Page 66 - Numerical Analysis and Modelling in Geomechanics

COMPRESSED AIR TUNNELLING 47
Laboratory tests
18
Donald carried out a series of shear tests on unsaturated fine sands and coarse silts
in a direct shear test box. The tests were run with atmospheric pore-air pressure
(zero gauge pressure) and negative pore-water pressure. The results of the tests
indicated that by increasing the matric suction, the soil strength increased to a
peak value and then decreased to a nearly constant value. Ho and Fredlund 19
studied the effect of matric suction, (u −u ), on the shear strength of soil. The
w
a
results can be applied to engineering problems in which only one stress state
variable (i.e. matric suction) is changed such as in slope stability and the swelling
of expansive soils. However, in compressed air tunnelling, a change in the
internal air pressure, u , will cause a change in matric suction and also a change
a
in net normal stress. Therefore, the stress path followed in this special problem is
different from the other problems studied so far in the literature.
The laboratory test associated with an engineering problem should closely
simulate the loading conditions that are likely to occur in the field. Therefore, a
suitable testing programme to study the effect of air flow in compressed air
tunnelling, should account for the variation of both independent stress state
variables.

Testing programme
Triaxial shear strength tests were carried out on a decomposed granitic soil
sample to study the effect of air flow on the shear strength of the ground in
compressed air tunnelling. The testing programme comprised three test series as
follows:

i. Simultaneous variation of both net normal stress and matric suction by
changing the air pressure,
ii Variation of only matric suction, keeping the net normal stress constant, and
iii Variation of net normal stress at constant (zero) matric suction (saturated
sample).

In order to eliminate the effect of variability of the soil sample, a multistage triaxial
testing programme was used, as proposed by Ho and Fredlund. 19 In this way,
considerable information could be obtained from a limited number of samples.
The tests were carried out in a modified triaxial apparatus. 11

Apparatus

Due to the presence of both air and water in the soil, the testing procedure for
unsaturated soils is more complicated than that for saturated soils. Special design
considerations are required for development of equipment for testing unsaturated
soils under different loading conditions. A suitable design should be able to