Engineering Rock Mechanics

Specimen geometry, loading conditions and environmental ekts 99

Axial

stress I (‘a

Uniaxial compression

Qa

Uniaxial tension

Normal

+<; Shear

Direct shear

Axial

• s:$=$;L UL y$* e3

Lateral

U ua stress u1

Biaxial compression Triaxial compression Polyaxial compression

Figure 6.13 Specimen loading conditi ons in general laboratory use.

We have discussed uniaxial compression; let us now consider uniaxial
or direct tension, and also the generation of a tensile stress through
compressive loading in indirect tensile tests.
The uniaxial tension test, as illustrated in Fig. 6.13, is not as a rule used
in engineering practice. There are two reasons for this: first, it is difficult to
perform; and, second, the rock does not fail in direct tension in situ. Through
the servo-controlled testing method, the complete stress-strain curve in
tension has been obtained using axial displacement as feedback in the closed
loop. This displacement is the most sensitive indicator of failure because a
single main crack develops laterally. However, this curve is really only of
academic interest because the essentially single crack failure mode leads to
ultra-brittle behaviour. Even for establishing the tensile strength of rock
itself, a state of pure tension with no applied or induced bending moments
is difficult to achieve. Some irregularity in the compression test can be
tolerated, but in tension any such irregularity leads to premature failure.
For these reasons, the tensile strength is normally measured by indirect
tests in which the tensile stress is generated by compressive loading. (The
tensile strength of the rock is very much lower than the compressive
strength, so that such indirect tests are possible; for the same reason, it is
not possible to have indirect compression tests.)
In Fig. 6.14, two indirect tensile tests are shown, with the point load test
being the most widely used test on intact rock. In each case, through the
testing configuration, the maximum tensile stress can be calculated from
elasticity theory as a function of the compressive force and specimen
dimensions. The tensile strength is, therefore, the maximum tensile stress
calculated to be present in the specimen at failure. Such a calculation is based
on ideal material assumptions and does not take account of different
critically stressed volumes in each test. As might be expected from our
earlier discussion, the tensile strength varies for a given rock type tested
in these different ways and hence is not an intrinsic material property.