Shear Strength of Unsaturated Soils - Unsaturated Soil Mechanics in Engineering Practice - page 568

Figure 11.62 Extended Mohr-Coulomb failure envelope based on direct shear test results.

(i.e., unsaturated soil shear strength function). The nature
of the shear strength versus net normal stress failure enve-
lope for saturated soils is explained in many soil mechanics
topics and is not given further consideration here.
Typical laboratory test results obtained from undisturbed
and compacted, unsaturated soil specimens are presented in
this section. Only results from initially similar undisturbed
or compacted soil specimens having the same dry density
and water content can be analyzed to obtain the φ b

Seventeen undisturbed specimens 63.5mm in diameter
and approximately 140mm in height were tested. The tests
were performed using the consolidated drained triaxial test-
ing procedure. The deviator stress was removed from the
soil specimen once a maximum value for deviator stress
was obtained for a particular stage (i.e., multistage testing
procedure). Then a new stress state was applied for the next
stage of testing. The strain rate used in the testing program
ranged from 1.7

shear

10 -5 %/ s. A 5-bar high-air-
entry disk (i.e., 505 kPa) was used for all tests. The effective
angles of internal friction, φ , were 33.4 ◦ and 35.3 ◦ for the
decomposed granite and rhyolite, respectively.
Figure 11.63 presents typical test results from a decom-
posed granite specimen using a cyclic multistage testing
procedure. The test was performed by maintaining a con-
stant net confining pressure σ 3 −

10 -5 to 6.7 × ×

strength parameter or function.
Triaxial test data are presented in the following sections
followed by direct shear test data. The triaxial test data
are categorized as (1) consolidated drained test results and
(2) constant-water-content test results. Both cases are used
to illustrate unsaturated soil behavior. Data are presented to
illustrate the linear and nonlinear relationship between shear
strength and matric suction. Limited data on undrained and
unconfined compression tests are also presented.

u a and varying the matric

suction u a − u w from one stage to the next. The test results

were interpreted based on the assumption that the effective
angle of internal friction remained as a constant value. Sim-
ilar multistage test results from two rhyolite specimens are
shown in Fig. 11.64.
The average φ b angles from all of the test results were
found to be 15.3 ◦ for the decomposed granite and 13.8 ◦
for the rhyolite. It was observed that the soil structure in
the soil specimens appeared to become disturbed to some
degree during the second and third stages of testing. As a
result, the measured peak deviator stress for the last stage
(i.e., stage No. 3) may actually be smaller than that which
would be obtained from a test on a single specimen under
the same stress conditions. Part of the reduction in strength
may also be due to nonlinearity in the shear strength versus
matric suction relationship.

11.8.1 Consolidated Drained Triaxial Tests
A series of multistage consolidated drained triaxial tests on
undisturbed specimens were performed by Ho and Fredlund
(1982b). The specimens were from two residual soil
deposits in Hong Kong, namely, decomposed granite and
decomposed rhyolite. The decomposed granite specimens
are mainly silty sand with an average specific gravity G s
of 2.65. The decomposed rhyolite specimens are sandy
silt with an average specific gravity of 2.66. The mineral
compositions of these two soils are similar. Both soils are
brittle and highly variable. Undisturbed specimens were
sampled tube samples from boreholes and block samples
from open cuts.