The objectives of this study are to determine strength and stiffness of rock salt under different stress paths in the laboratory and to evaluate the predictability of the existing failure criteria and their parameters. The effects of stress paths have been studied through the triaxial compression and triaxial extension tests. The prepared specimens are rock salt from the Maha Sarakham formation at depths ranging from 150 to 300 m. The specimens have the rectangular blocks with normal dimensions of 4.4´4.4´8.8 cm3. A true triaxial loading device is developed for this study. Four different stress paths have been implemented:
σ1 increases while σ2 and σ3 are equally maintained constant;
σ1 increases while σ2 and σ3are simultaneously decreases;
σ1 and σ2 equally increase while σ3 are maintained constant; and
σ1 and σ2 equally increase while σ3 decreases.
For the stress paths (I) and (III), the mean stress (σm) is increased during testing (conventional method) and for the other two, the mean stress is maintained constant (constant mean stress method). The results indicate that the strengths obtained for the triaxial extension is lower than those of the triaxial compression. The strengths of the constant mean stress method are lower than those of the conventional method. The elastic modulus and Poisson's ratio tend to be independent of the stress path. The modified Wiebols and Cook criterion can well predict the triaxial strengths under different stress paths with the coefficient of correlations of 0.91for path (I) and (III) and 0.89 for path(II) and (IV).
The reliable strength estimation of a rock salt is necessary to develop safe and economical designs for solution mining, compressed-air energy storage and underground salt mining. Rock salt is an inhomogeneous and anisotropic material with complex behavior. The effects of confining pressures at great depths on the mechanical properties of rocks are commonly simulated in a laboratory by performing triaxial compression testing. A significant limitation of these conventional methods is that the mean stress is not constant during the test. The actual in-situ rock is normally subjected to an anisotropic stress state where the maximum, intermediate and minimum principal stresses are different (σ1≠ σ2≠ σ3) and mean stress of this condition is always constant.