采煤工作面瓦斯涌出量LMD-BP神经网络建模预测研究  

Study on LMD-BP Neural Network Modeling and Prediction of Gas Emission from Coal Mining Face

作　　者：蔡振禹[1] 邵永华[1] 

机构地区：[1]河北工程大学经济管理学院,河北邯郸056038

出　　处：《煤炭工程》2014年第1期108-111,共4页Coal Engineering

基　　金："十一五"国家科技支撑计划"高原矿山采动地质灾害监控技术研究"(2007BAB18B01);华北科技学院科研基金项目(2011B046)

摘　　要：文章阐述了局部均值分解(LMD)非平稳信号处理方法,并通过对收集得到采煤工作面瓦斯涌出量数据进行LMD分解,得到多个PF分量。然后再用改进的神经网络方法对其分别进行预测,再把不同预测结果进行叠加重构合成,进而获得瓦斯涌出量预测值。通过对瓦斯实际监测数据进行分析,可以得出,此方法预测效果比常规的神经网络方法预测精度更高,其预测结果与实际监测结果相比具有极高的一致性。The paper explained the treatment method of the unstable signal with the local mean decomposition （LMD）. With the local mean decomposition （LMD） conducted on the gas emission data obtained from the coal mining face, several PF components were obtained and then the improved neural network method was applied to the prediction on the gas emission data individually. Superposition reconstruction composition was conducted on the different predicted results and therefore the predicted value of the gas emission quantity was obtained. With the analysis conducted on the actual monitoring and measuring data of the gas, the predicted effect of the method would have a higher predicted accuracy than the conventional neural network method and the predicted results would be highly consistent with the actual monitoring andmeasuring results.

关 键 词：采煤工作面 瓦斯涌出量预测 LMD 神经网络 

分 类 号：TD712.5[矿业工程—矿井通风与安全]