基于RBF核的多分类SVM围岩变形易发性评价模型研究  

Study on the Multi-Classification SVM Model of Surrounding Rock Deformation Susceptibility Evaluation Based on RBF Kernel

作　　者：左小伟 ZUO Xiaowei(China Railway 11th Bureau Group 4th Engineering Co.Ltd.,Wuhan Hubei 430073,China)

机构地区：[1]中铁十一局集团第四工程有限公司,湖北武汉430073

出　　处：《铁道建筑技术》2024年第8期158-162,共5页Railway Construction Technology

基　　金：湖北省技术创新专项(2022BEC002)。

摘　　要：在隧道、地铁等工程建设中,随着开挖深度逐步增加,地应力随之增加,围岩变形产生坍塌、岩爆风险显著提升,给工程建设和安全带来巨大挑战。采用支持少量样本且可以有效防止“过拟合”的支持向量机(SVM)用于隧道围岩变形易发性的多分类与评价中,用少量样本分析工程区域的围岩概况,解决易发性分类问题,不但可降低人力物力成本,而且使评价工作更加智能简便。以机场SG3标隧道施工为工程背景,结合研究区域实际情况对优化RBF函数核的SVM模型进行模型评价。结果表明:SVM模型具有高预测精度且适用于围岩变形易发性评价研究。In the construction of tunnel,subway and other projects,with the gradual increase of excavation depth,the ground stress gradually increases,and the risk of surrounding rock deformation resulting in collapse and rock burst significantly increases,which poses huge challenges to engineering construction and safety.The support vector machine(SVM),which supports a small number of samples and can effectively prevent“overfitting”,is used in the multi-classification and evaluation of the susceptibility of tunnel surrounding rock deformation.Using a small number of samples to analyze the general situation of surrounding rock in the engineering area and solve the problem of susceptibility classification,not only reduces the cost of manpower and material resources,but also makes the evaluation more intelligent and simple.In this paper,based on the engineering background of the construction of a tunnel in the SG3 standard space of an airport,the SVM model for the establishment and optimization of RBF core was evaluated by the actual situation of the study area and the results show that the SVM model has high prediction accuracy and is suitable for the susceptibility evaluation of surrounding rock deformation.

关 键 词：围岩变形 易发性评价模型 支持向量机 径向基核函数 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] U451.2[自动化与计算机技术—控制科学与工程]