基于PCA-SVM的岩爆预测

Rockburst prediction based on PCA-SVM

岩爆是深部地下工程开挖掘进过程中常见的地质灾害,具有显著的随机性、突发性和复杂性,随着深埋工程的增多,岩爆预测的重要性日益凸显。根据岩爆的影响因素、特点及成因,选取了围岩切向应力σθ,单轴抗压强度σc、单轴抗拉强度σt、脆性系数σc/σt、应力系数σθ/σc和冲击倾向性指数Wet等6个主要预测指标。首先用主成分分析法(PCA)对原始数据预处理,不仅消除了指标之间的相关性,而且降低了维度;然后使用粒子群算法(PSO)去优化支持向量机的惩罚c和核函数参数g,建立基于主成分分析和粒子群支向量机(PCA-PSOSVM)的岩爆预测模型,并将PCA-PSOSVM的预测结果与支持向量机(SVM)模型和人工神经网络(ANN)模型的预测结果进行比较。结果表明:PCA-PSOSVM模型的判别准确率比SVM模型和ANN模型高。

Rockburst is a common geological hazard during the excavation of deep underground engineering.It has significant randomness,suddenness and complexity.With the increase of deep buried projects,the importance of rockburst prediction is becoming increasingly important.According to the influencing factors,characteristics and causes of rockburst,the tangential stressσθof surrounding rock,uniaxial compressive strengthσc,uniaxial tensile strengthσt,brittleness coefficientσc/σt,stress coefficientσθ/σc,and impact propensity index Wet these 6 main forecast indicators.Firstly,the original data is preprocessed with principal component analysis(PCA),which not only eliminates the correlation between the indicators but also reduces the dimensions.Then,the particle swarm algorithm(PSO)is used to optimize the penalty c and kernel parameter g of the support vector machine.Principal component analysis and particle swarm branch vector machine(PCA-PSOSVM)rockburst prediction model,and the prediction results of PCA-PSOSVM are compared with the prediction results of support vector machine(SVM)model and artificial neural network(ANN)model.The results show that the accuracy of the PCA-PSOSVM model is higher than that of the SVM and ANN models.