Вы находитесь на странице: 1из 12

RD 2011

CFRD 05

CONSIDERATIONS ON THE SEISMIC DESIGN OF HIGH CONCRETE


FACE ROCKFILL DAMS (CFRDs)
Bayardo Matern1,, Gabriel Fernandez2
1, Director Bayardo Matern Associados Ltda, Av. Giovanni Gronchi, 5445- s/172, CEP- 05724-003, S.Paulo Brazil,
e-mail: bayardo.materon@terra.com.br
2, Independent Consultant, 207 Sherwin Circle Urbana, IL 61802 USA
e-mail: gfernand@illinois.edu

ABSTRACT
The height of Concrete Face Rockfill Dams (CFRDs) has been steadily increasing during the past forty years.
During this period the height of CFRDs have increased from 110m (Cethana, Australia, 1971) to 233m
(Shuibuya, China, 2008) and higher dams are under study. CFRD design has gradually evolved from a
predominately empirical approach based on precedent, as well as field observations and measurements, to the
recent use of numerical modeling. However, analytical models to discern static and dynamic behavior have been
absent. This paper focuses on the seismic design of CFRDs taking into account documented field observations
and basic principles of seismic response. The paper summarizes the observed behavior of two dams under
significant seismic loading, the Zipingpu CFRD of compacted rockfill in China shaken by the Wenchuan Quake
(5/12/2008) and the Ishibuchi CFRD of dumped rockfill in Japan affected by the Iwate Miyasi Quake
(6/14/2008); and discusses some aspects of CFRD dynamic response including a simplified approach to evaluate
seismic induced displacements based on the concept originally proposed by N. Newmark (1965). Suggestions
are also provided on design adjustments to minimize deleterious effects on embankment behavior subject to
large seismic loads that the authors are already applying in high CFRD structures under design and construction.

Key words: CFRDs, Seismicity, Dynamic Response, Design.


1. INTRODUCTION
The stability of embankment slopes subjected to seismic loads was initially treated as a pseudo-static
problem by estimating a dynamic factor of safety which included a dynamic force that was
approximated as a fraction (5% to 20%) of the weight of the sliding mass. The meaning of the
dynamic factor of safety however was somewhat nebulous, because the magnitude of the dynamic
forces is not constant and their direction changes throughout the duration of the quake. Furthermore, a
dynamic factor of safety of 1.0 would not preclude the development of seismic induced displacements.
An estimate of the seismic-induced slope displacements after shaking is required for the engineer to
make a judgment on the potential for damage to the dam, the likelihood for a decrease in the shear
strength of the embankment materials or an unacceptable loss of freeboard. A simplified approach to
evaluate seismic-induced displacement in embankment dam was developed by N. Newmark (1965). A
procedure to evaluate seismic response of embankment dams was subsequently developed by
Ambraseys and Sarma (1967) to estimate ground motions amplifications throughout the embankment
taking into account its fundamental period of vibration and the damping of the embankment materials.
Afterwards, a method for estimating embankment earthquake-induced deformations was proposed by
Makdisi and Seed (1978), incorporating the seismic response of the embankment and using the
concept originally proposed by N. Newmark (1965) to estimate seismic-induced displacements.
Numerical modeling is extensively used in current engineering practice to assess embankment
response during seismic loading. This methodology is a useful resource which provides for a
comprehensive evaluation of embankment performance. However, it is always prudent to have

analytical models (from shaking experiences) to calibrate and confirm the validity of the numerical
model results. Analytical models can also allow a readily assessment of key parameters controlling
embankment behavior. This paper describes a simplified procedure to assess the seismic response of
rockfill embankments using the concepts originally proposed by Ambraseys and Sarma and N.
Newmark, but with some modifications applicable to CFRDs. An objective of this paper is also to use
this simplified model to evaluate the beneficial effects of design adjustments currently being
implemented in the design of high CFRDs in seismically active zones and to explore the
implementation of additional design measures.
2. PRECEDENT OF CFRD SEISMIC BEHAVIOR
Prior to 2008 the experience of CFRDs behavior under heavy seismic loading was relatively limited
although the performance of central core rockfill dams under seismic loads had been extensively
documented (Marsal and Ramirez, 1967). One of the early documented cases of CFRD performance
under seismic loading, Cogoti Dam in Chile, was presented by Arrau, et al (1985). The dam is a
dumped rockfill concrete face dam, 85 m high, which was subjected to a PGA of 0.19 g triggered by
the Illapel earthquake on April 4th, 1943, with a magnitude of 7.9, and an epicenter at approximately
90 km from the dam. The crest of the dam is 8m wide and the slopes are 1.4 H:1V and 1.5 H:1V for
upstream and downstream respectively. The dam crest settled 40 cm or 0.47% of its height. The face
slab was designed and built with panels of 10 m x 10 m and with a variable thickness from 20 cm at
the crest to 80 cm at the bottom of the dam. Vertical joints had a 1 inch thick bituminous filler. No
major damage was observed in the face slab, although some deterioration of the joints and spalling of
the central joints have been observed during the life of the project.
In 2008 two major earthquakes occurred in China and Japan which were close to significant CFRD
structures. The China Wenchuan earthquake on May 12, 2008 with a magnitude of 8 had an epicenter
located at 17 km from the Zipingpu CFRD, 156 m high, resulting in a PGA of about 0.5 to 0.6 g. The
dam performance has been discussed by several authors including Chen (1990), Xu (2009) and
Wieland (2010). The crest of the dam is 12 m wide and the upstream slope is 1.4H:1V. The
downstream slope was designed with two inclinations, changing from 1.4H:1V at El.840, to 1.5H:1V
to the crest of the dam at El. 884m as a provision for an eventual earthquake shaking. Fig 1 shows a
cross section of the dam with the zoning following the international nomenclature. The crest settled 74
cm or 0.47% of its height and the maximum downstream displacement was reported to be close to 20
cm, at the crest of the dam.
Zipingpu Dam
12,0m

EL. 884m

EL. 877m
1.5
1.0

2B
3A Filter 3m thick

EL. 840m
EL. 830m

Concrete facing slab

1.4

1.4
1.0

1.0

3B Rockfill

Original ground surface

Dumped fill

3C Rockfill

3D Rockfill
1.4
1.0
EL. 763m

Foundation
EL. 728m

Curtain grouting
Interbedded sedimentary rocks

Fig. 1 Cross Section of Zipingpu CFRD showing the International Nomenclature.


The Second International Symposium on Rockfill Dams

At the time of the shaking the reservoir was low at El. 830 m about 47 m below the full supply level at
El. 877 m, and only 68% of the water head was applied
applied on the face slab (Fig 1). Thus the lower part of
the upstream slope was confined by the reservoir head while the upper zone of the slope (from El. 884
m to El. 830 m) was unconfined during the severe shaking. Recorded peak accelerations at the crest of
the dam near the center of the valley are presented by Ishihara (2010) as follows:
Parallel to axis
1.635 g

Perpendicular to axis
2.061 g

vertical up and down


2.065 g

Figures 2 to 5 show photographs of the slab above the reservoir level as well as of the downstream
slope. The construction joint between stage 2 and stage 3 at El. 845 m (Fig. 2) was seriously damaged.
It was also observed that the perimeter joint (Fig. 3), between the face slab and parapet was intensively
affected at the central part of the dam.

parapet
El.845
3 stage
2 nd stage

Fig. 2 Horizontal Joint Damaged at El.845

slab

Fig. 3 Perimetric Joint between Slab and Parapet

Some spalling is reported at the left abutment and at the center of the dam. The severe shaking
triggered cracking at the crest as seen in Fig. 4 and loosening of rock boulders in the downstream
slope Fig. 5, but the performance
ance of the dam from the view point of stability was adequate. The
Zipingpu dam was built in 2006, following the state of art of CFRDs in China. The dam had an
adequate performance under an earthquake loading higher than the original design, indicating the
th
resilience of well designed CFRDs under dynamic loading and emphasizing the importance of
including safety measures for the future higher dams.

Fig.4 Some Cracks at the Crest

Fig. 5 Rockfill Loosened at Downstream Slope

The Miyagi Inland earthquake with a magnitude of 7.2 had the epicenter located at 10 km of the
Ishibuchi CFRD, a dumped rockfill dam, 53 m high, as described by Matsumoto et al (2010),
Yamaguchi (2008) and Wieland (2010). The crest of the
the dam is only 6 m wide and the upstream slope
was steep, varying between 1.4 H:1V to 1.2 H:1V at the upper sections of the dam. The downstream
slope was constant with two berms and slope of 1.5 H:1V as shown on Fig 6.

The Second International Symposium on Rockfill Dams

Ishibuchi dam
6,0m
Max. Water level
EL. 318m
1

2
1,

EL. 323m
1.5
1.0
1,40

1.5
1.0

Min. Water level


EL. 300m
1

1:

1,4

EL. 313m

Dumped
rockfill

,3
:1

2,90

EL. 299m

1.5

Reinforced
concrete slab

1.0

EL. 270m

Drainage conduit

Base Rock
Grout Curtain
68,53m

83,53m
6,00m

Fig. 6 Cross Section of Ishibuchi CFRD


The average seismic-induced settlement of the crest was 60 cm or 1.13% and the corresponding
maximum downstream displacement was not clearly reported but it was indicated that the downstream
deformations were much higher than the upstream. During the earthquake the reservoir was also low
at El. 300, eighteen meters under the full supply level at El. 318, or 63% of the maximum water head
applied on the face slab.
During construction of the dam, 9 pillars of reinforced concrete were built to provide a bridge for
dumping rockfill. These pillars remained in the rockfill when the dam was completed, and the
settlement of the crest was more prominent between pillars reaching values up to 80 cm settlement or
1.5% deformation, almost 3 times larger than at the Zipingpu dam in China.

Fig. 7 Construction of Ishibuchi Dam by Dumping Rockfill from a Bridge Supported by Pillars
The peak acceleration perpendicular to the axis was reported as 0.95g at the crest of the dam. Several
boulders near the crest, on the downstream slope fell down. No damage was reported for the face slab
but cracks and distresses appeared on the crest as Zipingpu. The dam was not subject to problems of
stability. Figures 8 and 9 show distress in the dam crest.

Fig. 8 Longitudinal Crack on the Crest


The Second International Symposium on Rockfill Dams

Fig. 9 Distresses on the Crest


4

3.0

SEISMIC RESPONSE OF CFRD.

Concrete Face Rockfill Dams (CFRDs) are inherently well suited to accommodate significant seismic
loading because:
Embankment materials are generally sound, well compacted, dense rockfill with high friction angles
and adequate modulus of compressibility. In addition, the relatively tight concrete face, combined
with well executed grout curtains and proper embankment zoning maintains the water level at very
low elevations within the embankment materials, and thus pore pressures within the embankment
are nil; and
Deformations observed in well compacted rockfill subject to heavy shaking are nominal, lower than
one meter.
However, as CFRDs become higher, some aspects of their dynamic response deserve special attention
including the larger amplification of ground motions which might trigger the development of
significant seismic-induced displacements in the upper dam sections.
3.1

Ground Motions Amplification.

CFRDs do not behave as rigid bodies when excited into oscillation by strong earthquakes. Their
response depends on the nature of the ground movements, the properties of the rockfill materials, and
the geometry of the embankment. The combination of all these variables can result in accelerations
within the dam substantially larger than the maximum acceleration at the ground surface. Seismic
performance analyses, Ambraseys and Sarma (1967) as well as actual field observations indicate that
for a given dam, ground motion amplifications increase from the base to the crest of the dam; with
crest stations recording accelerations two to three times larger than those reported at the base of the
dam (Marsal & Ramirez, 1967) and (Resendiz and Romo, 1982). A simplified method to approximate
amplification factors at any location within the embankment was proposed by Ambraseys & Sarma
(1967), assuming that the structure would deform in simple shear, in one dimension. Under these
conditions, the fundamental period of the dam, To, can be approximated as:


To = 2.61



Where h is the height of the dam and Vs is the shear wave propagation velocity at strain levels
compatible with those induced by the ground shaking on the embankment materials. The Vs value can
be extrapolated from shear wave velocity measurements in the embankment materials. In our
experience, well compacted, dense rockfill materials with unit weights 2.2 T/m3 have Vs values in
the range of 1500 ft/sec (457m/sec) to 2000 ft/sec (610m/sec). The acceleration An at any point within
the embankment can be estimated as: An = Ag kn, where Ag is the maximum ground acceleration;
and kn is the amplification coefficient, which is a function of the fundamental period of vibration of
the dam, To, the degree of damping within the embankment materials, and the height above the base of
the dam. Relationships between the amplification coefficient, kn, and the period of vibration of the
dam, To, at various locations above the base of the dam, assuming a damping coefficient of 20% for
the embankment materials, are shown in Figure 10 from Ambraseys & Sarma (1967). As indicated in
Figure 10, the location within the dam at which the kn is estimated is specified by a single parameter
n = y/h where y is the depth of the point below the crest and h is the height of the embankment. For
embankment materials with damping coefficients other than 20% the amplification coefficients
obtained from Figure 10, need to be multiplied by a correction coefficient , which can be obtained
from Figure 11 (Ambraseys & Sarma, 1967).
The Second International Symposium on Rockfill Dams

Fig.10 Maximum Simultaneous Seismic Coefficient for 20% Damping

Fig. 11 Damping Correction Factor Corresponding to Figure 10


(From Ambraseys and Sarma, 1967)
For example, for a CFRD 600 ft high, with dense rockfill materials with a Vs of 1500 ft/sec to 2000
ft/sec the fundamental period, To, ranges from 0.8 to 1 seconds. The amplification coefficient kn of the
ground acceleration at a point located 120 ft below the crest (n = 0.2) can be estimated as 2.0 for a
20% damping coefficient. For a more realistic damping coefficient of 10%, this amplification factor
increases to 2.6. The kn versus To relationships in Figure 10 indicates that seismic-induced
displacements are more likely to take place along relatively shallow rock wedges, near the crest, where
ground motion amplification is likely to be higher.
3.2

Evaluation of the Dynamic Resistance.

Upstream slip surfaces are not likely to develop in CFRD at normal reservoir elevations, due to the
large resistance force generated by the hydraulic pressures against the tight concrete face. The factor
of safety against upstream sliding is generally an order of magnitude larger than that of equivalent
downstream slip surfaces and thus, the potential for seismically-induced upstream displacements is
concentrated in a shallow zone above normal pool elevation. Seismic-induced displacements however
The Second International Symposium on Rockfill Dams

can occur anywhere on the downstream slope, where the resistance to seismic-induced
seismic induced loads depends
on the inclination of the slope and the friction angle of the embankment materials.
The first step in the approach proposed in this analysis involves the calculation of the factor of safety
of the slope under static conditions. For the rockfill embankment in Fig. 12 the static factor of safety
against sliding along a shallow slip surface, AC making an angle 1 with the horizontal, can be
estimated using the analogy of the rigid body resting on an inclined surface as shown in Fig. 13. The
static equilibrium force polygon of the rigid body with a weight W in Fig. 13 equal to the weight of
the wedge ABC, resting on a slip surface with an inclination 1 can be established as:
 

FS =

Where N = W cos 1; replacing terms:

FS =




and where 1, the inclination of the slip surface is equal to the mobilized friction angle,
to maintain static equilibrium, and

m,

required

is the friction angle of the rockfill materials.

Fig. 12 Potential Sliding Wedge Geometry

Fig. 13 Force Polygon of Sliding Wedge

The magnitude of the static factor of safety of the potential sliding wedge provides an index of the
magnitude of the seismic-induced
induced displacements. In areas of high seismicity it is recommended to
providee a static factor of safety of at least 1.5 to maintain seismic-induced
seismic induced displacements within
acceptable values. Sliding wedges with static factors of safety ranging between 1.5 and 1.3 are likely
to experience considerable seismic-induced
seismic
displacements that
at might require extensive repair after
ground shaking; and sliding wedges with factors of safety lower than 1.3 are likely to experience
unacceptable displacement under dynamic loading.
After the static stability of the embankment has been determined, the
the next step in estimating seismicseismic
induced displacements of the embankment slope is to evaluate the dynamic resistance of the slope,
along potential sliding surfaces to identify the critical slip surface which yields the largest dynamic
displacements. The dynamic resistance is defined as the minimum seismic-induced
seismic induced force that will
trigger displacements along the slip surface, and is defined as a vector NW where N is a coefficient
and W is the weight of the potential sliding mass. The direction of the dynamic
dynamic resistance vector is
determined by the minimum reserve shearing resistance in addition to that required for static stability,
which can be mobilized to resist the effect of a dynamic load. Once the dynamic resistance is
determined it is possible to calculate
calculate the yield acceleration Ng, which represents the threshold
acceleration in the direction of NW which will just overcome the reserve resistance of the sliding
mass.
In the third step, the maximum acceleration, An, induced by ground shaking on the center of gravity of
the potential sliding mass can be compared with the corresponding yield acceleration Ng required to
The Second International Symposium on Rockfill Dams

overcome the reserve resistance along the assumed sliding surface. The magnitude of the maximum
earthquake-induced
induced acceleration, An, on any sliding mass on the slope can be estimated as proposed by
(Ambraseys and Sarma, 1967), and described above. If the maximum acceleration An on the potential
sliding mass is lower than Ng, then the slope is safe and
an no seismic-induced
induced displacements will take
place. On the other hand, if the induced acceleration, An, on any of the sliding masses considered,
exceeds the value of Ng, relative displacements will take place along the assumed failure surface. The
magnitude
ude of the accumulated displacements that take place along the sliding surface during these
periods of time when the seismic-induced
seismic induced acceleration exceeds the yield acceleration are calculated
using the simplified method proposed by N. Newmark (1965), described
des
below.
The magnitude and direction of the Dynamic Resistance Force NW can be determined using the force
polygon in Fig. 14 which shows the forces acting on a rigid block, resting on an inclined base
representative of the potential sliding mass in Fig. 13. In the force polygon of Fig. 14 the magnitude
and direction of the weight vector W are known. The mobilized friction angle,
m, required to
maintain static equilibrium along the sliding surface is also known as and is equal to 1, the inclination
of the assumed sliding surface. The direction of the resultant force, R, on the sliding surface is also
known, and is determined by the friction angle,
with respect to the normal, N, to the sliding plane.
The friction angle,
, of well compacted rockfill materials, generally ranges between 45 to 50
degrees, and can be approximated from well calibrated correlations (Leps, 1973), from precedent on
similar dams or from large scale triaxial tests if required.
ic resistance vector NW is represented by the force PQ in the force polygon which
The dynamic
corresponds to the minimum force required to mobilize the reserve friction angle ( 1) along the
sliding surface. The force, PQ, is the minimum force required
required to close the force polygon and has a
direction that makes an angle of 90 degrees with respect to the direction of the resultant R. The
magnitude of the dynamic resistance NW, can be estimated from the triangle OPQ as:

Fig. 14 Force Polygon of Sliding Wedge

sin ( - 1) =

=N

Thus, the yield acceleration Ng can be estimated as: Ng = sin ( - 1) g


3.3

Calculation of Seismic-Induced
Induced Displacements along Potential Failure Surface.

The method proposed by N. Newmark, 1965 to approximate


approximate the magnitude of the accumulated
seismic-induced
induced displacements is based on the estimates of the relative motion of a rigid block with
respect to the ground triggered by acceleration pulses with a magnitude larger than the yield
acceleration.

The Second International Symposium on Rockfill Dams

The results of this method are summarized in Fig. 15, which provides a relationship between the ratio
N/A, the coefficients of the yield acceleration over the seismic-induced acceleration on the sliding
wedge, and the corresponding seismic-induced displacements of the wedge. A review of the results
shown in Fig. 15 indicates that the magnitude of the seismic-induced displacements of the sliding
wedge can be classified in three different groups depending on the magnitude of the N/A ratio. For
those cases where the N/A ratio is equal to or greater than 0.5, the seismic-induced displacements are
likely to be relatively low, less that 6 in., and would not have a significant effect on the embankment
behavior. For cases, where the N/A values range between 0.5 and 0.20 the seismic-induced
displacement can be significant, up to 3 ft (1m), but still acceptable if adequate free board has been
provided. The potential off set of the filter materials caused by these displacement levels can be
tolerated, and the strength of the embankment materials is not significantly reduced by the shear strain
imposed by these displacements. However, if the N/A ratio drops below 0.20, the resulting seismicinduced displacements can be large, with a magnitude of several feet, and can trigger unacceptable
behavior, including loss of free board and overtopping, significant shear strength loss along the siding
surface, and/or unacceptable shear off sets within the filter layer. As a general design criteria, in
highly seismic zones, the N/A ratio should be maintained above 0.2.

Fig. 15 Standardized Displacement for Normalized Earthquakes, Unsymmetrical Resistance


(from Newmark, 1965)
The magnitude of the seismic-induced displacements for potential sliding wedges with N/A ratios
lower than 0.2 can be estimated as:
= 6 max 2/2 gN
where

seismic-induced displacements

max

peak particle velocity at the center of gravity of the sliding wedge, in


feet/sec

acceleration of gravity in feet/sec2; and

yield acceleration coefficient.

The Second International Symposium on Rockfill Dams

The magnitude of the peak particle velocity, max, can be estimated from the ratio of max/An = 150
cms/sec/g; where An is the seismic-induced sliding wedge acceleration. For example if the
acceleration, An, at the center of gravity of a shallow sliding wedge is equal to 0.7 g, the corresponding
peak particle velocity, max, would be equal to 3.88 ft/sec. If the sliding wedge has a yield acceleration
coefficient of 0.1 the seismic-induced displacements can be estimated as 11 ft.

4. DESIGN RECOMMENDATIONS
In highly seismic areas, the design criteria of large CFRD should include a minimum N/A ratio of
0.20. Because, the critical sliding wedges are located near the crest of the dam, special measurements
to achieve this criteria include:
a)
Flatten the slopes in the upper sections of the dam.
The downstream slope near the crest should be flatter than the average downstream slope as shown in
Fig.16. The upstream slope at the upper sections of the dam should be flatter in case of reservoir
fluctuating with high frequency.

8+2% Hm

PARAPET
.

Face Slab
Zone 1B

1.5

Flatter Slope
Depends of Dam height (25 - 30% H)
1.0

Zone 3B
Zone 3A
Zone 2B

Zone 3C

1.4
1.0

Zone 4

Zone 1A

1.4

1.0

DRAIN
Zone 2A

1A COHESIONLESS SOIL - COMPACTED BY CONSTRUCTION EQUIPMENT


1B RANDOM - COMPACTED BY CONSTRUCTION EQUIPMENT
2A PROCESSED MATERIAL ( MAX. = 3 4 ") - MANUAL COMPACTION
2B PROCESSED MATERIAL ( MAX. = 3" - 4 ") 4 - 6 PASSES OF 12 Ton VIBRATORY ROLLER
3A SELECTED SMALL ROCK PLACED IN SAME LAYER THICKNESS AS ZONE 2
3B QUARRY RUN ROCKFILL, ABOUT 0,60m TO 0,80m LAYERS, 4 - 6 PASSES OF 12 Ton VIBRATORY ROLLER
3C QUARRY RUN ROCKFILL, ABOUT 0,80m TO 1,00m LAYERS, 4 - 6 PASSES OF 12 Ton VIBRATORY ROLLER
4 DOWNSTREAM ROCKFILL - PLACED ROCKFILL

Fig.16 Suggested seismic design for CFRDs to 300m high.

b)
Widen the crest of the dam as the height increases.
Crest wide should be 8m minimum for dams up to 150m high. For dams higher than 150m to 300m
increase the crest width to : w = 8 + 0.02H where H: height of the dam in meters.
c)
Provide a zoning of the dam as shown on Fig.16
The zone 3B of well graded rockfill materials should be extended at the upper portion of the dam (25
30%H) to the downstream slope. This criterion is being applied in some CFRDs already under
construction.

The Second International Symposium on Rockfill Dams

10

d)

Provide good compaction to the rockfill

Specify compaction with vibratory rollers with a minimum of 12t over the cylinder and having a ratio
of static weight/cylinder width, equal or >5t/m. Restrict lift thicknesses to 0.8m and apply water with
monitors with a ratio of 200 l/mas a minimum.
e)
Provide a higher freeboard than conventional.
Besides the hydraulic requirements for the conventional calculation of freeboard an additional 0,3%H
should be added in areas of high seismicity.
f)

Limit height of parapet walls

Height of parapet walls should be limited to 3 to 4 m high.


g)
Provide external water stops with more capacity.
Besides the traditional copper water stop located internally at the perimetric joint, an external
corrugated or with additional capacity water stop should be designed
h)
Reduce the slab width in zones of steep topography.
The traditional width of face slab lanes of 15m, should be split to 7,50m in steeper zones where
differential settlement can occur during heavy shaking.
i)
Treat carefully construction joints between face slab stages.
Experience has demonstrated that some gaps can be developed behind the struded curb under the
construction joint of the face slab stages. These gaps should be well grouted to avoid serious
displacements during heavy shaking.
j)

Place compressible filler materials in the central longitudinal joints of the concrete
slab.
For narrow canyons where the A/H < 4, experience has demonstrated that it is necessary to place
compressible fillers to mitigate compression stresses during the settlement caused by reservoir filling
and severe ground shaking. A=Concrete face area in m; H=Height of the dam in meters.

REFERENCES
[1]

N.M. Newmark, Effects of Earthquakes on Dams of Embankments, Geotechnique 15:140141, 156, 1965.

[2]

N. N. Ambrasey and S. K. Sarma, The Response of Earth Dams to Strong Earthquakes,


Geotechnique 17:181-213, 1967.

[3]

F.I. Makdisi and H.B. Seed, Simplified Procedure for Estimating Dam and Embankment
Earthquake-Induced Deformations, Journal of the Geotechnical Engineering Division, ASCE,
Vol. 104, No. GT7, July, 1978, pp. 849-867.

[4]

J.B. Cooke and J.L. Sherard CFRD Design, Construction and Performance ASCE
Symposium, Detroit, Michigan, 1985.

[5]

L. Arrau, I. Ibarra, G. Noguera. CFRD Design, Construction and Performance CFRD book,
Performance of Cogoti Dam Under Seismic Loading, ASCE Symposium, Detroit, Michigan,
1985.

[6]

Chen Shengshui et al, Analysis of Seismic Permanent Deformation of Rockfill Dam, Journal of
Nanjing Hydraulic Research Institute, 1990(3): 277

The Second International Symposium on Rockfill Dams

11

[7]

Xu Zeping, Performance of the Zipingpu CFRD During the Wenchuan Earthquake, Hydropower
Dams, Issue three, 2009.

[8]

M. Wieland, Chairman ICOLD Committee on Seismic Aspects of Dam Design CFRDs in


Highly Seismic Regions Water Power and Dam Construction, April 2010.

[9]

P.T. Cruz, B. Matern and M. Freitas, Concrete Face Rockfill Dams Book, Released in the
Occasion of the 23rd ICOLD Congress, Braslia, Brazil, May 2009.

[10] N. Matsumoto et als, Performance of the Ishibuchi CFRD During the Miyage Earthquake,
Hydropower & Dams, Issue One, 2011.
[11] Y.Yamaguchi et al, Preliminary Investigation of Dams Stricken by the Iwate-Miyagi Nairiku
Earthquake, 2008.Article from Google.
[12] K. Ishihara, Performance of Rockfill Dams During Recent Large Earthquakes, Fifth
International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil
Dynamics, San Diego, California, 2010.
[13] R.J. Marsal and L. Ramirez, Performance of El Infiernillo Dam, 1963-1966, Journal of the
Soil Mechanics and Foundations Division, ASCE, Vol. 93, No. SM4, July, 1967, pp. 265-298.
[14] D. Resendiz, M.P. Romo and E. Moreno, El Infiernillo and La Villita Dams: Seismic Behavior,
Journal of the Geotechnical Engineering Division, ASCE, Vol. 108, GT1, January 1982.
[15] T.M. Leps, Flow Through Rockfill, Embankment Dam Engineering, John Wiley and Sons,
1973.

The Second International Symposium on Rockfill Dams

12

Вам также может понравиться