基于KPCA-ISSA-SVR的盾构施工诱导地面沉降预测模型研究  

Prediction Model of Shield Construction-Induced Ground Settlement Based on KPCAISSA-SVR

作　　者：刘育林 周爱红 姜礼涛 袁颖 LIU Yu-lin;ZHOU Ai-hong;JIANG Li-tao;YUAN Ying(Hebei GEO University,Shijiazhuang 050031,China;Hebei Technology Innovation Center for Intelligent Development and Control of Underground Built Environment,Shijiazhuang 050031,China)

机构地区：[1]河北地质大学城市地质与工程学院,河北石家庄050031 [2]河北省地下人工环境智慧开发与管控技术创新中心,河北石家庄050031

出　　处：《河北地质大学学报》2022年第5期42-49,共8页Journal of Hebei Geo University

基　　金：国家自然科学基金资助项目(41807231);河北省自然科学基金项目(D2019403182);河北省教育厅青年基金项目(QN2019196);河北地质大学科技创新团队项目(KJCXTD-2021-08)。

摘　　要：为了准确预测盾构施工诱发的地面沉降量,论文提出了核主成分—多策略融合的改进麻雀搜索算法优化支持向量回归机(KPCA-ISSA-SVR)预测模型。以73组地面沉降实例为总体样本集构建训练及测试样本,利用核主成分分析对影响地面沉降的地质因素及施工因素进行特征提取的基础上,采用ISSA算法优化参数C和g,建立KPCA-ISSA-SVR地面沉降量预测模型,并与核主成分-Tent混沌映射改进麻雀搜索算法优化支持向量回归机(KPCA-TentSSA-SVR)、核主成分—麻雀搜索算法优化支持向量回归机(KPCA-SSA-SVR)、麻雀搜索算法优化支持向量回归机(ISSA-SVR)模型进行对比。结果表明:KPCA能够剔除冗余信息,降低模型复杂度;ISSA全局寻优及局部探索能力强,能高效准确地确定模型参数;KPCA-ISSA-SVR预测精度更高,稳定性更强。In order to accurately predict the amount of ground settlement induced by shield construction,an improved sparrow search algorithm optimized support vector regression machine(KPCA-ISSA-SVR)prediction model with kernel principal component-multi-strategy fusion is proposed in this paper.Based on 73 sets of ground settlement instances as the overall sample set to construct training and testing samples,and using kernel principal component analysis to extract features of geological and construction factors affecting ground settlement,the ISSA algorithm is used to optimize parameters C and g to build the KPCA-ISSASVR ground settlement prediction model,and it is combined with the kernel principal component-Tent chaotic mapping improved sparrow search algorithm to optimize the support vector regression machine(KPCA-TentSSA-SVR),kernel principal componentsparrow search algorithm optimized support vector regression machine(KPCA-SSA-SVR),and sparrow search algorithm optimized support vector regression machine(ISSA-SVR)models for comparison.The results show that KPCA can eliminate redundant information and reduce model complexity;ISSA has strong global search and local exploration ability and can determine model parameters efficiently and accurately;KPCA-ISSA-SVR has higher prediction accuracy and better stability.

关 键 词：盾构施工 地面沉降 核主成分 多策略融合改进麻雀搜索算法 支持向量回归机 

分 类 号：P542[天文地球—构造地质学]