ABSTRACT

In the course of mining hard coal seams under rock burst hazard conditions one of the most important tasks to prevent the hazard is prediction of the time and place of occurrence of strong (Ew ≥ 104J) eismic tremors. The task is difficult and most often accomplished to an unsatisfactory degree. This contribution presents an attempt of applying neural networks with this end in view. The time series of seismic energy emission recorded while advancing a longwall face with caving at one of collieries was investigated. It hasbeen shown that the GRNN-type neural networks may fulfil such a task with an efficiency of about 50%.