BP神经网络在潘三煤矿突水水源判别中的应用  

Water-inrush source discrimination with BP neural network in Pansan Mine

作　　者：祝翠[1] 钱家忠[1] 周小平[1] 马雷[1] 

机构地区：[1]合肥工业大学资源与环境工程学院,合肥230009

出　　处：《安徽建筑工业学院学报（自然科学版）》2010年第5期35-38,共4页Journal of Anhui Institute of Architecture(Natural Science)

基　　金：国家自然科学基金(No:40872166);教育部新世纪优秀人才支持计划项目(NCET-06-0541)

摘　　要：矿井突水是影响煤矿安全生产的五大灾害之一,快速准确地判别突水水源是煤矿安全生产的保障,是矿井防治水工作的前提。本文从矿井突水的机理出发,以潘三矿为例,利用BP算法对训练样本进行学习,构建了BP神经网络判别模型,根据已训练好的神经网络对样本进行了判别。最后与模糊综合评判进行了比较,结果表明,运用此方法对矿井水源进行判别,能取得较好的效果。该研究使定性预测转为了定量预测,从而提高了突水预测的科学性和准确性。Water inrush is one of the top five disasters which affect coal mine safety production,discriminate water-inrush source exactly and quickly can ensure the mine safety production,and is the prerequisite of flood control of mining.In this paper,starting from the mechanism of water inrush,take Pansan mine as example,using BP algorithm to train samples,and constructing BP neural network identification model.Based on this model the unknown samples were discriminated,finally compared with the fuzzy comprehensive discrimination.The results show that using this method can achieve better results.The study converts the qualitative prediction into the dosing prediction.Accordingly,it can raise prediction scientificity and accuracy of the water-inrush source.

关 键 词：BP神经网络 判别模型 突水水源 潘三煤矿 

分 类 号：TD745.21[矿业工程—矿井通风与安全]