Gas outburst prediction based on the intelligent Dempster-Shafer evidence theory
Caixia Gao, Fuzhong Wang, Dan Xu · 2017
Predicting gas outburst in coalmine extraction face accurately is an effective method to prevent gas outburst disaster. Because there are the features of suddenness, unevenness, uncertainty and dynamic in gas outburst, the existing prevention method should be improved in accuracy and effectiveness. Thus in the paper the predicted model of gas outburst is built combining fuzzy neural network and Dempster-Shafer(D-S) evidence theory, and the model specifically introduces the overall structure design of gas outburst predicted model, the selection of gas outburst evaluation indicators, the design of fuzzy neural network unit and the design of D-S evidence theory unit. The eight key factors including the thickness of coal layer, the geological structure types of coal and the gas pressure of coal layer are selected as the evaluation indicators of gas outburst, and the preliminary judgment of gas outburst state in local point of mining working face, is made by fuzzy neural network, and then global judgment of gas outburst state in mining working face is made based on D-S evidence theory. The simulated result shows that this method can make accurate judgments of gas outburst state grade, and regarding with the judgments of the three kinds of gas outburst state, the accuracy error is less than 0.0048% and the uncertainty value approximates to 0.