Inverse Analysis Of Coefficient Of Subgrade Reaction And Its Application In Engineering

With the vigorous development of urban rail transit in China,the trend of subway construction is increasing.The rock and soil environment in the underground pit is complex and changeable,and the determination of rock and soil parameters is crucial.The inversion analysis method is another way to become a new idea to solve this problem.This paper relies on the deep foundation pit project of Dongli Station of Line 3 of Shijiazhuang Metro,and uses the numerical simulation software to analyze and study the deformation of the foundation pit during the foundation pit excavation and support.Based on the on-site monitoring data,a BP neural network was used to establish a correlation model to invert the elastic modulus and the coefficient of subgrade reaction of a typical soil in Shijiazhuang.The main content of the paper is reflected in the following aspects.(1)Summarize and analyze the determination method of the coefficient of subgrade reaction.It focuses on the analysis of the on-site in-situ test of the coefficient of subgrade reaction,the indoor test method,and the empirical formula method of the coefficient of subgrade reaction.The statistical analysis of the values of the coefficient of subgrade reaction of different soil layers in the Shijiazhuang area was conducted and the relevant formulas were fitted to guide similar projects.(2)Relying on the actual project of the well drainage section of the deep foundation pit at Dongli Station of Line 3 of Shijiazhuang Metro,the outline of the project is outlined,and the monitoring plan,monitoring requirements and contents of the pit drainage section of the foundation pit are introduced.Analyze the monitoring data obtained during the actual construction process and discuss the deformation law of the foundation pit.And provide data samples for inversion analysis of subsequent parameters.(3)Briefly introduce the constitutive model of FLAC3 D numerical simulation software,and the establishment method of numerical model,boundary condition setting and parameter selection.The numerical simulation of the excavation and support of deep drainage pit excavation in Dongli Station is carried out,and the simulation results are analyzed and researched.It also provides the corresponding analysis carrier for the inversion analysis.(4)Based on the basic principle of neural network,the MATLAB implementation of BP neural network is described.Orthogonal experimental design method was used to use FLAC3 D numerical simulation software as a carrier to establish a BP neural network learning and training sample.Inverse analysis of rock elastic modulus E and coefficient of subgrade reaction K in this area was conducted.