A material point method to simulate macro-scale problems in mining

Seismic hazard poses a major risk to mines, and is one of the biggest deciding factors of the future of a mine. It is not only capable of taking lives, but also guides the resources required to extract the minerals. Numerical modelling is an important tool available to mining engineers, to help them manage seismic risk. Modelling allows engineers to study how the rock mass will respond to the ever changing rock mass conditions, since they cannot physically observe it. Fortunately, there are a plethora of modelling tools available to mining engineers to assist in this task. However, most of these tools are specialised and are suited to only address certain problems. This study is aimed at researching and developing a tool that miners can use to address a broad spectrum of problems, with the ultimate goal of better understanding seismicity. This thesis is based on the Material Point Method (MPM), a novel numerical approach that combines the best aspects of Lagrangian and Eulerian techniques. The attraction of the MPM over many other numerical strategies is that it does not require repeated remeshing of the underlying computational grid when large deformations occur. Furthermore, it is relatively simple to construct intricate models of large mining layouts. This study makes four major contributions to the MPM. First, a new computational background mesh is developed for the MPM such that it can be used to simulate models of large mining layouts. Second, a novel method is developed to extract seismic events from elasto-plastic regions in the model. These events are not simply points in space, as commonly assumed in other numerical modelling packages that is currently available to the mining industry, but it also contains size as well as a source mechanism that may accurately resemble real seismicity that occur around a mine. Third, a novel algorithm of the coupled MPM is proposed to handle problems of fluid- and solid mixtures in mediums with heterogeneous permeability. Fourth, a new method for handling fractures as well as fracture propagation is proposed. This is used to simulate hydraulic fracturing preconditioning in mining where a dense fracture network is created within a small volume of rock such that stress is redistributed away from the area and become less seismically hazardous. These additions to numerical modelling in mining allows for addressing a wide range of problems, all with the goal to have better control over mining-induced seismicity.