A knowledge-based system for assessing spontaneous combustion risk in longwall mining

Abstract The application of computer technology has achieved considerable success in many parts of the mining industry. However, it has largely been confined to those operations which tend to consist of numerical data and exact algorithms. The knowledge-based system approach has made it possible to build applications which can aid decision making requiring empirical knowledge, subjective judgement, experience as well as uncertain information. The paper describes the development of an expert system for the assessment of spontaneous combustion risk in longwall mining. Previous studies on spontaneous heating risk are briefly reviewed. This application utilizes an expert system shell that combines a powerful inference engine with a user-friendly interface. The system quantifies the severity of self-heating risk in a specific longwall face using certainty factor techniques. The effect of alternative mining parameters on the risk of spontaneous combustion is evaluated such that the mining method can be modified to avoid dangerous combinations. The system also identifies the major factors contributing to the occurrence of self-heating risk and suggests precautionary measures against the potential heating risk. The system has been subjected to hypothetical case studies and the results are encouraging. Further developments of the system are discussed.