基于PCA-BP神经网络的瓦斯涌出量预测分析

Prediction Analysis on Gas Emission Volume Based on PCA-BP Neural Network

针对目前矿井回采工作面瓦斯涌出量预测准确率低、误差率大等问题,提出基于主成分回归分析以及BP神经网络原理和方法来预测回采工作面瓦斯涌出量,依据井下现场实测的数据,通过多元统计分析软件SPSS开展相关数据处理,分析影响工作面瓦斯涌出量11个因素之间的相互关系且提取主成分,来得到BP神经网络中的输入参数,并借助PCA-BP神经网络的方法建立回采工作面瓦斯涌出量预测模型。结果证明:使用PCA-BP神经网络方法的预测值与实际值最大相对误差为2.820%,最小相对误差为2.036%,平均相对误差为2.357%。

To resolve such problems as the low prediction accuracy and high error rate of gas emission volume in the working face of coal mines,the paper proposes to predict the gas emission volume of the working face based on the principle and methods of principal components regressive analysis and BP neural network.The underground test data is,the data processed by the multivariate statistical analysis(MSA)software SPSS,the interrelationship among the 11 influence factors on the gas emission volume is analyzed and the principal components are extracted for the input parameter of BP neural network.The prediction model for the gas emission volume is also built with the method of PCA-BP neural network.The results show that the maximum relative error is 2.820%between the prediction value by PCA-BP neural network and the actual value,the minimum relative error is 2.036%,and the average relative error is 2.357%.