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

net/publication/269557257

A state-of-the art review of 1-D heave prediction methods for expansive soils

Article  in  International Journal of Geotechnical Engineering · January 2012

DOI: 10.3328/IJGE.2012.06.01.15-41

CITATIONS READS

25 366

2 authors, including:

Sai K. Vanapalli
University of Ottawa
223 PUBLICATIONS   2,552 CITATIONS   

SEE PROFILE

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

Methods in Bio-Mediated CaCO3 Production for Improving the Liquefaction Resistance of Loose, Sandy Soils View project

Soil-pipeline system behavior extending the principles of saturated and unsaturated soil mechanics View project

All content following this page was uploaded by Sai K. Vanapalli on 11 December 2015.

The user has requested enhancement of the downloaded file.

 15

Sai K. Vanapalli1* and Lu Lu2

A state-of-the art review of 1-D heave

prediction methods for expansive soils

Abstract:  Expansive soils are typically in a state of unsaturated condition and are widely found in semi-arid and arid
regions of the world. The engineering properties of these soils are highly sensitive to changes in water content and suction
in the active zone depth. One of the key problems associated with expansive soils is their swelling characteristics due to an
increase in their natural water content. Practicing geotechnical engineers consider reliable prediction of the heave behavior
of expansive soils and understanding its impact on the structures as one of their greatest challenges. Several researchers and
practitioners from all around the world have made significant contributions over the past 60 years to better understand these
problematic soils. This paper synthesizes information of various techniques available in literature for estimating the swelling
pressure and the 1-D heave behavior of expansive soils. In addition, the limitations of using them in geotechnical engineering
practice are discussed.

Keywords: Unsaturated soils, expansive soils, suction, swelling pressure, 1-D heave prediction.

1.  INTRODUCTION
Conventional soil mechanics used for interpreting the engi-
neering behavior of saturated soils in practice cannot be
applied for expansive soils as they are typically in a state of
unsaturated condition. The engineering behavior of expan-
sive soils is sensitive to both the changes in water content and
as well as suction. Swell pressures associated with changes in
water content and suction of unsaturated expansive soils are
responsible to undesirable heave movements contributing to
significant damages to the structures. Expansive soils cause
damages to structures, pavements, pipelines, slopes, irriga-
tion channels and other constructions and contribute to
significant financial losses that far exceed other natural losses
such as earthquakes and tornados (Jones and Holtz, 1973).
The heave in an expansive soil deposit typically continues
until the degree of saturation is equal to 100%.
Practicing engineers rely on using a variety of design,
construction, and stabilization techniques to reduce losses
*Corresponding Author
Department of Civil Engineering, University of Ottawa, 161 Louis-Pasteur,
1
associated with expansive soils (Puppala et al. 2006). The
success of these techniques or procedures depends on key
information such as the reliable determination or estimation
of the swelling pressure and the soil heave. Reliable deter-
mination or estimation of the swelling pressure and the soil
heave however has proven to be difficult to date. Due to these
reasons, the design of foundations in expansive soils has been
one of the greatest challenges to the practicing geotechnical
engineers (Nelson & Chao 2003).
Reliable design of foundations on expansive soils should
include analyses of expected heave and consequences associ-
ated with the foundation movement over the design life of
the structure. The heave prediction or estimation techniques
in unsaturated expansive soils have been studied both using
the 3-D (for example, Vu & Fredlund 2004; Masia et al. 2004;
Wray et al. 2005) and the 1-D analysis (for example, Johnson
& Snethen 1978, Fredlund 1983, Hamberg & Nelson 1984,
Mitchell & Avalle 1984, Vanapalli et al. 2010a, b). The 1-D
heave is more commonly used in the estimation, prediction
and measurement in engineering practice applications. Most
of the available methods in the literature for the prediction of
heave generally make use of the linear relationship between

Ottawa, Ontario, Canada, K1N 6N5, telephone: (613) 562-5800 ext.

void ratio (vertical strain) and the net normal stress or soil
6638, fax: (613) 562-5173, e-mail: vanapall@genie.uottawa.ca suction (Fredlund 1983, Dhowian 1990 and Nelson et al.
Department of Civil Engineering, University of Ottawa, 161 Louis-Pasteur,
2 2006).
Ottawa, Ontario, Canada, K1N 6N5, telephone: (613) 562-5800 ext. Several millions of dollars have been invested towards
6197, fax: (613) 562-5173, e-mail: llu054@uottawa.ca collecting valuable field data to estimate heave in natural
expansive soils (for example, McKeen 1980; Snethen 1980;
International Journal of Geotechnical Engineering (2012) 6: (15-41) J. Ross Publishing, Inc. © 2012
DOI 10.3328/IJGE.2012.06.01.15-41
16  International Journal of Geotechnical Engineering
Fredlund 1983; Hamberg & Nelson 1984; Dhowian 1990;
McKeen 1992; Nelson & Miller 1992; Fityus & Smith 1998).
This paper provides a comprehensive summary of expansive
soil behavior in general and 1-D heave prediction and esti-
mation methods in particular that were developed during the
last 60 years and published in the literature.
2.  CONSTITUTIVE RELATIONSHIPS FOR
VOLUME CHANGE BEHAVIOR
Several constitutive relationships are available in the lit-
erature using the mechanics of unsaturated soils to interpret
the volume change behavior by using stress state variables
such as net normal stress, (σ–ua) and matric suction, (ua–uw)
(Matyas & Radhakrishna 1968, Alonso et al. 1990, Fredlund
& Rahardjo 1993). Figure 1 shows these constitutive relation-
ships which can be visualized in the form of volume-mass
constitutive surfaces on three-dimensional plots. The consti-
tutive surface can also be shown by plotting the volume-mass
parameters versus the logarithm of the stress state variables
(Nelson & Miller 1992). The constitutive surfaces exhibit
“uniqueness” (Fredlund & Rahardjo 1993). The “uniqueness”
of the constitutive surfaces demonstrates that there is only
one relationship between the deformation and stress state
variables (Chao 2007).
The basic constitutive relationships can be written in
three different forms; which include the elasticity param-
eters, the compressibility parameters and the volume-mass
form (Fredlund & Rahardjo 1993). These relationships are
summarized below.
2.1  Elasticity form
The elasticity form of the volume change constitutive rela-
tionships for an unsaturated soil (Fredlund & Rahardjo
1993) can be formulated as an extension of the equations
conventionally used for saturated soils. The soil is assumed
to behave as an isotropic, linear elastic material and the form
for soil structure is in below:
1–2µ 3 for an unsaturated soil, the volume change constitutive equa-
dεv = 3 d(σmean – ua) + d(ua – uw) (1)
E H
where, σmean = (σx + σy + σz) / 3 is the average total normal
stress; σx is total normal stress in the x-direction; σy is total
normal stress in the y-direction; σz is total normal stress in
the z-direction; εv is volumetric strain, dεv is volumetric strain
change for each increment, µ is Poisson’s ratio; E is modu-
lus of elasticity or Young’s modulus for soil structure; H is
modulus of elasticity for the soil structure with respect to a
change in matric suction, (ua–uw).
The water phase constitutive relationship proposed by
Fredlund & Rahardjo (1993) for a complete description of
the volume change behavior in an unsaturated soil can be
summarized as:
dVw 3 d(ua – uw)
= d(σmean – ua) + (2)
V0 Ew Hw
where, Vw is volume of water; V0 is initial overall volume of
an unsaturated soil element; Ew is the water volumetric mod-
ulus associated with a change in net normal stress, (σ–ua); Hw
is the water volumetric modulus associated with a change in
matric suction, (ua –uw).
2.2  Compressibility form
The constitutive equations can also be written in a compress-
ibility form which is more commonly used in conventional
soil mechanics. The equation for the soil structure of an
unsaturated soil to represent general three dimensional load-
ing is given below:
dεv = m1sd (σmean – ua ) + m1sd (ua – uw) (3)
where, m1s = 3(1 – 2μ) / E is the coefficient of compressibility
with respect to net normal stress; m2s = 3/H is the coefficient
of compressibility with respect to matric suction.
The compressibility form for the water phase in terms of
stress state variables (σ – ua) and (ua – uw) is:
dVw
= m1w d(σmean – ua) + m2w d (ua – uw) (4)
V0
where, m2s = 3/Ew is the coefficient of water volume change
with respect to net normal stress; m2s = 3/Hw is the coefficient
of water volume change with respect to matric suction.
2.3  Volume-mass form
The volume-mass form for the constitutive relationships for
an unsaturated soil involves the use of void ratio, e, water
content, w, and/or degree of saturation, S. Since the void
ratio change can be independent of the water content change
tion for three-dimensional loading is written in terms of void
ratio as below (Fredlund & Rahardjo 1993):
de = atd (σmean – ua ) + amd (ua – uw) (5)
where, at is the coefficient of compressibility with respect to
a change in net normal stress, d(σ – ua); am is the coefficient
of compressibility with respect to a change in matric suction,
d(ua – uw) (Figure 1).
For a complete volume-mass characterization, a second
constitutive relationship is required. Such a relationship can
 A state-of-the art review of 1-D heave prediction methods for expansive soils   17

3.  HEAVE PREDICTION METHODS

Void ratioe e
Void ratio (a)
Consolidation Lightly loaded structures such as pavements and residential
test plane
buildings constructed on expansive soils as discussed earlier
am are often subjected to severe distress subsequent to construc-
tion over a period of time as a result of changes in natural
water content and associated pore-water pressures (i.e., suc-
at tion) in the soil. The most commonly damaged structures
are near ground surface structures such as roadways, airport
runways, pipe lines, small buildings, irrigation canals and
spillway structures. This section provides a brief background
Matric suction (ua - uw)
of the heave prediction methodologies commonly used in
Net
Net normal (V- (uV
stress
normal stress a) - ua) engineering practice.
Heave prediction methods were first developed in the
late 1950s, and originated as an extension of methods used to
estimate volume change due to settlement in saturated soils
(b)
using results of one-dimensional oedometer (consolidation)
Water content w tests (Chao 2007). Heave prediction methodologies have
SWCC been refined continuously as knowledge and understanding
of unsaturated soil behavior has increased.
bm
Taylor (1948) proposed a mathematical model describ-
ing settlement of a layer of saturated expansive soil. The term
“swelling pressure” was coined and defined by Palit (1953),
bt as the pressure in an oedometer test required to prevent a
soil sample from swelling after being saturated. Jennings
& Knight (1957) first proposed the extension of settlement
theory to heave prediction using oedometer tests. Salas &
Matric suction (ua - uw )
Serratosa (1957) presented the oedometer heave prediction
Net normal stress (V- ua ) model in terms of the logarithmic pressure, and incorporated
the “swelling pressure” of a soil into the equation. Their
equation was of the same form as that presented by Taylor
Figure 1.  Constitutive surfaces for an unsaturated soil (a) Three- (1948).
dimensional void ratio constitutive surfaces (b) Three-dimensional water Aitchison (1973) was one of the earliest pioneers to
content constitutive surfaces (modified from Fredlund & Rahardjo 1993). propose a method for calculating moisture-related ground
movements taking account of change in pore-water pressure
be written as a water content relationship for three-dimen- (i.e., suction model). Fredlund et al. (1980) was probably the
sional loading as below: first of the investigators to provide a theoretical framework
to include soil suction into the prediction of heave in expan-
dw = btd (σmean – ua ) + bmd (ua – uw) (6) sive soils.
There are several techniques or procedures used in
where, bt is the coefficient of water content change with
geotechnical engineering practice to estimate the swelling
respect to a change in net normal stress, d(σ – ua); bm is the
pressure, swell potential and the 1-D heave in expansive soils.
coefficient of water content with respect to a change in mat-
These techniques can be divided into three main categories:
ric suction, d(ua – uw).
(i) empirical methods; (ii) oedometer test methods; and (iii)
The above relationships are rigorous and valuable in the
soil suction methods.
interpretation of the expansive soils; however, experimental
determination of such constitutive relationships are both
time consuming and difficult to develop. Therefore, the focus 3.1  Empirical methods
of research by several investigators from all over the world Empirical methods use soil classification parameters to pre-
has been directed towards developing simple 1-D heave pre- dict the swelling behavior of expansive soils. These methods
diction methods. are developed based on the limited data collected on local
18  International Journal of Geotechnical Engineering

soils or soils from a certain region. Table 1 summarizes the The CVS test procedure involves inundating the sample
list of several empirical methods from the literature. in the oedometer while seated under a nominal load, which
is typically 7 kPa (i.e., 1 psi). The load on the sample was
3.2  Oedometer test methods increased to prevent any volume increase or swelling of the
sample. The maximum applied stress required for main-
Prediction methods based on oedometer tests are more
taining constant volume condition was defined as the swell
widely used when compared to the other two methods. The
pressure. When the specimen no longer exhibits a tendency
swelling pressure determined from oedometer test methods
to swell, the applied load is further increased in a series of
is one of the key parameters used in the determination of the
increments in a manner similar to that of a conventional
1-D heave. Table 2 and Table 3 summarize these methods.
consolidation test. Once the recompression branch of the
Depending on the loading procedure, several methods
consolidation curve has been established, the sample was
are developed such as the free swell tests (FSV), the overbur-
allowed to rebound by complete load removal in order to
den swell tests, and the constant volume swell (CVS) tests
establish the swelling index, Cs.
using conventional oedometer test methods. The analysis
The test has two main measurements; namely, cor-
of the 1-D oedometer tests should also take into account
rected swelling pressure, P΄s and the swelling index, Cs. The
the loading and unloading sequence, surcharge pressure,
unsaturated undisturbed soil specimen collected from in-situ
sample disturbance and apparatus compressibility for reli-
condition is subjected to overburden stress (total stress). The
able determination of the swelling pressure. In many of the
total stress plane is a combination of total stress and matric
conventional consolidometer test procedures only the total
suction that provides an indication of the initial stress state
stress is controlled.
in the soil specimen. The volume change with stress state can
The evolution of heave prediction methodologies using
be computed by the change in stress state and the swelling
oedometer tests has been largely related to determination of
index, Cs.
the index parameters (i.e., swelling index, Cs; heave index
The matric suction is brought to zero during inundation,
Cρ or CH) and use them in the heave prediction equations.
but is not measured prior to this condition. However, the
Burland (1962) first proposed using the slope of the rebound
total stress on the specimen is increased to keep the specimen
portion of the consolidation swell curve for estimation of
from increasing in volume. The swelling pressure represents
heave. Fredlund (1983) indicated that the slope of the unload-
the sum of the in-situ overburden stress and the matric suc-
ing curve from consolidation-swell tests is approximately the
tion of the soil translated into the total stress plane (Figure 2).
same as the slope of the rebound curve determined from the
Therefore, the swelling pressure is dependent on the in-situ
constant volume tests.
soil suction.
Fredlund (1983) method and Nelson & Miller (1992)
The measured swelling pressure is underestimated
method used oedometer test results from both the consoli-
unless the effect of “sampling disturbance” and “apparatus
dation-swell test and the constant volume test to determine
compressibility” are taken into account (Fredlund 1969).
the index parameters (i.e., swelling index, Cs and heave index,
The interpretation of the CVS test must include a correction
Cρ). Nelson & Miller (1992) method (Equation (21)) uses
for the compressibility of the consolidation apparatus, the
the same equation as Fredlund (1983) method (Equation
compressibility of filter paper, and the seating of the porous
(19)). Feng et al. (1998) presented a comprehensive com-
stones and the soil specimen. Sample disturbance will result
parison study of swell pressure from different oedometer test
in a measured swelling pressure that is lower than the in-situ
methods. Nelson et al. (1998) and Bonner (1998) presented
value. Therefore, to determine the corrected swelling pres-
a method of estimating the index parameter, CH using test
sure, P΄s, the laboratory data should be corrected to account
results from only consolidation-swell tests. Nelson et al.
for the compressibility of the apparatus. The correction of
(2006) refined the analysis and developed the methodology
sampling disturbance is also used in order to establish P΄s.
for determining the percent swell as a function of the inunda-
The “uncorrected” swelling pressure is typically low to use in
tion pressure.
the prediction of total heave. Predictions using “corrected”
swelling pressures may often be twice the magnitude of those
3.3  Fredlund (1983) method: computed when no correction is applied (Fredlund 1983).
Fredlund (1983) proposed an equation that can be used to The following procedure is suggested for obtaining P΄s
calculate the 1-D heave in expansive soils using the constant from CVS test results. When interpreting the laboratory
volume swell (CVS) oedometer test results (see Equation (20) data, an adjustment should be made to the data in order to
in Table 3). account for the compressibility of the oedometer apparatus.
Desiccated, swelling soils have a low compressibility, and the

Table 1. Summary of the empirical methods
Author
Seed et al. (1962)
Van der Merve (1964)
Ranganathan & Satyanarayana (1965)
Nayak & Christensen (1971)
Vijayvergiva & Ghazzaly (1973)
Schneider & Poor (1974)
Chen (1975)
Weston (1980)
Picornell & Lytton (1984)
Dhowian (1990)
Vanapalli et al. (2010a)
A state-of-the art review of 1-D heave prediction methods for expansive soils   19
Equation/Description Eqn.
SP = 0.00216IP 2.44
(7)
SP = swelling potential, %;
Ip = plasticity index.
ΔH = Fe–0.377D(e–0.377H–1 (8)
H = volume change;
ΔH = total heave;
F = correction factor for degree of expansiveness;
D = the thickness of non expansive layer;
SP = 0.000413IS2.67 (9)
IS = shrinkage index, (LL-SL).
LL = liquid limit
SL = shrinkage limit
SP = 0.00229IP2.67(1.45c)/wi + 6.38 (10)
PS(psi) = [(3.58 · 10-2) IP1.12 c 2/wi2] + 3.79 (11)
wi = initial water content;
Ps = the swelling pressure;
c: clay content.
SP = 1/12 (0.4LL – wi + 5.55) (12)
log SP = 0.0526γd + 0.033LL – 6.8) (13)
LL = liquid limit.
log SP = 0.9(Ip /wi) – 1.19 (14)
SP = 0.2558e 0.08381Ip (15)
SP = 0.00411LLw 4.17σv–3.86wi–2.33 (16)
LLw = weighted liquid limit
n
(17)
ΔH = Σ f (V/V) H
1
i i
H = the stratum thickness;
Δv/vi = volumetric stain.
Fi = factor to include the effects of the lateral confinement;
ΔH = (SP%) H (18)
H KPf (19)
ΔH = Cs log
1 + e0 Cs Δw
Cw
10
H = thickness of the soil layer;
Pf ( = σy + Δσy – uwf) = final stress state;
Pʹs = corrected swelling pressure;
K = correction parameter;
Cs = swelling index;
σy = total overburden pressure;
Δσy = change in total stress;
uwf = final pore-water pressure;
e0 = initial void ratio.
Cw = suction modulus ratio;
Δw = change in water content.
20  International Journal of Geotechnical Engineering

Table 2. Oedometer tests used for heave prediction (modified and expanded after the original contribution of Nelson & Miller 1992)
Tests Location Description Reference
Double oedometer South Africa Two tests performed on adjacent samples; a consolidation-swell test under a small surcharge Jennings &
method pressure and a consolidation test, performed in the conventional manner but at natural mois- Knight (1957)
ture content. Analysis accounts for sample disturbance and allows simulation of various loading
conditions and final pore-water pressures.
Volumenometer South Africa Uses specialized apparatus, air-dried samples were inundated slowly under overburden pres- DeBruijin
method sure. (1961)
Sampson, Schuster & Colorado, Two tests performed on adjacent samples to simulate highway cut conditions; a consolidation- Sampson et
Budge method USA swell test under overburden surcharge, and constant volume-rebound upon load removal test. al. (1965)
Noble method Canada Consolidation-swell tests of remolded and undisturbed samples at various surcharge loads to Noble (1966)
develop empirical relationships for Canadian prairie clays.
Sullian and McClelland USA Constant volume test, samples initially under overburden pressure on inundation. Sullivan &
method McClelland
(1969)
Komornik, Wiseman & Israel Constant volume tests at various depths and swell-consolidation tests at various initial sur- Komornik et
Ben-Yacob method charge pressures representing overburden plus equilibrium pore water suction, used to develop al. (1969)
swell versus depth curves.
Navy method USA Swell versus depth curves determined by consolidation-swell tests at various surcharge pres- Navy (1971)
sures representing overburden plus structural loads.
Wong & Yong Method England Swell versus depth is determined as in Komornik, Wiseman & Ben-Yacob method and Navy Wong & Yong
method; but surcharge loads of overburden plus hydrostatic pore water pressures are used. (1973)
USBR method U.S.A. Double sample test, a consolidation-swell under light load, and a constant volume test. Gibbs (1973)
Direct model method Texas, USA Consolidation-swell tests on samples inundated at overburden or end-of-construction sur- Smith (1973)
charge loads.
Simple Oedometer South Africa Improved from double oedometer test. Single sample loaded to overburden, then unloaded to Jennings et al.
constant seating load, inundated and allowed to swell, followed by usual consolidation proce- (1973)
dure.
Mississippi State Mississippi, Consolidation-swell tests on remolded or undisturbed samples inundated at overburden sur- Teng et al.
Highway Dept. USA charge loads. (1972; 1973)
method Teng & Clisby
(1975)
Controlled strain test Colorado, Constant volume swell pressure obtained on inundation followed by incremental, strain-con- Porter &
USA trolled pressure reduction. Nelson (1980)
University of Canada Constant volume test. Procedure includes sample disturbance and apparatus deflection. Fredlund et al.
Saskatchewan (1980)
Sridharan, Rao & India Tests results from three methods, namely, a) conventional consolidation tests, b) equilibrium Sridharan et
Sivapullaiah method void ratios for different consolidation loads, and c) constant volume method are combined al. 1986
to study the swelling pressure of expansive soils. Results show that method a) gives a upper
bound value; method b) gives the least value; and method c) gives the intermediate value.
Erol, Dhowian & Saudi Arabia Assessment of the various oedometer test methods of ISO (improved swell oedometer test), Erol et al.
Youssef method CVS (constant volume swell test) and SO (swell overburden test) is used for heave prediction. (1987)
Shanker, Ratnam & India Studying the multi-dimensional swell behavior by testing cubic soil samples in oedometers. Shanker et al.
Rao method Swelling of samples is allowed to occur in 1-, 2- or 3- dimensions under a token surcharge. (1987)
Al-Shamrani & Saudi Arabia The stress path triaxial cell and oedometer are used to evaluate the vertical swell of expansive Al-Shamrani
Al-Mhaidib method soils under multi-dimensional loading conditions. Several series of triaxial swell tests were con- & Al-Mhaidib
ducted in which the influence of confinement on the predicted vertical swell was evaluated. (1999)
Basma, Al-Homoud & Jordan Two commonly used methods, zero swell test and the swell-consolidation test; and two rela- Basma, et al.
Malkawi method tively new techniques, “restrained swell test” and “double oedometer swell test” are using to (2000)
study the swell pressure of the expansive soil. The restrained swell test is believed to give more
reasonable results for swell pressure determination and thus is considered to more closely
resemble field conditions.
Subba Rao & Tripathy India One-dimensional oedometer is used to study the swell-shrinkage behavior of the compacted Subba Rao
method expansive soils. The compression-rebound tests were conducted on aged and un-aged com- & Tripathy
pacted specimens by incrementally loading them up to a certain surcharge and then unloading. (2003)
The cyclic swell-shrinkage tests were carried out in fixed ring oedometers with the facility for
shrinking the specimens at fixed temperature under constant surcharge pressure.
 A state-of-the art review of 1-D heave prediction methods for expansive soils   21

Table 3. Summary of oedometer tests methods

Author Equation/Description Eqn.
Fredlund (1983) H P (20)
ΔH = Cs log f
1 + e0 Pʹs
Hi = thickness of the ith layer;
Pf ( = σy + Δσy – uwf) = final stress state;
Pʹs = corrected swelling pressure;
Cs = swelling index;
σy = total overburden pressure;
Δσy = change in total stress;
uwf = final pore-water pressure;
e0 = initial void ratio.
Dhowian (1990) Cs P (21)
ΔH = H log s
1 + e0 P0
Cs = swell index;
Ps = swelling pressure;
P0 = effective overburden pressure.
Nelson & Miller (1992) Cρ σʹf (22)
ΔH = H log
1 + e0 σʹcv
Cρ = heave index;
σʹcv = swelling pressure from constant volume swell test;
σʹf = vertical stress at the midpoint of the soil layer for the
conditions under which heave is being computed.
Nelson et al., (2006) σʹcv (23)
ΔH = H CH log
σʹvo

%SA
CH = (24)
σʹcv
log
(σʹi)A
CH = heave index;
σʹcv = swelling pressure from constant volume swell test;
σʹvo = vertical stress at the midpoint of the soil layer for the
conditions under which heave is being computed.

compressibility of the apparatus can significantly affect the
evaluation of in situ stresses and the slope of the rebound
curve (Fredlund 1969). Because of the low compressibility
of the soil, the compressibility of the apparatus should be
measured using a steel plug substituted for the soil speci-
men. The measured deflections should be subtracted from
the deflections measured when testing the soil. The adjusted
void ratio versus pressure curve can be sketched by drawing
a horizontal line from the initial void ratio, which curves
downward and joins the recompression curve adjusted for
the compressibility of the apparatus (Fredlund & Rahardjo
1993). Second, a correction can be applied for sampling dis-
turbance after determining the equipment compressibility.
Sampling disturbance increases the compressibility of the
soil, and does not permit the laboratory specimen to return
to its in situ state of stress as its in situ void ratio. More
detailed testing procedures of this technique are available in
ASTM D4546-2000.
Casagrande (1936) proposed an empirical construction
on the laboratory curve to account for the effect of sampling
disturbance when assessing the preconsolidation pressure
of a soil. A modification of Casagrande’s construction was
extended for determining Pʹs. In this method, the point of
maximum curvature where the void ratio versus pressure
22  International Journal of Geotechnical Engineering

Void ratio, e
Void ratio

Final stress conditions

ef = 1.131
Actual stress path
Analysis stress path Shrinkage curve
e 0 = 0.962 (ua - u w) field S r <100%
(y - ua )
field
ef
Swelling pressure Matric suction
P' s = 627 kPa
Insitu
equation Cw
stress state
(ua - uw )e 'e
eo Saturation line

Rebound curve

a)
Matric suction, (ua- uw )
-u
, (
ss
stre
al
rm Shrinkage limit
t no 'w
Ne

Figure 2.  Stress paths representing swelling of Regina soil (modified wo wf

after Fredlund, 1983). Water content

Figure 4. Water content versus void ratio relationship and the deter-
mination of suction modulus ratio, Cw (modified after Hamberg 1985).

ment between the estimated heave and the in-situ measured

heave. More details of this case study are available in Yoshida
et al. (1983) and Fredlund & Rahardjo (1993).

3.4  Soil suction methods

Figure 3. Construction procedure to determine the corrected swelling
pressure incorporating the effect of sampling disturbance (modified
after Fredlund 1987).
curve bends downward onto the recompression branch is
determined (Figure 3). At the point of the maximum curva-
ture, a horizontal line and a tangential line are drawn. The
“corrected” swelling pressure is designated as the intersec-
tion of the bisector of the angle formed by these lines and
a line parallel to the slope of the rebound curve which is
placed in a position tangent to the loading curve (Fredlund
& Rahardjo 1993).
Fredlund (1983) method has been tested to predict 1-D
heave on Regina soil. This method provided a good agree-
The swelling pressure and the 1-D heave in expansive soils
can be more reliably measured or calculated using soil suc-
tion methods as they are based on the information of the
stress state (i.e., suction). In these methods, the influence
of suction is taken into account through the use of different
parameters. Several heave prediction formulations based on
soil suction methods proposed by various researchers are
summarized in this section.
3.5  Hamberg & Nelson (1984) methods
Hamberg & Nelson (1984) presented an approach (Equation
(41), see Table 3) that can be used to predict heave of expan-
sive soils using the relationship between water content and
volume change in the range of shrinkage limit to liquid limit.
The parameter, Cw in Equation (42) represents the variation
of volume of soil specimens with respect to water content
(see Figure 4) and can be obtained using the Clod test which
is the modified form of COLE (coefficient of linear extensi-
bility; Brasher et al. 1966) test. The COLE test is originally
developed to determine the heave beneath airfield pavements
(McKeen 1981, McKeen & Hamberg 1981).
 A state-of-the art review of 1-D heave prediction methods for expansive soils   23
The test procedure involves coating soil samples with a
liquid resin (i.e., DOW Saran F310) that allows for volume
measurements at different moisture conditions (Nelson &
Miller 1992). Once the resin dries on the soil sample, it acts
as a flexible membrane, containing the soil material with its
natural soil fabric intact. The resin is essentially waterproof
when exposed to liquid water for a short time, but it per-
mits gradual water vapor flow to and from the sample. The
volume of a soil sample of any shape may be determined by
weighing the soil clod while it is submerged underwater on a
balance. The reading of the balance, adjusted for the weight
of the pan and water, is a direct measurement of buoyant
force on the sample. Sample volume can then be determined
by Archimedes’ principle.
In the original COLE procedure, each resin-coated
sample was brought to 33 kPa suction in a pressure plate
device. The sample volume was determined at the initial
adjusted moisture condition using the weighing procedure
as described above. The samples were then oven dried for 48
hours, followed by another volume determination (Nelson &
Miller 1992).
A COLE value was defined as the normal strain that
occurs from the moist to the dry condition, defined with
reference to the dry dimension, as Equation (55) (Grossman
et al. 1968).
LM – LD LM γdM 0.33
COLE = = –1 = γ –1 (55)
LD L dD
where, LM is the length of moist sample at 33 kPa suction; LD
is the length of oven dried sample; γdM is the dry density of
moist sample at 33 kPa suction; and γdD is the dry density of
oven dried sample.
The basic difference between the Clod test and the COLE
test procedures is that in the Clod test, volume changes are
monitored along a gradually varying moisture change path.
This results in a smooth shrinkage or swelling curve for
each sample (Nelson & Miller 1992). The basic Clod test
procedure to develop a shrinkage curve is as follows: (i) coat
a sample with resin according to the COLE procedure speci-
fications and measure its volume; (ii) allow sample to dry
slowly in air; (iii) when the sample reaches a constant weight
under constant humidity conditions in a laboratory, volume
and weight measurements are taken; (iv) samples are oven
dried for 48 hours and final volume and weight measure-
ments are made.
The tests data provide void ratio and water content at
various points. Void ratio and water content are both directly
related to soil suction. This relationship is valid only at water
contents greater than the shrinkage limit; below the shrink-
age limit, changes in water content are not accompanied by
changes in volume, by definition (Hamberg 1985). For silty
clay soils, the Cw versus water content relationship shows
linear behavior for the water content greater than shrinkage
limit (Hamberg 1985; Figure 4).
The limitations of Equation (41) are: (i) the equation
does not take into account the effect of applied load; and (ii)
there are difficulties in estimating the Cw values for the water
contents close to shrinkage limit due to the nonlinearity in
that range. Table 4 lists the most commonly used soil suction
methods.
4.  DETAILS OF HEAVE PREDICTION
TECHNIQUES
Early swelling behavior studies were based on limited stud-
ies that focused on developing empirical relationships useful
for explaining the local expansive soil behavior. Presently,
oedometer tests methods (Table 3) and soil suction methods
(Table 4) are more commonly used in engineering practice.
Fredlund (1983) method and Hamberg & Nelson (1984)
methods were discussed in greater detail in the earlier sec-
tion. This section briefly provides the details of other heave
prediction methodologies in the literature.
4.1  Aitchison (1973) method
This is probably the one of the earliest methods published in
the literature that uses soil suction as a tool in the estimation
of soil suction heave (Jaksa et al. 1997). This method was
originally proposed for estimating heave in residential foun-
dations built on expansive soils. The surface movement, ΔH
can be estimated using Equation (25) (see Table 4).
The ratio of vertical strain to suction change, Δu was
experimentally observed and defined as instability index,
IPt by Aitchison & Woodburn (1969), Aitchison (1970), and
Lytton & Woodburn (1973). This index is equivalent to
the suction index, Cτ used by Snethen (1980) and Johnson
(1979), or suction compression index, γh used by McKeen &
Hamberg (1981). The Australian Standard for the design and
construction of residential slabs and footing, AS 2870-1996
(Standards Australia 1996), specifies three methods for the
estimation of instability index, IPt:
1. Laboratory tests. Three such tests are suggested;
the shrink-swell test, AS 1289.7.1.1 – 1992; the
loaded shrinkage test, AS 1289.7.1.2 – 1992; and
the core shrinkage test, AS 1289.7.1.3 – 1992
(Standards Australia 1992);
2. Correlations between the instability index, IPt and
other clay index tests;
24  International Journal of Geotechnical Engineering

Table 4. Summary of soil suction methods.

Author Equation/Description Eqn.
Hs
Aitchison (1973) 1 (25)
ΔH =
100 ∫I
0
Pt
Δu Δh

ΔH = surface movement;
IPt = instability index of the soil;
Δu = change in suction, in pF units, at depth z
z below the ground surface;
Δh = thickness of the soil layer under consideration;
Hs = depth of the design suction change
Lytton (1977) ΔH = γhH log10(hf /hi) – γσH log10(σf /σi) (26)
hf , hi = final and initial water potentials;
σf = applied octahedral normal stress;
σi = is the octahedral normal stress above which overburden
pressure restricts volumetric expansion;
γh , γσ = two constants characteristic of the soil.

Johnson & Snethen (1978) Cτ h0

ΔH = H log (27)
1 + e0 hf + ασf
Cτ = α GS /(100 B) (28)

log h0 = A – B w0 (29)

H = the stratum thickness;

Cτ = suction index;
α = compressibility index;
e0 = initial void ratio;
hf = final matric suction , kPa;
σf = final applied pressure, (overburden + external load), kPa;
h0 = matric suction without surcharge pressure, kPa
Snethen (1980) Cτ
ΔH = H (A – B w0 ) – log (τmf + α σf ) (30)
1 + e0
Cτ = α GS / 100 B (31)

log τm = A – B w0 (32)

Cτ = suction index;
τmf = final matric suction;
σf = final applied pressure (overburden + external load);
α = compressibility factor
A, B = constants (y-intercept and slope of soil suction versus
water content curve, respectively).
 A state-of-the art review of 1-D heave prediction methods for expansive soils   25

Table 4. (continued) Summary of soil suction methods.

Author Equation/Description Eqn.

McKeen (1992) hf
ΔH = – γh H log (33)
hi
Δv/vi
γh = – (34)
hf
log10
hi
ΔH = H Ch log Δτ f s (35)
Ch = (–0.02673)(dh/dw) – 0.38704 (36)
f = (1 + K0 ) /3 (37)
s = (1 – 0.01(%SP) (38)

γh = suction compression index;

hf , hi = final and initial weighted suction, respectively.
Δv/vi = volume change with respect to initial volume.
Ch = suction compression index, is the slope of volume change-
soil suction curve;
Δτ = the suction change in pF;
f = the lateral restraint factor;
K0 = the coefficient of earth pressure at rest, equal to 1;
s = the coefficient for load effect on heave;
SP = the percent of swell pressure applied.
Mitchell & Avalle (1984) ΔH = IPt Δu H (39)
ΔL/L Δw
IPt = . (40)
Δw Δu
IPt = instability index;
Δu = soil suction change.
Hamberg & Nelson (1984) Cw
ΔH = H Δw (41)
1 + e0
Δe
Cw = (42)
Δw
Cw
ΔH = H log (h)i (43)
1 + e0
Ch = Cw Dh (44)

Cw = suction modulus ratio;

Δw = change in water content;
Ch = suction index with respect to void ratio;
Dh = suction index with respect to moisture content.
26  International Journal of Geotechnical Engineering

Table 4. (continued) Summary of soil suction methods.

Author Equation/Description Eqn.

Dhowian (1990) Cψ ψ
ΔH = H log ψi (45)
1 + e0 f
α Gs
Cψ = (46)
100B
α Gs
ΔH = H (w – wi) (47)
1 + e0 f
ΔH = H Cw (wf – wi) (48)

Cw = α Gs / (1 + e0) (49)

Cψ = suction index;
ψi = initial suction;
ψf = final suction;
α = volume compressibility factor
B = slope of suction versus water content relationship;
Gs = specific gravity of solid particles;
Cw = moisture index
Fityus & Smith (1998) ΔH = H Iv α (w0i – w0f ) (50)

ΔH = H Cw (wf – wi) (51)

Iv = volume index;
α = empirical factor accounting for confining stress differences
in lab and field;
w0i , w0f = average initial water content and the average final
water content, respectively;
σv = vertical stress at the midpoint of layer.
Briaud et al. (2003) ΔH = Hf (Δw – Ew) (52)

Ew = Δw (ΔV/V0) (53)

f = (ΔH/H0 ) (ΔV/V0) (54)

Ew = shrink–swell modulus, slope of the water content versus

the volumetric strain line;
f = shrinkage ratio, ratio of the vertical strain to the volumetric
strain

3. Visual-tactile identification of the soil by an engi-
neer or engineering geologist having appropriate
expertise and local experience.
The method of estimating the instability index, IPt has
been widely used throughout the geotechnical engineering
community (Mitchell 1989, Jaksa et al. 1997, Fityus & Smith
1998). The method 3 was named as visual-manual method
(Mitchell, 1989). This technique involves a visual inspec-
tion of the soil and manually moulding and kneading the
soil in order to estimate its plasticity index, Ip (Jaksa et al.
1997). Mitchell (1979) presented an approximate relation-
ship between plasticity index Ip and instability index, IPt
which is often used in the visual-tactile method to estimate
the instability index, IPt. The relationship for estimating the
 A state-of-the art review of 1-D heave prediction methods for expansive soils   27

Soil Suction, h, log(kPa)

5
Soil Suction (kPa)

10

Log W = A - Bw 4

3 Slope, dh/dw = Suction Potential

1
2

1
22 24 26 28 30 0 10 20 30 40 50 60

Water Content (%) Water Content (%)

(a)
Figure 5. Soil suction versus water content relationships for Blue Hill
0.4
Shale (modified after Snethen 1980).

Volumetric Strain, H 0.3

instability index, IPt is detailed in the Mitchell & Avalle (1984)
method (see Section 4.6).
However, by its nature, the method is highly classifier-
0.2
dependent (Jaksa et al. 1997). In addition, the laboratory and
in-situ tests performed to obtain the required parameters
(i.e., instability index, IPt,; soil suction change, Δu) for the 0.1
approach are both time consuming, expensive and difficult.
Slope, dH/dh = Suction Compression Index

4.2  Lytton (1977) method 0.0

The method is used for estimating the heave and rate of
heave of foundations resting on expansive soil deposits. The
approach is based on simple laboratory tests and some field
observations of the shrinkage crack network. The ground
surface heave experienced by an elemental volume of soil
due to a change in water potential can be calculated using
Equation (26) (see Table 4). The volumetric strain can be
represented as Equation (56):
ΔV
= γhlog10 (hf /hi ) – γσ log10 (σf– σi ) (56)
V0
The first term (i.e., γhlog10 (hf /hi )) is the volumetric
strain due to a change in water potential. The contribution
of the second term (i.e., –γσ log10 (σf– σi ) ), which has the con-
trary sign, is only considered with increasing depth until the
strains become zero (Picornell & Lytton 1984).
4.3  Johnson & Snethen (1978) method
Soil heave is predominantly induced by suction change
within the active zone depth. When an expansive soil is wet-
ted, its soil suction decreases and the soil volume increases.
1 2 3 4 5 6
Soil Suction, h, log (kPa)
(b)
Figure 6.  Determination of the suction compression index (modified
after McKeen 1992).
This concept was used as a tool in the estimation of unit
heave which is expressed in Equation (27) (see Table 4).
The initial suction can be measured by thermocouple
psychrometer or filter paper method. Soil suction measure-
ments include preparing several undisturbed soil cubes and
modifying the moisture contents of the samples to fit the
range of field conditions of moisture variation (Snethen &
Huang 1992). The resulting soil suction versus moisture
content relationship is used to establish the required param-
eters. The parameters A and B, and compressibility factor
are determined from the plotted results of the soil suction
test procedure. A is the soil suction value (logarithmic scale)
at zero moisture content; B is the slope of soil suction versus
moisture content curve. The compressibility index, α is the
slope of specific volume, (1+e)/Gs, versus moisture content
28  International Journal of Geotechnical Engineering

Table 5. McKeen’s swell potential categories

20
(Olsen et al. 2000)
Suction
Swell Potential Suction Compression
Category Potential dh/dw Index, Ch 15

Strain 'L/L (%)

I Very High > -6 < -0.227
(McKeen
calls this
10
category
“Special ('L/L) / 'MC
Case”)
II High -6 to -10 -0.227 to -0.120 5
III Moderate -10 to -13 -0.120 to -0.040
IV Low -13 to -20 -0.030 to non-expansive
0
0 5 10 15 20 25 30 35
Water Content (%)
curve. The suction index, Cτ, reflects the rate of change of (a)
void ratio with respect to soil suction. The initial soil suction, 6.5
h0, is measured during suction testing and the final suction
profile is assumed based on one of the four criteria suggested 6.0
by Snethen (1980): (i) zero throughout the depth of active Total Soil Suction, u (pF)
zone; (ii) linearly increasing with depth through the active 5.5
zone; (iii) saturated water content profile; (iv) constant at
some equilibrium value. 5.0

4.4  Snethen (1980) method 4.5 Ipt = (('L/L) / 'MC) * ('MC / 'u)

This method uses soil suction data (Figure 5) to characterize

4.0
the expansive soil:
logτm = A – Bw (57) 3.5
0 5 10 15 20 25 30 35 40
where, τm is the soil suction without surcharge pressure (i.e., Water Content (%)
atmospheric pressure). The slope of the line, B is determined (b)
by calculating the inverse of the change in water content
over one cycle of the soil suction scale. The intercept, A is Figure 7. Determination of instability index IPt (modified after Mitchell &
calculated by applying Equation (32) at soil suction equal to Avalle 1984).
95.8 kN/m2. The volume change and swelling pressure of an
expansive clay stratum is estimated using Equation (30) (see
Table 4) and Equation (58): factor may be roughly estimated from the plasticity index,
log Ps = A – (100Be0 /Gs) (58) IP as: (i) IP < 5, α = 0; (ii) IP > 40, α = 1; (iii) 5 < IP < 40, α =
0.0275 IP -0.125.
The suction index, Ct reflects the rate of change of void The equation provides predictions of in-situ volume
ratio with respect to soil suction. The laboratory data neces- change of a soil stratum with respect to field conditions of
sary to apply the equation include Gs, e0, A, B, w0, and α, all of soil composition, structure, initial and equilibrium moisture
which (expect Gs) can be determined in the soil suction test profiles, and confining pressures.
procedure (Snethen 1980). The remaining two variables, τmf
and σf are functions of assumed depth of active zone and the 4.5  McKeen (1980, 1992) methods
assumed final soil suction profile. The compressibility factor,
α for highly compressible clays (i.e., CH) is commonly set McKeen (1980, 1992) used the suction compression index
equal to one, because the voids of these soils are filled with γh introduced by Lytton (1977) as a tool to propose a heave
water within a wide range of moisture contents (Snethen prediction technique as shown in Equation (34) (see Table 4).
1980). In the absence of measured data, the compressibility A measure of the physical phenomena was needed
to facilitate rating, classifying, and discussing the volume
 A state-of-the art review of 1-D heave prediction methods for expansive soils   29
8 200
Alluvial
RB3
Instability Index, Ipt (%)
RB5
6 150
Hindmarsh Clay
BE
4 100
2 50
0
0 20 40 60 80 100
Plasticity Index, Ip (%)
Figure 8.  Relationship between instability index and plasticity index
(modified after Mitchell & Avalle 1984).
change behavior of expansive soils. This was addressed by
suction compression index, γh which is volume-response
coefficient. The suction compression index, γh can be deter-
mined using different testing methods. The required data are
volume change measurement based on drying soil samples
and suction change over which the volume change occurred
(McKeen 1980). The data collected must cover the range
of moisture suction expected in the field environment. The
measurements may take place in a 1-D oedometer or a 3-D
configuration (i.e., unrestrained soil clods) (McKeen 1980).
One test routinely used for this purpose is the COLE (coef-
ficient of linear extensibility) method. All procedures for
determining suction compression index, γh require suction
measurements. For convenience, the empirical relationships
between this parameter and soil activity and cation exchange
activity, or percentage of clay are established by statistical
studies (McKeen 1981):
For clay content between 40% and 70% and high activity:
γh = 0.00179C – 0.041 (59)
For clay content between 25% and 70% and low activity:
γh = 0.00057C – 0.00057 (60)
where, C = percent < 2 μm (ASTM D422). Thus, the value
of the compression index γh can be determined and McKeen
(1980) method can be expressed in Equation (33) (see Table
4) and McKeen (1992) method can be shown in Equation
(35).
McKeen (1981, 1985 & 1990) proposed a classification
scheme by correlating the suction potential index, dh/dw
and the suction compression index, Ch (Figure 6) using a
In situ comulative heave (mm)
Depth interval (m)
(0 - 1) m 28 mm
(1 - 2) m 33 mm
(2 - 3) m 28 mm
(3 - 4) m 30 mm
>4m 41 mm
0
0 10 20 30 40 50 60
Time in weeks
Figure 9.  Development of heave with time for different depth intervals
(modified after Dhowian 1990).
database on Texas soils. In this correlation, a linear relation-
ship is suggested such that 85% of the compression index,
Ch value is larger than would be predicted by the relation-
ship (Olsen et al. 2000). The required information includes
suction potential, dh/dw, for which relatively simple suction
and water content measurements are needed. This method
provides a useful index for estimating the more fundamental
suction compression index, Ch which also requires the Clod
test (Olsen et al. 2000). McKeen (1992) proposed quantita-
tive criteria for using both the suction potential index, dh/
dw and suction compression index, Ch values to differentiate
five categories of swell potential ranging from very high to
non-expansive, as shown in Table 5.
McKeen’s suction potential, dh/dw and suction com-
pression index, Ch are recommended to be obtained on
undisturbed clods of soil. These values are dependent not
only by the type and amount of clay in a soil, but also by
the structure and pore-fluid chemistry of the soil, that result
from its geologic origin and history (Olsen et al. 2000).
4.6  Mitchell & Avalle (1984) method
A simple and relatively quick method is presented to enable
the prediction of expansive soil movements from soil suc-
tion changes. This method is assumed that vertical strain
of the expansive soil is linearly proportional to soil suction.
Therefore, the ground heave can be expressed as Equation
(39) (see Table 4).
The instability index, IPt is obtained from the core
shrinkage test. It involves the measurement of the linear
strain versus moisture content relationship, εv /Δw, and the
30  International Journal of Geotechnical Engineering

0 Table 6. well and suction parameters obtained from

Field laboratory tests (Dhowian 1990).
measurements
Clay Shale
Parameter Silty Shale
1 Lab Data Field Data
α (m3/κΝ) 0.090 0.029 0.085
Depth (m)

2 B 0.047 0.051 0.070

Cv 0.517 0.152 0.327

3

Predicted by:
4 instability index, IPt is calculated as the slope of linear dimen-
Free swell test data (ISO) sion change versus moisture content, Δεv/Δw, times the mois-
Swell pressure test data (CVS)
ture characteristic, c (Equation (40)).
5
0 5 10 15 20 Mitchell & Avalle (1984) shows the core shrinkage tests
results of 80 samples from 18 sites in South Australia and
Cumulative Heave (mm x 10)
three samples from one site in Victoria, Australia (Figure 8).
Figure 10. Measured and predicted heave based on oedometer tech- Considerable scatter exists in the relationship between the
nique (modified after Dhowian 1990). instability index, IPt and plasticity index, IP. A simple relation-
ship does not exist; however, IPt value is higher for soils with
higher IP. If the plasticity of a soil is assessed in a qualitative
30 way (i.e. visual-tactile method), then from experience with a
particular geological location, a reasonable accurate value of
25 instability index, IPt may be adopted without testing (Mitchell
& Avalle 1984).
20 Laboratory Tests Results
Swell (%)

4.7  Hamberg & Nelson (1984) method:

15
Heave is related to soil suction change and soil suction is
10 dependent on the moisture content of the soil; therefore,
heave may be predicted by measuring changes in moisture
5 content. Hamberg & Nelson (1984) extended this philosophy
Field Tests Results and proposed a method to determine the heave of expan-
0 sive soils using suction modulus ratio, Cw (Equation (42),
10 20 30 40 50 60 see Table 4) .The amount of heave can be calculated using
Suction (kPa x 10 ) 3 Equation (41) (see Table 4).
If the final soil suction profile is available instead of
Figure 11.  Suction during the course of swelling was measured in the moisture content change profile ΔH can be calculated using
field as well as in the laboratory (modified after Dhowian 1990). Equation (43) (see Table 4). Actual ground heave may be
adjusted depending on the confinement situation:

moisture characteristic, c (i.e., c = Δw/Δu), of unconfined ΔHact = f ΔH (61)

undisturbed core samples (
Figure 7; Equation (40)). Generally, the more clayey the
soil, the higher is its moisture characteristic (Morris & Gray
1976, Mitchell 1979). Several soil specimens are required for
obtaining information about different initial water contents.
The specimens are air dried for two days, during this period;
the length and mass of the shrinkage core are measured fre-
quently. It is then oven dried to obtain the moisture content.
The moisture characteristic, c, can be determined directly
through soil suction test using thermocouple psychrometer
or the filter paper methods on companion specimens. The
where, f is the correction factor, 0.33-1.0.
4.8  Dhowian (1990) method
Dhowian (1990) estimated heave of expansive shale forma-
tion based on soil suction change. The model described the
field volume change where the anticipated heave is compared
with direct measurement obtained from the field station.
For the purpose of obtaining in-situ heave measurements,
an instrumented field station was established in the central
region of Saudi Arabia (Figure 9). Expansive shale predomi-

0.75
Specific Volume, Jd (m /kNx10 )
-1
0.70 Clay Shale D = 0.091
3
0.65
-1
0.60
0.55
0.50
15 20
Water Content (%)
Figure 12.  Specific volume versus water content plot for clayey shale
(modified after Dhowian 1990).
nates near the ground surface in this area. Undisturbed shale
samples were obtained from the field and used for the free
swell test, the constant volume test, and the swell overburden
test. The swell parameters obtained from results of oedom-
eter tests are used to predict the anticipated field heaves using
Equation (21) (see Table 3). The method uses the same equa-
tion (Equation (20), see Table 3) as Fredlund (1983). In this
equation, Cs is the swelling index; Ps is the swelling pressure;
P0 is the effective overburden pressure.
Amongst the oedometer tests, free swell test gave the
highest swelling pressure value, swell overburden test gave
the least value, and the constant volume test is in between.
The swelling index, Cs value from swell overburden tests was
significantly higher than free swell test and constant volume
test. The estimated heaves calculated from Equation (20)
using tests results from three types of oedometer tests are
compared with the measured field heave values. The agree-
ment between the predictions based on constant volume test
and the field heave is quite well (Al-Shamrani & Dhowian
2002) (Figure 10).
Linear relationship is observed between swell and suc-
tion (Figure 11) for most homogeneous undisturbed shale
samples tested under controlled conditions. Such a trend is
analogous to the e–log P curve in the consolidation; hence,
the suction method can be presented as Equation (45) (see
Table 4). The equation tends to overestimate field heave
considerably. The discrepancy between the predicted and
measured heave may be attributed to the experimentally
determined parameters volume compressibility factor, α and
slope of suction versus water content relationship, B. For this
purpose, the average change in soil suction and water con-
tent coupled with the average ultimate swell measured at the
A state-of-the art review of 1-D heave prediction methods for expansive soils   31
0.020
Volume change index, Iv
0.015
0.010
Iv = 0.019 - 0.034log(Vv)
0.005
(1/Jd) = 0.035 + 0.091w
2
R = 0.86
0.000
10
25 30 35
Net normal stress (surcharge) (kPa)
Figure 13.  Estimation of the volume change index, Iv for Maryland clay
(modified after Fityus & Smith 1998).
end of the observation period are back-calculated to obtain
the parameters α and B. The values are presented in Table 6
with the laboratory data. The discrepancy between the field
and laboratory value of α is attributed to the measurement of
volumetric swell rather than the vertical swell because of the
lateral restraint in the oedometer chamber. Therefore, one
third of the experimentally obtained value is considered as
the field value of α and substituted in the suction method.
The slope of the linear part (Figure 12) is defined as the
compressibility factor α, hence
Δe = α Gs (wf – wi) (62)
where, wi, wf are the initial and final moisture contents,
respectively. The log-suction versus swell relationship has
been found out to be as
Δe = Cψ log(ψi /ψf) (63)
where, Cψ = suction index; and ψi and ψf are the initial and
final suction.
Thus, the relationship for Cψ is equal to
Cψ = α Gs (wf – wi )/log(ψi /ψf) (64)
Using the suction method, the heave prediction can
be estimated as Equation (47). Equation (48) indicates that
heave is linearly proportional to the change in moisture con-
tent. This equation is possible to determine the maximum
swell by determining the limits of moisture content variation.
It must be noted that the moisture index, Cw, should be deter-
mined using one third of the experimentally determined
compressibility factor, α.
32  International Journal of Geotechnical Engineering

4.9  Nelson & Miller (1992) method

0.00
The method presented in Nelson & Miller (1992) used test (a)
-0.02
results from both the consolidation-swell test and the con-
stant volume test to determine the index parameter. The -0.04
method uses the same equation (Equation (20), see Table 3)
-0.06
as Fredlund (1983).

H / H0
“Free-field” heave is the amount of heave that the ground -0.08
surface will experience due to wetting of the sub-soils with
no surface load applied. Because the surface load applied by -0.10

slab-on-grade floors is relatively smaller in comparison to -0.12

Porcelain Clay
the swell pressure generated by an expansive soil, the heave Bentonite Clay
of slabs is essentially the same as the free-field heave (Chao -0.14

2007). The general equation for predicting heave using the 0 10 20 30 40 50 60

oedometer methods can be presented as Equation (22) (see Time (hr)
Table 3). 0.00
(b)
-0.02
4.10  Fityus & Smith (1998) method f f
1
1
The surface movement prediction presented by Fityus and -0.04
Smith (1998) integrates volume changes which occur in sub- H / H0
-0.06
layers of soil over the depth of the active zone to provide an
estimate of the total movement at the ground surface. This -0.08

method requires the information of change in gravimetric -0.10

water content as a function of depth through the soil profile,
as well as an appropriate index to relate moisture change -0.12
Porelain Clay
directly to volume change. As the amount of volume change, -0.14
Bentonite Clay

ΔH is affected by the confining stress of the overburden, a

-0.30 -0.25 -0.20 -0.15 -0.10 -0.05 0.00
volume change index, Iv is used which takes this into account
(Equation (50), see Table 4). V / V0
The load dependent volume index, Iv was determined 0.4
from a series of simple, one dimensional swell tests, in which Porcelain Clay
Ew
Bentonite Clay
a clay sample is allowed to swell under a vertical confining
1
stress that was equal to the calculated vertical stress experi- 0.3

enced under field conditions. The tests were conducted on

Maryland clay (Figure 13). The approach described here is Ew
0.2
w

based on a number of assumptions (Fityus & Smith 1998): (i) wsh

1

wsh
it is assumed that a given change in water content will always
correspond with the same change in strain (ii) it is assumed 0.1
that the volume index, Iv determined from swelling clay tests
can be employed to predict both swelling and shrinking (c)
behavior (iii) it is assumed that the difference between one 0.0
and three dimensional volume changes can be reasonably -0.30 -0.25 -0.20 -0.15 -0.10 -0.05 0.00

accommodated using a factor, α of 0.33. V / V0

Figure 14.  Porcelain clay and bentonite clay shrink tests results (modi-
4.11  Briaud et al. (2003) method fied after Briaud et al. 2003): (a) ΔH/H0 versus t curve; (b) ΔH/H0 versus
Briaud et al. (2003) proposed a water content method to ΔV/V0 curve; (c) w versus ΔV/V0 curve.
estimate the vertical movement of the ground surface for soil
that swells and shrinks due to variations in water content,
Δw. In the field, the soil suction usually ranges from 30 to
3,000 kPa (McDowell 1956). Within that range, the soil water
characteristic curve can often be approximated by a straight
 A state-of-the art review of 1-D heave prediction methods for expansive soils   33

line. The soil water content change, Δw can be calculated by 0.25

(a)
using Equation (65), where β is the slope of the soil–water
characteristic curve. 0.20

Δw = log uf /ui (65)

0.15
Thus, for each layer, i the movement for swelling w
or shrinkage can be calculated by using Equation (52).
0.10
The ground surface movement can be calculated by using
Equation (66).
n
0.05
Sample without Overburden
ΔH = Σ ΔH
i=1
i
(66) Sample with Overburden

0.00
The procedures for the method include: 1) Determine -0.20 -0.15 -0.10 -0.05 0.00
the depth Zmax of water content fluctuation and break the
'V / V0
depth Zmax into an appropriate number of n layers, hi being
the thickness of layer i; 2) Collect samples at the site within 0.25
(b)
depth, Zmax; 3) Perform shrink tests on the samples; for each
sample, determine the shrink–swell modulus, Ew, which is 0.20
the slope of the water content versus the volumetric strain
line; and the shrinkage ratio f, which is the ratio of the verti- 0.15
cal strain to the volumetric strain; 4) Determine the change w
in water content, Δw as a function of depth within Zmax.
0.10
For remolded conditions, soils dryer than shrinkage
limit, wsh are unsaturated whereas soils wetter than swell
limit, wsw are saturated (Holtz and Kovacs 1981). For natural 0.05
Sample with Overburden
conditions, soils dryer than shrinkage limit, wsh are unsatu- Sample without Overburden

rated and soils wetter than swell limit, wsw are either saturated 0.00
or close to it (Fredlund et al. 2002). Thus, the water content of -0.08 -0.06 -0.04 -0.02 0.00 0.02
a soil ranges from shrinkage limit, wsh to swell limit, wsh. Most 'H / H0
of the volume change takes place between shrinkage limit 0.25
wsh and swell limit wsw and the relationship is linear within (c)
that; in other words, the volume change is clearly related to
0.20
the change in water content between this range. Therefore,
a direct relationship can be seen between the shrink–swell
modulus, Ew and the suction compression index, γh. This 0.15
w
water content method is similar to the McKeen (1992) soil
suction method (see Equation (35), Table 4), but it makes use 0.10
of the water content as the governing parameter.
The change in water content, Δw can be obtained from 0.05 Sample with Overburden
various possible sources including databases. The depth of Sample without Overburden
water content variation is estimated from a combination of
0.00
experience, databases, observations, and calculations. For -0.08 -0.06 -0.04 -0.02 0.00 0.02
the actual water content variation, it is as a function of time,
'D / D0
weather, vegetation and pipe leaks. There are several meth-
ods to predict it (i.e., Wray et al. 2005, Chao 2007). Figure 15.  Influence of vertical pressure on the shrink test results
The shrink tests were used for obtaining links between (modified after Briaud et al. 2003): (a) w versus ΔV/V0 curve; (b) w versus
ΔH/H0 curve; (c) w versus ΔD/D0 curve.
the water content w and the normal strain, εi = (ΔH/H0).
The test procedure which involves four steps are detailed in
Briaud et al. (2003): 1) Trim the sample into a cylinder (A 75
mm diameter × 150 mm high sample size is recommended
with the height corresponding to the vertical direction at the
34  International Journal of Geotechnical Engineering
STRAIN, H (%)
Consolidation
C POTENTIAL
swell test data
(+) HEAVE
Cc
%S
0 Cs
A
0 F
B D
(-) COMPRESSION
V'i V'cv V'cs
APPLIED STRESS, V' (log scale)
Figure 16.  Terminology and notation for oedometer tests (modified
after Nelson et al. 2006).
site). 2) The weight W0, the height H0, and the diameter D0
of the sample at time t = 0 is recorded. A minimum of three
heights and three diameter measurements for each height is
recommended at 120° intervals. 3) The sample is allowed to
air dry in a laboratory environment under constant humid-
ity conditions. The variation of weight W, the height H, and
the diameter D of the sample is collected a function of time,
t. The data is collected every hour for the first 8 hours. It
is recommended to record a minimum of three height and
three diameter measurements for each height at 120° inter-
vals at each reading. Typically, two days of data are usually
sufficient. 4) After the last reading is taken, dry the sample in
the oven and measure its oven-dried weight Wd. Data reduc-
tion consists of calculating the average height, the average
diameter, and the average volume for each time step as well
as the water content of the sample.
The plots of the water content w versus the volumetric
strain, (ΔV/V0) or the vertical normal strain, (ΔH/H0) are the
constitutive laws, which are used in this method. The plot of
(ΔH/H0) versus (ΔV/V0) gives the relationship between verti-
cal shrinkage and volumetric shrinkage. Examples of shrink
test results are shown in Figure 14. The shrink–swell modu-
lus Ew (Figure 14a), the shrinkage ratio f (Figure 14b), and the
shrinkage limit wsh (Figure 14c) can be calculated from the
shrink test by using the Equation (53) and (54).
The shrinkage described earlier is performed without
applying overburden pressure. However, the soil shrinks or
GROUND SURFACE
ZA (V'vo )A
ZB
(V'vo )B
DEPTH OF
HEAVE, ZC
(V'vo )C = V'cv
Figure 17.  Vertical overburden stress states at three different depths
(modified after Nelson et al. 2006).
swells under the overburden pressure in the field. The results
from the tests with applied vertical pressure indicated that
the vertical pressure applied to the sample does not influence
the shrink–swell modulus. In other words, this modulus is
independent of the vertical pressure imposed on the sample.
Examples of shrink tests results are shown in Figure 15. The
reason is twofold: 1) most of the volume change occurs when
the soil is saturated or near saturation (i.e., shrinkage limit
to swell limit); 2) for the tests results presented in Figure 15,
the volume change is simply due to water loss or water gain
which is independent of the pressure if water and soil grains
are considered incompressible (Briaud et al. 2003).
Therefore, if the soil is saturated, the shrink–swell mod-
ulus, Ew can be obtained using Equation (67); where γw is unit
weight of water, and γd is dry unit weight of the soil.
Ew = γw /γd (67)
While the soil is not completely saturated, Briaud et
al. (2002) developed an equation linking the change in
volume,(ΔV/V0) to the change in water content Δw that is
valid between the shrink limit wsh and the swell limit wsw.
ΔV/V0 = Sn(Δw/w) + 0.8n(S – 1)[Δw/(wsh – wsw)] (68)
where, S is degree of saturation; n is porosity; wsh is the
shrinkage limit obtained in the shrink test; and wsw is the
swell limit obtained in the free swell test (ASTM D4546). The
shrink–swell modulus, Ew can be calculated from Equation
(69):
Ew = [(Sn/w) + 0.8(S – 1)/(wsh – wsw)]–1 (69)
The first term in Equation (69) corresponds to Equation
(67) (i.e., γw /γd = w/Sn); the second term can quantify the
deviation from the saturated case. Parametric calculations
show that for degrees of saturation between 0.9 and 1, the Ew
 A state-of-the art review of 1-D heave prediction methods for expansive soils   35
VERTICAL STRAIN
CH
%SA %SB
C zero lateral strain with a vertical overburden stress equal to
(V'i)A (V'i)B V'cv V'cs B V'cs A
APPLIED STRESS, V' (LOG SCALE)
Figure 18.  Hypothetical oedometer test results for stress states (modi-
fied after Nelson et al. 2006).
value for the unsaturated case is within 10% of the Ew value
for the saturated case (Briaud et al. 2003).
The advantages of the proposed shrink test–water con-
tent method are as follows (Briaud et al. 2003). First, it is
based on the water content as a governing parameter that is
simpler than suction to obtain experimentally. Second, larger
databases exist for water contents than suction values; as a
result obtaining the magnitude of the actual water content
variation in the field and the depth of these variations from
such databases is much easier than in the case of suction
values. Third, the constitutive law is obtained experimentally
from a simple shrink test on site-specific samples instead of
correlations to index properties. Also, the constitutive law
(Equation (53), see Table 4) is independent of the stress level;
this is not the case for the suction constitutive law. However,
if the soil is highly fractured, shrink tests are difficult to per-
form; in this case, the Clod tests can be substituted and used
as an alternative shrink test.
4.12  Nelson et al. (2006) method
Nelson et al. (2006) describes a methodology to determine
heave using oedometer test data. An important constitutive
parameter used in the method is the heave index, CH, which
is the ratio of the percent swell observed in the oedometer
test to the vertical stress applied to the sample when it was
inundated (i.e., the inundation pressure).
When oedometer test data is plotted in a two dimen-
sional format, the entire stress path is projected onto the
plane defined by the (σ – ua) and εv axes (Figure 16). The
consolidation-swell test data follows the projected path
ABCF, whereas the constant volume test data follows the
projected stress path ABD. Example test data is shown as
below. The slope of the loading portion of the curve shown
in the figure is the compression index Cc, and of the rebound
portion of the curve is the rebound index Cs. The volumetric
strain experienced during inundation is the percent swell.
Figure 17 shows the vertical overburden stress states
at three different depths in a soil profile with similar soil
throughout. At all points all samples are in a condition of
σʹvo. If a consolidation-swell test is conducted on a sample
identical to that at depth, ZA, at an inundation stress, (σʹi)A
= (σʹvo)A, the sample will swell by an amount %SA as shown
in Figure 18. Similarly for a sample at depth ZB, the sample
would be subjected to an inundation stress, (σʹi)B = (σʹvo)B, and
the sample would swell by an amount %SB.
The general equation for predicting heave or settlement
in a soil stratum of thickness, H can be shown as Equation
(70):
Δe
ΔH = H= (70)
1 + e0
For uniform vertical strain throughout the stratum the
strain is equal to:
Δe
εv = (71)
1 + e0
Substituting Equation (71) into Equation (70):
ΔH = H εv = H %S (72)
At any depth in the soil, the percent swell will fall
along the line ABC. For all practical purposes the line can
be defined by a straight line connecting point A (the point
defined by the percent swell in a consolidation-swell test)
and point C (the point corresponding to the constant volume
swell pressure, σʹcv. The slope of that line is denoted by the
heave index, CH which is shown in Equation (24).
If values of heave index, CH and constant volume swell
pressure, σʹcv are known; the vertical strain, or percent swell,
that will occur during inundation at any depth z in a soil
profile can be determined from Equation (23). Equation (23)
can be re-written as Equation (73), when the soil at depth z
is inundated, the stress on the soil is the overburden stress,
(σʹvo)z. This value is therefore used for the inundation stress,
σʹi .in Equation (73).
σʹcv
εv = %S = CH log = (73)
(σʹvo)z
Therefore, heave prediction can be calculated using
Equation (23) (see Table 4).
36  International Journal of Geotechnical Engineering

K=0.0033e0.64('w)
Fredlund 1969
K=0.0039e0.64('w)
0.04 Hamberg & Nelson 1984
1000
K=0.0035e0.64('w)
Osman & Sharief 1987
Suction modulus, Cw

100 K=0.0035e0.64('w)

Correction parameter, K
0.03 Osman & Sharief 1987
10 K=0.0039e0.64('w)
Snethen & Huang 1992

0.02 1

0.1
Cw = 0.024 for IP > 30
0.01 0.01 KI = 0.0039e
0.64('w) 2
, R = 0.997
0.64('w)
KII = [-0.0018 ln(IP)+0.01]e
0.001
0.00
0.0001
0 20 30 40 60 80 0 5 10 15
Plasticity index, Ip (%)
Water content change, 'w

Figure 19.  Relationship between the suction modulus ratio, Cw and Figure 21.  Relationship between correction parameter K and water
plasticity index, IP using data published in the literature (publications content change Δw.
from which the data points are collected are summarized in the refer-
ences).

0.20 This method for determination of the heave index is

considered to be practical and rational but the actual value
Cs = 0.0193e
0.0343(IP) of λ to be used for different soils should be investigated on a
0.15
case by case basis. Generally, a value of λ in the range of 0.5
Swelling index, Cs

2
R = 0.965
to 0.7 may be reasonable.
0.10
4.13  Vanapalli et al. (2010a) method
Ching & Fredlund 1984
Nelson & Miller 1992
Vanapalli et al. (2010a) proposed a simple technique by
0.05
Vu & Fredlund 2004 deriving a new equation based on the Fredlund (1983) and
Clifton et al. 1984
the Hamberg & Nelson (1984) methods for estimating 1-D
0.00 heave. This method alleviates from the limitations of several
20 30 40 50 60 70 80 methods. Some of the limitations of the available methods
Plasticity index, Ip (%)
include: 1) they are not universally valid as they are proposed
using only limited soils data collected locally; 2) they do not
Figure 20.  Relationship between the corrected swelling index, Cw and
use the stress state variables approach that provides a rational
the plasticity index, IP.
basis for interpretation; 3) the various parameters required in
these approaches can only be obtained from time consuming
laboratory or in-situ tests that are expensive and difficult to
Nelson et al. (1998, 2006) indicated that an accurate be performed.
method to determine the heave index, CH would require Vanapalli et al. (2010a) technique requires three param-
several consolidation-swell tests at different inundation pres- eters; namely, corrected swelling index, Cs, suction modulus
sures and a constant volume test. This is neither practical ratio, Cw, and correction parameter, K which is a function of
nor economical. Therefore, the relationship between σʹcv and water content change, Δw and plasticity index, IP. Though
σʹf was proposed so that the value of the heave index, CH can the parameters Cs and Cw can be determined from laboratory
be determined from a single consolidation-swell test (Nelson tests, they proposed empirical equations to obtain the values
et al. 2006). A relationship between σʹcv and σʹcs exists that is of the parameters to avoid time-consuming or elaborate test-
of the form: ing procedures.
σʹcv = σʹi + λ(σʹcv – σʹi) (74)
 A state-of-the art review of 1-D heave prediction methods for expansive soils   37
The technique was tested on several case studies and
the results showed that the proposed technique can provide
reasonable 1-D heave estimations on natural expansive soils
from various regions of the world (Vanapalli et al. 2010b).
Fredlund (1983) method (Equation (20), see Table 3)
can be re-written as the Equation (75). The positive and the
negative sign in the equation indicate compression and heave
due to overburden and swelling pressure, respectively. The
heave calculated by using the second term is proportional to
the heave estimated using Hamberg & Nelson (1984) method
(Equation (41), see Table 4), and can be shown as Equation
(76).
H H
ΔH = Cs log Pf – Cs log Pʹs (75)
1 + e0 1 + e0
H H
Cs log ∞ Pʹs Cw Δw (76)
1 + e0 1 + e0
Equation (76) can be re-written as Equation (77) by
introducing a correction parameter K. The Equation (19) can
be obtained by substituting Equation (77) into Equation (20).
Cw Δw
10 Cs
Pʹs = (77)
K
Chen (1975) stated that the amount of swell in expansive
soils is governed by the change in water content, Δw, which
can be obtained from field investigation studies. This infor-
mation can be calculated using Equation (78) (Fredlund and
Rahardjo, 1993) based on the assumption that the soils attain
saturated (i.e., Sf = 100%) condition as well. Such an assump-
tion provides conservative estimations (i.e., maximum 1-D
heave).
Δw = Sf Δe/Gs + e0 ΔS/Gs (78)
The Cw value can be measured from Clod tests (see sec-
tion 3.3.1). For silty clay, clayey and expansive soils, the void
ratio linearly increases with increasing water content beyond
shrinkage limit (Hamberg, 1985; Tripathy et al. 2002). Using
this concept, an empirical relationship between suction mod-
ulus ratio, Cw and plasticity index, IP was developed using the
data published in the literature (see Figure 19).
Though the suction modulus ratio, Cw values show large
scatter for the IP values less than 30%, the relative constant
value (i.e., 0.024) was observed for the IP values greater than
30%. This indicates that Equation (79) can be used in the
estimation of 1-D heave of expansive soils since IP values of
typical expansive soils are generally greater than 30%.
Cw = 0.024  IP ≥ 30% (79)
The corrected swelling index, Cs can be determined
using the procedures detailed by Fredlund (1983) from 1-D
Constant Volume Swell (CVS) oedometer test. Determination
of the Cs is time-consuming and requires elaborate testing
equipment as discussed earlier (see section 3.2.1). It also
needs different corrections with respect to swelling pressure,
compressibility of the apparatus, and sample disturbance.
Hence, a simple empirical relationship between the corrected
swelling index, Cs and plasticity index, IP was developed using
the data available in the literature as shown in Equation (80)
and Figure 20.
Cw = 0.019e0.0343(IP) (80)
Figure 21 shows the empirical relationship between
correction parameter, K and Δw for the data from five case
studies in literature. The analysis in Figure 21 showed that K
is not a constant value but instead varies with IP (Equation
(81)). The best-fitting curve for the all data was estimated
as Equation (82) with relatively high R-squared value (i.e.
0.997). The correction parameters, K in Equation (81) and
(82) are hereafter referred to as KI and KII, respectively and
both were used in this technique.
KI = [–0.0018ln(IP) + 0.01]e 0.64(Aw) (81)
KII = 0.0039e 0.64(Aw) (82)
5.  SUMMARY
Empirical relationships that were developed using limited
studies for explaining the expansive soils behavior during the
period from 1950’s thru 1900’s were found to be not univer-
sally valid but only useful for explaining the local expansive
soil behavior. Aitchison (1973) was the earliest investigators
who introduced the soil suction models introducing the
concept of instability index (i.e., IPt) for calculating moisture-
related ground movements. This instability index, IPt has
been widely used by geotechnical researchers (Mitchell 1989,
Jaksa et al. 1997, Fityus & Smith 1998).
Several other techniques have been proposed using
oedometer test methods or soil suction methods for estimat-
ing the surface heave in expansive soil with lightly loaded
structures (see Table 4). Some techniques are based on
stress state variables approach (i.e., Snethen (1980) method,
Fredlund (1983) method) and others do not use the stress
state variables approach (i.e., Hamberg & Nelson (1984)
method, Nelson & Miller (1992) method). However, in all
these methods various parameters are required to estimate
the 1-D heave by time consuming laboratory and/or in-situ
tests. Most of the tests are expensive and difficult to be per-
formed by conventional geotechnical engineers. The method
proposed by Vanapalli et al. (2010a) is a much simpler and
inexpensive for estimating the 1-D heave that can be used
38  International Journal of Geotechnical Engineering
universally on expansive soils from all regions of the world
with a reasonable degree of accuracy. This method requires
only the information of plasticity index, IP and variation in
water content with respect to depth in the active zone.
REFERENCES
Aitchison, G. D. (1970). “A statement of the problems the
engineer faces with expansive soils.” Proceedings of the
2nd International Research and Engineering Conference
on Expansive Clay Soils, Texas A & M University,
College Station, pp. 33-51.
Aitchison, G.D. (1973). “The quantitative description of the
stress-deformation behavior of expansive soils – preface
to set of Papers.” Proceedings of the 3rd International
Conference on Expansive Soils, 2, 79 – 82.
Aitchison, G.D. & Woodburn, J.A. (1969). “Soil suction in
foundation design.” Proceedings of the 7th International
Conference on Soil Mechanics and Foundation engineer-
ing. 2. 7-8.
Alonso, E.E., Gens, A. & Josa, A. (1990). “A constitutive
model for partly saturated soils.” Géotechnique, 40(3),
405-430.
Al-Shamraani, M. A. & Dhowian, A.W. (2002). “Experimental
study of lateral restraint effects on the potential heave of
expansive soils.” Engineering Geology, 69, 63-81.
ASTM (2000). “Standard test methods for one-dimensional
swell or settlement potential of cohesive soils.” Annual
Book of ASTM Standards, D4546-2000, 04(8): 853-859.
AS 2870 (1996). “The Australian standard for the design and
construction of residential slabs and footings.” Standard
Australia.
Basma, A.A., Al-Homoud, A.S. & Malkawi, A.I. (2000).
“Swelling-shrinkage behavior of natural expansive
clays.” Appl. Clay Sci., 11, 211-227.
Brasher, B.R., Franzmeier, D.P., Valassis, V. & Davidson, S.
E. (1966). “Use of saran resin to coat natural soil clods
for bulk density and water retention measurements.” Soil
Science, 101-108
Briaud, J.-L., Hoffmann, S., Posey T. & Woodfin, G. (2002).
‘‘PREMISS: Phase relationship equation for mois-
ture induced swell in soils.’’ Proceedings of the 3rd
International Conference on Unsaturated Soils, Balkema,
Rotterdam, The Netherlands.
Briaud, J.-L., Zhang, X. and Moon, S. (2003). “Shrinkage test
– water content method for shrinkage and swell predic-
tions.” Journal of Geotechnical and Geoenvironmental
Engineering, 129(7), 590-600.
Bonner, J.P. (1998). “Comparison of predicted heave using
oedometer test data to actual heave.” Master’s Thesis,
Colorado State University, Fort Collins, Colorado.
Burland, J.B. (1962). “The estimation of field effective stresses
and the prediction of total heave using a revised method
of analyzing the double oedometer test.” The Civil
Engineer in South Africa. Transactions of South African
Institute of Civil Engineering.
Casagrande, A. (1936). “The determination of the preconsol-
idation load and its practical significance.” Proceedings
of the 1st International Conference Soil Mechanics and
Foundation Engineering. 3, 60-64.
Chao, K.C. (2007). “Design principles for foundations
on expansive soils.” Dissertation submitted in par-
tial requirement for the Ph.D. Degree, Colorado State
University, Fort Collins, Colorado.
Chen, F.H. (1975). Foundations on Expansive Soils. Elsevier
Scientific Pub. Co., Amsterdam, New York, NY.
Ching, R.K.H. & Fredlund, D.G. 1984. “A small Saskatchewan
town copes with swelling clay problems.” Proceedings of
the 5th International Conference on Expansive Soils, 306-
310. Adelaide, Australia.
Clifton, A.W., Yoshida, R.T. and Fredlund, D.G. (1984).
“Performance of dark hall, Regina, Canada, constructed
on a highly swelling clay.” Proceedings of the 5th
International Conference on Expansive Soils, Adelaide,
Australia, pp. 197-201.
Clifton, A.W., Wilson, G.W. & Barbour, S.L. (1999). “The
emergence of unsaturated soil mechanics: Fredlund
Volume.” NRC Research Press, National Research
Council Canada, Ottawa, Ont.
Croney, D. & Coleman, J.D. 1954. “Soil structure in relation
to soil suction (pF).” Journal of Soil Science, 5(1): 163-
177.
De Bruijin, C.M.A. (1961). “Swelling characteristics of a trans-
ported soil profile at Leeuhof Vereeniging (Transvaal).”
Proceedings of the 5th International Conference Soil
Mechanics Found Engineering 1, 43-49.
Dhowian, A.W. (1990). “Field performance of expansive
shale formation,” Journal of King Abdulaziz University
(Engineering Sciences), 2, 165–82.
Erol, A. O., Dhowian A. & Youssef, A. (1987). “Assessment of
oedometer methods for heave prediction.” Proceedings
of the 6th International Conference on Expansive Soils,
Technical Session III, 99-105
Feng, M., Gan, J.K.M. & Fredlund, D.G. (1998). “A labora-
tory study of swelling pressure using various test meth-
ods.” Proceedings of the Second International Conference
on Unsaturated Soils, Beijing, China, International
Academic Publishers, 6, 350-355.
 A state-of-the art review of 1-D heave prediction methods for expansive soils   39
Fityus, S. & Smith, D.W. (1998). “A simple model for the
prediction of free surface movements in swelling clay
profiles.” Proceedings of the 2nd International Conference
on Unsaturated Soils, Beijing, China, 473-478.
Fleureau, J.M., Verbrugge, J.C., Huergo, P.J. & Gomes
Correia, A. (2002). “Aspects of the behavior of compacted
clayey soils on drying and wetting paths.” Canadian
Geotechnical Journal 39: 1341-1357.
Fredlund, D.G. (1969). “Consolidometer test procedural fac-
tors affecting swell properties.” Proceedings of the 2nd
International Conference on Expansive Clay Soils, Texas,
435 – 456.
Fredlund, D.G., Hasan, J. U., & Filson, H. (1980). “The predic-
tion of total heave.” Proceedings of the 4th International
Conference on Expansive Soils. Denver, Colorado, 1 – 11.
Fredlund, D.G. (1983). “Prediction of ground movements
in swelling clays.” 31st Annual Soil Mechanics and
Found Engineering Conference. University of Minnesota,
Minneapolis.
Fredlund, D.G. (1987). “The prediction and performance
of structures on expansive soils.” Proceedings of the
International Symposium on Prediction and Performance
in Geotechnical Engineering, 51-60.
Fredlund, D.G. & Rahardjo, H. (1993). Soil Mechanics for
Unsaturated Soils. John Wiley & Son, Inc., New York,
NY.
Fredlund, M.D., Wilson, G.W. & Fredlund, D.G. (2002).
“Representation and estimation of the shrinkage curve.”
Proceedings of the 3rd International Conference on
Unsaturated Soils. 1, 145-150.
Gibbs, H. J. (1973). “Use of a consolidometer for measur-
ing expansion potential of soils.” Proceedings Workshop
on Expansive Clays and Shales in Highway Design and
Construction. University Wyoming, Laramie, May, 1,
206-213.
Grossman, R.B., Brashner, B.R., Franzmeier, D.P. & Walker,
J.L. (1968). “Linear extensibility as calculated from
natural-clod bulk density measurement.” Soil science of
Science American. 32(4), 570-573.
Haines, W. B. (1923). “The volume changes associated with
variations of water content in soil.” Jour. Agr. Sci. 13:
296-310.
Hamberg, D.J. & Nelson, J.D. (1984). “Prediction of floor slab
heave.” Proceedings of the 5th International Conference
on Expansive Soils, Adelaide, South Australia, 137-140.
Hamberg, D.J. (1985). “A simplified method for predicting
heave in expansive soils.” M.S. thesis, Colorado State
University, Fort Collins, CO.
Holtz, R.D. & Kovacs, W.D. (1981). An Introduction to
Geotechnical Engineering, Prentice–Hall, Englewood
Cliffs, N. J., 180.
Jaksa, M.B., Cavagnaro, R.L. & Cameron, D.A. (1997).
“Uncertainties associated with the visual tactile method
for quantifying the reactivity of expansive soils.”
Australian Geomechanics. 31, 84-91.
Jennings, J.E.B. & Knight, K. (1957). “The prediction of total
heave from the double oedometer test.” Proceedings of
Symposium on Expansive Clays, S. African Inst. Civil
Eng., Johannesburg. 7(9), 13 – 19.
Jennings, J.E.B., Firtu, R.A., Ralph, T.K. & Nagar, N. (1973).
“An improved method for predicting heave using the
oedometer test.” Proceedings of the 3rd International
Conference Expansive Soils, Haifa, Israel, 2, 149-154.
Johnson, L.D. & Snethen, D.R. (1978). “Prediction of poten-
tial heave of swelling soils.” Geotechnical Testing Journal,
1(3), 117-124.
Johnson, L.D. (1979). “Overview for design of foundations
on expansive soils.” U.S. Army Eng. Waterways Exp.
Sta., Vicksburg, MS, Misc. Paper GL-79-21.
Jones, D.E. & Holtz, W.G. (1973). “Expansive soils – hidden
disaster.” Civil Engineering: American Society of Civil
Engineers, 43(8): 49-51.
Komornik, G. Livneh, M and Wiseman, G. (1969). Pavements
on expansive clays. Jerusalem Academic press, Jerusalem,
Israel.
Likos, W.J., Lu, N. & Sharkey, K. J. (2005). “Laboratory char-
acterization of steeply dipping expansive bedrock in the
Rocky Mountain Front Range.” Journal of Geotechnical
and Geoenvironmental Engineering. 131(9): 1162-1171.
Lytton, R.L. (1977). “Foundations in expansive soils.”
Numerical Methods in Geotechnical Engineering. C.S.
Desai and J.T. Christian, (Eds.), McGraw Hill, N. Y.
427 - 457.
Lytton, R.L. & Woodburn, J.A. (1973). “Design and perfor-
mance of mat foundation on expansive clay.” Proceedings
of the 3rd International Conference Expansive Soils.
Haifa, Israel 1, 301-308.
Masia, M.J., Totoev, Y.Z. & Kleeman, P.W. (2004). “Modeling
expansive soil movements beneath structures,” Journal
of Geotechnical and Geoenvironmental Engineering,
130(6), 572–579.
Matyas, E.L. & Radhakrishna, H.S. (1968). “Volume change
characteristics of partially saturated soils.” Géotechnique,
18(4), 432-448.
McDowell, C. (1956). Interrelationships of loads, volume
change, and layer thickness of soils to the behavior
of engineering structures. Highway Research Board,
Proceedings of the 35th Annual Meetings, Publ. No. 426,
Transportation Research Board, Washington, D.C., 75th
pavement design. U. S. Department of Transportation,
Federal Aviation Administration, FFA Report No.
FAA-RD-78-59.
40  International Journal of Geotechnical Engineering
McKeen R.G. (1980). “Field study of airport pavements on
expansive clay.” Proceedings of the 4th International
Conference on Expansive Soils, Denver, Colorado, 242–
61.
McKeen, R.G. (1981). Design of Airport Pavements for
Expansive Soils. Federal Aviation Agency, U.S.
Department of Transportation, Washington,
McKeen, R.G. & Hamberg, D.J. (1981). “Characterization of
expansive soils.” Trans. Res. Rec. 790, Trans. Res. Board,
73 – 78.
McKeen, R.G. (1985). Validation of Procedures for Pavement
Design on Expansive Soils. U.S. Dept. of Transportation,
Federal Aviation Administration, Final Report.
McKeen, R.G. (1990). “Climate-controlled soil design param-
eters for mat foundations.” Journal Geotechnical Eng.,
116(7), 1073–1094.
McKeen, R.G. (1992). “A model for predicting expansive
soil behavior.” Proceedings of the 7th International
Conference on Expansive Soils, Dallas, Texas. 1, 1 – 6.
Miller, D.J., Durkee, E.B., Chao, K.C. & Nelson, J.D. (1995).
“Simplified heave prediction for expansive soils.”
Proceedings of the 1st International Conference on
Unsaturated Soils, 891–897. Balkema, Rotterdam.
Mitchell, P.W. (1979). “The structural analysis of footings on
expansive soil.” Research Rpt. No. L. K. W. G. Smith &
Assoc., Newton.
Mitchell, P.W. & Avalle, D.L. (1984). “A technique to pre-
dict expansive soils movement.” Proceedings of the 5th
International Conference on Expansive Soils, Adelaide,
Australia, 124-130.
Mitchell, P. W. (1989). Site investigation processes. Course
on Footings for Small Scale and Domestic Structures, Linn
Education and Training Services, 89-241.
Morris, P. O. & Gray, W. T. (1976). “Moisture conditions
under roads in the Australian environment.” Rep. 68,
Aust. Road Research Board.
Navy, Dept. of, Naval Facilities Engineering Command.
(1971). Design Manual-Soil Mechanics, Foundations and
Earth Structures. U. S. Naval Publications and Forms
Center, NAVFAC DM-7.
Nayak, N.V. & Christensen, R.W. (1971). “Swell charac-
teristics of compacted expansive soils.” Clay and Clay
Minerals. 19(4), 251–261.
Nelson, J.D., Durkee, D.B. & Bonner, J. P. (1998). “Prediction
of free-field heave using oedometer test data.” Proceedings
of the 46th Annual Geotechnical Engineering Conference,
University of Minnesota, St. Paul, Minnesota.
Nelson, J. D. & Miller, D. J. (1992). Expansive Soils: Problems
and Practice in Foundation and Pavement Engineering.
John Wiley & Sons, Inc., New York, NY.
Nelson, J.D. & Chao, K.C. (2003). “Design of founda-
tions for light structures on expansive soils.” California
Geotechnical Engineers Association Annual Conference,
Carmel, California.
Nelson, J.D., Reichler, D.K. & Cumbers, J. M. (2006).
“Parameters for heave prediction by oedometer tests.”
Proceedings of the 4th International Conference on
Unsaturated Soils. Carefree, Arizona, 951 – 961.
Noble, C.A. (1966). “Swelling measurements and prediction
of heave for a lacustrine clay.” Canadian Geotechnical
Journal 3(1): 32-41.
Olsen, H.W., Krosley, L., Nelson, K., Chabrillat, S., Goetz,
A., & Noe, D. C. (2000). “Mineralogy-swelling poten-
tial relationships for expansive shales.” Advances In
Unsaturated Geotechnics, ASCE, Denver, 361–378.
Osman, M.A. & Sharief, A.M.E. (1987). “Field and laboratory
observations of expansive soil heave.” Proceedings of the
6th International Conference on Expansive Soils, 105 –
110. New Delhi, India.
Palit, R.M. (1953). “Determination of swelling pressure of
black cotton soil.” Proceedings of the 3rd International
Conference on Soil Mechanics and Foundation
Engineering. Switzerland.
Picornell, M. & Lytton, R. L. (1984). “Modeling the heave
of a heavily loaded foundation.” Proceedings of the 5th
International Conference on Expansive Soils, Adelaide,
Australia, 104-108.
Porter, A.A. & Nelson, J.D. (1980). “Strain controlled testing
of soils.” Proceedings of the 4th International Conference
Expansive Soils, ASCE and Int. Soc. Soil Mech. Found.
Eng., Denver, pp. 34-44.
Puppala, A.J., Punthutaecha, K., and Vanapalli, S.K. 2006.
“Soil-water characteristic curves (SWCC) of stabilized
expansive soils.” Journal of Geotechnical Engineering
and Geo-environmental Engineering, American Society
of Civil Engineers, 132 (6): 736-751.
Ranganathan, B.V. & Satyanarayana, B. (1965). “A rational
method of predicting swelling potential for compacted
expansive clays.” Proceedings of the 6th International
Conference on Soil Mechanics and Foundation
Engineering, International Society for Soil Mechanics
and Geotechnical Engineering, London, 1, 92–96.
Salas, J.A.J. & Serratosa, J.M. (1957). “Foundations on Swelling
Clays.” Proceedings of the 4th International Conference
on Soil Mechanics and Foundation Engineering. London,
England, 1, 424-428.
Sampson, E., Schuster, R.L. & Budge, W. D. (1965). “A
method of determining swell potential of an expansive
clay.” Concluding Proceedings Engineering Effects of
Moisture Changes in Soils. Int. Res. Eng. Conf. Expansive
 A state-of-the art review of 1-D heave prediction methods for expansive soils   41
Clay Soils. Supplementing the Symposia in Print. Texas
A & M Univ. Press, College Station, TX, 255-275.
Schneider G.L. & Poor, A.R. (1974). “The prediction of
soil heave and swell pressures developed by an expan-
sive clay,” Research Report, No: TR-9-74, Construction
Research Center, Univ. Of Texas.
Seed, H.B., Woodward, R.J., Jr & Lundgren, R. (1962).
“Prediction of swelling potential for compacted clays,”
Journal of Soil Mechanics and Foundation Engineering
Division ASCE, 88, 53–87.
Shanker, N. B., Ratnam M. V. & Rao A. S. (1987). “Multi-
dimensional swell behavior of expansive clays.”
Proceedings of the 6th International Conference of
Expansive Soils, New Delhi, India.
Smith, A.W. (1973). “Method for determining the poten-
tial vertical rise, PVR.” Proceedings of Workshop on
Expansive Clays and Shales in Highway Design and
Construction. Univ. of Wyoming, Denver, CO, 245-249.
Snethen, D.R. (1980). Characterization of expansive
soils using soil suction data, Proceedings of the 4th
International Conference on Expansive Soils, Denver,
Colorado, 54–75.
Snethen, D.R. & Huang, G. (1992). “Evaluation of soil
suction-heave prediction methods.” Proceedings of the
7th International Conference on Expansive Soils, Texas,
12-17.
Sridharan, A., Sreepada Rao, A. & Sivapullaiah, P.V. (1986).
“Swelling pressure of clays.” Geotechnical Testing Journal,
GTJODJ, 9(1), 24-33.
Subba Rao, K.S. & Tripathy, S. (2003). “Effect of aging on
swelling and swell-shrink behavior of a compacted
expansive soil.” ASTM Geotechnical Testing Journal,
26(1), 36-46.
Sullivan, R.A. & McClelland, B. (1969). “Predicting heave of
buildings on unsaturated clay.” Proceedings of the 2nd
International Research Engineering Conference Expansive
Soils. Texas A & M Univ. Press, College Station, TX, 404-
420.
Taylor, D. W. (1948). Fundamentals of Soil Mechanics. John
Wiley & Sons, Inc., New York.
Teng, T.C.P. & Clisby, M.B. (1975). “Experimental work
for active clays in Mississippi. Transport. Engineering
Journal ASCE 101 (TEI), 77-95.
Teng, T.C.P., Mattox, R. M. & Clisby, M. B. (1972). A study
of active clays as related to highway design. Research
and Development Division, Mississippi State Highway
Department, Engineering and Industrial Research
Station, Mississippi State University, MSHD-RD-72-045,
134.
Teng, T.C.P., Mattox, R.M. & Clisby, M.B. (1973).
“Mississippi’s experimental work on active clays.” Proc.
View publication stats
Workshop on Expansive Clays and Shales in Highway
design and Construction, University of Wyoming,
Laramie, May, 2, 1-17.
Tripathy, S., Subba Rao, K.S. & Fredlund, D.G. 2002. “Water
content – void ratio swell – shrink paths of compacted
expansive soils.” Canadian Geotechnical Journal 39: 938-
959.
Van Der Merwe, D.H. (1964). “The prediction of heave from
the plasticity index and percentage clay fraction of soils,”
Civil Engineers in South Africa, 6, 337–42.
Vanapalli, S.K., Lu, L. & Oh, W.T. (2010a). ”Estimation of
swelling pressure and 1-D heave in expansive soils.”
Proceedings of the 5th International Conference on
Unsaturated Soils, Barcelona, Spain. September 6-8,
2010. eds: Alonso, E & Gens, A., CRC Press, Vol. 2: pp.
1201-1207.
Vanapalli, S.K., Lu, L. and Oh, W.T. (2010b). “Comparison
between the measured and the estimated 1-D heave of
expansive soils for seven case studies results using a
simple technique.” Proceedings of the 63rd Canadian
Geotechnical Conference. 13-15 September 2010, Calgary,
Canada. pp. 176-184.
Vijayavergiya, V.N. & Ghazzaly, O. I. (1973). “Prediction of
swelling potential for natural clays.” Proceedings of the
3rd International Conference on Expansive Soils. Haifa,
Israel, 1, 227-236.
Vu, H.Q. & Fredlund, D.G. (2004). “The prediction of one-,
two-, and three-dimensional heave in expansive soils.”
Canadian Geotechnical Journal 41(4), 713–737.
Weston, D. J. (1980). “Expansive roadbed, treatment for
Southern Africa.” Proceedings of the 4th International
Conference on Expansive Soils, 1: 339-360.
Wong, H.Y. & Yong, R.M. (1973). “A study of swelling and
swelling force during unsaturated flow in expansive
soils.” Proceedings of the 3rd International Conference of
Expansive Soils, Haifa, Israel, 1, 143-151.
Wray, W.K., El-Garhy, B.M. & Youssef, A.A. (2005). “Three-
dimensional model for moisture and volume changes
prediction in expansive soils,” Journal of Geotechnical
and Geoenvironmental Engineering, 131(3), 311–324.
Yoshida, R.T., Fredlund, D.G. and Hamilton, J. J. (1983).
“The prediction of total heave of a slab-on-grade floor
on Regina clay.” Canadian Geotechnical Journal, 20(1),
69-81.