See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/268430849

Prediction of subgrade resilient modulus for flexible pavement design

Article  in  Scientific research and essays · September 2011

DOI: 10.5897/SRE11.846

CITATIONS READS
2 550

2 authors:

Mehmet Ridvan Ozel Abbas Mohajerani

Hayat University RMIT University
3 PUBLICATIONS   8 CITATIONS    70 PUBLICATIONS   502 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Fibre length as a potential contributor to the shear strength gain in soft clays View project

All content following this page was uploaded by Mehmet Ridvan Ozel on 11 February 2015.

The user has requested enhancement of the downloaded file.

Scientific Research and Essays Vol. 6(21), pp. 4567-4576, 30 September, 2011
Available online at http://www.academicjournals.org/SRE
DOI: 10.5897/SRE11.846
ISSN 1992-2248 ©2011 Academic Journals

Full Length Research Paper

Prediction of subgrade resilient modulus for flexible

pavement design
Mehmet Ridvan Ozel1* and Abbas Mohajerani2
1
Department of Civil Engineering, School of Engineering and Architecture, Gediz University, Izmir, Turkey.
2
School of Civil, Environmental and Chemical Engineering, RMIT University, Melbourne, Australia.
Accepted 22 August, 2011

Resilient modulus of subgrade soils is an important input in mechanistic pavement design. The primary
objective of this work is to investigate the resilient modulus of four typical Victorian fine-grained
subgrade soils under traffic-like repeated loading and to suggest empirical predictive models
incorporating physical properties and/or strength of the soils along with the stress state. A repeated
load triaxial testing procedure was developed, which is capable of collecting resilient and permanent
deformation data from the same specimen. Stress levels for testing were defined as percentages of the
confined and/or unconfined soil static strengths. Stress dependency of resilient modulus was studied
through the models (such as bilinear model, power model, deviatoris stress model and octahedral
stress model) found in the literature and other possible combinations of deviator, confining and
octahedral stresses. A semi-logarithmic model was proposed for the prediction of resilient modulus of
the fine-grained subgrade soils. Calibration of model constants by soil properties was investigated. An
altervative prediction model was also developed based on unconfined compressive strength and
deviator stress. Resilient modulus values were back calculated using both the semi-logarithmic model
and the model based on unconfined compressive strength and deviator stress. Predicted values were
compared with the measured values. Predictive capability of the proposed models were proven for use
in flexible pavement design.

Key words: Resilient modulus, fine-grained subgrade soils, repeated loading, flexible pavement design.

INTRODUCTION

Subgrade soil characterization is an important stage of study. Resilient response of subgrades can be quantified
the mechanistic pavement design procedures for flexible by resilient modulus (Er) that is a stress-strain relationship
pavements (AASHTO, 1993; AUSTROADS, 1992). like modulus of elasticity. However, Er is determined from
These procedures are mainly based on analyzing the a repeated load triaxial compression test (RLTT) and is
response of the pavement materials under simulated based on only the recoverable portion of the strain (Yoder
traffic loads and environmental conditions. The complex and Witczak, 1975; Elliot and Thornton, 1988). It is
elasto-plastic deformational response of pavement expressed as the ratio of axial repeated deviator stress
materials is studied in two categories in order to simplify (σd) to the recoverable axial strain (εr) and considered as
the task: resilient strain (εr) and permanent strain (εp). an indication of load-carrying capacity of the subgrades
The resilient deformational response of pavement as given in Equation (1) (Seed et al., 1962):
subgrades under repeated loads is the scope of this
Er = σd / εr (1)

The Austroads pavement design guide recommends the

*Corresponding author. E-mail: [email protected], use of RLTT for determining the “modulus” of subgrades
[email protected]. Tel: 90-232-355-0000. Fax: 909-232- (AUSTROADS, 1992). However, in practice, due to the
425-6813. testing equipment needs and complexities involved in the
4568 Sci. Res. Essays

test, empirical relationships are generally used to Robnett (1979) showed that there is a bilinear
estimate the Er of the pavement subgrade soils. The relationship between Er and σd. This model is given as
empirical relationship based on California bearing ratio follows:
(CBR), which was first introduced by Heukelom and
Klomp (1962), is still the most common tool for pavement Er= k3 + k4 σ d when σ d < σ di, (5a)
engineers. The relationship is given below:
Er= k5 + k6 σ d when σ d > σ di, (5b)
Er = 10×CBR (MPa). (2)
Where σ di is the deviator stress where two fitted linear
Powell et al. (1984) developed another Er–CBR lines of the Er-σ d graph intersect and k3, k4, k5 and k6 are
expression in the Transport and Road Research material constants. Another form of relationship between
Laboratory (TRLL) as shown below: Er and σd was proposed in the study of Moossazadeh and
Witchzak (1981), in which three fine-grained soils were
Er = 17.6×CBR0.64 (MPa). (3) investigated. The model is known as power model or
deviatoric stress model and written as follows:
Nevertheless, these relationships should be used with
caution because of the quasi-static nature of the CBR Er= k7 σdk8, (6)
test and the absence of stress parameters in the
expressions. Furthermore, their accuracy has already Witczak and Uzan (1988) used octahedral stress
been examined and questioned in the works of other attributes for modeling the behavior of some granular
researchers like Thomson and Robnett (1976, 1979). soils and later other researchers (Houston et al., 1993;
Predicting a dynamic property of the soils from their CBR Puppala et al., 1996; Mohammad et al., 1999; Ozel and
values, which is actually a measure of static shear Mohajerani, 2001) adopted this model in their study and
strength of the material, may not always yield the correct proved its predictive capability for some fine-grained
results. soils. This model, which was originally derived by
The main objective of this study is to investigate the Er Shackel (1973), is considered more appropriate than the
of some Victorian fine-grained subgrade soils under models incorporating only the deviator stress, since it
traffic-like repeated loading and to suggest empirical accounts for both lateral and vertical stresses in three
predictive models incorporating physical properties dimensions. The general format of octahedral stress
and/or strength of the soils along with the stress state. A model can be expressed as:
laboratory repeated load triaxial testing program for Er,
which also allows collecting permanent strain data, has Er = k9 (σoct) k10 (τoct) k11 , (7)
been developed in this study. Four typical fine-grained
subgrade soils are chosen since most local research
Where σoct is octahedral normal stress; τoct is octahedral
works have been concentrated on granular soils despite
shear stress; and k9, k10, and k11 are material constants.
the fact that fine-grained soils have a wide surface
Drumm et al. (1990), on the other hand, examined the
coverage across Victoria and Australia (Nataatmadja and
resilient response of 11 Tennessee fine-grained soils and
Parkin, 1989; Symons and Poli, 1996; Moffatt et al.,
expressed the Er as below:
1998; Chen, 1999; Lo and Chen, 1999; Nataatmadja and
Tan, 2001; Bodhinayake, 2008).
Er = [k12 + k13 σ d] / σ d . (8)

All these models take only stress state into account and
RESILIENT MODULUS PREDICTION
the model constants need to be calibrated for local soil
conditions such as soil type, grading characteristics and
Stress dependency of resilient modulus (Er) of fine- moisture-density state.
grained soils has been defined by many researchers and There are also several investigations in which the Er
some mathematical models have been suggested. Er is (under a specific stress state) is directly modeled in terms
generally expressed in terms of applied deviator stress of the soil physical and strength properties (Thompson
(σd). Fredlund et al. (1977) adopted a semi-logarithmic and Robnett, 1979; Thompson and LaGrow, 1988; Elliot
model for Er and σd relationship and found consistent and Thornton, 1988; Drumm et al., 1990). In these
results for a fine-grained soil (moraine glacial till) of studies, regression equations were established between
Canada. The model is expressed as: Er and various soil properties such as liquid limit, plasticity
index, clay content, organic carbon content, degree of
log Er= k1 - k2 σ d, (4) saturation, dry density, percent passing 75 micron sieve,
moisture content, California bearing ratio and unconfined
Where k1 and k2 are material constants that depend on compressive strength. Nevertheless, the common
soil type and soil physical properties. Thompson and drawback of these models is that the stress dependency
 Ozel and Mohajerani 4569

Table 1. Basic properties of the experimental soils.

Soil physical properties S1 S2 S3 S4

Unified soil classification CH CL CH CH
Clay content (%) 70 46 71 54
Liquid limit (%) 66 34 77 66
Plasticity index 35 11 50 48
3
Max dry density (mg/m ) 1.45 1.72 1.32 1.53
Optimum moisture content (%) 29.5 18.5 33 27
Specific gravity 2.75 2.72 2.64 2.61

of Er is not included. in a standard proctor mould. Saturated porous stones were used
on top and bottom for the RLTT’s and UCS tests samples. A rubber
membrane was placed over the specimen and secured to the top
and bottom with o-rings to fully seal the specimen against
EXPERIMENTAL DESIGN
pressurized confining water during the RLTT’s.
Soils tested
METHODS
Four Victorian subgrade soils were selected for inclusion in the
study. These materials were obtained from a local street roadbed in
The experimental program requires the preparations of 12 RLTT
Narre Warren South, where it is known as the Baxter Sandstone of
samples for each soil at three different levels of moisture content,
Brighton Group soils (S1); Eastern Freeway extension project,
which is thought to be the most important physical property
Mitcham (S2); Geelong Road extension project, Altona (S3) and
affecting the deformational characteristics of fine-grained subgrade
from Pascoe Vale Road duplication project, Broadmeadows (S4).
soils. Optimum, 2% dry side of optimum and 2% wet of optimum
Basic soil tests such as specific gravity, Atterberg limits, particle
were chosen as testing moisture content levels. As Australian
size analysis and compaction were performed in accordance with
standard for repeated load triaxial test does not cover the testing of
the relevant Australian Standards. Compaction characteristics were
subgarde soils, an original procedure was developed in this study.
determined in the laboratory delivering the standard compactive
RLTT procedure was designed to allow εp and Er data being
effort. Some of the physical properties of the soils are presented in
collected from the same specimen (Ozel, 2003). This was achieved
Table 1.
by extending the conditioning stage up to 10,000 load repetitions, in
which the permanent strain data were collected and continuing with
the stress-stage test, in which various combination (15 stress
Testing equipment and sample preparation
levels) of confining (σ3) and deviator stresses (σd) were applied for
the investigation of resilient behavior of the materials (Table 2).
A Universal Testing Machine (UTM-5P) was employed to perform
Three levels of σ 3 were applied. The first conditioning stage of
repeated load triaxial tests (RLTT). The equipment was a close-
10,000 is called permanent strain (εp) test and the second phase is
loop, fully automatically controlled data acquisition system capable
called resilient modulus (Er) test here. Some findings of permanent
of applying repeated dynamic loads through a shaft of a pneumatic
deformation part of the study (the data collected and analyzed
actuator (Figure 1). The test specimens were accommodated in a
during the first 10,000 cycles of the RLTT’s) were published
triaxial pressure cell, which is made of perspex cylinder and suitable
elsewhere (Ozel and Mohajerani, 2002). Because different deviator
for testing specimens having dimensions of 200 mm height by 100
mm diameter. The triaxial pressure cell is fitted for water as a stress levels (DSL) were applied during the εp test, a transition
confining medium. The loading pulse duration and rest period were stage, which had DSL of 0.7, was placed between the εp and Er
selected 0.5 and 1 s respectively and axial deformations were tests, in order to bring the specimens to the same deformational
measured using three linear variable differential transducers (LVDT) status before Er test begins. Stress-strength ratio in this study is
mounted on clamps outside the triaxial chamber. Three-stage called DSL and can be defined as the ratio of the σd of RLTT to the
unconsolidated undrained (UU) triaxial tests (Head, 1982) were soil strength obtained from UU triaxial or UCS test. There are two
performed by using a computer controlled triaxial testing main advantages of defining stress levels through the DSL
equipment. Tests were conducted on the cylindrical specimens of approach: first, keeping the stresses always under the soil failure
38 by 76 mm having the same initial conditions with the RLTT envelope and the second, testing the compacted soil close to its
specimens. UTM-5P test machine was also used to determine the working stresses. The stress levels developed is given in the Table
unconfined compressive strength (qu) of the specimens. The 2. After completing Er test, confining pressure released and
specimens for RLTT’s and unconfined compressive tests (UCS) specimens were tested under quasi-static loading to determine their
were 101.5 mm in diameter and 202 mm in height. qu’s (after RLTT qu values).
The disturbed soils were processed in accordance with the Since the σd for RLTT in the experimental program are defined by
relevant Australian standard. Representative samples obtained soil strengths, three-stage unconsolidated undrained (UU) triaxial
from the processed soil were first mixed with water and allowed to tests (Head, 1982) were first conducted for Narre Warren soil. Soil
cure for 48 h before compaction, so that the water could completely static confined compressive strengths (σss) were determined as the
wet up the clayey soil. Mixtures were kept in sealed plastic bags achieved deviator stresses for σ3 of 15, 30 and 45 kPa. The
during waiting periods. All samples were compacted by delivering deviator stresses for RLTT were then calculated as certain
standard compaction energy. Static triaxial test specimens of Narre percentages of the soil static strengths, which are shown in the
Warren soil were produced by extruding out of the soil compacted Table 2. However, having found not much difference between σss
4570 Sci. Res. Essays

Figure 1. Repeated loading triaxial cell and soil specimen.

(for relatively low values of applied σ3 values) and qu values for this study. Although the models provided reasonably
Narre Warren soil (Table 3), UCS test was decided to employ for 2
good agreement having coefficient of determination (R )
Mitcham, Altona and Broadmeadows soils in determining soil
ranging from 0.60 to 0.90, different combinations of σd,
strengths and subsequently calculating the σd’s for RLTT. UCS test
results of these soils can be seen in Table 4. σ3, σoct and τoct with Er were also investigated before
adopting any of the available models for the Er data of
this study. These further analyses have been performed
ANALYSIS OF RESULTS due to the known deficiencies of the models found in the
literature, which are the absence of σ3 effects in the case
Model for resilient modulus with stress state and soil of σd models and collinearity problem for σoct and τoct
properties models. The models with the only variable of σd
(Equations 4, 5 and 6) are criticized elsewhere for not
The influence of stress state on Er and stress state including the effect of lateral pressure on Er (Houston et
models for fine-grained soils found in the literature are al., 1993; Puppala et al., 1996; Muhanna et al., 1998;
summarized previously. Regression analyses were Ozel and Mohejerani, 2001). These researchers revealed
performed using these models for the data obtained in the effect of σ3 on Er for fine-grained soils. Therefore this
 Ozel and Mohajerani 4571

Table 2. Stress levels for RLTT.

¶ *
Test name Stress levelσ 3 DSL§ (σ
σd†/σ
σss #) σd/qu±)
DSL (σ N
Permanent strain 30 0.3~0.9 0.3~0.9 10,000
Transition stage 30 0.7 0.7 200
1 45 0.15 0.3 200
2 45 0.30 0.4 200
3 45 0.45 0.5 200
4 45 0.60 0.6 200
5 45 0.75 0.7 200
6 30 0.15 0.2 200
7 30 0.30 0.3 200
Resilient modulus 8 30 0.45 0.4 200
9 30 0.60 0.5 200
10 30 0.75 0.6 200
11 15 0.15 0.1 200
12 15 0.30 0.2 200
13 15 0.45 0.3 200
14 15 0.60 0.4 200
15 15 0.75 0.5 200
¶ § † # ±
confining stress (kPa); deviator stress level; deviator stress (kPa); confined compressive strength for S1 as given in Table 3 (kPa); unconfined
*
compressive strength for S2, S3 and S4 as given in Table 4 (kPa); number of load repetitions.

Table 3. Soil strengths for S1.

UU test UCS test

RLTT sample number Moisture state σss *
σ3 ±
qu#
01 to 04 Dry 30 330 330

45 280 270
05 to 08 Optimum 30 255
15 220

09 to 12 Wet 30 190 210

± * #
confining pressure used in UU triaxial test (kPa); unconsolidated undrained triaxial compressive strength (kPa); unconfined compressive
strength (kPa); all values are the average of three test results.

Table 4. Unconfined compressive strengths, qu, for S2,S3 and S4 soils.

qu (kPa)
RLTT sample number Moisture state
S2 S3 S4
01 to 04 Dry 200 220 170
05 to 08 Optimum 175 195 145
09 to 12 Wet 120 165 105
All values are the average of three or four tests.

parameter should be in the model. stresses and the model has been found to be the best in
The octahedral stress model, which is given in representing the behavior of the fine-grained soils
Equation 7, includes the effect of σ3 through octahedral reported in this study. However, the constants of
4572 Sci. Res. Essays

Table 5. Coefficient of determinations (R2)based on Equation 9.

Sample number S1 S2 S3 S4
1 0.94 0.75 0.94 0.81
2 0.93 0.71 0.95 0.82
3 0.86 0.64 0.94 0.96
4 0.94 Failed 0.92 Failed
5 Not available 0.43 0.84 0.91
6 0.93 0.64 0.93 0.89
7 0.93 0.52 0.89 0.95
8 0.97 0.54 0.93 Failed
9 0.97 0.75 0.91 0.66
10 0.99 0.48 0.95 0.89
11 0.93 0.56 0.90 0.99
12 Failed Failed 0.91 0.88

octahedral stress model are determined by multiple as given in Equation (9) has been adopted to model the
regression analysis, which might be subject to collinearity resilient behavior of the fine-grained soils in this study on
problem. The collinearity becomes the major concern the basis that it overcomes the aforementioned
when the “independent” variables of a multiple regression deficiencies of the models found in the literature, namely
equation are correlated with each other. In the case of the absence of σ3 effect and collinearity problem.
octahedral stress model, normal (σoct) and shear stress By overcoming these two and having obtained
2
(τoct) parameters (two variables of regression equation) acceptable R values with the proposed model, it has
are both derived from the same major and minor principal been adopted to model the resilient behavior of the fine-
stresses, resulting strong correlation between them. This grained soils in this study. Table 5 presents the
means they are not actually independent variables as coefficient of determinations (R2) of multiple regression
assumed. The collinearity is evaluated by variance analyses based on Equation (9).
inflation factor (VIF). Most of the VIF values obtained in
this study ranged from 5 to 15 for octahedral model.
Possible collinearity is suspected with the VIF values The constants k1, k2 and k3
above 4 and VIF of 10 or above indicates that there is a
high risk of collinearity (Belsley et al., 1980). Therefore, Since the need for quantification of the material constants
the adoption of the octahedral stress model for this study is obvious, an extensive multiple regression analyses
2
has been omitted despite of yielding highest R values were also performed between the material constants and
among the models in the literature. some soil physical properties including moisture content
(w), dry density (ρd), clay content (C) plasticity index (PI),
activity (A) and degree of saturation (S). In regression
Having examined the possible combinations of stresses
analyses, the more the variables the higher the
(σd, σ3, σoct and τoct) with Er, the following model is
coefficient of determination (R2), however, “the more the
adopted in this study:
variables” also mean less practical and less statistically
sound expressions. Therefore, the most important and
log (Er/Pa) = k1 + k2 (σ d/Pa)+ k3 (σ 3/Pa), (9) practical factors were included in the final regression
equations. In this regard, for instance, degree of
Where Pa is atmospheric pressure (100 kPa), k1, k2 and saturation (S) is taken into consideration to represent the
k3 are material (model) constants. This semi-logarithmic combined effect of w- ρd condition. Another advantage of
model is an improved form of Equation (4) with the the use of S in the multiple regression analyses is that
inclusion of the parameterσ 3 and it has not been used the risk of collinearity would be eliminated as w and ρd
elsewhere to predict Er. The model was normalized with are interrelated. The range of moisture contents used in
respect to Pa to make it dimensionally consistent. Having the derivation of the equations was 2 percent below and
compared the model with the ones in the literature, the above the optimum. This moisture content range
model developed in this study has two main superiorities: resulted the degree of saturation to have changed from
it includes the effect of σ 3 and it does not subject to the 77 to 98%. It is believed that this falls into the range of
collinearity among the variables (σ d/Pa and σ 3/Pa) as VIF in-situ moisture conditions of Victorian subgrades. The
values were smaller than 4 (statistically meaning no values from optimum to the upper end of the range
possibility of collinearity). Therefore the predictive model represents the equilibrium moisture content of the
 Ozel and Mohajerani 4573

Table 6. The regression constants of Equation 9.

k1 k2 k3
Soil name
Range Average Range Average Range Average

S1 2.95 to 3.15 3.09 -0.27 to -0.07 -0.15 0.16 to 0.37 0.29

S2 2.34 to 2.88 2.66 -0.62 to –0.32 -0.43 0.29 to 0.75 0.43
S3 2.75 to 2.92 2.85 -0.18 to –0.09 -0.14 0.08 to 0.18 0.14
S4 2.72 to 2.86 2.78 -0.43 to –0.18 -0.25 0.09 to 0.40 0.25
All soils 2.34 to 3.15 2.87 -0.62 to –0.07 -0.23 0.08 to 0.75 0.27

subgrade, which is eventually reached in the subgrade capable of predicting acceptable values of Er for the
depending upon the water table level, rain fall, clay type experimental soils for a given stress state through the
and drainage conditions, while the lower end of the range calibration of model constants with respect to w, S, C and
could be encountered in the road subgrades of semi-arid PI as given in Equations 10, 11 and 12.
region of the state. The soils with more moisture
contents (beyond the range used in this study) are
unlikely to be found, as good drainage system is Model for resilient modulus with unconfined
desirable for modern roads. The use of following compressive strength
equations for moisture contents other than the ones used Since UCS test is one of the simple soil strength tests
in the derivations of the equations is not recommended, that can be done almost in any soil laboratory, several Er-
as the extrapolations may not be valid. qu correlations have been developed in an attempt to
For the four soils investigated in the study, the material facilitate the Er prediction by performing one simple UCS
constants k1, k2 and k3 can be best defined in terms of test. The relevant studies, for example, can be found in
some soil physical properties as follows: the works of Thompson and Robnett (1979) and Lee et
al.(1997). However, the correlations found in the
Log k1=0.019–0.468 × log w–59 × log C+0.073 × log PI, literature were developed for a specific stress level. In
R2 = 0.70, SEE = 0.017, (10) other words, the role of stress state on Er is not included
in the models. Here, an Er - qu correlation development
k2=0.16-1.14×log S+0.856×log C+0.195×log PI, has been attempted, which also accommodates a stress
R2 = 0.84, SEE = 0.056, (11) parameter.
To accomplish this, the measured Er values under five
k3=0.37+0.0023×S–0.0015×C–0.0058×PI DSL’s were first averaged for each moisture content
R2 = 0.48, SEE = 0.11. (12) group within each soil. In other words, the Er values
shown in Table 7 were obtained from the stress level of
The minimum, maximum and average values of the 6, 7, 8, 9 and 10 during the RLTT’s (Table 2). Then,
constants are also given in Table 6. It should be noted simple regression analyses were carried out between qu
that extreme values of the constants were excluded results and each set of average Er values and the
before establishing Equations10, 11 and 12 and the relationships were established in the form of following
values in Table 6. equation (The results can be seen in Table 8):
To prove the predictive capability of Equation (9)
through the calibration of model constants as given in Er = a + b qu . (12)
Equations 10, 11 and 12; material constants (k1, k2 and
k3) were first calculated using the measured soil physical To be able to include the stress affect into the model, the
properties for a total of 46 samples. Then, Er values were relationship between the constants a and b with
back calculated for the stress levels of 6, 7, 8, 9 and 10 corresponding DSL’s have been examined. Both
(Table 2). And the measured and predicted Er values constants have been found to be greatly dependent on
were averaged for each moisture content group. The DSL. These correlations were formulated by means of
results were compared as seen in Figure 2. It is known simple regression analyses as follows:
that the chosen stress levels and corresponding σ 3 of 30
kPa represent a meaningful range of stresses that a = -26.36 + 29.37×DSL, R2= 0.82,
subgrades are likely to be subjected to during the service. SEE=2.54; (13)
As can be seen from Figure 2, most values are found
close to the line of equality with a coefficient of variation b = 0.632 – 0.5458×DSL, R2= 0.98,
20. Therefore, it can be concluded that Equation 9 is SEE=0.013. (14)
 4574 Sci. Res. Essays

Table 7. Er, UCS and CBR test results.

S1 S2 S3 S4
Dry Optimum Wet Dry Optimum Wet Dry Optimum Wet Dry Optimum Wet

Er @DSL=0.2¶ 153 135 91 86 64 38 85 81 68 68 60 50

¶
Er @DSL=0.3 139 120 83 67 42 29 80 76 64 63 54 44
¶
Er @DSL=0.4 119 104 65 55 40 20 74 69 62 59 48 43
¶
Er @DSL=0.5 106 89 54 50 34 19 70 63 57 50 44 39
¶
Er @DSL=0.6 94 80 49 49 33 18 66 60 51 45 39 35
∗
qu (kPa) 330 270 210 200 175 120 220 195 165 170 145 105
CBR† 19 9 7 24 7 4 17 12 6 10 7 5
¶ †
Measured Er values (MPa) under the given DSL’s; ∗unconfined compressive strength; unsoaked California bearing ratio.

Table 8. Er - qu relationship.

a b R2 SEE
DSL=0.2 -18.05 0.5186 0.95 7.79
DSL=0.3 -21.15 0.4836 0.91 10.08
DSL=0.4 -14.39 0.4038 0.87 10.43
DSL=0.5 -11.06 0.3504 0.85 9.71
DSL=0.6 -8.41 0.3123 0.86 8.22

If a and b in Equation 12 are replaced with the Equations were also performed for each soil at the same three
13 and 14, Equation 15 given as follows is obtained: moisture content levels like RLTT and UCS samples.
Table 7 also gives the results of these tests. Since the
Er = 29.370×DSL + 0.632×qu – 0.546×σ d - 26.360, (15) Er–CBR expressions (like Er= 10×CBR in Austroads
pavement design guide) are recommended in the
Where Er is resilient modulus in Mpa, qu is unconfined pavement design procedures, the veracity of such
compressive strength in kPa, σd is deviator stress on top correlations has been investigated for the soils in this
of the subgrade (kPa) and DSL is the ratio of σd to the qu. study. Having compared Er values of various stress
This resultant equation is a very easy way of predicting Er levels with their CBR’s, it has been found that Er changes
through two simple parameters, which are qu and σd. by 2 to 15 times of the CBR values depending on the soil
To satisfy the usefulness of the Equation 15, Er values type, moisture content and stress level of the subgrade. If
were back calculated using qu of each condition and the the comparison is limited for the samples at wet side of
corresponding testing σd’s. Figure 3 shows that there is a optimum moisture content, however, Er values fall into a
good agreement between measured and predicted Er range where they deviate up to 30 percent from the
values through the established model. The same “10xCBR” prediction. Even in this case, Er prediction
coefficient of variance of 20 has been obtained for this through CBR is quite controversial. Therefore it was
comparison. Note that the Equation 15 was derived for concluded from the experimental data that Er cannot be
four fine-grained subgrade soils whose qu values range correlated to the CBR of the soils tested. If the “Er=
from 145 kPa to 275 kPa at optimum moisture content. 10×CBR” expression is utilized as suggested in several
The predictive capability of the model might lessen or not pavement design guides and still widely used by
be valid for fine-grained soils outside the range of qu practitioners, under- or over-estimation of the Er values
values used in this study. would be inevitable. Having checked in the same
manner, Equation 3 has also been found unsatisfactory
for Er prediction.
Resilient modulus – California bearing ratio
relationship Conclusions

As mentioned previously, California bearing ratio tests On the basis of the present study, the following
 Ozel and Mohajerani 4575

Figure 2. Comparison of measured and predicted resilient modulus values based on Equation 9, 10, 11 and 12.

Figure 3. Comparison of measured and predicted resilient modulus values based on Equation 15

conclusions can be drawn through the evaluation of (i) A new laboratory-testing program for the repeated load
resilient modulus of four Victorian fine-grained subgrade triaxial testing of subgrade soils was developed in which
soils: stress levels were defined as the ratios of soil strengths
4576 Sci. Res. Essays

(deviator stress level approach). Heukelom W, Klomp AJG (1962). Dynamic testing as a means of
controlling pavements during and after construction. Proceedings of
(ii) Unconsolidated undrained static triaxial (UU) or the international conference of the structural design of asphaltic
unconfined compressive strength (UCS) test results can pavements, Ann Arbor, MI.
both be used for establishing the stress levels for Houston WN, Houston SL, Anderson TW (1993). Stress state
repeated load triaxial tests as it has not been found much considerations for resilient modulus testing of pavement subgrade.
difference between the results of the two tests due to the Trans. Res. Record., 1406: 124-132.
Lo S-CR, Chen K (1999). Granular base materials subject to repeated
relatively low confining stresses applied in UU tests. loading along constant stress increment ratio path. Proceedings of
(iii) Octahedral stress model for the prediction of resilient th
the 8 Australia New Zealand Conference on Geomechanics, Hobart,
modulus (Er) has been found to be subjected to pp. 753-759.
Mohammad LN, Huang B, Puppala AJ, Allen A (1999). Regression
collinearity problem as the variables of the model are
model for resilient modulus of subgrade soils. Trans. Res. Record.
interrelated. 1687: 47-54.
(iv) A semi-logarithmic model, which expresses the stress Moffatt MA, Sharp, KG, Vertessy NJ, Johnson-Clarke JR, Vuong BT,
dependency of Er for fine-grained soils, has been Yeo REY (1998). The performance of in-situ stabilised marginal
sandstone pavements. Austroads Pavement Research Group,
introduced and its predictive capability has been
Melbourne. p. 22.
developed. This model incorporates both deviator and Moossazadeh J, Witczak MW (1981). Prediction of subgrade moduli for
confining stresses. The model constants, k1, k2 and k3, soil that exhibit nonlinear behavior. Trans. Res. Record., 810: 9-17.
can be calibrated with respect to some soil physical Muhanna AS, Rahman MS, Lambe PC (1998). Model for resilient
properties so that the proposed model can/could be used modulus and permanent strain of subgrades soils. Trans. Res.h
Record., 1619: 85-93.
to estimate the Er of other fine-grained soils. However, Nataatmadja A, Parkin AK (1989). Characterization of granular
confirmatory repeated load triaxial tests are materials for pavements. Canadian
recommended before adopting the proposed model for Geotech. J., 26: 725-730.
Nataatmadja A, Tan YL (2001) Resilient response of recycled concrete
other soils.
road aggregates. J. Trans. Eng., 127(5): 450-453.
(v) An alternative prediction model has also been Ozel MR, Mohajerani A (2001). Resilient modulus of a stabilised fine-
developed, which accommodates deviator stress (σd) and grained subgrade soil. Austr. Geomech. 36(3): 75-86.
unconfined compressive strength (qu). This model is Ozel MR, Mohajerani A (2002). Permanent deformation behaviour of
two Victorian subgrade soils under repeated loading. Austr.
superior over the earlier “Er – soil strength” expressions Geomech., 37(3): 41-50.
with the inclusion of the deviator stress attribute. Ozel MR (2003). Deformation response of fine-grained subgrade soils
(vi) The comparison of California bearing ratio (CBR) test for the design of flexible pavements. PhD Dissertation. RMIT
results with the Er values indicated that Er has changed 2 University, School of Civil and Chemical Engineering, Australia.
Powell WD, Potter JF, Mayhew HC, Nunn ME (1984). The structural
to 15 times the CBR depending upon the stress level, design of bituminous roads. Transport and Road Research
moisture content and the soil type. Therefore, it was Laboratory. Report LR 1132.
concluded that there can not be a unique Er – CBR Puppala AJ, Mohammad LN, Allen A (1996). Engineering behaviour of
expression. lime-treated louisiana subgrade soil. Trans. Res. Record., 1546: 24-
31.
Seed HB, Chan CK, Lee CE (1962). Resilient characteristics of
subgrade soils and their relation to fatigue failures. Proceedings of an
REFERENCES International Conference: Structural Design of Asphaltic Pavements,
Ann Arbor, pp. 611-636.
AASHTO (1993). Guide for Design of Pavement Structures. AASHTO, Shackel B (1973). The derivation of complex stress–strain relations.
th
Washington, D.C., USA. Proceedings of the 8 International Conference on Soil Mechanic and
AUSTROADS (1992). Pavement Design: A Guide to the Structural Foundation Engineering, Moscow, pp. 353-359.
Design of Road Pavements, Austroads, Sydney. Symons MG, Poli DC (1996). Road rehabilitation by recycling:
Belsley DA, Kuh E, Welsch RE (1980). Regression diagnostics: stabilisation of pavement soils from
Identifying influential data and sources of collinearity. John Wiley & Victoria, University of South Australia, Quensland, Australia.
Sons, New York, USA. Thomson MR, Robnett QL (1976). Final report: resilient properties of
Bodhinayake BC (2008). A study on nonlinear behaviour of subgrades subgrade soils. FHWA-IL-UI-160, University of Illinois, Urbana.
under cyclic loading for the development of a design chart for flexible Thomson MR, Robnett QL (1979). Resilient properties of subgrade
pavements. PhD dissertation, University of Wollongong, Australia. soils. Transportation Engineering J. ASCE 105.
Chen K (1999). Repeated Load Triaxial Testing on Unbound Pavement Thomson MR, LaGrow T (1988). A proposed conventional flexible
th
Materials. Proceedings of 8 Australia New Zealand Conference on pavement thickness design procedure. Transportation Engineering
Geomechanics, Hobart, pp. 747-751. Series (55), University of Illinois at Urbana-Champaign.
Drumm EC, Boateng-Poku Y, Pierce TJ (1990). Estimation of subgrade Witczak MW, Uzan J (1988). The universal airport pavement design
resilient modulus from standard tests. J.Geotech. Eng., 116(5): 774- system, Report I of V: Granular material characterization, University
789. of Maryland, USA.
Elliot RP, Thornton SI (1988). Resilient modulus and AASHTO Yoder EJ, Witczak MW (1975). Principles of pavement design, A Wiley-
pavement desing. Trans. Res. Record.1192: 1-7. Interscience Publication, Newyork, USA.
Fredlund DG, Bergan AT, Wong PK (1977). Relation between resilient
modulus and stress conditions for cohesive subgrade soils. Trans.
Res. Record., 642: 73-81.
Head KH (1982). Manuel of soil laboratory testing. Volume 2:
Permeability, shear strength and compressibility tests. Pentech
Press, London.

View publication stats