Validity of continuous-failure-state unloading triaxial tests as a means to estimate the residual strength of rocks

Abstract

The residual strength of rocks and rock masses is an important parameter to be constrained for analysis and design purposes in many rock engineering applications. A residual strength envelope in principal stress space is typically developed using residual strength data obtained from compression tests on many different specimens of the same rock type. In this study, we examined the potential for use of the continuous-failure-state testing concept as a means to constrain the residual strength envelope using a limited number of specimens. Specifically, cylindrical specimens of three rock types (granodiorite, diabase, and Stanstead granite) were unloaded at the residual state such that a full residual strength envelope for each individual specimen was obtained. Using a residual strength model that introduces a single new strength parameter (the residual strength index, or RSI), the results of the continuous-failure-state unloading tests were compared to conventionally obtained residual strength envelopes. Overall, the continuous-failure-state residual strength data were found to be consistent with the conventional residual strength data. However, it was identified that the primary factor limiting an accurate characterization of the residual strength for a given rock type is not the amount of data for a given specimen, but the variety of specimens available to characterize the inherent variability of the rock unit of interest. Accordingly, the use of continuous-failure-state testing for estimation of the residual strength of a rock unit is only recommended when the number of specimens available for testing is very limited (i.e. < 5).