1944
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
Table 1. Classification of swell potential (after USACE 1983).
Classification of swell potential
Plasticity index, PI (%)
Low
< 25
Marginal
25 – 35
High
> 35
The depth where the seasonal soil moisture variations occur
below the ground surface is called the active zone as shown in
Figure 1. The depth of active zone is influenced by soil
permeability, precipitation and evaporation amounts, seasonal
temperature fluctuations, and presence of tree roots. Active zone
depths in several U.S. cities were reported by O’Neill and
Poormoayed (1980) as: 1.5 to 3.0 m in Houston, Texas; 2.1 to
4.2 m in Dallas, Texas; 3.0 to 4.6 m in Denver, Colorado; and
3.0 to 9.0 m in San Antonio, Texas.
The degree of swell potential of expansive soils can be
classified by the soil’s liquid limit,
LL
, or plasticity index,
PI
.
As the liquid limit or plasticity index increases, the swell
potential of the soil increases. The classification used by the
U.S. Army Corps of Engineers (USACE) (1983) based on the
plasticity index is given in Table 1.
1.2
Conventional sheet pile wall design
The design of sheet pile walls is based on active and passive
earth pressures which are concerned with the failure condition
using the Mohr-Coulomb failure criterion. For a typical wall
section, the lateral earth pressures and the resulting forces acting
on the wall are shown in Figure 2, where
P
A
and
P
P
are
resultant effective active and passive earth forces, respectively;
d
A
and
d
P
are moment arms with respect to the anchor elevation;
and
FS
is factor of safety. The factor of safety is applied to the
passive loads during wall design (NAVFAC 1986; USACE
1994). The safety factors are used to take into account the
uncertainties in soil conditions, method of stability analysis,
loading conditions, as well as to restraint soil movements at an
acceptable level (Potts and Fourie 1984).
Figure 2. Typical sheet pile wall section and forces acting on the wall.
Wall penetration depth required below the bottom of the
excavation is determined by considering the moment
equilibrium about the anchor elevation. Because the water level
is assumed to be at the same elevation behind and in front of the
wall during this study, hydrostatic forces cancel each other.
Once the wall penetration depth is determined, the anchor force,
A
P
, is calculated from horizontal force equilibrium. Based on
the active and passive pressure distributions and the calculated
anchor force, maximum wall bending moment is determined.
The design moment is calculated by applying the moment
reduction factor (Rowe 1952) to the calculated maximum
bending moment. The steel sheet pile section is then selected
based on the design moment, and the wall design is completed
by selection and design of an anchorage system. This
conventional design approach does not take into account any
swell pressures that may affect walls when they are installed at
locations where expansive soils are present.
2 METHOD OF APPROACH
The effect of swelling pressures on anchored sheet pile wall
behavior has been investigated through a range of expansive
soil activity. The swell pressures were calculated for a range of
plasticity index values covering soils from low to high swell
potential based on a study performed by Erzin and Erol (2007).
Using these swell pressure potentials and moisture change
profile in the ground within the active zone, the swell pressure
distribution was developed. The swell pressure distribution
developed was then applied on the anchored sheet pile wall as
potential swell pressure, additional to the lateral earth pressures
presented in Figure 2.
A parametric study for a range of plasticity index values, i.e.
expansive soil activity and swell potentials, have been
performed using the free earth support design method to
investigate the effect of swell pressures on anchored sheet pile
walls. Design of the wall was first performed for non-expansive
soils, i.e. using only the traditional lateral earth pressure
distributions, as a baseline case. Then the swell pressures, based
on the varying plasticity index values, have been applied and
the wall was re-analyzed.
3 SWELL PRESSURES
There are many factors that govern expansive behavior of soils.
The primary factors are availability of moisture, amount and
type of clay particles, and initial condition of soil in terms of
dry density and moisture content (Day 1994). Several earlier
studies (e.g., Snethen 1980, Erzin and Erol 2007) indicate that
soil suction is the most relevant soil parameter for the
characterization of swell behavior of expansive soils.
Multiple regression analyses carried out by Erzin and Erol
(2007) revealed that the soil suction relates to the plasticity
index and water content as
log 2.02 0.00603 0.0769
s
PI
w
(1)
where
s
=soil suction (in bar),
PI
=plasticity index (in percent),
and
w
=water content (in percent). The study performed by Erzin
and Erol (2007) also showed that the swell pressure, for
pressures between 0 and 100 kPa, can be given as
3.72 0.0111 2.077
0.244 log
s
dr
PI
s
y
(2)
where
s
=swell pressure (kg/cm
2
),
PI
=plasticity index (in
percent),
dry
=dry density (g/cm
3
), and
s
=soil suction (in bar).
For this study, the plasticity index was used as the only
variable to determine swell pressures. The plasticity index
values considered ranged from 10% to 50%, which covers low
to highly expansive soils as presented in Table 1. A constant
value of 15% for the moisture content and a constant value of
1.65 g/cm
3
for the dry density were used. These selected values
represent average values of the ranges considered by Erzin and
Erol (2007) in their study. The swell pressures calculated using
Eq. 2 for the range of plasticity index values studied, with the
moisture content of 15% and dry density of 1.65 g/cm
3
, are
shown in Figure 3.
3.1
Distribution of lateral swell pressures behind the wall
As shown in Figure 1, seasonal variation of soil moisture
content is the highest at the ground surface and it diminishes as
the depth from the ground surface increases. The change in the
moisture content with increasing depth is not linear. However,
the variation is assumed to be linear in this study. This
