基于决策树模型的岩爆烈度预测  

Prediction of rockburst intensity based on decision tree model

作　　者：陈顺满[1,2] 吴爱祥[1,2] 王贻明[1,2] 许梦国[3] 

机构地区：[1]北京科技大学金属矿山高效开采与安全教育部重点实验室,北京100083 [2]北京科技大学土木与环境工程学院,北京100083 [3]武汉科技大学资源与环境工程学院,湖北武汉430081

出　　处：《武汉科技大学学报》2016年第3期195-199,共5页Journal of Wuhan University of Science and Technology

基　　金："十二五"国家科技支撑计划项目(2012BAB08B02);教育部长江学者和创新团队发展计划项目(IRT0950)

摘　　要：综合分析了岩爆的主要影响因素,选取岩石应力系数σ_θ/σ_c、岩石脆性系数σ_c/σ_t和弹性能量指数Wet作为评价指标,采用决策树方法进行岩爆烈度预测。以国内外比较典型的32个岩石地下工程实例作为基础样本数据,结合ID3算法建立了判定岩爆烈度的决策树模型。再选取15个工程实例,运用建立好的决策树模型对其岩爆烈度进行分级,并与实际岩爆等级以及采用模糊灰关联法、距离判别法、ν-SVR算法的评判结果进行对比。结果表明,本文提出的决策树方法具有计算简单、准确可靠、预测效率高的特点。On the basis of analyzing the major influencing factors in rockburst,decision tree method is applied to predict the rockburst intensity with rock stress coefficientσθ/σc,rock brittleness coefficientσc/σtand elastic energy index Wetas evaluation indicators.Thirty-two typical domestic and overseas underground rock projects are chosen as the basic sample data,and the decision tree model is built using ID3 algorithm.Rockburst intensities of another fifteen projects are graded by this model.The results are compared with the actual rockburst grades as well as the evaluation results by fuzzy grey incidence theory,distance discriminant analysis andν-SVR algorithm,which indicates that the proposed decision tree method is simple and reliable,and has higher accuracy and efficiency.

关 键 词：岩爆 烈度 决策树 预测 地下工程 

分 类 号：TD32[矿业工程—矿井建设] TU452[建筑科学—岩土工程]