318
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
shear device, the magnitudes of the intermediate and minor
principal stresses are not known and are governed by the
vertical stress applied and the strength properties of the soil
being tested.
Progressive failure is the condition where the peak shear
strength is not mobilized in every point of the failure plane at
the same instance. This is caused by the non-uniform
distribution of strain in the failure plane combined with the
strain-softening characteristic of the soil. Some locations on the
failure plane will mobilize the peak shear strength while others,
having achieved more or less displacement, will mobilized a
shear strength below the peak shear strength. For the direct
shear device, investigators have reached different conclusions
about the effect of stress concentrations in this device. Hvorslev
(1960) measured horizontal displacements along the failure
plane in a direct shear device and found that the displacements
are not uniformly distributed, thereby causing progressive
failure to occur. Alternatively, a finite element study of the
direct shear box presented by Potts et al. (1987) showed that
although stress concentration exist on the failure plane of the
direct shear box, at the moment of failure the stresses on the
failure plane are more or less uniform and the peak shear
strength measured is not affected by progressive failure. For the
triaxial device, stress concentration caused by the end restraints
can influence the results. Research performed by Taylor (1941)
showed that if the ratio of length to diameter is between 1.5 and
2.5, the effect of the stress concentration is negligible.
In the direct shear device, the orientation of the principal
stresses on the failure plane varies during the shearing stage of
the test and the final orientation is unknown. In the triaxial test,
the major and minor principal stresses act on the horizontal and
vertical planes and this orientation does not change during
shear.
The orientation of the failure plane in the direct shear device
is predetermined as being near the midpoint between the upper
and lower halves of the shear box. In the triaxial device, the
orientation of the failure plane is governed by the soil structure
and the strength properties of the soil.
A literature review was undertaken to locate previous
comparisons of the shear strength obtained using the triaxial and
direct shear apparatuses. Skempton (1964) stated that the same
effective stress shear strength parameters were obtained from
tests conducted on eight specimens of Boulder Clay using the
direct shear and triaxial devices. Casagrande and Poulos (1964)
presented the results of CD direct shear and CD triaxial tests
performed on compacted specimens of a lean clay which
showed that about the same shear strength envelope was
obtained with the two tests. Moon (1984) performed CD direct
shear and CU triaxial tests on undisturbed samples of a fat clay
and found differences of less than one degree for the effective
stress friction angle and less than 8 kPa for the effective
cohesion intercept obtained from these two tests using the
maximum principal stress ratio as the failure criterion. Thomson
and Kjartanson (1985) performed CD direct shear tests and CD
and CU triaxial tests on undisturbed samples of a lean clay and
a fat clay and found that the results plotted on the same failure
envelope. Abdel-Ghaffar (1990) compiled results from the
literature where direct shear and triaxial tests were performed
on undisturbed samples of the same soil. He concluded that the
direct shear and triaxial devices provide comparable values for
the effective stress friction angle and cohesion intercept.
Maccarini (1993) performed CD direct shear and CD triaxial
tests on a residual soil from Rio de Janeiro. For these tests, the
tests specimens were oriented so that the failure plane in both
devices coincided with the direction of stratification of the soil.
The stratification of the soil had a dip angle of 25
°
. Based on the
results obtained, Maccarini concluded that similar values of
effective stress cohesion and friction angle are obtained from
both tests if the stratification is taken into account.
Based on the results presented above, it can be seen that the
available information in the literature is divided on whether
these two devices will provide the same effective stress shear
strength parameters. Investigators that have found agreement in
the results did not state whether the soil tested had a preferred
particle orientation or layering that could influence the results.
3 DATA COLLECTION AND RESULTS
3.1
Undisturbed test specimens
A subset of 63 CU triaxial test series and 146 CD direct shear
test series was selected from the test results available from the
New Orleans investigation. Each test series normally consisted
of three individual tests conducted at different confining
pressures. Only high quality tests were selected. The test results
selected had to comply with the following requirements: 1) At
least two test specimens were consolidated to stresses that were
higher than the preconsolidation stress; 2) The end of primary
consolidation was achieved during the consolidation stage of the
test, 3) Index properties were available; 4) A peak deviator
stress was reached for CU triaxial tests in less than 15% strain
and a peak shear stress was reached for CD direct shear tests at
a horizontal displacement less than 0.4 inches.
Since all the samples used were normally consolidated, the
shear strength interpretation assumed that the effective stress
cohesion intercept was zero. The least-squares method was used
to obtain the corresponding effective stress friction angle.
Shown in Figure 1 are the values of the effective stress
friction angles as a function of the Plasticity Index (PI) for the
undisturbed New Orleans area test specimens for CD direct
shear and CU triaxial tests. The trend lines shown on this figure
are based on a statistical analysis performed by the authors.
These lines are presented to better show the trend of the data
and are not intended to be used as a correlation to predict
friction angles. From this figure, it can be seen that the effective
stress friction angle measured with the CU triaxial device is
generally higher than that measured with the direct shear device.
This difference was found to increase with increasing plasticity
index of the soil. In general, the difference in friction angle
ranged from about 2 degrees to 5 degrees.
Plasticity Index (%)
0
20
40
60
80
100
120
Effective Stress Friction Angle (deg)
10
15
20
25
30
35
40
CD Direct Shear
CU Triaxial
Figure 1. Relationship between effective stress friction angle and
plasticity index for CD direct shear tests and CU triaxial tests on
undisturbed samples.
3.2
Remolded test specimens
Three remolded clays were tested to allow a comparison of the
fully softened shear strength measured in CD direct shear tests
and CU triaxial tests. The index properties of these clays are
shown are in Table 1. NOVA clay was obtained from a site in
Northern Virginia in Fairfax County. Vicksburg Buckshot Clay
(VBC) was obtained from the stockpile maintained at the
Engineering Research and Development Center of the U.S.
Army Corps of Engineers. It has been the subject of many
research projects (Ladd and Preston 1965; Mitchell et al. 1965;
Al-Hussaini and Townsend 1974). Colorado Clay is a lean clay
(CL) from Silverthorne, Colorado.
To prepare the remolded samples, the soils were first soaked
in water for at least 48 hours, and then processed through a No.
