中文摘要：

矿井突水严重威胁矿区安全生产,快速准确地判别突水水源,对有效防治矿井突水灾害有重大意义。根据鹤壁矿区6个主要含水层水化学成分的差异性,选取了Ca2＋、Mg2＋、Na＋＋K＋、CO2-3、HCO-3、Cl-、SO2-47个指标作为突水水源判别的评价因子,并选用鹤壁矿区294个水样作为学习样本,以Matlab为平台,先采用主成分分析法提炼出三个主成分Z1、Z2和Z3。然后把这3种主成分的值作为贝叶斯判别指标,建立了鹤壁矿区突水水源判别模型。利用该模型对鹤壁矿区随机选取22个突水水样进行水源判别,判别正确19个,错误3个,精度达到86%。判别错误的原因,主要是因为有的含水层水化学类型相似或者几个含水层之间存在一定的水力联系。研究结果表明,基于主成分分析的矿井突水水源Bayes判别模型能够满足矿井生产要求,可为防治水工作提供决策依据。

英文摘要：

Mine water inrush is a serious threat to the safety of mine work,so determining the source of waterinrush quickly and accurately makes a great significance to the effective prevention of such disasters.According to the hydrochemistry diversity in six aquifers of Hebi coal mine,we chose seven indexes（Ca2 ＋,Mg2 ＋,Na＋＋ K＋,CO2-3,HCO-3,Cl-,SO2-4） as the evaluation factors and used 294 water samples as training representatives in water bursting source determination.With Matlab platform,three principal components Z1,Z2,Z3 were extracted using principal component analysis.Then taking them as factors of Bayesian discrimination we established a models for determining the source of water bursting in Hebi mines.Using this model,22 randomly selected samples of water inrush were discriminated.The result shows that 19 were correct,and only 3 were inaccurate,which means the accuracy reaches 86%.There are 3 failed discriminations mainly because some aquifers have similar types of hydrochemistry.Also the existence of hydro relationships among several aquifers is another possible reason.Therefore,our Bayesian discriminant model for determining the source of mine water inrush which based on principal component analysis can satisfy the requirements of mine safety,and thus provides a decision-making basis for prevention and controlling of water inrush.