Laboratory Tests - Rock Mechanics Based on an Anisotropic Jointed Rock Model - page 427

For large deformation of the specimen the axial strain ε a must be calculated from the

differential displacement of axial measuring points d related to their actual spacing .
Integration yields the so-called “natural strain” or “logarithmic strain”:

(14.39)

where 0 is the initial spacing of axial measuring points.
Creep tests are often carried out with several load stages, including loading and unload-
ing (Fig. 5.5).
The results of creep tests are usually represented in the form of strain-time curves
as illustrated in Figs. 5.4 and 5.5 for rock salt. The plots of measured strain rates
as a function of time and strain are often used for the representation of creep test
results as well (DGGT 1994a).
The evaluation of creep tests is normally carried out by means of fi tting the measured
strains to a specifi ed constitutive law or creep law, respectively (Sections 5.2 and 5.3,
Figs. 5.4 and 5.5).

Relaxation tests
Relaxation tests are uniaxial compression tests in which axial stress changes of a speci-
men subjected to a constant axial displacement are measured as a function of time.
Recommendations for a corresponding test procedure are given in DGGT (1994b).

True triaxial test
The testing procedure described in Section 14.4.2 is known as a “conventional triaxial
test”. Conventional means that only two principal normal stresses in the specimen are
different. In contrast, a triaxial test in which all three principal normal stresses are dif-
ferent is referred to as a “true triaxial test”.
Most true triaxial testing devices were designed to apply three independently variable
principal normal stresses to the opposing faces of cubical shaped specimens (Hojem &
Cook 1968, Mogi 1971, Paterson 1978, Sture & Desai 1979, Gau et al. 1983, Michelis
1985a, Esaki & Kimura 1989, Takahashi & Koide 1989, Hunsche & Albrecht 1990,
Hunsche 1992a, Hunsche 1994, Wawersik et al. 1997, Haimson & Chang 2000). Lubri-
cation of the specimen's end faces is necessary using these testing devices (Cristescu &
Hunsche 1998). Otherwise the friction between the loading platens and the specimen's
ends may lead to an overestimation of the intact rock's strength (Fig. 14.6).

Triaxial cells that allow application of true triaxial stress states to thin-walled hollow
cylindrical specimens (Paterson 1978) or to cylindrical specimens (Smart 1995, Smart et
al. 1999) are also reported in literature.