Numerical analyses of pillar behavior with variation in yield criterion, dilatancy, rock heterogeneity and length to width ratio

Abstract

Abstract With recent advances in numerical modeling, design of underground structures increasingly relies on numerical modeling-based analysis approaches. While modeling tools like the discrete element method (DEM) and the combined finite-discrete element method (FDEM) are useful for investigating small-scale damage processes, continuum models remain the primary practical tool for most field-scale problems. The results obtained from such models are significantly dependent on the selection of an appropriate yield criterion and dilation angle. Towards improving its capabilities in handling mining-related problems, the authors have previously developed a new yield criterion (called progressive S-shaped criterion). The focus of the current study is to demonstrate its use in modeling rock pillars through a comparative analysis against four other yield criteria. In addition to the progressive S-shaped criterion, only one out of the four other criteria predicted a trend in strength consistent with an empirical pillar strength database compiled from the literature. Given the closely-knit relationship between yield criteria and dilation angle in controlling the overall damage process, a separate comparison was conducted using a mobilized dilation model, a zero degree dilation angle and a constant non-zero dilation angle. This study also investigates the impact of meso-scale heterogeneity in mechanical properties on the overall model response by assigning probability distributions to the input parameters. The comparisons revealed that an isotropic model using a combination of progressive S-shaped criterion and mobilized dilation angle model is sufficient in capturing the behaviors of rock pillars. Subsequently, the pillar model was used to assess the effect of L/W (length/width) ratio on the peak strength.