基于KPCA-IPSO-SVM模型的地下洞室围岩变形预测研究  

Prediction of Underground Cavern Surrounding Rock Deformation Based on KPCA-IPSO-SVM Model

作　　者：向文俊 陈新[1] 张榴梅 XIANG Wenjun;CHEN Xin;ZHANG Liumei(College of Water Resource and Hydropower,Sichuan University,Chengdu 610065,Sichuan,China)

机构地区：[1]四川大学水利水电学院,四川成都610065

出　　处：《水力发电》2023年第1期58-62,共5页Water Power

基　　金：国家重点研发计划(2018YFC0406803)。

摘　　要：为了提高地下洞室施工期围岩变形预测的准确性和运算速度,研究了KPCA-IPSO-SVM预测模型在地下洞室施工期围岩变形预测中的运用,并对粒子群算法(PSO)进行改进。核主成分分析法(KPCA)可将输入参数映射到高维空间处理,依据累计贡献度提取模型输入的非线性主元,降低输入参数的维度;改进粒子群算法(IPSO)对支持向量机(SVM)的核函数参数和惩罚因子进行优化。以向家坝水电站地下洞室施工期围岩变形监测数据作为研究对象,对该模型进行训练和预测,并将KPCA-IPSO-SVM模型与单一SVM、PSO-SVM、IPSO-SVM预测模型进行比较。结果表明,KPCA-IPSO-SVM模型具有更高的预测精度和运算速度。To improve the accuracy and computing speed of surrounding rock deformation prediction during underground cavern construction period, the application of KPCA-IPSO-SVM prediction model in the prediction of surrounding rock deformation is studied, and the particle swarm algorithm(PSO) is improved. The kernel principal component analysis(KPCA) could map input parameters to high-dimensional space processing, extract the nonlinear pivot element of model input based on the cumulative contribution degree, and reduce the dimension of input parameters. The improved particle swarm algorithm(IPSO) is used to optimize the kernel function parameter and penalty factor of support vector machine(SVM). Taking the monitoring data of surrounding rock deformation of Xiangjiaba underground cavern construction as the research object, the model is trained and predicted, and the KPCA-IPSO-SVM model is compared with the SVM, PSO-SVM and IPSO-SVM prediction models. The results show that the KPCA-IPSO-SVM model has higher prediction accuracy and computing speed.

关 键 词：核主成分分析 改进粒子群算法 支持向量机 地下洞室 围岩变形 变形预测 

分 类 号：TV223.1[水利工程—水工结构工程]