基于MPSO-RBF的瓦斯涌出量预测研究  

Study on Gas Forecast Based on MPSO Algorithm for RBF Neural Network

作　　者：付华[1] 舒丹丹[1] 荆晓亮[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：《计算机测量与控制》2012年第10期2625-2627,共3页Computer Measurement &Control

基　　金：国家自然科学基金(70971059);辽宁省科技攻关项目(2011229011)

摘　　要：我国煤矿的重大灾害事故中70%以上是瓦斯事故,煤矿瓦斯是影响煤矿安全生产的重要因素;针对瓦斯煤尘爆炸和煤与瓦斯突出给煤炭矿山带来的危害极大的问题,引入了基于改进PSO算法的RBF神经网络的混合优化算法(MPSO-RBF算法),即将PSO算法的全局搜索能力和RBF神经网络局部优化相结合,并建立了瓦斯预测模型;仿真与实际数据验证表明,优化算法所求的最优解具有良好的收敛能力,瓦斯涌出量的预测结果与实际值的误差在+1.44%至-0.63%之间,改进的粒子群算法优化的RBF神经网络对瓦斯涌出量预测能达到良好的效果。Our country coal mine disaster in 70% above are to gas accident, coal mine gas is the important factor affecting coal mine safety production. In view of the harmful effects of gas and coal dust explosion and gas outburst in coal mine, a kind of hybrid optimization algorithm based on the improved PSO algorithm and RBF neural network neural network technology （MPSO--RBF） are introduced to solve the prediction of gas emission, it combined the PSO algorithm global search ability with local optimization of RBF neural network to establish the prediction models of gas emission. Simulation and actual data show that the optimal solution of optimization algorithm has good convergence ability, the error of gas emission prediction result and the actual value ranges from ＋1.44% to -0. 63%, improved particle swarm optimization algorithm for RBF neural network to the prediction of gas emission can achieve good results.

关 键 词：RBF神经网络 改进的PSO算法 瓦斯预测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]