In this paper a new empirical failure criterion is proposed to predict rock strength under polyaxial stress conditions. The proposed strength criterion is a three-dimensional extension of the popular two-dimensional criterion by Bieniawski (1974). The criterion has two empirical constants that can be determined using standard triaxial compression tests. Using these same empirical constants, failure under polyaxial conditions can be predicted. The accuracy of the failure criterion has been established by conducting a series of triaxial compression and hollow cylinder tests on Apache Leap tuff. First, the results of the triaxial compression tests are used to determine the two empirical constants for Apache Leap tuff. Secondly, using these two constants, the failure of the hollow cylinders of Apache Leap tuff under various conditions are predicted and compared with the actual experimental results under the same conditions. The new empirical failure criterion matches the experimental results extremely well.
Many strength criteria have been recently developed to describe the failure of geological materials, e.g., Wiebols and Cook (1968), Franklin (1971), Bieniawski (1974), Hoek and Brown (1980), Yudbir et al. (1983), Kim and Lade (1984), Johnston (1985), Sheorey et al. (1989) and Fuenkajorn and Daemen (1992). Most of these criteria predict that the intermediate principal stress has no effect on rock strength. However, experimental studies have indicated that the intermediate principal stress does influence rock strength in many instances (Murrell, 1966; Jaeger and Hoskins, 1966; Mogi, 1967; Handin et al., 1967).
In this paper a new empirical failure criterion is proposed to predict rock strength under polyaxial stress conditions (i.e., all three principal stresses can be differen0. The proposed strength criterion is a three-dimensional extension of the popular twodimensional criterion by Bieniawski (1974). The criterion has two empirical constants that can be determined using standard triaxial compression tests (i.e., 02=03). Using these same empirical constants, failure under polyaxial conditions can be predicted. The accuracy of the failure criterion has been established by conducting a series of triaxial compression and hollow cylinder tests on Apache Leap tuff. First, the results of the triaxial compression tests are used to determine the two using these two constants, the failure of the hollow cylinders of Apache Leap tuff under various conditions are predicted and compared with the actual experimental results under the same conditions. This same procedure has been repeated for published test results on limestone by Hoskins (1969). In both cases the new empirical failure criterion matches the experimental results extremely well.
The following section of the paper introduces the new failure criterion and the various attributes of this criterion. Section 3 then describes the results of standard triaxial and hollow cylinder tests that were used to verify the criterion, and Section 4 compares the predictions of the criterion with the experimental results. Conclusions and future work are described in Section 5.
