The Luo Ping Tunnel Construction Monitoring And Intelligent Back Analysis Of Rock Parameters

The thought of information-based design and construction in tunnel engineering have been accepted by people and adopted more and more. How to get the right rock parameter has always been a thorny problem for a long time. That using back analysis method to get the rock parameter is provided a more valid path. Based on the Luoping tunnel of Wufeng highway in Chongqing city to the carrier of study, to the scene measured data as the basis, and to establish a intelligent back analysis system of surrounding rock parameters based on the MATLAB neural network toolbox, it is applied to actual projects similar to the design, construction, monitoring, has important guiding significance.The main content and research result of this article are:Firstly, this paper introduced the engineering situation, engineering geological conditions, hydrogeology condition of the Luoping tunnel in detail, and the subject of tunnel engineering design methods, etc. According to the actual situation of the reasonable monitoring measurement implementation plan, Through the site monitoring measurement of Luoping tunnel's dynamic construction process, obtain the first-hand measurement data on the excavation process, master the changes of the surrounding rock, stress and deformation of lining on the tunnel excavation process, using the regression analysis of mathematical method to find the inner displacement variation, determine the stable time and total deformation of displacement convergence, provide a reasonable support time for second-lining. Secondly, based on the characteristics of tunnel engineering, A discussion was made on the various factors influencing the displacement of tunnel surrounding rock, preliminarily determined the inversion parameters of tunnel surrounding rock, establish a reasonable parameters inversion verified model of surrounding rock by using ANSYS numerical simulation software.This paper introduced the concept of artificial neural network, the concept of neurons and its characteristics, the model of artificial neuron. Then mainly recommends the learning behavior and learning algorithms of neural network, the classification of neural network model. This article introduces the programming process and design methods of BP network, and points out the deficiency and related improvement measures of BP network.To design different inversion parameters, such as elastic modulus and poisson's ratio, using orthogonal design, uniform design method and ANSYS numerical model for the group simulation, based on the BP artificial neural network versus nonlinear function which have powerful mapping ability, generalization ability and self-learning ability, constructed the learning samples and the test specimen of artificial intelligence algorithm, compared with the field measurement data, and get the optimal solution of the inversion parameters.Finally, based on the already determined inversion parameters, all the parameters of surrounding rock parameters is substitution in the inversion verified model, the vault crown settlement and horizontal convergence displacement value is calculated by forward simulation, is basically coincide with practical measurement values. Then the convergence value of tunnel surrounding rock deformation which is not excavation can be predicted, provide a reasonable basis for the later design, construction, monitoring measurements of tunnel.