地铁隧道土体参数优先级表和神经网络反分析

发布时间：2018-03-03 00:05

本文选题：土体力学参数　切入点：优先级表　出处：《中国安全科学学报》2014年10期 　论文类型：期刊论文

【摘要】：为获取准确的土体力学参数,保障隧道施工过程的安全可靠,提出一种优先级表和神经网络相结合的地铁隧道土体力学参数反分析方法。首先,采用优先级表的实时调度算法确定待反演土体力学参数;然后,采用正交试验设计方法,寻求反演参数取值的最优组合,正演获取训练和测试样本;其次,用混合遗传算法优化的BP神经网络,找出待反演参数与土体变形量的非线性映射关系;最后,将工程现场土体变形信息作为输入样本,对土体力学参数进行反演分析和灵敏度检查。结果表明,土体位移模拟值与实测值最大绝对误差为0.018 mm,最大相对误差为1.070%,符合安全规范及设计要求。
[Abstract]:In order to obtain accurate soil mechanical parameters and ensure the safety and reliability of tunnel construction process, a method of back analysis of soil mechanical parameters of subway tunnel is proposed, which combines priority table and neural network. The real-time scheduling algorithm of priority table is used to determine the mechanical parameters of soil to be retrieved. Then, the orthogonal design method is used to find the optimal combination of inversion parameters to obtain training and test samples. The BP neural network optimized by hybrid genetic algorithm is used to find out the nonlinear mapping relationship between the parameters to be retrieved and the deformation of the soil. Finally, the deformation information of the soil is taken as the input sample. The results show that the maximum absolute error and the maximum relative error between the simulated and measured soil displacement values are 0.018 mm and 1.070 mm respectively, which meet the requirements of safety specification and design.
【作者单位】：大连交通大学土木与安全工程学院;大连交通大学隧道与地下结构工程技术研究中心;
【基金】：国家自然科学基金资助(U1261121/E0422) 辽宁省科技厅公益基金资助(2014004027) 中国铁路总公司科技研究开发计划重点课题(2014X012-D)
【分类号】：U231.3;U452.11

【参考文献】