Dynamic prediction model of gas emission in Tangshang Mine

To improve the prediction accuracy of gas emission, a BP neural network was applied to establish a dynamic prediction model of gas emission under the MATLAB environment by using BP neural networks' characteristics of self-learning, self-organizing and self-adapting. The model was trained and tested by analyzing the real-time monitoring data of gas signals from Tangshan Mine. Test results show that the model has higher prediction speed and accuracy. By using the model the dynamic prediction of gas emission in the working face can be realized, the safety state and the potential hazard can be synthetically estimated to provide security for safety production.