Research on mechanisms of roof pressure prediction

An intelligent recognizing method of coal mining roof pressure real-time prediction was presented, based on SVM(support vector machine) which has good generalization ability and adaptive multi-mode control algorithm of adaptive characteristics. Firstly, the measured roof pressure data was processed with adaptive algorithm, and a set of intrinsic mode functions(IMF) was obtained based on the empirical mode decomposition(EMD). Then mechanical research was conducted on those IMF. Finally, regression prediction was carried out using SVM. The application example shows that this model can reflect the regular changes of roof pressure, obtain satisfactory prediction results and can meet the needs of safe production.