Macau and to the district of Gongbei, Zhuhai in
mainland China, see Figure 1.
Figure 1. Location of the HZM project in south-eastern China.

The link is 36 km in total length of which 6 km comprises the
immersed tunnel. The remainder consists of two artificial tran-
sition islands and low bridges some 30 km in total length.
The whole connection has the capacity of a dual three lane
highway.
Provisions for two possible future 570 m wide navigation
channels are planned along the immersed tunnel alignment with
proposed design dredging levels some 15-20 m below existing
seabed level.
The particular challenges for the design of the immersed
tunnel are:
the presence of very soft clays requiring extensive
dredging profiles and soil improvement,
very deep foundation level of the tunnel in order to al-
low for future navigation channels 570 m wide over
the central part of the tunnel,
up to 23 m sedimentation load over the central part of
the tunnel,
potential of differential settlements due to the highly
varying loading and ground stiffness conditions,
the need for mixed foundation solutions with end bear-
ing or settlement reducing piles near the artificial is-
lands and direct foundation for the central part.

In order to provide the structural designers with the requisite
input for proper soil structure interaction analysis for Detailed
Design, a very extensive site characterisation was required. The
scope and findings of this site characterisation are described in
this paper.
The Project Owner is the HZM Bridge Authority, and the de-
sign and construction is being undertaken by a Joint Venture
headed by the contractor China Communications and Construc-
tion Company (CCCC) Ltd.
624
th
Figure 2. Simplified geological model along the immersed tunnel alignment.
2 GEOLOGICAL CONDITIONS
The project area is located in the Pearl River drainage basin,
which historically has been shaped as a result of the uplift of the
Tibetan Plateau during the Tertiary and Quaternary Periods,
forming the present-day Pearl River Delta with its network sys-
tem and estuarine bays (see Figure 1). The river delta is one of
the most important and complex large-scale estuarine systems in
China.
The Holocene development of the delta has been controlled
and affected by the variations in the deposition of sediments,
sea-levels and groundwater levels.
The soil deposits in the present-day Pearl River delta over-
lying weathered basement rock can be traced back to the Late
Pleistocene and Holocene periods.
These deposits consist of three cycles of upward fining se-
quences of delta deposits, namely one Holocene and two Pleis-
tocene delta cycle, which have been divided by two previously
exposed and subsequently eroded surfaces.
Based on the described regional geology and the findings of
the site investigations carried out for the project, the soil depos-
its and rock formations encountered along the alignment of the
immersed tunnel, and in the locations of the artificial islands,
can be grouped into five main units for soil deposits, and two
main units for rock formations:
Marine deposits of clays and sands formed during the Holo-
cene period,
Continental deposits of clays and sand from a once exposed
surface formed during the late Pleistocene period,
Marine alluvial deposits of clays and sands formed during
the Mid to Late Pleistocene period,
Fluvial alluvial deposits of clays and sands formed during
the Early to Mid Pleistocene period,
Residual soils formed during the Early Pleistocene period,
Highly to completely migmatic schists formed during the
Sinian period,
Moderately to completely weathered migmatic granites
formed during the Sinian period.

A simplified geological model is shown in Figure 2.
3 SCOPE OF INVESTIGATIONS
Three geotechnical investigation campaigns have been carried
out for the project:
Feasibility Study investigations carried out in 2004 and
2008: Only 16 Nos. boreholes were carried out in the vicin-
ity of the immersed tunnel.
Preliminary Design investigations carried out in 2009:
151 Nos. boreholes were carried out for the artificial islands
and 115 Nos. boreholes, 29 Nos. CPTUs and seismic P-S
suspension logging (in 10 Nos. boreholes) was carried out
along the immersed tunnel alignment.

Supplementary Soil investigations were carried out in 2010-
2011: 80 Nos. boreholes, 364 Nos. CPTUs, 20 Nos.
CPTUDs and seismic P-S suspension logging (in 6 Nos.
boreholes) was carried out along the alignment of the im-
mersed tunnel and at the locations of the artificial islands.

The Supplementary Soil investigations formed the main basis
for Detailed Design, and the scope of and specifications for
these investigations were defined by COWI as being a member
of the design and construction Joint Venture. Site and laboratory
works were followed closely by means of inspections carried
out by COWI's geotechnical engineers, in order to ensure that
all works were carried out in accordance with applicable stan-
dards.
The boreholes for the Supplementary Soil investigations
were split into two types of boreholes: the GITB-series where
geotechnical in-situ testing was carried out and disturbed sam-
ples were retrieved, and the TCB-series that were used entirely
to retrieve undisturbed samples of fine grained soils. Most of
the boreholes were carried out in pairs, each pair consisting of
one GITB borehole and one TCB borehole, and as a general rule
the GITB and TCB boreholes were drilled within five meters of
each other, in order to produce mirror boreholes displaying
similar geological and geotechnical properties. The drilling
depths varied from 29 to 107 m below existing seabed level.
The general distance between boreholes (and borehole pairs)
was on average approx. 200 m in the longitudinal direction.
In general the CPTUs were carried out along three lines par-
allel to the tunnel alignment at distances of 0 m, +25 m and
-25 m from the tunnel axis. The probing positions were stag-
gered (cf. Figure 3), in order to effectively allow for one CPTU
carried out at 25 m spacing along the projected centreline of the
entire immersed tunnel alignment. Furthermore, additional
CPTUs were carried out near the artificial islands. The CPTUs
were carried out to penetration depths varying from 28 to 43 m
below existing seabed level (basically to refusal in the fluvial al-
luvial sands and clays underlying soft deposits of marine clays).
A typical arrangement of investigations along the immersed
tunnel alignment is shown in Figure 3.
The complete results of the Supplementary Soil investiga-
tions were provided by the geotechnical sub-contractors, Fourth
Harbour Design Institute (FHDI) and Fugro, in native AGS 3.1
format.

Figure 3. Typical arrangement of investigations along immersed tunnel
alignment.
625
4 DRILLING AND IN-SITU TESTING
Drilling for the Supplementary Soil investigations was carried
out from five drilling barges equipped with suspended rotary
top drive drilling rigs and passive heave compensation.
Three different passive heave compensation systems were
installed on the five drilling barges used for the investigations:
A strictly mechanical weight load system on one barge,
A spring loaded mechanical system on three barges, and
A hydraulic piston system on one barge.

The above mentioned systems were able to be operated with
good results (in terms of heave compensation) at maximum
heave of approx. 0.7 to 1.0 m.
Undisturbed samples (fine grained soils) were primarily re-
trieved with a 76 mm diameter thin walled stationary piston
sampler with stainless steel seamless sampling tubes of length
1.0 m.
Undisturbed samples were sealed with wax and taped-shut
end caps immediately after retrieval. Storage and transportation
were carried out vertically in wooden boxes filled with shock
absorbing material (coarse sawdust).
SPT testing in coarse grained soils was generally carried out
at 1.5 m intervals, and the hydraulic head in the boreholes was
as a minimum kept at a level corresponding to sea level. The
SPT-N Energy Transfer Ratio (ETR) was determined by carry-
ing out PDA tests of the equipment used from three different
barges.
In situ shear vane testing was performed at 1 m intervals in
fine grained soils using the Chinese electrical vane equipment
with cruciform vanes of dimensions 75 mm x 150 mm for the
softer clays.
CPTU testing was carried out using underwater seabed pie-
zocone penetration systems deployed from barges where the po-
sition was maintained by means of 4 heavy anchors. Two dif-
ferent CPTU systems were used, the Wheeldrive Seacalf with
200 kN thrust and the ROSON system with a 100 kN thrust. All
CPTU testing was carried out in accordance with the ISSMGE
(2001) standard.
5 LABORATORY TESTING
Classification testing for the Supplementary Soil investigations
consisted of natural moisture content, bulk and dry density, par-
ticle density, Atterberg limits, particle size distributions, maxi-
mum and minimum dry densities and organic content.
Incremental loading (IL) oedometer testing was carried out on
both undisturbed fine grained soil samples and reconstituted
coarse grained soil samples in accordance with BSI (1990a).
The specific schedule for the IL oedometer tests on fine
grained samples was designed to take into account the in-situ
and pre-consolidation stress together with the anticipated stress
history imposed by the construction activities.
The maximum net stress increments under the tunnel ele-
ments were not expected to lead to exceedance of the in-situ
stresses neither along the middle part of the immersed tunnel
alignment nor towards the artificial islands.
In view of the above, special attention was paid to determine
reliable estimates of the values of the pre-consolidation stress
and the reloading stiffness. The IL oedometer tests carried out
on fine grained samples were performed in two batches:
Batch I IL oedometer tests: Mainly carried out to provide an
estimate of the pre-consolidation pressure (and the virgin
compression index),
Batch II IL oedometer tests: Carried out to provide an esti-
mate of the reloading stiffness from varying unloading
stress levels below the pre-consolidation stress estimated
from the Batch I tests.
Initial unloading/reloading steps from/to the presumed in-situ
stress were included for both the Batch I and II IL oedometer
tests in an attempt to quantify and reduce the sample distur-
bance resulting from sample retrieval, transportation and extru-
sion. The application of the this initial branch of unload-
ing/reloading conceivably improved the apparent sample quality
significantly, as e.g. evaluated in accordance with NORSOK
(2004), on average from poor to very good/excellent sample
quality.
Triaxial testing of fine grained undisturbed samples was car-
ried out as Consolidated Anisotropic Undrained (CAU) triaxial
tests in accordance with BS1(1990b).
The triaxial tests allowed for site specific calibration of the
N
kt
cone factor for determination of realistic undrained shear
strengths based on CPTUs. Secondly, they allowed the value of
s
u
/'
pc
for the normally consolidated condition (often referred to
as the c/p ratio) to be determined. In this way a site specific
SHANSEP relation could be established allowing determination
of the undrained shear strength variation from actual unload-
ing/reloading cycles as a consequence of construction activities.
6 CPTU CORRELATIONS
For the purpose of establishing a detailed geological and geo-
technical model of the subsurface conditions, a combination of
cored boreholes and closely spaced CPTU soundings was se-
lected as the primary method of investigating the project site.
The CPTUs and boreholes were generally carried out as de-
scribed. The locations of the boreholes were arranged to provide
a total of 68 Nos. pairs of boreholes and CPTUs along the
alignment. This allowed for a site specific correlation between
the stratigraphy as encountered within the boreholes and the
corresponding principal CPTU properties with respect to cone
resistance, friction ratio and pore pressure. The boreholes and
CPTUs carried out in pairs were generally positioned within a 5
m distance from each other.
Initially, two approaches were investigated to find the most
appropriate correlation model for the site investigation data,
namely a conventional method developed by Robertson et al
(Lunne et al 1997) and a site specific approach based on pairing
the CPTU and borehole data.
The depiction of the site CPTU results categorised into the
different main geological units and using the Robertson classi-
fication chart is shown in Figure 4.

Figure 4. CPTU results superposed on soils classification chart (Lunne
et al 1997).

Instead, the CPTU data were analysed statistically, yielding rep-
resentative ranges and frequency distributions of each geo-
logical unit with respect to cone resistance, friction ratio and
excess pore pressure. In this way a unique "foot print" was pro-
duced for each geological unit as e.g. shown in Figure 5.

626
th

(1)
where q
tM
is the stress-adjusted cone resistance and
r
is a refer-
ence stress (=100 kPa).
Based on the modulus number from the oedometer tests and
the stress adjusted cone resistance from CPTU testing, the
modulus modifier, a, was derived or each soil deposit from (1).

Figure 5. Example of q
c
frequency distribution "foot print" for Marine
Clay.

Based on the frequency distributions, representative ranges were
established for the three principal CPTU properties, which in
turn were used as filter criteria for a template predicting the
geological unit.
The modulus modifier is plotted in Figure 7 assessing all oe-
dometer results for fine grained samples. The results shown in
this figure indicate relatively little data scatter and a general
grouping of fine grained soils around 2 to 5 and 60 to 90 for the
coarse grained soils (the latter values are not shown in Figure
7).

The pore pressures varied greatly within each geological unit
and were not used as a criterion for the geological interpreta-
tion, but merely as a guide when visually cross checking the re-
sults.
The interpretation template set up in this way worked on the
premise that if a data set fell within the established "foot print"
criteria, the template would subsequently yield the associated
geological unit. The criteria were regarded as a key to a typical
geological interpretation, not an unambiguous analysis. The fi-
nal geological interpretation was therefore based on additional
factors such as the combined appearance of the qc, Rf and u2
distributions combined with cross referencing to nearby bore-
holes.
Approximately 400 Nos. CPTUs (including those carried out
during the Preliminary Design investigations) were interpreted
using this method. This allowed for a 3D stratigraphical model
to be set up for the geotechnical interpretation of the subsurface
conditions surrounding the tunnel alignment, see e.g. Figure 6.
Figure 7. Modulus modifier, a, for selected geological units as derived
from oedometer and CPTU testing results.

7 GEOTECHNICAL INTERPRETATION
The interpretation of the results of the oedometer tests carried
out yielded the modulus number, m, recompression modulus
number, m
r
, secondary compression index, C
, secondary re-
compression index, C
r
, coefficient of consolidation, c
v
and ex-
cess preconsolidation pressure, '
pc
(= '
pc
- '
v0
).
The recompression branch of the oedometer tests on fine
grained soils indicated a linear correlation rather than a log-lin-
ear correlation. Further, the recompression modulus number, m
r
,
resulting from the reloading branches was found to vary with
load for the fine grained soils. A reasonable approximation was
achieved by applying different m
r
values above and below an in
situ stress of 100 kPa.
The resulting recompression modulus modifier, a
r
, was
therefore defined for in situ stress below and above 100 kPa.
The use of CPTUs was a key element in the evaluation of the
settlement/stiffness variation along the alignment of the Having
established the modulus number, m, for a range of soil deposits
through laboratory oedometer testing, the modulus modifier, a,
can be determined based on the formula:
Relatively little data scatter was observed in the a
r
values,
with a general grouping of a
r
values for fine grained soils
around 14 to 25 and 14 to 33 for in situ stress above and below
100 kPa, respectively.

Figure 6. Example of contour plot generated based on the compiled 3D stratigraphical model showing top of Continental/Marine Alluvial deposits in
the location of the East Artificial Island.

627
The SHANSEP concept derives from the empirical observation
that the ratio of the undrained shear strength, s
u
, to the effective
confining stress, '
v
, is approximately constant for a given Over
Consolidation Ratio (OCR) and varies linearly with OCR
:
(2)
where S is the proportionality constant (also referred to as the
c/p ratio) and is the memory exponent. These values were es-
timated from the CAU triaxial testing carried out on undisturbed
samples.
The S (or c/p-ratio) value was determined based on CAU
tests loaded anisotropically to >150% of the assumed precon-
solidation stress (as determined from the Batch I IL oedometer
tests) and then sheared. The S-value thus determined was used
for the determination of the value for tests loaded anisotropi-
cally to below the assumed preconsolidation stress. Due to rela-
tively high uncertainty with regards to the determination of the
preconsolidation pressure, the memory exponent was found dif-
ficult to determine with accuracy.
For the clay deposits found along the alignment of the im-
mersed tunnel average S and values shown in Table 1 were
found.

Table 1. Average values of S and for clay deposits found along the
mmersed tunnel alignment. i
Soil deposit
Nos. of
tests
S (avg.) avg.)
Marine clay 2 0.31 0.7
Continental clay 2 0.40 NA
Marine alluvial clay 7 0.31 1.0
Marine alluvial clay
with sand laminae
4 0.36 0.7
Notes: NA = Not Applicable

The results of the CAU triaxial tests were also used to provide a
correlation to results of CPTU testing, and thereby for providing
an estimate of the N
kt
cone bearing factor as used in the fol-
lowing equation (e.g. Lunne et al 1997):
(3)
where
v0
is the overburden pressure at the cone tip and q
t
is the
cone resistance corrected for pore pressure.
For the clay deposits found along the alignment of the im-
mersed tunnel, the N
kt
values were found to be 17 on average
for the four deposits referenced in Table 1.
8 SETTLEMENT/SPRING STIFFNESS CALCULATION
Based on the geotechnical interpretation of the geology and set-
tlement characteristics of soil deposits, the settlement and spring
stiffness was calculated for each individual CPTU location.
The settlement analysis was carried out using the Janbu
(1963) tangent modulus method, which accounts for the general
non-linear load deformation relationship of soils. The settlement
equations differ between coarse grained (sandy) and fine
grained (clayey and silty) soils, and whether or not the precon-
solidation stress is exceeded.
All in all four different equations were established.
Eq (4) for coarse grained soils below and above the preconsoli-
dation stress:
(4)
and Eq (5) for fine grained soils below and above the precon-
solidation stress:
(5)
Here is the vertical strain, '
v
is the increase in effective ver-
tical stress from the tunnel ('
1
-'
0
), '
p
is the preconsolidation
pressure, '
0
is the in-situ vertical stress prior to loading, '
1
is
the final vertical effective stress and '
r
is a reference stress of
100 kPa.
The secondary settlement was calculated from (Terzaghi et
al. 1996):
(6)
where C
is the secondary compression index, and t/t

p
is the ra-
tio between the lifespan of the structure and the time for pri-
mary consolidation (t/t
p
= 100 was conservatively assumed).
When the final load was lower than the preconsolidation
stress, the secondary recompression index, C
r
, was used in-
stead of C
.
The calculation of settlement was terminated at the top of
rock, and due to the limited penetration of the CPTUs into the
fluvial alluvial deposits of sand and gravel, the settlement calcu-
lations were based on SPT N data between the bottom of the
CPTUs and the top of rock. An empirical q
c
/N correlation de-
pendent on the grain size distribution was used (Kulhawy &
Mayne 1990):
(7)
where p
a
is a reference stress of 100 kPa, d
50
is the mean grain
size in mm and q
c
is given in kPa.
The spring stiffness was then calculated as:
(8)
The settlement/spring stiffness calculations were carried out in
purposefully set up Excel spreadsheets.
The settlement/spring stiffness calculations were carried out
for some 400 Nos. CPTUs, and considering that each CPTU
could contain up to 6,000 measurement points, running the en-
tire series of calculations could take up to 2 hours.
The variation of calculated settlement and spring stiffness
along the immersed tunnel alignment is shown in Figures 8 and
9, respectively.

Figure 8. Calculated settlement along immersed tunnel alignment centre
line and lines at 25 m distance from centreline.

628
th

Figure 9. Calculated spring stiffness along immersed tunnel alignment centre line and lines at 25 m distance from centreline.

9 CONCLUSION
The design of the 6 km world record long immersed tunnel with
highly variable soil and loading conditions poses significant
challenges to both the geotechnical site characterization and the
soil-tunnel interaction.
The structural tunnel design is very sensitive to differential
settlements and rotations of individual tunnel elements and
segments and thus to variation is soil stiffness along and across
the tunnel alignment. Rather than resolving to empirical rules
for handling the soil stiffness variation (Monte Carlo simulation
or additional sinusoidal variation around the mean stiffness) the
variation was handled directly by the tight mesh of CPTU prob-
ing points along and across the alignment.
Thus, the CPTUs provided a strong tool for clear geological
unit delineation and allowed for very detailed settlement and
soil stiffness assessment along the entire tunnel. The CPTU data
were correlated with results from oedometer and CAU triaxial
test results to provide site specific correlations regarding stiff-
ness and undrained shear strength.
The geotechnical site characterization thus facilitated the
tool for interaction between geotechnical and structural design
of the tunnel elements and allowed for a robust and safe design.
10 ACKNOWLEDGEMENTS
The authors gratefully acknowledge the permission by COWI to
publish the paper.
11 REFERENCES
COWI 2011. Hong Kong-Zhuhai-Macao Link Immersed Tunnel De-
tailed Design, Final Geotechnical Interpretative Report.
ISSMGE 2001. International Reference Test Procedure for the Cone
Penetration Test (CPT) and the Cone Penetration Test with pore
pressure (CPTU).
BSI 1990a. British Standard Methods of test for Soils for civil engineer-
ing purposes. Part 5. Compressibility, permeability and durability
tests, BS1377:Part 5:1990.
BSI 1990b. British Standard Methods of test for Soils for civil engineer-
ing purposes. Part 8. Shear strength tests, BS1377:Part 8:1990.
NORSOK 2004. Standard. G-001. Rev. 2. Marine Soil Investigations.
Lunne, T., Robertson, P.K., Powell, J.J.M. 1997. Cone Penetration Test-
ing in Geotechnical Practice, First Edition.
Massarsch, K.R., Fellenius,B.H. 2002. Vibratory compaction of coarse
grained soils. Canadian Geotechnical Journal 39, 695-709.
Janbu, N. 1963. Soil compressibility as determined by oedometer and
triaxial tests. III European conference on soil mechanics and foun-
dation engineering, Wiesbaden, Vol. 1, pp. 19-25 and Vol. 2, pp.
17-21.
Terzaghi, K, Peck, R.B., Mesri, G. 1996. Soil Mechanics in Engineering
Practice, Third Edition.
Kulhawy, F.H. and Mayne, P.W. 1990. Manual on estimating soil prop-
erties for foundation design. EPRI EL-6800, Cornell University.

629
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris2013
Controversial and Contradictory Evaluations in Analyses of Ground Vibrations
from Pile Driving
valuations controverses et contradictoires dans l'analyse des vibrations de terre
par suite de l'enfoncement de pieux
Svinkin M.R.
VIBRACONSULT, Cleveland, USA
ABSTRACT: Pile driving operations are powerful and wide-spread sources of construction vibrations which may detrimentally
affect adjacent and remote structures, make obstacles for operating sensitive processes and devices, and disturb people. A number of
factors can affect ground vibration from pile installation. Wave propagation from pile driving is a complicated problem, and
different approaches are utilized to analyze this phenomenon. A paper presents several controversial and contradictory issues in
assessment of ground vibrations generated by pile driving such as connections between wave propagation in piles and ground
vibrations, the relationship between pile impedance and intensity of ground vibrations, effects of the hammer energy on ground
vibrations and a possible correlation between pile capacity and ground vibrations. Also, it is important to show the inadequate roles
of condition surveys of structures and monitoring of ground vibrations and the necessity to properly assess crack changes in
condition surveys. Analysis and clarification of various approaches are important for practical applications.
RSUM: Les oprations denfoncement de pieux sont des sources puissantes et trs rpandues de vibrations de construction qui
pourraient affecter de faon nuisible des structures adjacentes et loignes, faire obstacle lutilisation de procds et dappareils
sensibles, et perturber des gens. De nombreux facteurs peuvent affecter la vibration du sol par suite de l'installation de pieux. La
propagation dondes cause par lenfoncement de pieux est un problme compliqu, et diverses mthodes sont utilises pour
analyser ce phnomne. Cet article prsente plusieurs points controverss et contradictoires dans lvaluation des vibrations de sol
gnres par lenfoncement de pieux, telles que des connections entre la propagation des ondes dans les pieux et les vibrations de
sol ; les rapports entre l'impdance des pieux et lintensit des vibrations de terre ; des effets de lnergie percutante sur des
vibrations de sol, et une corrlation possible entre la capacit des pieux et les vibrations de sol. De plus, il est important de montrer
le rle inadapt des conditions de surveillance des structures, du contrle des vibrations de sol et la ncessit d'valuer les
modifications des fissures par les oprations de contrle. L'analyse et la clarification de diverses mthodes sont importantes pour des
applications pratiques.
KEYWORDS: pile driving, ground vibrations, stress wave theory, impedance, energy, survey
1. INTRODUCTION
Installation of driven piles creates soil vibrations and
displacements which may affect adjacent and remote
structures, people and sensitive equipment. Therefore, various
approaches are used for evaluation of vibration effects of pile
driving.
There is a trend to connect stress-wave propagation in piles
during pile driving with prediction or calculation of the peak
particle velocity (PPV) of ground vibrations from pile
installation. However, there are ambiguous problems in using
of this approach for assessment of ground vibrations.
Pile impedance affects force and velocity at the pile head in
opposite ways at the same time. Therefore, the pile impedance
effect on the intensity of ground vibrations is not obvious.
Pile driving generates ground vibrations due to the hammer
energy applied to a pile, but some case histories demonstrate
no correlations between the hammer energy and the maximum
velocity of ground vibrations. Other factors such as the depth
of pile penetration into the ground and soil resistance to pile
driving should be taken into account.
The relationship between pile capacity and ground
vibrations is not clear. Moreover, pile capacity and ground
vibrations are outcomes of pile driving and only an accidental
correlation between them is possible. For sure, both variables
are dependent on the hammer energy.
On the one hand approximate calculation of expected
ground vibrations and even vibration monitoring yield relative
information on vibration effects on structures, and these results
could be inconclusive. On the other hand condition surveys of
structures before, during, and after pile driving provide complete
information on structural responses to vibration excitations and
this information can be much beneficial than vibration assessment
and measurements.
Clarifications of different ways used for analyses of pile
driving as the source of construction vibrations, ground vibrations
generated by pile driving and various effects of these ground
vibrations on structures, people and sensitive devices are
important to understand the problem and prevent harmful
consequences of pile driving operations.
2. STRESS-WAVES PROPAGATION IN PILES AND
GROUND VIBRATIONS
For about forty years, the stress-wave theory is successfully
used for driveability analysis of driven piles and also for
determination of pile capacity at the time of testing, for
example Proceedings of IS-Kanazawa 2012 (2012). In recent
years, there is a trend to connect stress-wave propagation in
630
th
th
piles during pile installation with prediction or calculation of
PPV of ground vibrations generated by pile driving, Robertson
(2006) and Massarsch & Fellenius (2008).
The first such an attempt was made by Svinkin (1996) in
favor of the Impulse Response Function Prediction (IRFP)
method for prediction of ground and structure vibrations from
pile driving; the method was developed toward prediction of
complete time-domain vibration records on existing soils,
buildings, and equipment prior to installation of impact
machine foundations, Svinkin (2002). In the application of this
method to pile driving, wave equation analysis was used to
assign a movement of the pile top, but its necessary to
underline that the top pile movement can be assigned
arbitrarily, for example as a damped sinusoid, because ground
vibrations at some distance from a dynamic source depend only
on the dynamic force transmitted on the machine support and
soil properties, Svinkin (2002).
It is necessary to point out that a connection of the stress-
wave theory with ground vibrations from pile driving has few
problems. First, there are several different programs for signal
matching techniques which produce diverse results for the
same piles and their outcomes depend on variety of soil
conditions and pile types, Svinkin (2012). It is not clear what
software should be used. Second, stress waves in piles
obviously generate internal forces in driven piles. Third,
according to Saint Venants principle, wave propagation in piles
does not affect dynamic field at some distance from a driven
pile.
It is known that impact hammers for pile driving and forge
hammers released comparable amounts of the energy and they
generate similar vibration records of ground vibrations
(Steffens 1974). Therefore, it is reasonable to compare both
dynamic sources and their effects on ground vibrations.
A forge hammer foundation is considered as a rigid body
which transfers impacts loads from a hammer onto the ground.
Dynamic forces in the machine foundation itself are internal
forces generated by stress-waves propagated in the machine
foundation under forge hammer impact. The duration of internal
forces is substantially smaller than the duration of dynamic
forces transferred from a machine foundation onto the ground,
and these two kinds of dynamic forces work in different time
frames. Consequently, internal dynamic forces in hammer
foundations are not taken into account in determination of
dynamic loads transferred from a hammer foundation on the
ground and consideration of ground vibrations generated by
oscillations of forge hammer foundations (Barkan 1962 and
Richart et al. 1970).
In prediction of ground vibrations from operating forge
hammers, ground vibrations depend on the impulse dynamic load
applied to a hammer foundation, the damping coefficient and the
natural frequency of vertical foundation oscillations, and also the
impulse response functions of the considered dynamic systems.
The latter represent the soil medium where wave propagate from
the hammer foundations to destination locations. The
experimental studies showed that at some distances from the
source, ground vibrations become dependent only on the impulse
load transmitted to a hammer foundation and the soil medium
where waves disseminate from the source (Svinkin 2002). These
results are in agreement with a dynamic version of Saint
Venants principle (Timoshenko & Goodier 1951and Karp &
Durban 1997).
A similar picture of a dynamic load transfer from a forge
hammer on its foundation and the ground can be represented for
pile installation. Piles also can be considered as rigid bodies in
which stress-waves propagate from hammer ram impacts and
generate internal forced in piles which are the causes of pile
movement and vibrations. Besides, a pile-soil load transfer is
released by means of both concentrated loads from the pile toe
and distributed loads generated along pile shaft. Similarly to
hammer foundations, at some distances from a pile, as the
dynamic source, ground vibrations become dependent only on the
dynamic load applied to a pile and the soil medium where waves
propagate from the source. It is known that velocities of wave
propagation in piles are about 4000 m/s in concrete piles and
about 5100 m/s in steel ones (PDA 1991). Velocities of shear
wave propagation in the ground are shown in Table 1.
Velocities of surface waves are equal about 0.92-0.96 of the
velocities of shear waves, Barkan (1962).
Table 1. Velocity of shear waves in soils, Savinov (1979)
Soil Velocity
m/s
Sand 120 150
Sand with gravel 150 250
Loess with natural
moisture
130 160
Plastic clay 150 - 400
It can be seen that that wave propagation in piles under
impact load is much faster process in comparison to wave
propagation in the ground. Therefore, dynamic loads
transferred from driven piles onto the ground for practical
purposes can be considered as the point impulse load at some
distance from the source, Svinkin (2000).
It can be expected that this conception is correct at distance
derived from an assumption that the time of surface wave
propagation with velocity, c
s
, in the ground at distance, D, from a
driven pile is 5-10 times larger than the time of stress wave
propagation with velocity, c, in the pile with length, L (Svinkin
2000).
c / Lc ) 10 5 ( D
s
(1)
Minimum distances from a driven concrete pile as the point
vibration source are shown in Table 2 (coefficient 10 was used).
Table 2. Minimum distance from pile as point vibration source
c
s
/c
150/4000 300/4000
Pile
Length
m
Lc
s
/c
m
10Lc
s
/c
m
Lc
s
/c
m
10Lc
s
/c
m
10 0.375 3.75 0.75 7.5
15 0.5625 5.63 1.125 11.25
20 0.8438 8.44 1.5 15.0
30 1.125 11.25 2.25 22.5
40 1.6875 16.88 3.0 30.0
It can be expected at distances determined by equation (1),
that only dynamic forces transferred to piles during pile driving
and soil medium where waves propagate from driven piles will
affect ground vibrations generated by pile driving.
It is important to point out that calculation of expected
ground vibrations during the time of pile installation is irrelevant.
For example, Massarsch & Fellenius (2008) tried to connect
stress-wave propagation in piles under the hammer ram impact
with ground vibrations, but they eventually suggested the old
empirical equation to calculate attenuation of PPV of ground
vibrations generated by surface waves, which contain more
than 2/3 of the total vibration energy, from pile installation
without any connection with the stress-wave theory. Ground
vibrations have to be measured during pile driving operations.
631
th
3. PILE IMPEDANCE
During pile installation, an impulse load from the hammer ram
is applied to the pile top, and dynamic longitudinal force in the
pile is transferred to the surrounding soil. According to Peck et
al. (1974) and Woods (1997), pile impedance affects the force
transmitted down the pile. Pile impedance characterizes the
pile ability to overcome the soil resistance to pile penetration
and develop required capacity.
One of pile impedance, Z, definition can be presented as
c / EA Z (2)
where E is modulus of elasticity of pile material; A is pile
cross-section area; and c is longitudinal stress wave velocity. It
can be seen that impedance depends only on the pile material
and dimensions.
Recognizing the importance of pile impedance for
assessment of to the ground vibration, Heckman and Hagerty
(1978) proposed the equation for the peak particle velocity of
ground vibrations from pile driving as a function of the rated
hammer energy, W
r
, and the distance, D, from a driven pile
with the coefficient, k, which is dependent on pile impedance.
D
W
k v
r
(3)
The coefficient, k, is inversely proportional to pile impedance.
It means that driven piles with higher impedance generates
lower PPV of ground vibrations and vice versa.
Svinkin (2000) derived equations for PPV of pile
vibrations, V, and the maximum force, F, measured at the pile
head as
ZL
cW 2
V
t
(4)
and
L
cZW 2
F
t
(5)
where W
t
is the energy transferred to a pile.
Similarly to equation (3), equation (4) shows that the
velocity triggered by the hammer ram impact is an inversely
proportional function of pile impedance. However, equation (5)
displays that the force is proportional to the root square of pile
impedance. It means that pile impedance affects force and
velocity at the pile head in opposite ways.
Case histories presented in a number of publications, for
example Svinkin (2000), demonstrate higher ground vibrations
triggered by installation of high soil displacement piles
(concrete piles and steel pipes with closed ends) in comparison
with low soil displacement piles (H-piles and steel pipes with
open ends). A practical experience is the evidence that pile
impedance affects ground vibrations in the proximity of driven
piles, but this pile property does not affect the dynamic field at
some distance from driven piles in accordance with Saint
Venants principle.
4. HAMMER ENERGY
Pile installation generates ground vibrations due to the hammer
energy applied to a pile. Obviously, PPV of ground vibrations
have to be a function of the hammer energy transferred on a
pile. However, some case histories demonstrate no correlations
between the hammer energy and PPV of ground vibrations,
Hope and Hiller (2000). It happens due to the effects of soil
conditions, the pile penetration depth, and the soil resistance to
pile penetration into the ground. Nevertheless, the hammer
energy is the major cause of ground vibrations because without
the hammer energy there are no pile penetration into the
ground and ground vibrations.
5. PILE CAPACITY AND GROUND VIBRATIONS
Some authors, for example Robinson (2006), found enormous
scatter of PPV of ground vibrations as a function of the
hammer energy. For example, PPV of ground vibrations
changed between about 0.4-21.6 mm/s at the rated energy of
135 kJ and between about 0.9-17.8 mm/s at the transferred
energy of 40 kJ. It happened because other factors mentioned
above affected ground vibrations and in consequence that data
measured at various construction sites with different soil
conditions, pile types and pile driving implementations were
considered together. However, Robinson (2006) suggested a
correlation between ground vibrations and pile capacity
determined during pile driving. He believes that pile-soil
interaction, not energy, is the major influence in the generation
of ground vibrations from driven piles. Obtained conclusions
are not accurate because ground vibrations and pile capacity
are outcomes of the same pile driving process and only an
accidental correlation between them is possible.
It is necessary to say that ground vibrations and pile
capacity for sure depend on the hammer energy because pile
capacity cannot be mobilized without the sufficient hammer
energy. Moreover, during pile driving, the static pile capacity
is determined by signal matching software on the basis of force
and velocity measurements at the pile head. Unfortunately,
different software produces different results. It means that PPV
of ground vibrations are dependant on signal matching
technique used for analysis of testing data. Besides, during pile
installation, ground vibrations should be measured not
calculated because of possible detrimental effects of pile
driving operations and also measured ground vibrations are
more reliable than calculated ones.
6. CONDITION SURVEYS AND VIBRATION
MEASUREMENTS
Approximate calculation of expected ground vibrations and
even vibration monitoring yield relative information on vibration
effects on structures, and these results could be inconclusive.
Moreover, there is uncertainty in application of the existing
vibration limits for assessment of pile driving effects on soils and
structures. Therefore, it is imperative to perform condition
surveys of structures before, during and after pile installation
which provide complete information on structural responses to
vibration excitations. Obtained information can be much
beneficial for analysis of causes of damage to structures than
vibration assessment and measurements. Dowding (1996)
pointed out the necessity of professional performance of a
preconstruction survey.
Condition surveys during pile installation and after the
completion of pile driving are significant for analysis of
possible causes of damage to structures. Each construction site
632
th
th
is unique and even similarity of soil deposits does not mean the
same condition of the dynamic settlement development.
Physical evidences of damage to structures from dynamic
sources are very important. Therefore, much attention is
provided for measurement of crack width at condition surveys
of structures during pile driving.
Changes of crack dimensions are the major evidences of
vibration effects on structures. Micrometers are used to
determine changes of crack widths. It is necessary to keep in
mind that each structure has its own breathing because of
microseisms in the earth and human activities inside and
outside structures. Hence, it is typical that crack widths may
vary in time. If crack widths increase without increasing of
crack lengths, it is a safe situation. However, if variations of
crack widths trigger increasing of crack length, it becomes
dangerous for structures. Thus, changes of crack widths alone
are not the indicators of damage to structures from pile driving.
Therefore, it is necessary to measure crack widths together
with assessment of crack length enlargements.
7. CONCLUSIONS
Ground vibrations from pile driving may harmfully affect
structures, people and sensitive devices, and these effects
should be evaluated before and during pile driving operations.
The paper presents several controversial and contradictory
issues in assessment of ground vibrations generated by pile
driving. Analyses of various approaches are important for
practical applications.
A connection of the stress-wave theory with ground
vibrations from pile driving has few problems. There is no
unique solution of stress-wave propagation in the pile because
different signal-matching software provides different
outcomes. Internal forces in piles may somewhat affect ground
vibrations in the proximity of the pile. However, according to
Saint Venants principle, wave propagation in piles does not
affect dynamic field at some distance from a driven pile.
Pile impedance affects ground vibrations in the proximity
of driven piles, but this pile property does not affect the
dynamic field at some distance from driven piles in accordance
with Saint Venants principle.
Pile installation generates ground vibrations due to the
hammer energy applied to a pile. Missing correlation between
PPV of ground vibrations and the hammer energy in some case
histories occurred on account of the effects of soil conditions,
the pile penetration depth, and the soil resistance to pile
penetration into the ground. Nevertheless, the hammer energy
is the major cause of ground vibrations because without the
hammer energy there are no pile penetration into the ground
and ground vibrations.
Pile capacity and ground vibrations are outcomes of the
same pile installation and only an accidental correlation
between them is possible.
Condition surveys should be performed before, during and
after pile driving. Assessment of crack length enlargements has
to accompany measurements of crack widths because changes
of crack widths alone are not the indicators of damage to
structures from pile driving.
Clarification of different views on the problems would be
helpful in practice for assessment of pile driving effects on
surrounding structures.
8. REFERENCES
Barkan, D.D. 1962. Dynamics of Bases and Foundations. New York:
McGraw Hill Co.
Dowding, C.H. 1996. Construction Vibrations. Prentice Hall, Upper
Saddle River.
Heckman, W.S. & D.J. Hagerty 1978. Vibrations associated with pile
driving. Journal of the Construction Division, 104(CO4): 385-394.
ASCE.
Hope, V.S. and Hiller, D.M. 2000. The prediction of groundborne
vibration from percussive piling. Canadian Geotechnical Journal,
37, 700-711.
Karp, B and Durban, D. 1997. Towards a dynamic version of Saint
Venants principle. Modern Practice in Stress and Vibration
Analysis. M.D. Gilchrist (Ed.): 251-255. Rotterdam: Balkema.
Massarsch, K. R., and Fellenius, B. H. 2008. Ground vibrations
induced by impact pile driving. SOAP 3, Proceedings of the Sixth
International Conference on Case Histories in Geotechnical
Engineering: 1-38, Arlington, Virginia: OmniPress.
PDA (1991). Pile Driving Amalyzer
TM
- Manual. Pile Dynamic, Inc.
Cleveland, Ohio.
Peck, R.B., Hanson, W.E. and Thornburn, T.H. 1974. Foundation
Engineering, 2
nd
ed., New York: John Wiley & Sons, Inc.
Proceedings of IS-Kanazawa 2012. Testing and Design Methods for
Deep Foundations. Proceedings of the 9
th
International Conference
on Testing and Design Methods for Deep Foundations, Kanazawa,
Japan, 18-20 September 2012.
Robinson, B.R. 2006. Models for Prediction of Surface Vibrations from
Pile Driving Records. A thesis submitted in partial fulfillment of the
Requirements for the degree of Master of Science, NC State
University.
Richart, F.E., Hall, J.R. and Woods, R.D. (1970). Vibrations of soils and
foundations, Prentic-Hall, Inc., Englewood Cliffs, NJ.
Savinov, O.A. 1979. Modern construction of machine foundations and
their calculations. Second Ed. Stroiizdat, Leningrad.
Steffens, R.J. 1974. Structural vibration and damage. Building Research
Establishment Report, HMSO.
Svinkin, M.R. 1996. Overcoming soil uncertainty in prediction of
construction and industrial vibrations. Proceedings of Uncertainty in
the Geologic Environment: From Theory to Practice, C.D.
Shackelford, P. Nelson, and M.J.S. Roth (Eds.), Geotechnical
Special Publications No. 58, ASCE, 2: 1178-1194.
Svinkin, M.R., Roth, B.C. and Hannen, W.R. 2000. The effect of pile
impedance on energy transfer to pile and ground vibrations.
Proceedings of the Sixth International Conference on the Application
of Stress-Wave Theory to Piles. S. Niyama &J. Beim (Eds.): 503-510,
Rotterdam: Balkema
Svinkin, M.R. 2002. Predicting soil and structure vibrations from impact
machines. Journal of Geotechnical and Geoenvironmental
Engineering., 128(7): 602-612. ASCE.
Svinkin, M.R. 2012. Engineering evaluation of static capacity by dynamic
methods. Proceedings of the 9
th
International Conference on Testing
and Design Methods for Deep Foundations, Kanazawa, Japan, 18-20
September 2012: 179-186. Kanazawa University.
Timoshenko, S.P. and Goodier, J.N. 1951. Theory of Elasticity. New
York: McGrawHill Book Co.
Woods, R.D. (1997). Dynamic Effects of Pile Installations on Adjacent
Structures, NCHRP Synthesis 253, Transportation Research
Board, National Research Council, Washington, D.C.
633
CPT/PCPT- Based Organic Material Profiling
Matire organique - Le profilage bas sur le CPT/PCPT
Tmay M.T.
Louisiana State University, Baton Rouge, LA, USA and Boazii University, stanbul, Turkey
Hatipkarasulu Y.
The University of Texas at San Antonio, San Antonio, TX, USA
Marx E.R.
Fugro Consultants, Inc., Baton Rouge, LA, USA
Cotton B.
Fugro Consultants, Inc., Kenner, LA, USA
ABSTRACT: Cone and Piezocone Penetration Test (CPT and PCPT) based analysis and modeling is a popular and handy tool for
geotechnical engineers for subsurface investigations and soil characterization. However, effective identification and extent of organic
content proves to be a challenge based on traditional CPT and PCPT data and methodologies. This paper presents a comprehensive
CPT/PCPT-based organic content identification method using Zhang and Tumay (1999) probabilistic soil classification method. The
probabilistic method employs a non-traditional modeling approach that takes the uncertainty of correlation between the soil
composition and soil behavior into account. The method is based the conformal mapping of the Douglas and Olsen (1981)
classification chart which results in the soil classification index (U) and in-situ behavior index (V). The organic content identification
method proposed in this paper uses the in-situ behavior index (V) in combination with the compositional soil classification index (U)
to estimate the organic content. A detailed description of the proposed methodology and a discussion of effective applications are
included in the paper.
RSUM : Le pntromtre quasi-statique et le pizocne (CPT et PCPT) constituent des outils danalyse populaires et pratiques
pour la reconnaissance gotechnique des sites et la caractrisation des sols. Cependant, lidentification des sols organiques ainsi que
lvaluation de la teneur en matire organique partir des donnes classiques obtenues au CPT et au PCPT se sont rvles tre un
challenge. Cet article prsente une mthode dvaluation complte de la teneur en matire organique base sur le CPT et la mthode
de classification des sols probabiliste de Zhang et Tumay (1999). Cette mthode probabiliste utilise une approche de modlisation non
conventionnelle qui prend en compte lala sur la corrlation entre la composition du sol et son comportement mcanique. La
mthode est base sur labaque de classification des sols de Douglas et Olsen (1981) qui permet de dfinir lindice de classification
des sols (U) et lindice de comportement in situ (V). La mthode dvaluation du contenu en matire organique propose dans larticle
utilise lindice de comportement in situ (V) combin lindice de classification (U) pour estimer la teneur en matire organique du
sol. Une description dtaille de la mthodologie propose et une discussion de ses applications sont aussi prsentes dans larticle.
KEYWORDS: CPT, PCPT, Cone Penetration, Soil Classification ,Organic Soils

1 INTRODUCTION
Cone and Piezocone Penetration Test (CPT and PCPT) (ASTM
D5778-12) based analysis and modeling is a popular and handy
tool for geotechnical engineers for subsurface investigations and
soil characterization. Since the 1960s, several modeling
approaches have been developed for soil classification and
evaluation of different soil properties such as the strength and
consolidation characteristics of geomedia. However, effective
identification and extent of organic content has proved to be a
challenge based on traditional CPT and PCPT data and
methodologies. Although some models identify organic
materials as a separate soil class, they do not provide a
continuous profile (for example, Schmertmann, 1978;
Robertson et al, 1986; Robertson, 1990). Considering the
likelihood of having different levels of organic content in any
soil type, a continuous profile will provide additional
understanding and evaluation of the subsurface.
This paper presents a comprehensive CPT/PCPT-based
organic content identification method using Zhang and Tumay
(1999) probabilistic soil classification method. The organic
content identification method proposed in this paper uses the
Zhang and Tumay methods in-situ behavior index (V) in
combination with the compositional soil classification index (U)
to estimate the organic content. A detailed description of the
proposed methodology and a discussion of effective
applications are included in the paper. The paper also presents
analyses of two well-documented test sites to illustrate the
effectiveness of CPT/PCPT-based profiling and their correlation
to laboratory test results.
2 CPT-BASED PROBABILISTIC SOIL CLASSIFICATION
Unlike the traditional chart-based two-dimensional
classification methods, the Zhang and Tumay method uses a
probabilistic region estimation method to address the
uncertainty in misclassifying the soil layers. This statistical
based method provides a profile of the probability or the chance
of having each soil type (clayey, silty, and sandy) with depth.
This method is similar to the classic soil classification methods
which are based on soil composition. The probability of
incorrectly identifying soil type using the tradition CPT
classification charts, especially in transition zones, motivated
the development of the probabilistic region estimation method.
This CPT classification method addresses the uncertainty of
correlation between the soil composition and soil mechanical
behavior.
In the Zhang and Tumay probabilistic method, conformal
mapping was performed on the Douglas and Olsen (1981) chart
to transfer the chart axis from the CPT data (q
c
, R
f
) to the soil
classification index (U) and in-situ behavior index (V). The
conformal transformation is accomplished using the following
equations:
634
th
35 . 3 log 8870 . 0 1539 . 0
c f
q R x (1)

37 . 0 log 4617 . 0 2957 . 0
c f
q R y (2)

The soil classification index (U) and in-situ behavior index
(V) are given as:

2
2 1 2
2
1 2 1
1 2 1 1 2 1
d y c x c d y c x c
d y c x c b y a x a
U

2
2 1 2
2
1 2 1
2 1 2 2 1 2
d y c x c d y c x c
d y c x c b y a x a

(3)

2
2 1 2
2
1 2 1
2 1 2 1 2 1
) ( ) (
) )( (
10
d y c x c d y c x c
b y a x a d y c x c
V

2
2 1 2
2
1 2 1
2 1 2 1 2 1
) ( ) (
) )( (

d y c x c d y c x c
d y c x c b y a x a

(4)

The coefficients in equations 3 and 4 are defined as: a
1
=-11.345,
a
2
=-3.795, b
1
=15.202, b
2
=5.085, c
1
=-0.269, c
2
=-0.759,
d
1
=-2.960 and d
2
=2.477.
A statistical correlation was then established between the U
index and the compositional soil type given by the Unified Soil
Classification System (USCS) (ASTM D2487-11). A normal
distribution of U was established for each reference USCS soil
type (GP, SP, SM, SC, ML, CL, and CH). Each U value
corresponds to several soil types with different probabilities.
Boundary values were used to divide the U axis into seven
regions as described in Figure 1.
Soil types were further rearranged into three groups: sandy
and gravelly soils (GP, SP, and SM), silty soils (SC and ML)
and clayey soils (CL and CH). Figure 1 also gives the
probability of having each soil group within each region. The
original method gives constant probability of each soil type
(represented by the step lines) regardless of the U value within
the same region (R1 to R7 in Figure 1). This allows for the
sudden drop in the probabilities as the U value crosses the
border from one region to another. This method was further
modified to allow smooth transition of probability (curved
lines) with U values, and hence to provide a continuous profile
of the probability of soil constituents with depth.

U Value
0
10
20
30
40
50
60
70
80
90
100
P
r
o
b
a
b
i
l
i
t
y

(
%
)
R
7
R
6 R
5
R
4
R
3
R
2
R
1
-0.14 0.61 1.33 2.01 2.7 2.91
GS, SP, SM
SC, ML
CL, CH

Figure 1. Regional boundaries and the corresponding probabilities of
each soil group.

In-situ behavior index (V) provides a profile of soil behavior
and, in combination with the compositional soil classification
index (U), estimates of soil organic content and of soil
rigidity/stiffness (indirectly OCR) can be determined (Tmay
et.al, 2012).
3 ORGANIC MATERIAL PROFILING
The proposed profiling method utilizes a combination of the soil
classification index and the in-situ behavior index values where
the [(V-U) > 3.0] indicates significant organic content. The
organic content indicator (V-U) makes it possible to profile the
changes in organic content continuously while acknowledging
the likelihood of having organic material in any given soil
composition.
To illustrate the proposed method, two well documented
data sets from the Mississippi River Long Distance Sediment
Pipeline study located near Barataria Waterway in Jefferson
Parish, Louisiana are considered. Tables 1, 2, and 3 present the
laboratory test results for data sets B-7 and B-28 including
moisture content percentage (ASTM D2216-10), Atterberg
Limits (ASTM D4318-10), and material content finer than No.
200 sieve (ASTM D1140-06). Table 4 presents the organic
content percentages (ASTM D2974-07) for the same data sets.

Table 1. Laboratory testing values for B-7 data set.

Depth
(m)
Moisture
Content %
Liquid
Limit %
Plastic
Limit %
Plasticity
Index %
1.2 125 133 35 98
3.0 42 38 22 16
4.9 93 99 32 67
9.1 74 85 27 58

Table 2. Laboratory testing values for B-28 data set.

Depth
(m)
Moisture
Content %
Liquid
Limit %
Plastic
Limit %
Plasticity
Index %
2.1 72 44 20 24
2.8 81 78 24 54
4.3 80 86 28 58
16.8 50 69 23 46

Table 3. Material passing No. 200 sieve for B-7 and B-28 data sets.

Data Set Depth (m) Material Passing No.200 Sieve %
1.2 89 B-7
9.1 100
2.1 96
3.7 27
4.2 98
6.1 87
8.8 27
10.7 12
B-28
16.5 100

Table 4. Organic content values for B-7 and B-28 data sets.

Data Set Depth (m) Ash Content % Organic Content %
0.6-0.9 85.42 14.58
1.8-2.1 96.85 3.15
B-7
3.0-3.4 96.85 3.15
0.6-1.2 85.95 14.05
2.4-2.8 84.78 5.22
3.7-4.0 91.73 8.27
B-28
7.2-7.6 97.10 2.90

Figures 2 and 3 illustrate the CPT sounding values (q
c
, f
s
and
R
f
) (ASTM D 5778-12), Zhang and Tumay (1999) probabilistic
soil classification soil results with U and V index values,
organic content indicator (V-U), and corresponding lithology
obtained from the boring logs for B-7 and B-28 data sets.

635

% Cl ay
% Si l t
% Sand
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
0.00 1.00 2.00
D
E
P
T
H

(
m
)
MPa
TIP RESISTANCE (qc)
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
0.00 0.02 0.04
MPa
SLEEVE FRICTION (fs)
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
0.00 10.00 20.00
%
FRICTION RATIO (Rf)
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
-5.0 0.0 5.0
U INDEX
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
0.0 2.0 4.0
V INDEX
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
-3.0 3.0 9.0
Inorganic | Organic
Organic Content (V-U)
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
%0 %50 %100
ZT99 Probabilistic
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
%0 %50 %100
Lithology
ORGANIC CLAY (OH),
very soft, bl ack
LEAN CLAY (CL), very
soft to soft, gray and
dark gray
FAT CLAY (CH),
very soft, dark
gray
soft, gray, wi th shel l
fragments, and sand
FAT CLAY (CH),
soft, gray, wi th sil t
seams
and lenses
FAT CLAY (CH),
soft to fi rm, gray,
wi th sil t
seams and l enses

Figure 2. CPT-Based organic material profile and lithology B-7 Data Set.

% Cl ay
% Si l t
% Sand
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
0.00 6.00 12.00
D
E
P
T
H

(
m
)
MPa
TIP RESISTANCE (qc)
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
0.00 0.05 0.10
MPa
SLEEVE FRICTION (fs)
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
0.00 5.00 10.00
%
FRICTION RATIO (Rf)
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
-5.0 0.0 5.0
U INDEX
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
-2.0 0.0 2.0 4.0
V INDEX
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
-3.00 3.00 9.00
Inorganic | Organic
Organic Content (V-U)
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
%0 %50 %100
ZT99 Probabilistic
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
16.0
18.0
%0 %50 %100
Lithology
ORGANIC CLAY (OH),
very soft, dark gray,
with roots, peat, and
shell s
FAT CLAY (CH),
firm, gray, wi th
organi cs and
roots
soft, gray
SILTY SAND (SM),
very l ooseto
medi um-dense, gray
LEAN CLAY (CL), gray
- with sand seams
FAT CLAY (CH), w/organics
ORGANIC CLAY (OH)
FAT CLAY (CH), w/organics
LEAN CLAY (CL), w/organi cs

Figure 3. CPT-Based organic material profile and lithology B-28 Data Set.

As illustrated in Figures 2 and 3, the organic content
indicator (V-U) identified the significant organic content in the
sample data sets while providing a continuous profile. When
this information is combined with the CPT-based soil
classification, it provides a better understanding of the
subsurface conditions. For example, laboratory testing values
indicate a fine grained soil with high plasticity index (98%) for
the first meter of the B-7 data set where the organic content test
resulted in 14.58%. The test values show a significant drop after
2.0 meters for the plasticity index (16%) and the organic content
(3.15%). This profile change is clearly illustrated in Figure 2.
Similarly, as shown in Figure 3, the test results show organic
content over 5% for the B-28 data set between 2.5 and 4.0
meters. This value decreases to 2.90% at 7.2 meters.
636
th
The data capture frequency of the CPT soundings (2 cm/sec)
also allows for the identification of the thinner seams of sandy
and silty sand layers as well as the increases in the organic
content. For example, the soil classification of the B-28 data set
between 8.0 and 12.0 meters shows thin layers of clayey
materials. At these layers, the organic content appears to be
higher that the surrounding silty sand.
4 CONCLUSION
Although there are several CPT-based soil classification models
effectively used in subsurface investigations, accurate
identification of organic materials using CPT soundings
provides a challenge. The organic content indicator (V-U)
proposed in this study offers a continuous profile for the organic
content based on the soil classification and in-situ behavior
indexes as defined by the Zhang and Tumays probabilistic
method.
For the examples provided in this paper, the organic content
indicator shows a strong agreement with the test results and [(V-
U) > 3.0] indicates an approximate level of 5% organic
material. The CPT-based indicator may provide a useful tool
considering the importance of identifying organic materials
which could lead to problems concerning stability, compaction,
compressibility and usability.
Soil classification effectiveness and accuracy of the Zhang
Tumay (1999) method has been evaluated with several data sets
under different conditions and test sites, and compared with
other classification charts, for subsoil stratigraphy (Tmay et.
al., 2011), multi model comparison (Hatipkarasulu and Tmay,
2011(1)), tip resistance value correction (Tmay and
Hatipkarasulu, 2011), effective area ratio assumption
(Hatipkarasulu and Tmay, 2011(2)), and data reduction effects
(Tmay and Hatipkarasulu, 2012). The method uses a non-
traditional approach which takes into account the probability of
having each soil type with depth. The use of soil classification
(U) and in-situ behavior (V) indexes for organic profiling
extends the same concept of accounting for the chance of
having organic material at any given depth. The organic content
indicator (V-U) profiling shows reliable performance in clayey
materials and its further evaluation is imperative for non-clay
organic soils.
5 ACKNOWLEDGEMENTS
This study rests on the interpretations of the field and laboratory
investigations conducted for the Mississippi River Long
Distance Sediment Pipeline Project designed and undertaken by
the Louisiana Department of Transportation and Development
(LADOTD). The contents of this paper reflect the views of the
authors, who are responsible for the facts and the accuracy of
the data presented herein, and do not necessarily indicate
official assessments of the agencies, firms and institutions with
which the authors are affiliated.
6 REFERENCES
ASTM D5778-12. 2012. Standard Test Method for Electronic Friction
Cone and Piezocone Penetration Testing of Soils. ASTM
International, West Conshohocken, PA. DOI: 10.1520/D5778-12.
http://www.astm.org/Standards/D5778.htm
ASTM D2487. 2011. Standard Practice for Classification of Soils for
Engineering Purposes (Unified Soil Classification System). ASTM
ASTM D2216-10.2010. Standard Test Methods for Laboratory
Determination of Water (Moisture) Content of Soil and Rock by
Mass. ASTM International, West Conshohocken, PA. DOI:
10.1520/D2216-10.
ASTM D4318-10. 2010. Standard Test Methods for Liquid Limit,
Plastic Limit, and Plasticity Index of Soils. ASTM International,
West Conshohocken, PA. DOI: 10.1520/D4318-10.
ASTM D2974-07a. 2007. Standard Test Methods for Moisture, Ash,
and Organic Matter of Peat and Other Organic Soils. ASTM
ASTM D1140-00. 2006. Standard Test Methods for Amount of
Material in Soils Finer than No. 200 (75-m) Sieve. ASTM
Douglas, J. B., and Olsen, R. S. 1981. Soil Classification using Electric
Cone Penetrometer, Symposium on Cone Penetration Testing and
Experience, Geotechnical Engineering Division, ASCE, St. Louis,
MO, USA, (1981), 209-227.
HatipKarasulu, Y. and Tumay, M.T. 2011(1). Practical visual
presentation approach for CPT-based soil characterization and
modelling. Proceedings of the ASCE Geo-Frontiers Conference:
Advances in Geotechnical Engineering, Dallas, TX, pp. 2387-2396
Hatipkarasulu, Y. and Tmay. 2011(2). Impact of effective area ratio
assumption on PCPT-based soil classification. Proceedings of the
Fourth International Conference on Site Characterization, ISC4:
Geotechnical and Geophysical Site Characterization, Porto de
Galingas Pernambuko, Brazil. pp. 275-282
Robertson, P. K., Campanella, R. G., Gillespie, D., and Greig, J. 1986.
Use of Piezometer Cone Data. Proceedings of the ASCE Specialty
Conference on In Situ86: Use of In Situ Tests in Geotechnical
Engineering, Blacksburg, Virginia, pp. 1263-1280.
Robertson, P. K. 1990. Soil Classification using the Cone Penetration
Test. Canadian Geotechnical Journal, Vol. 1, No. 27, pp. 151-158.
Schmertmann, J.H. 1978. Guidelines for Cone Penetration Test,
Performance and Design. Report No. FHWA-TS-78-209, U.S.
Tmay, M. T. and HatipKarasulu, Y. 2011. Impact of Using Measured
v. Corrected Tip Resistance Values in PCPT-Based Soil
Characterization and Modeling, Proceedings, of the ASCE Geo-
Frontiers Conference: Advances in Geotechnical Engineering,
ASCE Special Publication No. 211, Dallas, TX, pp. 2544-2553.
Tmay, M. T., HatipKarasulu, Y., Mynarek, Z., and Wierzbicki, J.
2011. Effectiveness of CPT-Based classification methods for
identification of subsoil stratigraphy. Proceedings of the 15th
European Conference on Soil Mechanics and Geotechnical
Engineering, Athens, Greece, pp.91-98.
Tmay, M. T. and HatipKarasulu, Y. 2012 Effects of data smoothing
and reduction on CPT-based probabilistic soil classification.
Proceedings of the Fourth International Conference on Site
Characterization, ISC4: Geotechnical and Geophysical Site
Characterization, Porto de Galingas Pernambuko, Brazil. pp. 843-
850
Tmay, M.T., Hatipkarasulu, Y., Marx, E.R and Cotton, B. 2012. Multi
Model Subsurface Evaluation for Louisiana I-10 Bridge
Replacement Using Electronic CPT. Proceedings of the Fourth
International Conference on Site Characterization, ISC4:
Geotechnical and Geophysical Site Characterization, Porto de
Galingas Pernambuko, Brazil. pp. 1281-1288
Zhang, Z., and Tumay, M.T. 1999. Statistical to Fuzzy Approach
toward CPT Soil Classification. ASCE Journal of Geotechnical and
Geoenvironmental Engineering, Vol. 125, No. 3, pp. 179 186.
637
Geotechnical Challenge for Total Cost Reduction related to Construction
of Connecting Bridge with Pile Foundations
Dfi gotechnique pour la rduction totale des cots lis la construction du pont de liaison
avec les fondations sur pieux
Yasufuku N., Ochiai H.
Kyushu University, Fukuoka, Japan
Maeda Y.
West Nippon Expressway Company Limited, Osaka, Japan
ABSTRACT: Changes of geotecnical engineering profile are briefly mentioned based on the density of in-situ investigations and
laboratory tests. Then, the method used for evaluating the vertical bearing capacity of driven piles in the actual design is presented..
The applicability is also verified by comparing the predicted results with the results from the full-scale pile load tests, whose results
were linked with the reduction of the safety factor for design. Finally, the significance of geotechnical investigations including in-situ
and laboratory tests and full scale pile load tests are discussed in terms of the cost performance of the construction of pile foundations
for supporting the connecting bridge. It is concluded that in-situ and laboratory investigation with reasonable geotechnical
considerations can reduce the total cost of the construction of the bridge with pile foundations for New-Kitakyushu airport.
RSUM : Dans ce papier, la politique de base et des concepts pour des tudes gotechniques et de conception fondation sur pieux du
pont qui relie pour la Nouvelle-Kitakysh sont introduits. Les changements de profil gotechnique sont brivement mentionns bass
daprs la densit du terrain (in-situ) et des essais au laboratoire. Ainsi, la mthode utilise pour lvaluation de la capacit portante
des pieux battus conu selon la mthodologie actuelle est prsente sur la base des considrations gotechniques. L'applicabilit est
galement vrifie en comparant les rsultats prdits avec les rsultats des essais en vraie grandeur de chargement de pieux. Les
rsultats ont t compars en termes de rduction du facteur de scurit utilis au dimensionnement. . Enfin, l'importance des tudes
gotechniques y compris les essais in situ et en laboratoire et les essais en vrai grandeur de chargement de pieux sont discuts en
termes de performance des cots de la construction des fondations sur pieux pour soutenir le pont de liaison.
KEYWORDS: cost reduction, field investigations, pile foundations design, bearing capacity

1 INTRODUCTION
Connecting Bridge
2.1km, 24 piers
Aug. 2002
Airport site

Figure 1. Overview of connecting bridge under construction
A connecting bridge has been constructed on the sea as an
access road for New Kitakyushu airport, which will be opened
in 2005. The length of the bridge is about 2km and 24 piers are
mounted for supporting the bridge. An overview of the
connecting bridge under construction is shown in Figure 1. In
order to clarify the geological and mechanical characteristics of
the ground for supporting the bridge and the manmade airport
island, a large number of in-situ and laboratory tests had been
performed for five years from 1991 to 1995.
In this paper, the basic policy and concepts for geotechnical
investigations and design of this project are introduced. The
changes of geotecnical engineering profile are briefly
mentioned based on the density of in-situ investigations and
laboratory tests. The process of producing a model ground for
design is also made clear, which is used for estimating the
bearing capacity of driven piles. Further the method used for
predicting the vertical bearing capacity of driven piles is
presented based on the geotechnical considerations. The
applicability is also verified by comparing the predicted results
with the results from the full-scale pile load tests, whose results
are linked with the reduction of the safety factor for design.
Finally, the significance of in-situ investigations and full scale
pile load tests are discussed in terms of the cost performance of
the construction of pile foundations for supporting the
connecting bridge.
Design of foundation
Assessments based on
Geotechnical considerations
Investigation design
Select of possible
models
Decision of sort and
number of field & lab.
tests
Modeling of ground
Careful selection of soil parameters
Determination of adequate model
Reconsideration of
safety factors et al.
Implementation of rational and
Economical design in total
Verification by
site investigations
Full scale load tests
Field observations
(Collaboration)
Implementation of
site investigation
(Feedback) (Reassessment)

Figure 2. Collaboration of geotechnical investigations with design
2 GEOTECHNICAL INVESTIGATIONS AND DESIGN
Figure 2 shows the policy and concept of geotechnical
investigation and design for constructing the connecting bridge
638
th
T
0
-1
-2
-3
-4
-5
-6
-7
-8
-9
Sch
Sch
Ac
Ac Ac
DLc
DLc
Schw
Ttc
Ttc
DLs
DUs
DUc
DLs
DLc
TP
0
-10
-20
-30
-40
-50
-60
-70
-80
-90
(c) 1994-95
P22
Airport ar ea Sea ar ea Land area
E
l
e
v
a
t
i
o
n

(
m
)
A2 P12
P1
Unconformity
Figure 3. Final geotechnical engineering profile obtained
for New-Kitakyushu airport. As shown in this figure, the field
and laboratory investigations and the engineering design are
conducted based on the clear policy, which includes that:
1) The strong collaboration between geotechnical investigators
and designers should be made for a rational design and
construction in pile foundations.
2) The design parameters should be determined based on the
geotechnical considerations, which reflect the results obtained
from the geotechnical investigations and laboratory soil tests.
The model for estimating the bearing capacity of piles in design
should be based on the geotechnical considerations.
3) A rational bearing stratum should be carefully selected based
on the geological and geotechnical investigations.
4) The predicted performance in design should be checked by a
full-scale model tests as much as possible. The results are
reflected to the reduction of factor of safety for design.
Such policy seems to be strongly linked with the
performance based design, which may become the mainstream
in foundation design near future.
3 GEOTECHNICAL ENGINEERING MAP FOR DESIGN
REFLECTED THE SOIL PROPERTY
3.1 Geological profile with increases of site investigation
Figure 3 shows the final geotechnical engineering profile
mainly by the field investigations from 1992 to 1995, which
covers the land, sea and airport areas. Figure 3 was drawn by
adding the boring data in each pier of the access road, where the
total number of borings became more than 65 with 3500m in
total length, and the geological investigations on the diatom
earth and also volcanic ash deposit with the results of the
seismic exploration. The boring densities of each area in 1992,
1993 and 1995 are roughly 350m, 180m and 70m respectively.
It is judged that the geotechnical engineering profile becomes
more precise with the increasing boring density and quality of
in-situ investigations. The accuracy of geotechnical
investigations is believed to lead to the economical and rational
design and construction, even if the percentage of investigation
cost to the total one might be somewhat increased (see Table 2).
Figure 3 clarified that 1) the investigated ground consists of
alluvial clayey layers with 7-9m thickness and Pleistocene
layers laminated by sandy and clayey soils with 20-60m
thickness below the alluvial layers and also weathered
crystalline schist as the base layer. The corresponding ground is
therefore roughly divided into 3 layers. 2) The undulation of
base layer is extremely high in which the difference becomes
more than 45m. 3) The structure of Pleistocene layers is
complicated and the continuities in horizontal direction are not
so clear, and so the lens shape layers are found here and there.
4) The surface of unconformity in Pleistocene layers is clearly
found from sea area to airport area of which inclination is about
15 degrees in the longitudinal direction.
Embankment
Ac
DLs
1
DLc
1
DLs
2-5
DLc
2-6
Sch
DUs
DUc
Dutf

Figure 5. Model geotechnical engineering map for design
3.2 Model geotechnical engineering map for design
0 50
a) DLs layer
150 200 0 50 100 150 200
-10
-20
-30
-40
-50
-60
30 20
b) DLc layer
E
v
e
l
a
t
i
o
n
,
T
.
P
.
100
Bearing stratum Bearing stratum
Measured N-values
Averaged N-values
Range of dispersion
N-blow values N-blow values

Figure 4. Distribution of N-values in DL layer against depth

When determining a good bearing stratum for pile foundation,
Japanese design code by Japan Road Association recommends
that the N-values of sandy or sand-gravel layers are grater than
30 blow counts, and also N-values of clayey layers are more
than 20. Figure 4 shows the characteristics of N-values in
Pleistocene sandy and clayey layers obtained from the SPT. The
N-values of both layers tend to become more than 30 in average
when the depth is roughly deeper than 30m T.P. level. Based on
the results, the following guideline for pile foundation design
was determined such that: 1) The layer at 30m T.P level was
judged as an effective bearing stratum for driving the pile
foundation. A steel pipe sheet-pile foundation was selected as a
type of pile foundation in this project, where, all of pile tips are
set up in Pleistocene laminated ground at around 30m T.P.
levels. 2) As shown in Figures 3 and 4, the scatters of N-values
seems not to be small and also it is not easier to distinguish
from the sandy and clayey layers from N-values obtained
because the site consists of the complicated laminated sandy
and clayey layers. In this circumstances, the uniform and
empirical method based on the N-values is not rational and
precise to evaluate the pile bearing capacity. Thus, a method for
evaluating the pile vertical bearing capacity should be
introduced together with a proper geotechnical engineering map
for foundation design, which is derived by geotechnical
Table 1. Soil constants of each layer
c' ' 'cv
(tf/cm
3
) (tf/m
2
) (degs.) (degs.)
Ac 0.0 0.53 0.292 0.0 33.0 1
Volucanic Dutf 11.0 0.66 0.6 30.0 33.7 1-6
Sandy DUs 30.4 0.90 0.0 37.0 34.6 1-2
Clayey DUc 0.0 0.53 8.1 24.0 36.9 1-6
DLs
1
17.0 0.90 2.6 35.4 35.7 1-2
DLs
2-5
40.0 0.53 5.5 32.6 35.7 1-2
DLc
1
27.0 0.94 2.6 34.8 35.5 2-8
DLc
2-6
32.4 0.97 4.4 29.6 36.4 2-8
Gravel DLg 47.5 0.99 0.0 36.0 36.0 1
Sch-w 29.7 0.90 5.7 22.7 - 1
Sch 98.3 0.90 5.7 22.7 - 1
* OCR i t l d d t th d th
Clayey
Metamorphic rocks
OCR*
'
Sandy
N-value
Strength parameters
Alluvial clay
Pleistocene
(Upper)
Pleistocene
(Lower)
639
considerations based on the results of the large numbers of in-
situ and laboratory tests. The resultant geotechnical engineering
map and the soil constants of each layers as characteristic values
are summarized in Figure 5 and Table 1, in which the soil
constants are mainly obtained by the standared consolidation
and triaxial undrained and drained compression tests. 3) Full-
scale pile load tests are conducted to confirm the validity of the
predicting method used for foundation design. The possibility of
reducing the safety factor for design to 2.5 from 3.0 is
considered through the geotechnical point of view based on the
field investigations, laboratory test results and the accuracy of
the predicting method with full scale pile load tests.
4 EVALUATION OF VERTICAL BEARING CAPACITY
OF DRIVEN PILES
4.1 Basic idea
Specification for Highway Bridge gives a following equation as
an estimating method of the ultimate pile bearing capacity based
on the results of the field and laboratory investigations (JRA,
1996):
A q f L U R
d i i u
- (1)
Where R
u
: ultimate bearing capacity of pile, A: pile tip area, q
d
:
pile end bearing capacity, U: pile circumference, L
i
: thickness in
each layer, f
i
: maximum skin friction of pile. The first and
second terms are related to the skin friction of pile and pile-tip
bearing capacity, respectively. However, the main part of the
vertical bearing capacity of a pile is often mobilized from the
skin friction in practical designs within the limits of allowable
displacement, because relatively large displacements are needed
to mobilize the end bearing capacity. In addition, as a normal
open-end pile is used as a type of pile foundation, the end
bearing capacity strongly depends on the degree of the blockade
effect and thus the precise prediction of the end bearing capacity
was considered to be quite difficult. Then, as shown in Figure 6,
the skin friction mobilized through the internal face of the pile
under the bearing stratum was assumed as the equivalent end
bearing capacity in the design. Therefore, the second term q
d
A
is expressed as UALf
i
.
Skin friction
mobilized
Steel pile with diameters of 1m
Bearing stratum
Penetration
depth AL
Skin Friction mobilized here
is assumed as pile-tip resistance

Figure 6. Basic idea of pile bearing capacity
4.2 Evaluation of skin friction
4.2.1 Basic equation
The following basic equation is therefore used for calculating
the skin friction of piles which is determined as the sum of pile
to soil adhesion and friction components:
' ' '
tan
o o
o o
h
c f - (2)
c
o
and o'
o
are the adhesion and friction parameters between pile
and soil, and o'
h
is the effective lateral stress acting on the pile.
4.2.2 Soil constants as characteristic values
An idea that the adhesion between pile and soils is roughly
equal to the apparent cohesion of soils c is widely used for a
practical design. It is mentioned that the applicability of this
idea is effective, irrespective of type of soils such as clay and
sand (e.g. Tomlinson 1980). Therefore, c
o
in eq. (4) was
assumed to be equal to the apparent cohesion c of soils. In
practical design, the axial pile capacity is estimated for the
settlements of approximately 10% of the pile diameter. The
10% settlements usually exceed those for mobilizing the
maximum skin friction of pile. Further, when considering that
the mobilized mechanism of skin friction between pile and soils
surrounding the pile, it is reasonable to use the friction angles at
the critical state corresponding to sufficiently large
displacement o
cv
as

o
o
(Yasufuku et al. 1997). Here o
o
is
assumed to be conservatively two-third of o.

o
o
is thus given
by
'
3
2
' o o
o
(3)
where, o: effective friction angle at peak strength state.
4.2.3 Coefficient of lateral effective stress K
The mobilization of the skin friction is dependent on the lateral
effective stress o'
h
and thus in turn is dependent on the
overburden pressure o'
v
. When considering o'
h
is given by
, Eq.(2) is rewritten by
v
K ' o
' ' '
tan
o o
o o
v
K c f - (4)
K is a coefficient of lateral effective stress and o
v
is vertical
effective stress. The coefficient of lateral effective stress K was
estimated from the previous research findings related to the K
0
-
value. K-values in Pleistocene clayey layers were determined by
the following equation (Mayne and Kulhawy, 1982).
( )
' sin
' sin 1
o
o OCR K - (5)
where, OCR is over-consolidation ratio defined as the ratio of
the consolidation yield stress p
c
to the overburden pressure o'
v
.
Values of OCR, o in average and the calculated K-values in
Eq.(5) are measured against elevation. We can say that applying
this equation into the Pleistocene clayey layers, most of K-
values became more than 1.0. Based on the experimental
evidence, K-value for design was decided as 1.0, irrespective of
type of Pleistocene layers. Thus, the presented model for
1.0m
D
r
i
v
e
n

P
i
l
e

(
L
=
2
9
m
)
P12 site
DLs
DLs
Acl
(a)
D
e
p
t
h

z

(
m
)
0
5
10
15
20
25
30
DLc
DLc
DLs
Strain
gage
(b)
S1
S2
S3
S4
S5
S6
S7
S8
1.0m
D
r
i
v
e
n

P
i
l
e

(
L
=
2
9
m
)
P12 site
DLs
DLs
Acl
(a)
D
e
p
t
h

z

(
m
)
0
5
10
15
20
25
30
DLc
DLc
DLs
DLs
DLs
Acl
(a)
D
e
p
t
h

z

(
m
)
0
5
10
15
20
25
30
DLc
DLc
(a)
D
e
p
t
h

z

(
m
)
0
5
10
15
20
25
30
D
e
p
t
h

z

(
m
)
0
5
10
15
20
25
30
0
5
10
15
20
25
30
DLc
DLc
DLc
DLc
DLs
Strain
gage
(b)
Strain
gage
(b)
S1
S2
S3
S4
S5
S6
S7
S8
0 2 4 6 8 10
Axial force
(MN)
(d)
0
5
10
15
20
25
30
0 20 40 60
N-Values
D
e
p
t
h

(
m
)
(c) (a)

Figure 7. Soil profile, N-values and measured axial force in pile
load test at P12 site
evaluating the vertical bearing capacity is expressed as
i i i u
Lf U f L U R A - (6)
640
th
'
3
2
tan 0 . 1
' '
z c f
i
(7)
where, z is an arbitrary depth from the surface and L is a
penetration depth from the bearing stratum (see Figure 6).
5 FULL-SCALE PILE LOAD TESTS AND THE
REDUCTION OF FACTOR OF SAFETY
In order to verify the applicability of the presented model and to
confirm the characteristics of the pile bearing capacity of each
layer, full scale pile load tests were conducted at two
representative sites, which locate at 12P and 22P sites shown in
Figure 3. As an important engineering judgment in this project,
the reduction of the factor of safety from 3.0 to 2.5 for pile
foundation design was discussed through comparing the
predicted results with the results of full scale pile load tests.
Figure 7 shows the soil profiles and N values with depth for
12P site. N-values can be seen to widely change with depth
from nearly zero to more than 20 and also N-values at pile tips
are roughly 30. The steel piles with a diameter of 1.0m were
carefully driven using vibration and hydraulic hummers. The
effective length of each pile was about 30m. Tests were
conducted based on the multi-cycles method, which is
recommended by the JGS (1993). Four strain gauges were
located at each of the cross sections as shown by the dots in
Figure 7.
Figure 8 shows the comparison of the estimating total
vertical bearing capacities with those of full-scale pile load tests
at 12P and 22P sites, in which Eqs. (6) and (7) was used to
calculate the predicted values. The bearing capacity calculated
by the empirical model based on the measured N-values
recommended by JRA is also depicted in this figure. The model
used here can reasonably estimate both total skin friction and
pile tip resistance at both sites, comparing with those from JRA
recommendation. As shown in Table 1 and Figure 3, we have a
clear grasp of the soil characteristics values for each layer and a
practically efficient geotecnical profile. Therefore, the model
can apply very well to evaluate the pile bearing capacity
according to the ground profile at each site, with the
consequence that the accuracy of the prediction clearly
increased and these facts became an important evidence to
reduce the factor of safety for pile foundation design from 3.0 to
2.5.
6 EFFECT OF A REDUCTION IN TOTAL COSTS
The comparison of the cost performance in terms of the
construction of pile foundations driven in P1 to P24 sites is
summarized in Table 2, which is a result of trial calculation.
Note that the cost is normalized by the cost obtained by the
standard manner for evaluating the pile bearing capacity using
N-values (JRA, 1993) without any full scale pile load tests. For
comparison, the layer of the bearing stratum for each case was
assumed to be same, however, the penetration depth L was
considered to depend on the calculation manner. Total cost are
divided by 2 parts, in which one is the part for the cost related to
the geotechnical investigations which include in-situ and
laboratory soil tests, and full scale pile load tests, and the other
is related to the normalized total pile construction cost in terms
of P1 to P24 piers. The presented manner used here is expected
to cut the cost more than 15% comparing with the total cost by
the standard approach using N-values. Thus even if the cost of
the geotechnical investigations became roughly two times
higher comparing with the general manner, the appropriate in-
situ and laboratory investigation with a reasonable
considerations can reduce the total cost in the project. This is
due to the highly accurate ground profile and the proper
evaluation method of pile bearing capacity with the results of
the full scale pile load tests which reflected the decrease of
safety factor from 3.0 to 2.5. It is believed that the geotechnical
considerations and manner treated here can give an important
information for the geotechnical investigators, structural
designers and construction engineers.
(b) P22
T
o
t
a
l

s
k
i
n

f
r
i
c
t
i
o
n
4.6MN
0
2
4
6
8
10
M
e
a
s
u
r
e
d
V
a
l
u
e
s
P
r
e
d
i
c
t
e
d
V
a
l
u
e
s
N
-
V
a
l
u
e
s
T
o
t
a
l

b
e
a
r
i
n
g

c
a
p
a
c
i
t
y
9.5MN
8.4MN
P
i
l
e
-
t
i
p

r
e
s
i
s
t
a
n
c
e
10.0MN
(a) P12
T
o
t
a
l

s
k
i
n

f
r
i
c
t
i
o
n
9.19
0
2
4
6
8
10
M
e
a
s
u
r
e
d
V
a
l
u
e
s
P
r
e
d
i
c
t
e
d
V
a
l
u
e
s
N
-
V
a
l
u
e
s
T
o
t
a
l

b
e
a
r
i
n
g

c
a
p
a
c
i
t
y
6.5MN
16.2MN
Pile-tip resistance
6.4MN
9.2MN

Figure 8. Comparison of predicted total bearing capacities with
those of pile load tests
Table 2. Total cost benefit
Method by N-
values
Method
proposed here
Cost for geotechnical
investigations*
1
2.11
Construction cost for piles
1
0.82
Total cost
1
0.84
* the cost includes full scale load tests
7 CONCLUSIONS
An importance of integrating the geotechnical investigations
with pile foundation design was clarified through a case study
in terms of connecting bridge for New-Kitakyushu airport. The
following major conclusions were drawn:
1) A rational method for evaluating the pile bearing capacity
was presented which reflected the soil characteristic values and
geological environmental history. In addition, the applicability
of the presented method was confirmed through full-scale pile
load tests, with the consequence that the safety factors for pile
foundation design were reduced from 3.0 to 2.5.
2) In-situ and laboratory investigation with reasonable
geotechnical considerations can reduce the total cost of the
construction of the bridge for New-Kitakyushu airport.
8 REFERENCES
JSSMFE Standards for Vertical Load Tests of Piles. 1993. Japanese
Geotechnical Society, 113-121.
Japan Road Association. 1996. Specifications for Highway
Bridge Part IV, 330-337(in Japanese).
Mayne P.W. and Kulhawy F.H. 1982. K
0
-OCR relationships in soils, J.
Geotech. Eng. Div., ASCE, 108(GT6), 851-872.
Ochiai H. and Yasufuku N. 2003. Investigation, design and construction
of the connecting bridge for New-Kitakyushu airport. 9
th
Annual
Meeting of Chinese Soil Mechanics and Geotechnical Engineering,
1, 214-22

641
Dynamic CBR as a method of embankment compaction assessment
Dynamique CBR comme une mthode d'valuation de compactage du remblai
Zabielska-Adamska K., Sulewska M.J.
Bialystok University of Technology, Bialystok, Poland
ABSTRACT: In engineering practice, earth construction requires suitable soil compaction, usually relating to the Proctor methods.
Materials of the built-in embankment and the subgrade have their own specifications, dependant on the kind of earth structure and soil
plasticity characteristics. Care should be taken not to use compaction degree (% of maximum compaction) as the only parameter to
assess soil compaction. This applies to both cohesive soil and to fly ash whose permeability and mechanical properties dependent on
moisture content at compaction. Therefore, for these types of soils California Bearing Ratio could be used as a method of compaction
assessment being an indicator of soil bearing capacity. The CBR research was done for both static (classic) and dynamic methods on
fly ash samples without soaking them to replicate field conditions. A load of 2.44 kPa was applied to all the samples subjected to
penetrations. The dynamic CBR tests were conducted by using Light Weight Deflectometer consisting of a falling weight to produce a
defined load pulse of the CBR piston. The CBR test could be used for running compaction control during embankment erection, which
specially refers to dynamic CBR test due to the speed of research execution.
RSUM: Dans la pratique dingnierie, la construction en terre ncessite un compactage du sol adapt, concernant en gnral les
mthodes Proctor. Les matriels encastrs du remblai et de la plate-forme ont leurs propres spcifications, dpendant du genre de la
construction en terre et de caractristiques de plasticit du sol. Il faut prendre soin de ne pas utiliser le degr de compactage (% de
compactage maximum) comme le seul paramtre pour valuer la compactage du sol. Cela sapplique aux sols cohsifs et cendres
volantes dont la permabilit et des proprits mcaniques dpendent de la teneur en humidit au compactage. Donc, pour ceux types
de sol l'indice portant californien pourrait tre utilis comme une mthode dvaluation du compactage tant un indicateur de la
capacit portante. Les recherches CBR ont t effectues pour les mthodes statiques (classiques) et dynamiques sur les chantillons
de cendres voltantes sans les faire tremper reproduire les conditions de terrain. Une charge de 2,44 kPa a t appliqu toutes
chantillons soumis des pntrations. Les tests de dynamique CBR ont t effectus a laide de dflectomtre constitu par la masse
tombante pour produire une impulsion de charge dfinie du piston CBR. Le test CBR pourrait tre utilis pour excuter le contrle du
compactange lors de lrection de remblai, qui se rfre en particulier l'essai dynamique de CBR en raison de la rapidit d'excution
de la recherche.
KEYWORDS: compaction, California Bearing Ratio (CBR), dynamic CBR (CBR
d
), fly ash, compaction assessment.
1 INTRODUCTION
In engineering practice, earth construction requires suitable soil
compaction, usually relating to the Standard and Modified
Proctor methods. Materials of the built-in road embankment and
the subgrade have their own specifications, dependant on the
kind of earth structure and soil plasticity characteristics. Care
should be taken not to use compaction degree (% of maximum
compaction) as the only parameter to assess compaction of
material in embankments. This applies to both cohesive soil and
fly ash. The permeability and mechanical properties of
compacted fly ash are dependent on moisture content present
during compaction, as are properties of cohesive mineral soils
(Turnbull and Foster 1956, Mitchell et al. 1965, Zabielska-
Adamska 2006 and 2011). Consequently different values of
geotechnical parameters are obtained for water content on either
side of the optimum water content on the compaction curve, for
the same dry densities. Thus for these types of soils California
Bearing Ratio, CBR, may be used as a method of compaction
assessment, since it is an indicator of ground bearing capacity
broadly used in the design of civil engineering.
The laboratory CBR tests by means of both static (classic)
and dynamic methods were carried out to establish relationship
between bearing ratio and fly ash compaction. Samples,
compacted by the Standard or Modified Proctor methods, were
prepared without soaking them to replicate field conditions
during earth structure erection. The dynamic CBR, CBR
d
, tests
were done by using impact generator and guide rod, which are
the parts of Light Weight Deflectometer (LWD), and additional
equipment in a CBR piston. A falling weight is to produce a
defined load pulse of the CBR piston that can be used both in
laboratory and field tests. The aim of this study was to prove
that CBR tests could be used as the methods of road
embankment or subgrade compaction assessment. This refers
especially to CBR
d
test which may be used for running
compaction control during embankment erection due to the
speed of research execution, as well as Light Weight
Deflectometer (Sulewska 2012).
1 LITERATURE REVIEW
California Bearing Ratio, CBR, is expressed as the percentage
ratio of unit load, p, which has to be applied so that a
standardized circular piston may be pressed in a soil sample to a
definite depth with a rate of 1.25 mm/min and standard load,
corresponding to unit load, p
s
, necessary to press the piston at
the same rate into the same depth of a standard compacted
crushed rock.
(1) % 100
s
p
p
CBR
CBR value is used for evaluation of the subgrade or subbase
strength, and may be applied to assess the resistance to failure
642
th
or indicate the load-carrying capacity. It should be noted here
that CBR values in pavement design do not reflect the shear
stresses that are generated due to repeated traffic loading. The
shear stress depends on many factors; none of them is fully
controlled or modelled in CBR test (Rico Rodriquez et al. 1988,
Brown 1996).
In laboratory, CBR penetration test is performed on material
compacted in a specified mould and placed in loading machine
equipped with a movable base that rises at uniform rate used in
forcing the penetration piston into the specimen. Tested
specimens are penetrated directly after compaction or are to be
previously soaked. CBR test in-situ is carried out with a
mechanical screw jack for continuous increase of the applied
load to the penetration piston. A reaction forcing the penetration
piston into the soil is provided by a lorry equipped with a metal
beam and attachments under its rear.
The dynamic CBR, CBR
d
, test can be performed both in
laboratory and in situ. The test can be conducted as an
alternative to the static CBR test, especially due to the short
period of time required. CBR
d
advantage, compared with the
classic CBR, is the elimination of a loading frame necessary in
static loading. The CBR
d
test is carried out with the use of Light
Weight Deflectometer, where a falling weight is used to
generate a defined load pulse on the CBR piston. CBR
d
is
calculated on the basis empirical formula (Zorn 2002) as:
(2)
where 87.3 is the number standing as a value of dynamic
loading including empirical coefficient, and s is the settlement
in millimetres. CBR
d
is recommended to specify when it is
greater or equalled 20% and is equalled or lower than 150%.
Turnbull and Foster (1956) carried out broad studies on CBR
for compacted mineral soils. They determined penetration
resistance of unsoaked samples of lean clay, compacted by
means of four different energy values and at different moisture
contents. It was proved that the CBR value for compacted clay
is a function for both water content as well as dry density.
Compacted samples reached higher CBR values when higher
energy values were applied. Moisture increase of compacted
samples decreased CBR value and in cases of compacted
samples with moisture contents greater than optimum water
content, penetration resistance was close to zero. Soaking of
samples caused the decrease of CBR value, quite significant in
compacted samples dry of optimum, less significant at
optimum water content. The smallest decrease was observed in
samples compacted at wet of optimum. Rodriguez et al. (1988)
described CBR dependence on compaction parameters
moisture contents and dry densities, as well as on conditions of
compaction energy and methodology of compaction. The
authors point to the fact that the CBR value of the soil
compacted with higher energy value may be lower than that
resulting from the compaction with lower energy value. CBR
dependence on moisture in the process of compaction was
confirmed in the course of studies conducted by Faure and
Viana Da Mata (1994). The authors straightforwardly claim that
dry density resulting from the compaction of a sample does not
have any impact on CBR value which, on the other hand is
influenced by moisture present in the process of compaction.
CBRs relationship with moisture content was also observed in
the case of compacted marl from Saudi Arabia (Aiban 1995),
where marl was subjected to tests at moisture optimum and
moisture on the dry and wet sides of optimum. Moisture
density curves and CBR(w) dependency curves were said to be
similar; the highest CBR values were obtained at optimum
moisture. The studies of the samples tested immediately after
compaction and the soaked samples confirmed that the effect of
soaking is decreased when the samples are compacted at
moisture greater than optimum.
Zabielska-Adamska (2006 and 2011) concluded that the
highest CBR values for unsoaked samples of fly ash (class F)
appear in modified compaction in case of moisture level
below optimum, and in standard compaction in case of
moisture level within or slightly below optimum. In saturated
samples, the highest values for bearing ratio CBR are present in
moisture level equal optimum for both compaction energy
levels. Once optimum moisture is exceeded, CBR value drops
dramatically, regardless of the compaction energy and method
of preparation of samples, soaked or unsoaked. High moisture
results in the loss of contact among fly ash grains. Hence CBR
value dependence on moisture level of fly ash is quite apparent.
CBR of samples compacted by means of modified method for
optimum moisture is almost twice as high than in the case of
optimum compaction by standard method, which points to a
significant influence of compaction energy and dry density. It is
interesting how compaction energy influences CBR in samples
of the same level of moisture, compacted, however, with the use
of different energies. Ash samples with moisture value w,
compacted by Proctor modified compaction, where w>w
opt1
,
show far lower CBR than samples of the same moisture level w,
but compacted by standard method where w<w
opt2
. The lowest
CBR in the analysis of various samples of fly ash was obtained
in case of fly ash of the finest graining which influences
increase of optimum moisture and decrease of density of solid
particles. Zabielska-Adamska and Sulewska (2009) studied
relationships between CBR and analysed parameters of various
samples of fly ash by means of Artificial Neural Networks
(ANNs) and as a result concluded that the most relevant
variables were
d
and relation w/w
opt
, which confirms the fact
that optimum water content and moisture content at compaction
are the most significant parameters in CBR. Dry density, as
another significant parameter, should be considered as dominant
when comparing CBR values for different fly ash shipments
compacted with the use of different energies.
(%
The results of the dynamic CBR are extremely poorly
represented in the literature, which is probably due to a low
prevalence of this method in the world. The first study of CBR
d
,
done on the road mineral materials, were presented by
Weingard et al. (1986). A good correlation between test results
was obtained using static and dynamic method. A study
conducted by Schmidt and Volm (2000) is the only one known
to the authors of this paper which presents results of research
with CBR
d
carried out on cohesive soil with different
compaction. The studies were conducted for silty clay with
moisture content grade from 11 to 18%, and optimum water
content established as 15.6%. As a result of laboratory studies,
the researchers obtained two curves CBR
d
(w) and CBR(w),
shifted in relation to each other by approximately 57%. In case
of moisture content greater than optimum, the difference
between static values and dynamic values changed to approx.
9%. Higher bearing ratio was obtained in dynamic studies.
CBR
d
is recommended for control research in embankment
erection with the use fine grained soils compacted at moisture
contents lower than optimum.
2 LABORATORY TESTS
All the tests were conducted on the basis of fly ash from hard
coal burning in Bialystok Thermal-Electric Power Plant, stored
at a dry storage yard. The fly ash shipment corresponded in
graining to sandy silt. Physical parameters are shown in Tab. 1.
The laboratory CBR tests were carried out to establish
relationship between bearing ratio and fly ash compaction. The
tested samples were compacted by two methods: the Standard
Proctor and the Modified Proctor at moisture contents within
the range of w
opt
5% for each compaction method. The fly ash
samples were saturated 24 hrs prior to the test so that their
moisture content could increase by approx. 2.5%. After that,
they were deposited in sealed containers. Each compaction
curve point was designated on a separate sample. During the
compaction tests, individual samples of fly ash were used only
)
3
9
CBR
. 87
5 . 0
s
643
once, otherwise they could not be regarded as representative
(Zabielska-Adamska 2006). The CBR tests were conducted on
unsaturated samples. All the samples subjected to penetration,
tested both methods static and dynamic, and were loaded with
ASTM 1883-73 recommended load of 2.44 kPa. The static
(classic) CBR research was done on fly ash samples directly
after compaction. Higher CBR value was accepted as a result
D
50
(mm)
s
(g/cm
3
) C
U
= D
60
/D
10
C
C
=D
30
2
/D
60
D
10
0.0550.065 2.110.01 3.894.25 0.941.03
Modified Proctor method Standard Proctor method
1 opt
w (%)
max1 d
(g/cm
3
)
opt2
w (%)
max2 d
(g/cm
3
)
37.00 1.068 45.50 1.009
Table 1. Geotechnical parameters of tested fly ash shipment.
Tested fly ash parameters
Figure 1. CBR research (from the left): static test;
changed mould basis and prepared mould extension for
dynamic CBR; specimen ready for dynamic test (photo
by Zabielska-Adamska).
calculated on the basis of pressing piston resistance, represented
in a given depth 2.5 or 5.0 mm. Next, after levelling off the
surface of the same specimen and replacement of basis of the
mould, dynamic CBR was carried out. The CBR
d
tests were
conducted using Light Weight Deflectometer (LWD) consisting
of a falling weight (7.07 kN) vertically movable along the guide
rod to produce a defined load pulse (3.6 MN/m
2
) of the CBR
piston. Electronic measurement system gauged the depth of the
pistons penetration in the tested soil after a single impact. CBR
tests are shown in Figure 1.
Figure 2 represents the results of standard and dynamic CBR
testing, depending on moisture content at compaction, in
relation to compaction curves of fly ash, obtained by means of
two Proctor methods. Static CBR results confirm earlier results
obtained by the author. CBR of unsaturated samples of fly ash
reaches the highest values in the case of samples compacted at
the moisture content lower than optimum. The samples
compacted above optimum water content have still lower CBR
values simultaneously with an increase of moisture content.
These relationships can be observed in both methods of
compaction standard method and modified method. However,
samples compacted with the use of modified Proctor method,
the curve CBR(w) definitely reaches maximum. The shape of
the curves CBR
d
(w) is similar to that obtained according to the
standard method CBR(w). In the case of modified compaction,
curves CBR
d
(w) and CBR(w) are characterised by a similar
scope of moisture content; from w
opt1
5% to optimum moisture
content, w
opt1
(difference in relation to CBR up to about
2.5%). Once curve CBR
d
(w) exceeds w
opt1,
it also exceeds
standard curve, passing CBR by 16% at w
opt1
+5%. In the case of
standard compaction, at moisture level w
opt2
5%, CBR
d
value
equals CBR value. After this, as the moisture content increases
the difference also increases and when the moisture level is
equal to w
opt2
, the CBR difference is exceeded by 5%.

CBR
MP
= 1026.64 + 59.09w 0.82w
2
CBR
d MP
= 467.44 + 27.46w 0.37w
2
CBR
SP
= 297.17 + 16.96w 0.22w
2
CBR
d SP
= 519.42 + 25.88w 0.30w
2
(R
2
=0.8751)
(R
2
=0.7768)
(R
2
=0.8047)
(R
2
=0.8468)
Figure 2. CBR test results versus moisture content at compaction in
comparison with compaction curves: MP Modified Proctor method,
SP Standard Proctor method, CBR static test results, CBR
d

dynamic test results.
With further increase of moisture content, the difference may be
as high as 13%. Significant differences in the results of the
studies carried out by means of static and dynamic methods, at
moisture level exceeding w
opt
originate from the differences in
speed of loading and lack of possibility of pore pressure
644
th
dissipation in the case of impact loading. Similar observations
can be made during studies on the influence of penetration ratio
on the resistance of saturated clayey soils in cone penetration
tests (Kim et al. 2008).
Figure 3 presents dependence of static and dynamic CBR on
dry density. It can be seen in Figure 3 that there are points
standing out, with the coordinates (
d
, CBR) obtained in the case
of standard method at moisture content higher than optimum by
at least 2.5%, and in modified method higher by at least 5%.
This is the result of dependence of mechanical parameters of fly
ash on moisture content in the process of compaction. Once
these points are excluded, statistically valid relationships -
CBR(
d
) can appear, especially in the case of CBR
d
values,
where for value CBR
d
(
d
) coefficient of determination
R
2
=0.8675 was obtained (Fig. 4). CBR
d
dependence on CBR is
also statistically valid. Equation CBR
d
=17.28+0.52CBR
explains 84.9% of variance in the value of statistic CBR.
Figure 3. Relationship between CBR value and dry density with an
indication the points obtained at moisture contents at compaction
w=w
opt
+(2.55%): MP Modified Proctor method, SP Standard
Proctor method, CBR static test results, CBR
d
dynamic test results.
Figure 4. CBR value versus dry density excluding the points obtained at
moisture contents at compaction w=w
opt
+(2.55%), along with 95%
confidence interval.
3 CONCLUSIONS
1. The dynamic CBR method, as well as static (classic)
method can be used to assess compaction of fly ash and
cohesive soils embedded in subgrade or layers of
embankment. The results of studies of CBR
d
, and CBR, are
closely connected with the characteristics of compaction.
2. The current compaction quality control of fine grained
anthropogenic ground conducted through CBR
d
tests with
the use of Light Weight Deflectometer producing a defined
load pulse of the CBR piston is recommended in the cases
of embedded material at moisture contents equal optimum
or lower. CBR
d
studies of anthropogenic ground compacted
at moisture levels exceeding optimum water content may
lead to overstating of the test results due to lack of pore
pressure dissipation after impact ground loading.
3. Dynamic CBR test, using Light Weight Deflectometer,
should be widely used due to its speed and ease of research
as an alternative method to classic method of quality
control in compaction process or assessment of subgrade
bearing capacity.
4 ACKNOWLEDGEMENTS
This work, carried out in 2012 at the Bialystok University of
Technology, was supported by Polish financial resources on
science. The authors gratefully acknowledge the assistance and
cooperation of M. Piasecki and D. Tymosiak who performed the
laboratory tests.
5 REFERENCES
Turnbull W.J. and Foster Ch.R. 1956. Stabilization of materials by
compaction. Journal of Soil Mechanics and Foundations Division
82 (SM2), 1-23.
Mitchell J.K., Hooper D.R., Campanella R.G. 1965. Permeability of
compacted clay. Journal of Soil Mechanics and Foundation
Division 91 (SM4), 41-63.
Zabielska-Adamska K. 2006. Fly ash as a material for constructing
sealing layers. Publishing House of BTU, Bialystok. (in Polish)
Zabielska-Adamska K. 2011. Fly ash as a barrier material. In:
GeoFrontiers 2011 Advances in Geotechnical Engineering,
ASTM STP 211, ASTM Int., PA, 947-956.
Sulewska M.J. 2012. The control of soil compaction degree by means of
LFWD. Baltic Journal of Road and Bridge Eng. 7(1), 36-41.
Brown S.F. 1996. Soil mechanics in pavement engineering.
Gotechnique 46 (3), 383-426.
Zorn. 2002. Operating Manual. Light Drop-Weight Tester ZFG 05 for
the dynamic CBR test and the dynamic plate loading test. Gerhard
Zorn Mechanische Werkstatten, Stendal.
Rico Rodrigues A., del Castillo H. and Sowers G.F. 1988. Soil
mechanics in highway engineering. Trans Tech Publication,
Clausthal-Zellerfeld.
Faure A.G. and Viana Da Mata J.D. 1994. Penetration resistance value
along compaction curve. Journal of Geotechnical Engineering 120
(1), 46-59.
Aiban S.A. 1995. Strength and compressibility of Abqaiq marl, Saudi
Arabia. Engineering Geology 39 (3-4), 203-215.
Zabielska-Adamska K. and Sulewska M.J. 2009. Neural modelling of
CBR values for compacted fly ash. In: Proc. 17th Intern. Conf. on
Soil Mechanics and Geotechnical Engineering, Eds. M. Hamza, M.
Shahien, Y. El-Mossallamy, IOS Press - Millpress, Vol. I, 781-784.
Weingart W., Hanebutt J. and Rummert W. 1990. Dynamic laboratory
and field testing device for determination of the CBR value of
mineral concrete. Die Strasse 26 (2), 48-51. (in German)
Schmidt H.-H. and Volm J. 2000. Dynamic CBR test new method of
embankment quality control. Geotechnik 23 (4), 271-274. (in
German)
Kim K., Prezzi M., Salgado R. and Lee W. 2008. Effect of penetration
rate on cone penetration resistance in saturated clayey soils. Journal
of Geotechnical and Geoenvironmental Eng. 134 (8), 1142-1153.
dynamic
static

International Soil Mechanics and Geotechnic Conference Proceedings

Загружено:

Сведения о документе

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

International Soil Mechanics and Geotechnic Conference Proceedings

Загружено:

Авторское право:

Доступные форматы

Technical Committee 102

Ground Property Characterisation

diagram with the extra rheological

technique, or by the ROTOSTAF

dynamic penetrometer to improve its

penetrometer to assess the

penetrometer in order to characterize

, equal to 20% of the random field standard

Figure 1. Photograph of the Probe

The chamber was filled up to 1m mark with the silty sand

Figure 3. Extrapolation of the Stream Power to Assess

is the secondary compression index, and t/t

Вам также может понравиться