深基坑工程的三维数值模拟及敏感性分析

Parameters' sensitivity analysis of deep foundation pit excavation based on RBF neural network

为研究深基坑开挖过程对支护结构的变形以及周边环境的影响,通过FLAC3D数值分析软件和RBF神经网络,建立了深基坑开挖过程的三维数值模拟和基坑位移与参数之间的映射关系,分析了基坑开挖过程中不同内支撑的施加对基坑的变形和周边环境的影响,以及基坑参数的敏感性。分析结果表明:基坑开挖过程的三维数值模拟分析可以较真实地反映土层条件及施工工况;内支撑的施加方式不同,将引起基坑不同的变形规律,对周边环境的影响也不同;基坑在稳定情况下的变形主要与支护结构的形式有关;基坑开挖过程中,不同土层的各参数敏感性系数不同,且随着基坑的开挖有减小的趋势。

In order to study the deformation of retaining structure and surrounding environment influenced by deep foundation pit excavation, a three-dimensional numerical simulation for deep foundation pit excavation and a mapping relation between the displacement of foundation pit and the parameters are established with three-dimensional numerical analysis software FLAC3D and RBF neural network, respectively. The influence of different supports during foundation pit excavation on the de- formation of the pit and surrounding environment is analyzed, as well as the sensitivity of the param- eters of the pit. The analysis results show that the 3 D numerical simulation of the excavation reflects soil conditions and construction conditions; that different support ways will lead to different deforma- tion of foundation pit and different impact on the surrounding environment; that the deformation of foundation pit under stable condition is mainly related to the supporting structure; that the sensitivity coefficients of parameters of different soils are different, and the coefficients decrease with the excavation.