Soil-Water Characteristic Curves for Unsaturated Soils - Unsaturated Soil Mechanics in Engineering Practice - page 254

This pattern can be observed when solving all partial differ-
ential “field” equations.
The variable that is of primary interest when solving a
geotechnical engineering problem has a significant influence
on howaccurately the unsaturated soil property functions need
to be determined. A crude approximation of the SWCC (and
subsequently the unsaturated soil property function) is all that
may be required for the computation of a particular variable
(e.g., the independent variable). On the other hand, a more
accurate assessment of the soil properties may be required for
the prediction of some other variables (e.g., some dependent
variables). Possibly no single factor more strongly influences
the engineering solution than an understanding of whether
it is the dependent or independent variable that is of primary
importance for solving the problem at hand. An understanding
of the primary variable that is of greatest interest for the prob-
lemat handmay also influence whether hysteresis effects need
to be taken into consideration in the assessment of saturated
soil property functions.
The above rationale applies for both saturated and unsat-
urated soil systems. There are also other factors such as the
level of risk that may influence the accuracy with which the
unsaturated soil property functions need to be determined.
For example, when the level of risk is high, the unsaturated
soil properties need to be assessed with greater accuracy.
Unsaturated soil properties are an extension of the satu-
rated soil properties. In many cases it is the saturated soil
properties that have the most significant influence on the
computed engineering solution.

SWCC. These estimation techniques are attractive, but the
user must be aware of the assumptions and limitations asso-
ciated with each of the estimation procedures. The volume
of voids in a soil is estimated and then used to estimate a
likely desorption curve for the soil. The effects of stress his-
tory, soil fabric, confinement, and hysteresis are difficult to
address when using estimation procedures associated with
the grain-size distribution curve.
Several models have been proposed for the estimation
of the SWCC from the grain-size distribution curve. The
starting point for most methods involves the representa-
tion of the grain-size distribution curve as shown by M.D.
Fredlund (2000) (Fig. 5.99). The methodology proposed by
M.D. Fredlund (2000) hypothesized that the grain-size dis-
tribution curve can be viewed as having incremental parti-
cle sizes ranging from the smallest particles to the largest
particles.
Each of the particle sizes can be assembled to build a
SWCC. Small increments (on a logarithmic scale) of uniform-
sized particles can be transposed to form an SWCC. Once the
entire grain-size distribution curve has been analyzed as a
series of incremental particle sizes, the individual SWCCs
can be combined using a superposition technique to give the
overall SWCC for the soil.
The SWCC for each uniform particle size range is assumed
to be relatively unique when building the overall SWCC using
the M.D. Fredlund (2000) methodology. Typical SWCCs for
various mixtures of sand, silt, and clay have been studied.
Representative SWCCs for various soils can be fitted using
the Fredlund and Xing (1994) equation. There are approxi-
mate curve-fitting parameters representative of various effec-
tive grain-size diameters.
The shape of an estimated SWCC is predominantly con-
trolled by the grain-size distribution and secondarily influ-
enced by soil density (i.e., initial porosity). The M.D. Fred-
lund (2000) unimodal and bimodal fitting of grain-size dis-
tribution curves can be used as the starting point for the
estimation of the SWCC.

5.12.1 Use of Grain-Size Distribution
for Estimation of SWCC
The grain-size distribution curve provides information on
the distribution of solids in a soil (i.e., percent passing cer-
tain particle sizes). The SWCC is related to the distribution
of voids in the soils and the amount of water in the voids.
Various studies have been undertaken that show that the
grain-size distribution curve can be used to estimate the

100

80

Fredlund unimodal

Experimental

60

USCS
% clay

40

Particle-size
log PDF

20

USCS
% sand

0

0.0001 0.001 0.01 0.1 1 10 100

Particle size, mm

Figure 5.99 Fit of grain-size distribution curve for uniform silt (data from Ho, 1988) with M.D.
Fredlund (2000) grain-size equation.