The mining industry can greatly benefit from automation. A great deal of work has been done on this subject and is still ongoing. With automation comes the possibility for optimization, because more information is available, and actions can be repeated with more accuracy. Many factors in an underground environment make mining automation a challenging prospect. These factors include the difficulty and cost of installing the needed infrastructure. The work described in this dissertation focuses on a mining setup where vehicles such as LHDs and trucks are used to collect and transport ore underground. Considerable progress has been made in automating underground vehicles, and successful tests have been done underground. The next obvious step is to find ways of using the increased data to optimize the decisions that are made with regards to the dispatching of the vehicles. Possible solutions to the problem of optimizing the autonomous vehicle dispatch system in an underground mine are investigated. Possible optimization strategies are evaluated using a simulated environment. In the simulated environment a block cave mine is modelled, and the simulation setup is discussed in detail. The operation of a block cave mine as it is operated currently is simulated to obtain a benchmark for the evaluation of further results. The simulation results for the developed strategies are evaluated against specific criteria, and indicate definite improvements on current methods used in mines. Some important things that must be kept in mind for the physical implementation of the dispatching strategies, as well as mining automation in general, are also discussed.
Dissertation (MEng (Electronic Engineering))--University of Pretoria, 2007.