基于Fisher判别分析和质心距评价法的矿井水源判别

发布时间：2018-02-21 19:24

本文关键词： 矿井突水水化学分析 Fisher判别分析法质心距评价法　出处：《煤田地质与勘探》2017年01期 　论文类型：期刊论文

【摘要】：矿井突水是影响煤矿安全生产的灾害之一,快速准确地判别突水水源对于水害治理至关重要。Fisher判别分析法在判别混合水时,只能将此确定为最相像的某一分组,从而发生误判,因此在Fisher判别分析理论的基础上引入了质心距评价法,剔除可能为混合水样的样本数据,可以提高判别结果的准确率。为此选取13种水化学离子作判别因子,30个有代表性的水样作为训练样本建立判别函数模型,应用判别函数模型及质心距辅助评价指标对6个取自特定含水层的单一水样和4个取自水仓的混合水样验证分析,引入质心距评价分析后判别准确率从60%提高到83.3%。研究结果表明,在Fisher判别分析后进行质心距评价分析可提高判别结果的准确度。
[Abstract]:Mine water inrush is one of the disasters that affect the safety of coal mine production. It is very important to judge water inrush water source quickly and accurately. Fisher discriminant analysis method can only determine this as one of the most similar grouping when judging mixed water, thus misjudgment occurs. Therefore, based on the Fisher discriminant analysis theory, the centroid distance evaluation method is introduced to eliminate the sample data that may be mixed water samples. For this reason, 13 kinds of hydrochemical ions are selected as discriminant factors, and 30 representative water samples are used as training samples to establish discriminant function model. The discriminant function model and centroid distance index were used to verify and analyze six single water samples from a specific aquifer and four mixed water samples from a silo. The accuracy of centroid distance evaluation is improved from 60% to 83.3. The results show that the centroid distance analysis after Fisher discriminant analysis can improve the accuracy of the discriminant results.
【作者单位】：山东科技大学地球科学与工程学院;山东科技大学山东省沉积成矿作用与沉积矿产重点实验室;山东省地矿工程勘察院;
【基金】：国家自然科学基金项目(41372290,51509148) 教育部博士点基金项目(20133718110003)~~
【分类号】：TD745

【相似文献】