基于BP神经网络的采煤工作面瓦斯涌出量预测

Gas emission prediction of coal mining face based on BP neural network

在建立基于BP网络采煤工作面瓦斯涌出量预测模型的基础上,结合样本数据,采用经验公式与MATLAB程序相结合的方法,自动选取了预测模型隐含层的最佳节点数;分别采用3种改进的BP算法对网络模型进行了训练,结果表明Levenberg-Marquardt算法的训练误差最小,收敛速度最快;通过对预测模型的训练及仿真,证明了该模型精度高,完全能够满足工程实际应用的要求。

Through the establishment of the coal mining face gas emission prediction model based on BP neural network,which combines with sample data,uses the method of the empirical formula and MATLAB program,to automatically select the best nodes of the hidden layer in prediction model.Three kinds of improved BP algorithm on the network model are used to train.The results show that the Levenberg-Marquardt algorithm has the minimum training error and the fastest convergence rate.Through training and simulating of prediction model,the high precision of the model is proved,which can fully meets the requirement of engineering application.