【摘要】 突水是影响煤矿安全生产的重要灾害之一,及时确定突水水源是进行水灾治理的关键,采用核主成分分析(Kernel Principle Component Analysis,KPCA)结合朴素贝叶斯分类方法(Naive Bayes Classifier,NBC)对煤矿突水水源进行判别.首先采用核主成分分析方法对淮南孔集煤矿78条水化学成分数据进行降维处理,选取其中5种成分数据进行分析,然后采用朴素贝叶斯方法对突水水源进行识别.通过与模糊综合评价方法进行比较,所得结果除了老空水外其它含水层的识别结果均优于或者等于模糊综合评价法所得结果,表明基于KPCA-NBC的煤矿突水水源判别方法具有较高的实用性和有效性.更多还原

【Abstract】 Water inrush is one of the main disasters in the safety production.It is the key to discriminate the water inrush source in the process of flood control. KPCA and NBC were used to identify water inrush sources based on hydrochemistry in this paper. First, 78 groups of different aquifers’ data from Kongji Coalmine, Huainan were analyzed.Then KPCA was used to reduce the dimensions of the data, and 5 composition data were analyzed by NBC.After comparing with the fuzzy comprehensive evaluation method, it showed that except the discrimination result of goaf water data the others were all better than or equal to the FCE. which showed that the discriminate model of water inrush source based on KPCA-NBC method has a good practicability and validity.更多还原