This paper presents a methodology to design jointed hard rock pillars transacted a clay-filled shear zone at Ngezi mine, Portal 4 (P4), in Zimbabwe. The nature of the problem is highly complex due to the large number of joints and possible shear zone configurations. On the mine, the shear zone can be found either above, below or within the ore zone. It was observed underground that large footwall heaves occur where the shear zone is in the footwall, while hanging wall fallouts were recorded where the shear zone is on the hanging wall. Tensile fracturing and large pillar lateral displacements were also observed and associated with the existence of the shear structure.
An investigation into the behaviour of pillars under these conditions were carried out using 2D numerical modelling. A discrete element approach in the software program UDEC was used to account for the large number of joints within the rock mass. Tensile fracturing observed in the failed pillars was simulated by using a voronoi tessellation background and calibrating the shear strength properties at the boundaries of the voronoi blocks in order to replicate macroscopic responses of the intact rock measured in the laboratory, i.e. unconfined compressive strength (UCS) test and indirect tensile test – Brazilian (UTB). This calibration exercise was conducted by replicating the actual UCS, UTB, Young’s modulus and Poisson’s ratio in UDEC and adjusting the contact properties until a satisfactory correlation between the modelled and laboratory results was obtained.
A discrete fracture network (DFN) approach was utilised to assimilate structural data from a mapping exercise conducted on the mine. The built-in DFN generator in the program 3DEC was used to generate stochastically several 3D DFNs from which, 2D sections were cut and imported into the 2D program, UDEC. Due to time constraint, only two DFNs (DFN 1 and DFN 2) were effectively examined and were selected such that DFN 1 incorporated a low angle random joint set. Four shear zone configurations were examined for each DFN including, shear in the footwall, in the hanging wall, in the orebody and a case without shear structure.
Based on the current mine design, three pillar sizes were examined, 2, 4 and 6 m or a total of 24 UDEC simulations. In the end, it was found that the models replicated satisfactorily the pillar behaviour and mechanisms observed underground together with the expected pillar peak strength. The findings of these study should constitute a vital input into the evaluation and assessment of future mining options at NGEZI, P4, while the methodology followed may be applied where highly jointed rock mass are intersected by a shear structure.
