Вы находитесь на странице: 1из 3

The 5th PSU-UNS International Conference on Engineering and

Technology (ICET-2011), Phuket, May 2-3, 2011

Prince of Songkla University, Faculty of Engineering
Hat Yai, Songkhla, Thailand 90112


Puttapon Thongindam1*
Rajamangala University of Technology Thanyaburi, Faculty of Engineering, Thailand
*Email: th.puttapon@gmail.com

Abstract: In order to deal with subgrade stiffenning with artificial bedrock should be inserted at the depth result to
depth and the stiff layer beneath pavment structure in deal with the above mentioned problems. To verify the
backcalculation process based on multi-layered elastic CSCM based on the field data, FWD-tests have been
theory, we need the appropriate pavement model. Since conducted on the selected highways in Chiang Rai
2009 an heuristic method for determining depth to Province, Thailand. The concept of the CSCM and the
artificial bedrock (DTAB) coined as the “consistent test procedure are briefly explained in the next sections.
slope changing method (CSCM)” has been proposed by
the author. To perform a field verification of the CSCM, 2. CONSISTENT SLOPE CHANGING METHOD
the road 1020 in Chiang Rai province was selected as (CSCM)
test sections. Deflection basins of the pavement surface Based on the classical pavement mechanistic theory
are collected by the falling weight deflectometer. The and the results from the related investigations, an
DTAB results are calculated using CSCM. Discussion is alternative method for determine DTAB has been
made on using these results for modelling pavement developed by Thongindam [2] in 2009. The main
sturcture. presumption is that the area above the nearly conical
stress zone is not affected by the acting load. Implying
Key Words: Consistent slope changing method, Pavement that the outer surface deflections are mostly originated
model, Backcalculation, Layer moduli from the subgrade soil under this stress zone. In other
words, the position of zero surface deflection should be
1. INTRODUCTION strongly related to the depth beneath the pavement at
The in situ layer elastic moduli are needed to be which no deflection occurs, i.e. an apparent stiff layer or
known for analyses of pavement rehabilitation projects. bedrock. Therefore, if the deflection basin is able to be
One of the most popular method to implement that extrapolated radially with respect to subgrade behavior
known as “bakcalculation” in which the modulus of till the point that surface deflection equal to zero, the
elasticity of pavement materials are determined by appropriate depth for inserting an artificial bedrock
working mulit-layered elastic theory (MLET) backward should be determined from that point.
using measured surface deflections and appropriate Since the stiffness of real subgrade materials is
pavement model parameters such as Poisson’s ratio and governed by many factors (soil type, stress state, water
layer thicknesses. Since the subgrade thickness is usually content, bedrock etc.), the effective character of these
unknown, this might be lead to at least two problems in materials should be somehow related to the change of
modeling of pavement structures. Firstly, since the slope of deflection curve especially at its outer part.
bottom layer in most of backcalculation schemes is Therefore, the slope changing rate of the extrapolated
assumed to be semi-infinite in depth with a constant deflection curve should be consistent with those of the
elastic modulus but in the real world an apparent stiff measured one, this procedure is for this reason named as
layer or “bedrock” can exist at any depth which may lead “Consistent Slope Changing Method” or “CSCM”. More
to a very high modulus value. Secondly, it is well-known in detail, the deflection curve should be extrapolated to
that most subgrade soils are “non-linear” in stiffness the next point which has a radial distance from the
behavior and the modulus in most subgrade materials former one equal to the average of in-use FWD sensors
often increases with depth [1]. Therefore, some attempts spacing by keeping slope changing rate consistent. This
have been done to deal with subgrade stiffening with concept can be implemented by using least square
depth in backcalculation analysis based on MLET by technique with all default slope values. The radial
adding artificial bedrock at some depth under the distance to the point where the deflection is zero presents
subgrade layer [3]. In 2009 the CSCM was purposed by the desired DTAB. This depth can be, therefore,
the author for determining DTAB from in situ surface determined by using the intercept of radial distance axis
deflection [2]. In setting up the pavment model, the in term of inverse value as shown in Figure 1.
Deflection vs Inverse offset distance
slope at some
measured and
extended curve
Deflection [m m]




Intercept yielded Inverse of DTB

0.000 0.001 0.002 0.003 0.004 0.005

Inverse Distance [1/mm]

Fig. 2. FWD test on the highway 1020, Chiang Rai

Fig. 1. Mesured and extrapolated deflection curve using Province
CSCM plotted against inverse offset [2]
2.2 Testing procedure
The process of extrapolation of the deflection curve During the tests, the following FWD important
has to be done until either the horizontal axis intercept is configurations are used: a 30 cm diameter load plate,
found, which can also be observed in the spread sheet with 50 kN weight set and the deflection sensors placed
data from the change of deflection value from positive to at radial distances of 0, 200, 300, 450, 600, 900, 1200,
negative, or the radial surface distance is greater than 1500 and 1800 mm. The air and asphalt temperature
14.0 m where the influence of bedrock on backcalculated were measured. Especially for the upper asphaltic course,
moduli can be considered as insignificant [2]. the temperature was measured at surface and at the
middle of its layer. As shown in figrue 3, the tests were
2. FIELD TESTING PROCEDURE conducted on both lanes of carriage way. The first testing
2.1 Description of test section point started at the distance of 50 m from the first station.
As mentioned above, an in-service two traffic-lanes The next and others locate at 100 m spacing distance (i.e.
highway 1020 in Chiang Rai province was selected as sta. 0+050, 0+150, 0+250, 0+350…end station). For
the test sections. It locates in northern region of another lane, backward testing, the testing was further
Thailand. The road surface is asphaltic concrete with the conducted at the middle point of the adjacent tested
carriage way dimension of 8.0 m width included 1 m of points of the forward testing (i.e. sta. 15+000, 14+900,
verge at both side. The detail of pavement structure at 14+800…end station). With this testing method, the
initial state is summarized in Table 1 and the history of testing points are well distributed on both carriage way.
surface overlaying with asphaltic concrete, which is The testing point at each station is always conducted at
usually conducted with thickness of 5 cm is also the outer wheel path.
registered. However, the real thickness of the uppper
layers are also explored using bore scope equipment.




Forward:  0+450 
Table 1. Detail of pavement structure of the selected site Lane 1

Road Station Pavement structure at initial state Backward:  

No. Lane 2 

1020 120+000 50 mm Asphaltic Concrete*


– 150 mm Crushed rock base






135+000 150 mm Soil aggregate subbase

150 mm Selected material
embankment Fig. 3. Testing outline at road with two traffic-lanes
*Overlaying History (5 cm) : Year 2004 at station
121+600 - 126+500 3. RESULTS AND DISCUSSION
Using the program developed with MAPLE
language, the DTAB results are automatically
Figuer 2 shows the testing site highway 1020 and the determined at each test station from the FWD in-situ
operating of the FWD equipment. The type of FWD used deflection data as shown in figure 4. These results can be
in this study is Dynatest 8000 which is usually used in used to model the pavement structure in backcalculation
pavement survey projects of the department of highway process. Hence, the usual unknown thickness of subgrade
Thailand. The test section distance is approximately layer can be determined by deducting the total thickness
15.0 km, i.e. sta. 120+000 to 135+000. of the surface and base from the depth to the artificial
Moduli Using Genetic Algorithms”, Dissertation,
Gottfried Wilhelm Leibniz University of Hannover,
Hannover, 2009.
[3] Ullidtz, P. “Modelling flexible Pavement Response
and Performance” Lungby: Polyteknisk Forl, 1998 –
ISBN 87-502-0805-5.
[4] DOH: Department of Highway Thailand, FWD
Testing Procedure, Personally Contact (Not
Published), Ministry of Transport, Thailand,

Fig.4 Calculated DTBs at Road no. 1020 using CSCM

For stations at which there is no bedrock or it could

not be found within the depth of 14.0 m beneath the
pavement surface, the value of 14.0 m is assigned
automatically to the DTAB result since it has been found
that adding the fictive stiff layer at this depth has very
little influences on backcalculated moduli [2].
Considering the results in the figure 4 regardless the
stations at which their DTABs are 14.0 m, the average of
3.54 m can be calculated. The more interesting is the
compatibility of the stations which have no bedrock (i.e.
bedrock presence is deeper than 14.0 m or the behavior
of subgrade stiffening with depth is inconsiderable) in
both traffic-lanes. This occurrence can be found between
the station 128+000 till the end station of this testing site.
This congruity of the results at both traffic-lanes should
be used to affirm the reliability of CSCM in application
with FWD deflection data in actual field case.

There has been a report that inserting a fictive
bedrock at suitable depth could be deal with subgrade
stiffening with depth and real bedrock beneath the
pavment structure. This study focused on verifying the
method, CSCM, using FWD maximum deflection data.
The DTAB results from test section have illustrated the
congruity of DTAB on both testing traffic-lanes. This
compatibility should be used to affirm the reliability of
CSCM to apply with FWD deflection data. Hence,
inserting an artificial bedrock at the depth results
calculated from CSCM should be considered as an
alternative for modeling pavement structure for
backcalculation process. The usual unknown thickness of
subgrade layer can be determined by deducting the total
thickness of the surface and base from the depth to the
artificial bedrock.

[1] Rohde, G.T.; Smith, R.E. “Determining Depth to
Apparent Stiff Layer from FWD Data” Texas
Department of Transportation, Report No.
FHAW/TX-91-1159-1, Texas, 1991
[2] Thongindam, P. “Enhancement of Backcalculation
Technique for Assessing Flexible Pavement Layer