基于Fisher判别理论的尾矿库安全评价模型  

Safety Evaluation Model of Tailings Based on the Theory of Fisher Discriminant

作　　者：江雅勤 田亚坤[2] 张志军[2] 汪敏[2] 张振园 JIANG Yaqin TIAN Yakun ZHANG Zhijun WANG Min ZHANG Zhenyuan(School of Mechanics ＆ Civil Engineering, China University of Mining ＆ Technology Beijing 100083)

机构地区：[1]中国矿业大学(北京)力学与建筑工程学院,北京100083 [2]南华大学核资源工程学院,湖南衡阳421001

出　　处：《工业安全与环保》2017年第1期47-50,共4页Industrial Safety and Environmental Protection

基　　金：国家自然科学基金(51004067);湖南省研究生科研创新项目(2015SCX13);南华大学"铀矿山岩土工程灾害预测与控制"校级创新团队计划项目

摘　　要：结合多元统计分析中的判别分析思想,建立尾矿库安全等级评价的Fisher判别模型。以湖南某尾矿库为例,合理选取一星期内观测的44组尾矿库数据进行判别分析,其中4组为检验样本,不参与判别函数的训练。研究结果表明,安全超高与标准判别函数之间的汇聚组间绝对相关性最大;各类协方差矩阵F检验的显著性概率Sig.<0.05,判别分析显著;训练样本回判的正确率高达95%,检验样本判别的正确率达到100%,判别结果与实际尾矿库的安全情况基本吻合,为尾矿库的安全评价提供了一条更加系统、精准的途径。Based on actual characteristics of the tailings,a Fisher discriminant analysis model is established to evaluate the safety grade of tailings. Regarding a non-ferrous tailings in Hunan Province as the case,44 sets of tailings data are selected to conduct discriminant analyses,among which four of them are testing samples and they don＇t participate in the training process. The results prove that it has the largest absolute correlation between free height and standardized discriminant function,in other words,the significant probability of F test Sig. 0. 05 and so the discriminant analysis is obvious; the discriminant rates of training samples and testing samples reach 95% and 100% respectively,according with the actual conditions,which will provide a more systematic and precise way for safety evaluation of tailings.

关 键 词：尾矿库 安全评价 FISHER判别 预测 

分 类 号：TD926.4[矿业工程—选矿]