基于正交设计与主成分回归的矿柱稳定性分析  

Analysis of pillar stability based on orthogonal design and principal component regression

作　　者：罗周全[1] 彭东[1] 苏汉语 王婷玉[1] 

机构地区：[1]中南大学资源与安全工程学院,湖南长沙410083

出　　处：《中国地质灾害与防治学报》2015年第4期50-55,共6页The Chinese Journal of Geological Hazard and Control

基　　金：国家自然科学基金资助(51274250)

摘　　要：以湖南某金属矿为研究背景,通过矿柱安全系数的点安全系数计算方法与正交设计实验方法获得各因子在同一水平的数据矩阵。将数据矩阵进行主成分与回归分析,得到各影响因素与稳定性系数之间的线性模型。实现了各影响因素之间以及各影响因素与矿柱稳定性之间的定量分析。另外,通过将正交极差实验方法获得的各影响因素敏感性分析结果与主成分回归分析获得的线性模型的结果进行比较,由相关系数矩阵与共线性诊断指标可知,矿柱稳定性各影响因素之间具有极强的相关性,在进行敏感度分析是应该考虑交互作用。基于正交设计的主成分回归分析方法与以前的各种分析方法相比,要更加简单、准确、全面。Taking a metal mine in Hunan as the research background,through the calculation method of pillar＇s point safety factor and orthogonal design experiment method to obtain the data matrix. The data matrix conduct principal component analysis and regression analysis,getting the linear model between the factors and the stability coefficient of each influence. To realize the quantitative analysis among the influencing factors and between the influence factors and the stability of pillar. In addition,by comparing the range of factors sensitivity analysis result that obtained from the experimental method of orthogonal range and the results of linear regression model obtained from Principal component regression analysis,we can know that there is a close correlation between the various factors influencing stability of pillar from the correlation coefficient matrix and collinearity diagnosis index and the interaction in the pillar stability influence factor sensitivity analysis should be considered. Therefore,compared with Principal component regression analysis method based on orthogonal design and the previous analysis methods,it should be more simple,accurate,comprehensive.

关 键 词：矿柱稳定性 正交设计 主成分回归分 

分 类 号：TD353[矿业工程—矿井建设]