Geoscience Reference
In-Depth Information
6.3.2.3.
Triaxial tests with measurement of capillary pressure (TTu)
In most cases, capillary pressure was measured separately from the modulus
measurements.
A usual triaxial cell is equipped with a semi-permeable porous stone (with
1.5 MPa air entry pressure), located within the pedestal in order to prevent air from
passing into the water circuit. The pressure in the water phase is measured by means
of an absolute pressure transducer, while a relative pressure transducer connected to
the upper part of the specimen is used to measure the pressure in the air phase.
For water contents above the modified Proctor optimum water content,
corresponding to pore-water pressures higher than -50 kPa, the device is used as a
tensiometer (i.e. with
p
a
= 0 and
p
w
< 0). For water contents below the optimum, the
axis translation technique was used: a positive air pressure was applied to the sample
until positive pore-water pressure was obtained; thus, cavitation problems in the
measuring devices are avoided so long as air is prevented from passing through the
porous stone, which was the case in all these tests.
Volume changes are derived from the water volume coming in or out of the cell,
measured by the confining pressure controller.
Considering errors as a consequence of expansion of the Perspex envelope, the
compression of the latex membrane or air bubbles during isotropic loading, this
method was only used during application of the stress deviator, when the confining
pressure is constant.
The consolidated undrained tests with measurement of pore-water pressure and
control of air pressure were performed in two steps:
(i) isotropic consolidation under stresses of 7, 14, 28, 52 and 79 kPa;
(ii) compression under 79 kPa confining stress at a constant strain rate
(≈ 10
-6
s
-1
), up to failure.
During the consolidation phase, the changes in capillary pressure were recorded
until equilibrium was reached.
6.3.3.
Normalization of data
The results of SPTT and LPTT were all corrected to eliminate the effect of the
different initial void ratios (
e
0
) of the specimens tested and a “normalized” value
corresponding to
e
ref
= 0.5 was derived from the expression in Iwasaki
et al.
[IWA 78]:
Search WWH ::
Custom Search