Numerical computation of average pillar stress and implications for pillar design

Abstract

A number of issues relating to the computational aspects of pillar design are addressed in this paper. The computation of average pillar stress values is important when attempting to establish criteria for pillar design and in the analysis of the stability of tabular pillar layouts. One of the default ‘classic’ numerical methods that are used to determine pillar stresses is the displacement discontinuity method. In many instances it is not clearly understood that this approach does suffer from some limitations, particularly in relation to the fact that in coarse element simulations, the simulated average pillar stress (APS) can depend on the chosen mesh size. The nature of this error is highlighted in this paper and some strategies are suggested to bound the magnitude of these errors. It is demonstrated as well that the popular linear stiffness approximation to pillar or seam compressibility does appear to allow reasonably accurate estimates of the average pillar stress when either the pillar height is varied or when the seam modulus differs from the host rock modulus. A practical implication of this study is that if the seam modulus is noticeably lower than that of the host rock, such as for coal seams, it is important to use a linear stiffness constitutive model for the pillars rather than a ‘rigid’ pillar assumption. This added complexity seems unnecessary, however, when simulating hard rock pillars in mines where the seam modulus is very similar to that of the surrounding rock.