| As a new method for determining the calculation parameters of geotechnical engineering,the displacement back-analysis method has attracted much attention since its advent because of its strong practicability,and it was applied in many projects.However,the traditional displacement back-analysis method also has problems such as complicated calculation and low inversion efficiency.In order to overcome these shortcomings of traditional method,a intelligent displacement back analysis method is proposed.By introducing the artificial intelligence method into the traditional method,this method can not only solve the problem of rock mechanics parameter estimation,but also solve the problem of constitutive model identification,which has the characteristics of simplicity,practicality and high precession.Therefore,based on the neural network and particle swarm optimization algorithm,this thesis takes Liyanlong tunnel as background and uses Midas/GTS finite element simulation to carry out the following work on the intelligent displacement back-analysis method and its application:(1)After the research of artificial neural network and particle swarm optimzation algorithm,it is found that BP neural network and PSO algorithm are easy to fall into a local optimum.In response to this deficiency,the standard PSO algorithm is improved by dynamically changing the inertia weight and adding particle adaptive mutation,and the network’s weights and thresholds are optimized by the improved PSO algorithm.Furthermore,the effectiveness of the improved algorithm is verified by a non-linear function fitting experiment.(2)Based on the improved PSO algorithm and neural networks,A new intelligent displacement back-analysis method is proposed.this method uses orthogonal and uniform experimental design methods to determine the experimental scheme,the finite element simulation model is established by Midas/GTS to construct training samples,Combining the non-linear relationship which is established by BP network and the network is trained by improved PSO algorithm between the inversion parameters and the measured displacement,the optimal inversion parameters are searched by improved PSO algorithm.Then this method is applied to the back-analysis of Liyanlong tunnel to verify the feasibility and reliability of the intelligent displacement back-analysis method.(3)Based on the results of displacement back-analysis of liyanlong tunnel,the safety factor of tunnel under different shape and span of drift heading and middle drift heading block scheme is calculated by the FEM strength reduction method,and the optimal design scheme of double side drift heading of liyanlong tunnel is determined.In this thesis,a new and practical intelligent displacement back-analysis method is establish based on the neural network and particle swarm optimization algorithm,which lays a technical foundation for the application of the back-analysis method in parameter estimation,engineering forecasting,dynamic feedback design and reliability evaluation. |
