Effect of Various Unloading Criteria on Rock Failure Parameters From Multi-Stage Triaxial Test - a Comprehensive Study

This paper compared various unloading criteria used in multi-stage triaxial tests for determining the Mohr-Coulomb failure envelope. Results from single-stage triaxial tests formed the baseline for comparison. Three types of rock outcrop were used: Berea sandstone, Eagle Ford shale, and carbonate rock. Mineralogy, porosity, grain and bulk density were also measured, which are used to determine the sample heterogeneity and interpret the discrepancy of test results. Examples illustrated both zero and maximum volumetric strain criteria were not always applicable, especially for brittle rocks. The criterion of radial-strain gradient, defined as the ratio of change of radial strain and change of time, is generally suitable for any type of rocks. Irrespective of the applied confining pressure, samples were observed to break at almost the same radial-strain gradient for the same type of rocks. Failure envelopes from the radial-strain gradient method matched very well with those from the maximum volumetric strain for all samples tested. Compared to results of single-stage triaxial tests, multi-stage tests yielded a very good approximation of failure envelopes for Berea sandstone, while discrepancy was observed for Eagle Ford shale and carbonate rock because of the heterogeneity of samples.