LS-SVM变形预测在雪峰山1号隧道的运用

以沪昆高铁雪峰山1号隧道出口段为例,借助最小二乘支持向量机的反分析技术,探索该方法在大断面隧道位移反分析预测中的应用。首先雪峰山隧道的勘察设计相关资料和相关规范,通过正交设计方法获得多组参数组合,用来做数值模型的参数输入。再通过有限差分方法进行数值分析获得各组参数对应的位移信息,从中提取支持向量机的训练样本和测试样本,获得支持本隧道的向量机模型。最后以实际的拱顶监测位移值带入模型获得反演计算的围岩相关参数,并且通过数值模拟进行验证,取得了很好的效果,获得了与实际近似的数值模型,进而通过最小二乘支持向量机进行隧道拱顶变形预测分析,指导施工。

Deformation prediction of the 1th Xuefengshan tunnel based on the LS-SVM

Due to the fact that the conventional displacement back analysis methods have disadvantages of complicated computing,hard to master,and not high precision solutions,someone applied Least Squares Support Vector Machine to solve the displacement prediction problem.In this paper,the new method is applied in the Xuefengshan tunnel.The main point of the method is to use LS-SVM and FDM program FLAC3D to get train sample and test sample,and then through the parameter optimization method to get the best SVM model.in the forward process and to map the nonlinear relation of samples and reduce computed data.Finally,the back analyses result is used to FLAC model of the tunnel,the simulation results show that the LS-SVM method is not only feasible but effective in the field of tunnel engineering.