Design and construction of pavements on
swelling subgrade may induce sever problems due to volume changes of
the subgrade. Variations in water content in a subgrade may induce differential
deformations which result in cracking of
the pavement leading to a reduction in life expectancy of the pavement and/or
higher maintenance costs. The design and construction of pavements on subgrade
materials with inferior characteristics such as low strength, high potential
for volume change, high water content and
poor workability is not trivial. Improvement of the subgrade properties
could involve a combination of the following solutions: replacement of the
problematic soils, installation of impermeable membranes to guard against
changes in moisture content or the use of stabilizers such as cement based,
lime based, fly ash, polymer or liquid based stabilizers. Stabilization of the
subgrade material enhances the engineering properties of the subgrade layers
which produces a structurally sound pavement and will allow for the design of a
thinner overall pavement or alternatively
extended life span and reduction in required maintenance.
The primary objective
of this study is to gain insight to the effects of ANSS stabilizer on the behavior and
performance of a fine grained soil of high swell potential. The specific tasks
included determination
of engineering properties influenced by the stabilization such as vertical and horizontal swell, resilient modulus, permanent
deformation, volumetric strain and bulk modulus. The performance of the stabilized clay with ANSS was compared to the
performance of clay stabilized with
conventional stabilizers; cement and lime.
The investigation is
based on following sets of laboratory tests:
 (a) Laboratory swell tests aimed to
investigate the effect of ANSS content on the vertical volume change of swelling
clay as a function of initial water content and vertical pressure. A secondary set
included the definition and evaluation of the coefficient of swell (α_{ω}) as a
function of ANSS content and initial water
content. The coefficient of swell is defined as the ratio of vertical strain in response to a unit change in water content.
The coefficient of swell is an
indicator of the potential activity of the stabilized clay.
 Resilient modulus tests were performed
in order to investigate the resilient response of the stabilized clay. The resilient modulus
test is a repeated load triaxial test performed under constant confining
stress. The resilient modulus is defined as the ratio of the maximum axial
cyclic stress to the maximum recoverable axial strain of the specimen.
Resilient modulus of the unstabilized and stabilized clay was tested also after
permanent deformation and wetting process.
 Permanent deformation tests were
performed in order to evaluate the influence of stabilization
with
ANSS, lime and cement on unrecovered accumulated deformation during repeated loading and unloading. The permanent
deformation test was performed also after wetting process.
Soil
characterization
The soil examined in the research was
sampled from the Yizrael Valley in northern Israel. The tested clay is
classified as A76 according to AASHTO classification and as CH according to
the USCS classification system. Atterber limits of the clay are as follows:
Liquid Limit (LL), 77 %, Plastic Limit (PL), 28 % and Plasticity Index (PI) 49
%. The Specific gravity (Gs) of the clay is 2.72,
the free swell is 150% and the California bearing ration (CBR) is 3.5%. The
above properties indicate that the
soil is of high swell potential and low strength.
Swell tests
The aim of measuring the vertical swell
under an applied vertical surcharge is to examine the change of the vertical
dimension of the specimen (swelling percentage) over time, as the clay specimen
is allowed to absorb an unlimited amount of water. The apparatus used in this research
was a standard rigid ring consolidometer.
The specimen is laterally restrained and can absorb water freely through
two porous stones placed at the upper and lower boundaries of the specimen. The
change in height of the specimen is
monitored during the entire test.
Tests for the
determination of the coefficient of swell
Determination of the coefficient of swell
was accomplished by measuring the vertical strain per unit water content change
as the specimen wets up. The tests were performed under a vertical pressure of
1 kPa. Upon reaching a desired percent swell the specimen was quickly removed
from the consolidometer ring and the water
content measured. The resulting water content is the average water content since water absorbed into the specimen
is not uniform from all sides. The tests included 3
sets, one for each nominal water content; each set included natural clay specimens and specimens
stabilized with 2, 4 and 6% of the ANSS stabilizer. Every individual test was built from 4 points, each point representing the
vertical swell that the specimen reached up until
the point in time (percent swell) that the test was stopped and the water
content determined. A linear
regression of the four points going through the origin was computed. The slope of
the regression represents
the coefficient of swell per unit change in water content (α_{ω})
.
Resilient Modulus
and permanent deformation tests
The resilient modulus
tests were performed in accordance with the LTPP P46 Protocol test method. The procedure consists of applying 15 stress sequences
using a cyclic haversine shaped waveform with duration of 0.1 seconds and a
rest period of 0.9 seconds (fixed cycle duration of 1.0 sec). The haversine shaped load is considered the optimal
waveform to simulate the induced load in pavement layers. For
each sequence of the applied load, vertical and horizontal displacements were recorded by three
linear variable differential transducers (LVDT) and a clip gage to measure horizontal deformation.
Unstabilized and
stabilized specimens were tested for resilient modulus under the following
conditions:

after compaction
and curing at initial nominal water content of the plastic limit (28%).

after permanent deformation testing

after wetting of the specimen. All the
stabilized specimens were prepared with 4% stabilizer based on the dry weight of the soil.
The results were correlated to a
nonlinear loglog model (Uzan, 1985). This model was selected because it is
incorporated to the new MEPDG for unbound materials. In this model resilient modulus is expressed as a
function of bulk stress and deviator stress.
The same specimen tested for resilient
modulus was retested in the permanent deformation tests. Unstabilized and stabilized specimens were
tested for permanent deformation. The permanent deformation test was performed
at a single deviator stress level of 70 kPa and a single confining pressure of 20 kPa. The loads are of haversine
shape and were applied over 100,000 repetitions.
Permanent deformation tests enabled to
compute the bulk modulus (K) of the unstabilized and stabilized specimens.
The bulk modulus (K) is a measure of the material stiffness to volume change. The bulk modulus can be obtained by linear
elastic theory and is related to the elastic modulus (E) and
Poisson ratio (ν).
FINDINGS AND
CONCLUSIONS
Swell tests under
various vertical loads
 In general,
addition of ANSS stabilizer to the tested clay reduces the vertical swell. The reduction
is noted at all pressures and for all nominal water contents.
 The effect of
ANSS content on the vertical swell is not linear. The effect of the stabilizer content on the vertical swell for
stabilizer content of up to 4% is significant. This is true for different
pressures and nominal water contents. On the other hand, stabilizing beyond 4% produces minimal additional effect beyond that
produced by the 4% stabilizer content.

Stabilizing the soil under
consideration at contents greater than 4% ANSS does not yield any additional engineering benefit. The
vertical swell at high stabilizer contents, i.e. 4% and 6% are close to zero.
For vertical pressures between 150 kPa at all nominal water contents the vertical swell was limited to 0.041.1%.
Coefficient of
Swell tests

(a) Addition of ANSS stabilizer to the
soil produces a reduction in the coefficient of swell. The coefficient of swell
for the unstabilized soil reaches a value of 1.0. Specimens with 6% stabilizer led to a coefficient
of swell close to zero.

(b) The effect of stabilizer content on
the coefficient of swell is more significant than initial water content.

(c) The engineering
implication from the coefficient of swell tests is that the value of the coefficient of swell is a descriptor of
the activity of the clay, and the sensitivity of the soil to changes in water
content. Higher values of the coefficient of swell (α_{ω}) indicate that the soil is more sensitive to changes in water
content. It can be seen that at higher ANSS contents the soil is less prone to
swell in response to increase in water content.
Resilient Modulus
tests

Resilient modulus
increases as a result of stabilization. All three stabilizers improve the
resilient modulus of the clay tested. The
lime stabilized soil showed the highest improvement in resilient modulus at all
different stages of the test. For example, the resilient modulus for the
natural clay after compaction and curing was 63.0 MPa. Addition of 4% ANSS
increases the resilient modulus by a factor of about 2.5, the resilient modulus
reached a value of 141.0 MPa. Adding lime and cement to the clay increases the
resilient modulus by a factor of 7 and 2.5
respectively to resilient modulus of 438.0 MPa with lime and 150.0 MPa with
cement.

All samples exhibit a decrease in
resilient modulus after the wetting process. The degree of reduction in resilient modulus values
varied with stabilizer type. Still, resilient modulus of stabilized specimens
is less sensitive to moisture variations. Comparing resilient modulus before wetting to the resilient modulus after
wetting yields a factor of 6.6 in unstabilized clay, 3.5 with clay stabilized
with ANSS and 2.3 and 0.8 with specimen stabilized with cement and lime respectively.

Regression equations were developed to
estimate the resilient modulus. Predicted values were well
correlated to measured values for all specimens and at all experimental stages.

The deviator and confining pressure
had little effect on magnitude of the resilient modulus for either
stabilized or unstabilized specimens.
Permanent
deformation tests
The permanent deformation parameters were obtained by two methods as follows:

Vesys model, (1978) for which the rutting parameters α and μ are determined. μ is a parameter representing the constant proportionality between permanent
and elastic strain, α is a
parameter indicating the rate
of decrease in
permanent deformation as the number of load applications increase.

(2) A model
that connects between accumulated plastic strain
(ε_{p}) at N repetitions of load to the resilient strain (εr),
(ε_{p}/ε_{r}=EPER) by a second order polynomial equation.
The conclusions from the test results are as followed:

The permanent deformation parameters
by Vesys model for the unstabilized and stabilized
specimens at all
stages of the tests were unrealistic. For example, the Vesys parameters for the
stabilized materials after curing show
that as a result of the stabilization α decreases and μ increases, which
implies that rutting increases as a result of stabilization.

The development of EPER with load
cycles is similar to the accumulated permanent strain. Comparison
between computed EPER and measured EPER yield a good correlation.
Bulk modulus
results

(a) Bulk modulus increases as a result of
stabilization. The lime stabilized soil showed the highest improvement in bulk
modulus. For example, the bulk modulus of the natural clay after compaction and curing was 57.0
MPa. Addition of 4% ANSS increases the bulk modulus by a factor of about 2.3, the bulk modulus reached a value
of 129.0 MPa. Adding lime and cement to the
clay increases the resilient modulus by a factor of 6 and 1.7 respectively.

(b) After the wetting process the ANSS
stabilized specimens show a reduction in bulk modulus (from a value of 129 MPa in the dry
condition to 59 MPa in the wet condition). A fact that indicates the sensitivity of the material tested to variations in water
content.
In conclusion, a well designed stabilized
subgrade can have a significant structural benefit in the design of a flexible
pavement. A stabilized subgrade can improve the workability, reduce the swell potential, reduce the sensitivity to moisture
variations and improve the support of the pavement foundation by increasing the stabilized layer's resilient modulus and
strength.