基于GP-DE的拱盖法车站位移-应力联合反分析研究

A Displacements-stress Jointed Back Analysis Based on GP-DE for Subway Station Constructed by Arch-cover Method

在上软下硬的地质条件下采用拱盖法进行大跨度暗挖车站施工时,由于土岩复合和跨度较大的特点,使得车站的围岩参数难以准确获取。针对这一问题,提出了一种适用于拱盖法车站的智能反分析方法:高斯过程-差异进化协同优化算法(GP-DE算法)。以大连地铁五号线石葵路车站为工程背景,在车站典型断面处安装传感器,采集施工中应力位移数据;同时建立FLAC^(3D)有限元模型,设计25组正交方案进行计算,并得出结果,通过敏感性分析找出关键参数;联合传感器实测数据与正交方案所得监测数据,运用该算法进行位移应力联合反分析,并使用FLAC^(3D)正算验证反演参数的精度。结果表明,反分析所得参数代入有限元模型正算结果与实际值较为吻合,反分析结果精度高。将反演结果运用到车站设计中,反演精度满足工程需要。

Due to the characteristics of soil-rock composite and large-span,the surrounding rock parameters of the station are difficult to obtain accurately under the geological condition of upper soft and lower hard,arch cover method is used to carry out the construction of large-span underground excavation station.To solve this problem,an intelligent back analysis methodGauss Process Differential Evolutionary cooperative optimization(GP-DE)algorithm,is proposed.Taking Shikui road station of Dalian metro line 5 as the engineering background,sensors are installed at the typical section of the station to collect the stress anddisplacement data during construction.At the same time,the FLAC^(3D) numerical model was established,25 orthogonal schemes were designed for calculation and the results were obtained,and the key parameters were found through sensitivity analysis.Combining the measured data of the sensors with the monitoring data obtained from the orthogonal schemes,the algorithm was used to carry out the jointed back analysis of displacement-stress,and FLAC^(3D) positive calculation was used to verify the accuracy of the inversion parameters.The results show that the parameters obtained from the reverse analysis are in good agreement with the actual values,and the accuracy of the back analysis is high.The inversion results are applied to the station design,and the inversion precision meets the engineering needs.