Engineering Rock Mechanics

98 Intact rock

Slenderness increases,

strength decreases

Sample

shape

&

Figure 6.12 The shape effect in uniaxial compression.

To predict the strength of the rock in situ, and avoid the influence of the

shape effect, we can proceed in one of two ways: improve the laboratory

test procedures; or use empirical formulae to account for the shape effect.

The main laboratory method is to use platens which reduce the confining

effect, either through brush platens (which effectively load the specimen

ends over a number of small zones, thereby reducing the volume of rock

in triaxial compression) or flatjacks (which prevent shear stresses being

transmitted between the platen and the specimen). Other laboratory

techniques involve specimen geometries which reduce the effect, for

example, axial loading of hollow cylinders. Empirical formulae are the main

engineering approach, where a numerical relation is utilized to take into

account the shape effect. In fact, these formulae can incorporate the

diameter and the length separately and hence it may not be directly clear

from the formula how to separate the size and the shape effects, should

one wish to do so.

6.4.3 Loading conditions

We have seen, with reference to the shape effect, how the loading

conditions can affect the rock behaviour in uniaxial compression. Let us

now consider the many possibilities for rock testing and illustrate some of

the terms in general use. The sketches in Fig. 6.13 show the loading

conditions in the six main testing configurations. A particular point to note

is the difference between triaxial and polyaxial compression. Over the

years, triaxial compression has come to mean a test conducted using a

pressure vessel, with the consequence that o2 = 03. This is not true triaxial

compression in the sense that all three principal stresses can be

independently applied: for this latter condition we use the term polyaxial

compression. The application of three different principal stresses is quite

difficult to achieve in practice, and hence the test is not used routinely in

rock mechanics.