基于Fisher判别分析理论的地下开采安全评价模型  

Safety Assessment Model of Underground Mining based on the Fisher Discriminant Analysis

作　　者：史秀志[1] 崔松[1] 黄敏[1] 邱贤阳[1] 孙磊[1] 周健[1] 

机构地区：[1]中南大学

出　　处：《金属矿山》2010年第8期152-155,182,共5页Metal Mine

基　　金：“十一五”国家科技支撑计划项目(编号:2006BAB02A02);中南大学学位论文创新资助项目(编号:2009ssxt230)

摘　　要：矿山安全是矿山企业稳定、持续、高速发展的根本保证。应用Fisher判别分析理论并结合矿山安全生产的标准规范,根据矿山人、机、环境、管理4个因素的实际特点,综合考虑到地下矿山的开采工艺和指标数据的可靠性,选取23项矿山安全评价因素作为评价影响因子,建立矿山地下开采安全评价的F isher判别分析模型（FDA）。对18组地下矿山的特征参数值作为学习样本进行训练和检验,回判估计的误判率为0。利用该模型对另外4组实测数据作为预测样本进行测试,预测结果与实际情况吻合较好。研究结果表明,该方法回判估计的误判率低,判别性能良好,对地下矿山安全管理现状及水平评价提供了可操作的方法,进而为矿山有关部门提供了科学安全管理的依据。Mine safety is the ultimate guarantee to keep mining corporation stable,persistent and fast developing.Based on the Fisher discriminant theory,and combining with the standard of mine safety production,according to actual four mine characteristics of person,machine,environment,management,the main factors of index system of mine safety appraisal,such as mine underground mining processes and indicators of the reliability,twenty-three factors influencing of underground mine were defined as the criterion indices for safety evaluation in the proposed model,and the Fisher discriminant analysis model for safety assessment of underground mining was established.Taking 18 sets of parameter data of underground mine as training and testing samples,the ratio of mis-discrimination is 0.Another 4 groups of measured data are tested as forecast samples,the predicted results are consistent with the measured ones.And it is indicated that this method have a feature of low mis-discrimination ratio and good discrimination performance.So it is a new approach to provide advisable methods for the assessment of underground mine safety management status,and basis for scientific and safe management of the relative departments.

关 键 词：地下矿山 矿山安全 安全评价 FISHER判别分析 预测 

分 类 号：TD163.1[矿业工程—矿山地质测量]