Parameters Inversion And Numerical Simulation Of Surrounding Rock In Tunnel Close Construction Based On PSO-BP Algorithm

In 21 century,the utilization of urban road resources in our country tends to be saturated.A large number of subways are constructed to reduce the traffic volume of road traffic in various cities.Because the buried depth of subway is shallow,it is sensitive to excavation disturbance of soil.The excavation causes surface subsidence,even serious it may cause ground collapse.The surrounding original buildings are affected by construction of subway,so that the settlement occurs.It can't be ignored how construct a new subway tunnel at the same time to ensure the stability of original building.In practical engineering,according to mechanical parameters of surrounding rock,it can ensure the excavation mode of subway construction and material parameters of support structure.But rock is a kind of heterogeneous,nonlinear and discontinuous material.It is particular important how obtains accurate mechanical parameters by simple and efficient method.Through obtaining mechanical parameters of surrounding rock,it can predict accurately and take corresponding measures to control settlement caused by construction.It has guiding significance for practical construction.Based on practical project of Dalian Metro Ling Shui Yi Hao bridge interval,BP neural network optimization algorithm based on particle swarm optimization is proposed.Through site monitoring data,mechanical parameters of surrounding rock are obtained.It is certain significance for practical project application.The main contents of the paper are as follows:(1)Through finite element analysis software ABAQUS,the subway tunnel construction with full section excavation method,step method,CD method and CRD method were numerical simulated respectively.In four different construction methods,surface settlement,stress field and displacement field of surrounding rock,and internal force of anchor and lining were obtained.Through comparative analysis,the results show that tunnel control effect of CRD method is better,but tunnel control effect of full section excavation method is relatively poor.(2)According to principle of orthogonal experimental design,the effect of various experimental factors on displacement settlement of surrounding rock was analyzed by range method and variance method.The results show that the elastic modulus is the main affecting factor relative to other factors.The range of physical mechanical parameters of rock mass were chosen,constructed with learning samples of displacement back analysis,then calculated BP neural network algorithm and PSO-BP algorithm respectively.Test value and learning value were contrasted and analyzed.By self validating method,the error of PSO-BP algorithm is smaller and prediction ability is better.(3)Combined with practical monitoring data of Ling Shui Yi Hao bridge interval,through orthogonal experimental design,learning samples of PSO-BP algorithm were constructed.Then mechanical parameters of surrounding rock were put into ABAQUS.Considering approaching construction and simulation dynamic excavation with step method of subway,the effect of approaching construction on surrounding rock and pile stability were analyzed.Contrasting with the surface settlement,crown settlement,side wall clearance convergence value,pile settlement value and simulation value,the maximum error is 18.4%,so numerical simulation analysis is reliable.