Experimental Evaluation Of A Mine Electrical Failure Prediction System

The research and development of a pattern recognition system which can predict incipient failures and safety hazards in the mine power network is described. An examination of mine electrical safety hazards revealed they are often related to the electrical equipment down-time experienced on mine sections. Based on this, a method is proposed that can improve both production and safety, without compromising either. Equipment failure modes are examined to illustrate how the ultimate failure can be predicted for enough in advance to allow a convenient and safe repair, followed by a discussion of actual monitoring system performance. Finally, the paper is concluded by projections of the system's utility to industry.