人工鱼群投影寻踪回归在洞室岩爆预测中的应用

A Prediction Method of Rock Burst in Cavern Based on Projection Pursuit in Artificial Fish-Swarm Algorithm

通过对国内外地下工程施工中所发生岩爆的实例进行分析,探讨了岩爆发生的机理,得出了影响岩爆发生的主要因素为:围岩最大切向应力σθ、岩石单轴抗压强度σc、岩石单轴抗拉强度σt和岩石冲击倾向指数Wet。针对传统岩爆预测分析方法存在的问题,提出了基于人工鱼群算法岩爆预测的投影寻踪方法,并建立了相应的岩爆预测分析模型。选取15个工程实例作为训练样本、3个工程实例作为预测样本,以σθ/σc、σc/σt及Wet3个指标作为影响岩爆发生主要参数,进行了地下洞室的岩爆预测实例检验。研究表明,用投影寻踪回归分析法进行岩爆预测,评价过程中无需人为确定权重,避免了传统预测方法由于主观原因造成的误差,预测结果具有较高的准确率、方法简单,为岩爆的预测提供了新途径。

Rockburst mechanism was discussed by analyzing the case of rock burst occurred in underground project construction at home and abroad,the main affecting factors of rockburst were obtained：they are the maximum tangential stress of surrounding rock σθ,uniaxial compressive strength σc,uniaxial tensile strength σt,and impact tendency index Wet.A rock burst prediction method of Projection Pursuit based on Artificial Fish-Swarm Algorithm was proposed for the existing problems found in traditional forecasting methods.The prediction analysis model was established.15 projects were selected as the training samples,and 3 projects as the prediction samples,three indicators were selected as the main parameters of rock burst,and the model is applied to actual engineering examples.The prediction results are accurate,this method is simple,and some errors were eliminated because of avoiding the subjective factors for traditional methods.The projection pursuit is a new prediction method of rock burst,it has a wide application prospect.