矿井瓦斯爆炸超压值分布的神经网络预测方法  

Overpressure Distribution Prediction of Mine Gas Explosion Based on Artificial Neural Networks

作　　者：李江涛[1,2,3] 

机构地区：[1]煤炭科学研究总院沈阳研究院,辽宁沈阳110016 [2]煤矿安全技术国家重点实验室,辽宁抚顺113122 [3]东北大学资源与土木工程学院,辽宁沈阳110819

出　　处：《煤矿安全》2013年第2期157-160,共4页Safety in Coal Mines

基　　金：国家重点基础研究发展规划(973计划)资助项目(2010CB735500)

摘　　要：瓦斯灾害是影响煤矿安全的重要问题之一,而爆炸后的主要危害之一是冲击波的伤害。而且,煤矿防隔爆措施是否能起到有效作用也依赖于冲击波超压值的测量和预测。在前人实验分析的基础上,应用神经网络理论,分别用BP神经网络和RBF神经网络对瓦斯爆炸后的冲击波超压值和测点之间的关系进行了预测。结果表明,BP神经网络的预测误差最小,应用神经网络进行预测可以明显的减小预测的误差,适合煤矿企业实际应用。Gas disaster is one of important problems concerning the safety of mine. The shock wave after explosion is the main damage for human. Moreover, whether the anti - explosion - proof measures play an effective role is depended on the measurements and fore-casts of blast overpressure. Based on previous experimental analysis, the paper uses BP and RBF ANN to predict the relationship of gas explosion overpressure between point measurement and overpressure value. The results show that application of BP ANN prediction can significantly reduce the forecast error, it is suitable for mining enterprises to operate.

关 键 词：瓦斯爆炸 超压 神经网络 预测 BP RBF 

分 类 号：TD712.7[矿业工程—矿井通风与安全]