摘要
针对BP神经网络在煤层瓦斯含量预测中的局限性,如收敛速度慢和可靠性差等缺点,根据煤层瓦斯含量与其影响因素之间相互作用和耦合的特点,建立了粒子群算法和BP神经网络相结合的煤层瓦斯预测模型.在采用BP网络对煤层瓦斯含量进行预测的基础上,采用粒子群算法优化隐含层神经元个数和网络中的连接权值,并根据现场的实测数据,提出了粒子群神经网络训练和检验样本集,对预测模型进行训练和检验.仿真结果表明,该预测模型加快了网络收敛的速度,克服了易陷入局部极小的问题,具有可靠性强和预测精度高等特点.
According to the coupling interaction characteristics between the coal seam gas content and its influence factors,BP neural network prediction method based on particle swarm optimization algorithm is proposed for the limitations of BP neural network in coal seam gas content prediction such as slow convergence rate and poor reliability,which organically combines the BP neural network and particle swarm optimization algorithm.On the basis of predicting coal seam gas content with BP neural network,particle swarm optimization algorithm is applied to optimize the weights and thresholds of the network.According to the measured data,a particle swarm neural network training and testing samples are established to train and check the prediction model.The simulation results indicate that the adopted prediction model has the characteristic of high reliability and prediction precision,which speeds up the network convergence rate and overcomes the local optimum of a BP neural network.
出处
《河南理工大学学报（自然科学版）》
CAS
北大核心
2014年第6期724-727,共4页
Journal of Henan Polytechnic University（Natural Science）
基金
河南省科技计划项目(102102210203)
关键词
瓦斯含量
粒子群
神经网络
预测模型
gas content
particle swarm optimization
neural network
prediction model
