A automating potash miner bit replacement using neural networks

An important factor in the productivity and profitability of continuous mining machines used in Saskatchewan potash operations is the replacement of worn cutting tools. This activity is commonly performed either once per shift, or when the operator judges the machine is not producing at a sufficient level. Research is in progress in conjunction with the University of Saskatchewan to investigate new methods of detecting bit wear and automating machine outage scheduling using neural networks. A successful solution would further enable industry to achieve more advanced tele-operated mining. The neural network method may also demonstrate improved average profitability compared to conventional methods. This paper describes the neural network methods being employed and the current status of simulation testing of the technique.