True triaxial compressive strengths of Maha Sarakham (MS) salt are determined by using a polyaxial load frame. The salt specimens are cut and ground to obtain rectangular blocks with a nominal dimension of 5.4×5.4×10.8 cm3. The load frame equipped with two pairs of cantilever beams is used to apply the constant lateral stresses (σ2 and σ3) to salt specimen while the axial stress (σ1) is increased at 0.5–1.0 MPa/s until failure occurs. The deformations induced along the three loading directions are monitored and used to calculate the tangent elastic modulus and Poisson's ratio of the salt. The results indicate that the elastic modulus and Poisson's ratio of the MS salt are averaged as 21.5±2.6 GPa and 0.40±.04. For the Coulomb criterion the internal friction angle determined from the triaxial loading condition (σ2= σ3) is 50°, and the cohesion is 5 MPa. The effect of σ2 on the salt strengths can be best described by the modified Wiebols and Cook criterion with the mean misfit = 3.5 MPa. The empirical (power law) Mogi criterion tends to underestimate the salt strengths particularly under high σ3 values. The modified Lade criterion overestimates the actual strengths at all levels of σ3, showing the mean misfit = 15.4 MPa. The Coulomb and Hoek and Brown criteria cannot describe the salt strengths beyond the condition where σ2 = σ3, as they cannot incorporate the effects of σ2. Both circumscribed and inscribed Drucker-Prager criteria severely underestimate σ1 at failure for all stress conditions, showing the largest mean misfit of 19.5 and 34.7 MPa, respectively.
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 of cylindrical rock core specimens. A significant limitation of these conventional methods is that the intermediate and minimum principal stresses are equal during the test while 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). It has been commonly found that the compressive strength obtained from conventional triaxial testing cannot represent the actual insitu strength where the rock is subjected to an anisotropic stress state [1–6]. From the experimental results on brittle rocks obtained from Haimson [2], Colmenares & Zoback [7], it can be generally concluded that in a σ1 - σ2 diagram, for a given σ3, σ1 at failure initially increases with σ2 to a certain magnitude, and then it gradually decreases as σ2 increases. The effect of σ2 is more pronounced under higher σ3. This states that the intermediate principal stress confines the rock in such a way that fractures can only be initiated and propagated in the direction parallel to σ1 and σ2. The effect of σ2 is related to the stress-induced anisotropic properties, and the end effect at the interface between the rock surface and loading platen in the direction of σ2 application.
