Application of artificial intelligence for improving longwall mine stability

摘要: Longwall mining is now well accepted by the coal mining industry due to its high productivity and safety. However, unexpected threats, such as roof falls, floor failure, and shield damage due to instability remain major concerns. Prediction of these potentially dangerous conditions requires an understanding of the interactions among face supports, the immediate and main roofs, the overburden strata, the floor and their combined behaviour as the face advances. Micro-computer technology has now made it practical to monitor responses of longwall face supports and to interpret the data in real-time utilizing artificial intelligence techniques. This paper describes the detailed framework of a computer controlled data acquisition system, where both historical information and current data is subsequently interpreted in near-real time to provide a warning system for imminent instability problems. Neural network and tree classification techniques are employed to evaluate current and future consequences of ground-stability. 12 refs., 7 figs., 1 tab.