基于改进BP神经网络的矿井瓦斯浓度预测算法  

Prediction Algorithm of Gas Concentration in Coal Mine Based on Improved BP Neural Network

作　　者：姚青华[1] 邱本花[1] YAO Qing-hua eural Network QIU Ben-hua Mine Based Oil(Zhengzhou Institute of Science and Technology, Zhengzhou 450064, Chin)

机构地区：[1]郑州科技学院,郑州450064

出　　处：《煤炭技术》2017年第5期182-184,共3页Coal Technology

摘　　要：针对传统煤矿瓦斯预警的可靠性差和误差大等问题,提出了一种基于改进BP神经网络的矿井瓦斯浓度预测算法。提出的新型算法在传统BP神经网络算法的基础上,将遗传算法与BP神经网络算法有效结合,采用优化连接权方法对BP神经算法进行优化。该方法降低了瓦斯浓度预测模型的迭代次数和绝对误差。In view of problems of poor reliability and larger error fi）r traditional coal mine gas early warning, presents an algorithm of mine gas concentration prediction based on improved BP neural network. Based on the traditional BP neural network algorithm, the new algorithm combined genetic algorithm with BP neural network algorithm, and the optimization of BP neural algorithm is optimized by the method of optimizing the connection weights. This method reduces the iteration number and absolute error of the prediction model of gas concentration.

关 键 词：矿井瓦斯浓度 BP神经网络 遗传算法 

分 类 号：TD712[矿业工程—矿井通风与安全] TP277[自动化与计算机技术—检测技术与自动化装置]