Bayes discriminative analysis was adopted to predict rockburst grades, and based on 108 rockburst cas-es, a Bayes discriminative model was built with six basic indicators such as maximum tangential stress of the sur-rounding rockσθ, uniaxial compressive strength σc , uniaxial tension strength σt , stress coefficient σθ/σc , brittle-ness coefficient σc/σt and elastic energy indicator Wet. In light of the Bayes discriminative model, the concept of odds ratio was proposed and used in the analysis of relationships between rockburst indicators and rockburst grades, and the analysis of the sensitivity of rockburst grades to indicators, then the reasonableness of the Bayes discrimina-tive model with different indicator combinations was evaluated. The results show that the Bayes discriminative mod-el with the combination of indicatorsσθ/σc and Wet was the most reasonable with high prediction effects, and the ob-tained expression for rockburst grade probability can be a reference for predicting rockburst grades in underground engineering.%文章将Bayes判别法应用到岩爆等级预测分析中,选择了围岩的最大切应力σθ、单轴抗压强度σc、单轴抗拉强度σt、应力系数σθ/σc、脆性系数σc/σt和弹性能量指数Wet作为基本指标,以108组岩爆实例作为研究对象建立Bayes判别模型.针对Bayes判别模型,提出了优势比概念并将其用于分析岩爆指标与岩爆等级之间的趋势性关系以及岩爆等级对指标的敏感性,在此基础上评价不同指标组合下的Bayes判别模型是否合理.分析结果表明:σθ/σc、Wet指标组合建立的Bayes判别模型是最合理的模型,其预测效果好,所得出的岩爆等级概率表达式可用于地下工程的岩爆等级预测.
展开▼
