Development Of On-Line Evaluation Of Longwall Roof Behavior Using Neural Networks And Fuzzy Sets

Micro-computers have made it practical to install a high-speed acquisition system to collect data from a longwall face in near real-time. By simultaneous data analyses using neural networks and fuzzy sets, information regarding longwall mining performance and underground behavior can be provided and forecasting longwall instability and detecting geological anomalies is feasible. This paper presents details of the application of neural networks and fuzzy sets in developing on-line evaluation of longwall roof behavior. Techniques for interpreting longwall shield pressures and characterizing roof conditions are discussed.