Research On Intelligent Back-Analysis Based On BP Neural Networks And Immuney Programming Of Tunnel Surrounding Rock Displacement

How to obtain the accurate mechanical parameters of tunnel surrounding rock thought the relatively simple methods, is the active research topic for the modern scientific and technological workers and engineering and technical personnel of the tunnel. Back-analysis is better one way to gain the parameters of rock mechanics. This paper first had a more in-depth analysis of geotechnical engineering system identification theory, then Obtained the computing model of back analysis on the tunnel surrounding rock and its parameters. Second , optimizing the initial weights and bias of the neural network with algebraic algorithm's principle could reduce the randomness of neural network and fast the rate of convergence. Also it could determine the number of neuron with the theory of algebraic algorithm, and build the BP neural network forecasting system. On the basis of the general immune programming algorithm, the intelligent artificial immune system was built with the Random q competition mechanism, the mechanism of thickness control and the mechanism of entropy control. Then a system of intelligent back analysis of tunnel surrounding rock displacement with the BP neural network forecasting system and intelligent artificial immune system, and was used to back analyse the mechanical parameters of tunnel surrounding rock. The Rock displacement measured of the highway tunnel named Shui Tian Wan Two was used to back analyse the mechanical parameters of tunnel surrounding rock. The parameters get by this way were consistent with the engineering.