矿井突水水源的PCA-GRA判别模型研究  

Research on PCA-GRA Discrimination Model of Water Inrush Source in Mines

作　　者：杨中元 丁风帆 黄平华[1] YANG Zhong-yuan;DING Feng-fan;HUANG Ping-hua(School of Resources and Environment Engineering,Henan Polytechnic University,Jiaozuo 454000,China)

机构地区：[1]河南理工大学资源与环境学院

出　　处：《煤炭技术》2019年第12期84-87,共4页Coal Technology

基　　金：中国博士后科学基金（2017M612395）;河南省高校科技创新团队支持计划（15IRTSTHN027）

摘　　要：为了快速准确地识别突水水源,选取Ca2+、Na++K+、Mg2+、HCO3-、SO42-、Cl-共6个判别指标,结合主成分分析理论和灰色关联理论,建立了PCA-GRA突水水源判别模型。先采用主成分分析法对38个水样的水质指标进行降维处理,然后利用灰色关联分析法根据关联度对未知水样进行判别。将该模型应用于11组待判水样的判别,判别结果正确率为100%。而PCA-Bayes判别模型识别精度是81.8%。因此,PCA-GRA突水水源判别模型能够有效地提高识别精度,迅速准确地判断突水水源,为矿井安全生产提供保障。In order to identify the water source of water inrush quickly and accurately,six discriminative indexes of Ca2+, Na++K+, Mg2+, HCO3-, SO42- and Cl- were selected. The PCA-GRA water inrush water source discrimination model was built by principal component analysis theory and grey relational theory. Principal component analysis was used for dimensionality reduction of water quality indicators of 38 water samples, and then the gray correlation analysis method was used to discriminate the unknown water samples according to the correlation degree. The constructed model is used to prognosis of 11 samples, and the accuracy of discriminant result is 100%. On the contrary, the recognition accuracy of the PCA-Bayes discriminant model of water inrush is only 81.8%.This reveals that the constructed PCA-GRA discrimination model of the water inrush source in mines can increase the recognition accuracy effectively, thus providing guarantee to the safety production of mines.

关 键 词：矿井突水 水源识别 主成分分析 灰色关联理论 

分 类 号：TD745[矿业工程—矿井通风与安全]