The triaxial test has usually been employed to characterize the mechanical properties of rock-like materials in a laboratory. This conventional method, called the single-stage triaxial test (SST) provides only a pair of peak and residual strengths from a single specimen Therefore, it is very inconvenient in achieving the whole spectrum of mechanical characteristics of rock specimens whose amount is so limited due to the geological conditions, etc..
Kovari and Tisa (1975) have proposed a test method, designated the multiple-stage triaxial test (MST) which is capable of deducing more sets of strength parameters from a single specimen and comprehending the full range of strength envelopes by using a few specimens. It has been proved that their MST method is a suitable substitute for the conventional SST when a dry and hard rock specimen such as marble and sandstone is used. However, their test procedure may not be adequate for a saturated porous soft rock.
A new triaxial test method was developed to determine the mechanical behaviors of saturated soft rocks under various confining pressures. The new one is tentatively called as MST for soft rock in order to differentiate from the SST and the MST for hard dry rocks. In this method, a set of peak and residual strengths are obtained after each consolidation-loading- unloading cycle. Loading is normally done by the axial strain rate control.
When the specimen is subjected to lower confining pressure, there exist some difficulties in loading-unloading control of the MST for peak-strength, especially with a brittle specimen. As the gradient change near the peak strength of the axial load-displacement curve is insignificant, a correct prediction of the peak point is usually very difficult and misjudgement causes an abrupt failure of the test specimen.
To improve the accuracy of peak-strength MST under lower confining pressure, a new test procedure has been added. This involves the measurement of lateral strain while the test specimen is in axial compression. This is beneficial as the lateral strain can sensitively reflect the volume change due to the development of microcracks within the specimen. Including the monitoring of the axial stress-lateral strain curve, the correct position of peak strength can more accurately be predicted.
In addition, a peak strength MST method with lateral strain rate control has also been studied. The mechanical behaviors of the specimen, obtained from the results under lateral strain rate control, are compared with those under axial strain rate control.
In the study, two kinds of rock specimen were used, namely sandy silt rock and tuff (Nikko rock). From the physical properties listed in Tables 1, one can readily presume that all of specimens were very porous and comparatively isotropic and homogeneous.
Cylindrical specimens were prepared in the same direction from rock block and carefully trimmed up to 5 cm in diameter and 10 cm in height, with core cutter and diamond-bladed saw.
