基于小波神经网络的煤岩破裂电磁辐射预测模型研究  

Research on Electromagnetic Emission Prediction Model in Fracture of Coal and Rock Based on Wavelet Neural Network

作　　者：杨桢[1] 付华[1] 李鑫[1] 

机构地区：[1]辽宁工程技术大学电气与控制工程学院,辽宁葫芦岛125105

出　　处：《煤矿机电》2012年第5期19-21,共3页Colliery Mechanical & Electrical Technology

基　　金：国家自然科学基金项目(50874059);辽宁省教育厅基金项目(L2010172);辽宁省科学技术计划项目(2011229011)

摘　　要：针对电磁辐射信号的特点,提出用小波神经网络建立煤岩破裂电磁辐射预测模型,用最大最小蚁群算法来训练网络初始参数,改善网络性能,并引入扰动因子和惩罚因子来解决算法的局部收敛和收敛速度慢的问题。以开滦煤样为例,应用结果表明,该模型精度高、响应快、实时性较好,具有良好的应用前景。Aiming at the features of the electromagnetic emission signal, an electromagnetic emission prediction model in fracture of coal and rock is established by using wavelet neural network. The max-rain ant colony algorithm is used to train the initial parameters of network and to improve the performance of network, and by introducing the disturbance factor and punishment factor to solve local and slow convergence problems of the algorithm. With coal samples of Kailuan as example, the application results show that the model has high accuracy, fast response, better real-timelines and good application prospect.

关 键 词：小波神经网络 电磁辐射 最大最小蚁群算法 预测模型 

分 类 号：TD326[矿业工程—矿井建设]