The Effect of Natural Fractures on the Mechanical Behavior of Limestone Pillars: A Synthetic Rock Mass Approach Application

Abstract

In general, underground limestone mines have inherently strong rock and experience good ground stability. Also, modern pillar design guidelines developed by National Institute for Occupational Safety and Health (NIOSH) have improved the design of stable layouts for modern limestone mines. However, ground control-related incidents are still an important problem. In underground limestone mines, previously mined sections stay open for the life of the mine which may be many years, and it is possible for travel ways to working faces to pass through these old sections. In a recent massive pillar collapse in an old section of a mine in Pennsylvania (Pa), three miners were injured outside of the mine due to an air blast. Also, frequent reports are indicating pillar sloughing, spalling and roof falls. These incidents highlight the potential safety impact on the miners in underground limestone mines. In the pillar design guidelines published by NIOSH, pillars are mostly examined for the existence of one-large discontinuity crossing completely through the pillar. However, the influence of multiple joint sets and natural fractures on the insitu pillar strength prediction and localized failures of the pillar are not covered by the guidelines. In this thesis, the influence of naturally exiting joint sets and fractures on the mechanical behavior (i.e. strength and failure mechanisms) of underground stone pillars is studied. In order to investigate pillar mechanics, a systematical methodology is developed based on the novel approach, the Synthetic Rock Mass (SRM) by utilizing the two-dimensional Universal Distinct Element Code (UDEC). In order to form the first component of SRM, the Bonded Particle Model (BPM), the mechanical properties of the standard size laboratory rock specimen scaled up to the upper-limit of the Hoek-and-Brown Scaling Equation. Then, Voronoi-Trigon Discretization Logic is used to model the intact rock matrix of the stone mine pillars. Later, field data is used to stochastically generate Discrete Fracture Networks (DFNs), and SRM models are established by integrating the BPM and DFNs. Then, rock specimen sizes are increased from laboratory size to field size by sampling the generated DFNs. In the up-scaling operation (i.e. specimens’ size increase), the homogenization process is applied that the estimated strength properties of the pillars by SRM are captured with a new BPM. By doing so, the numerical simulations calibrated against the empirical stone mine pillar strength equation established by NIOSH. Finally, the predicted strength parameters are used to examine the pillar failure mechanics with various