BACK ANALYSIS ON ROCK MECHANICS PARAMETERS FOR SIDE-EXPANDING SECTION OF HIGHWAY TUNNEL BASED ON BP NEURAL NETWORK METHOD

Tunnel surrounding rock is a complicated comprehensive system, various parameters have the very big indetermination, and therefore, the reasonable obtainment of its mechanics parameters is the important factor for promoting the development of tunnel engineering. Based on BP neural network, combined with the actual project, two monitoring cross-sections are set at side-expanding section to monitor the deep displacement of surrounding rocks, and acquires the numerical data for arch crown settlement and the horizontal convergence of the tunnel inner perimeter resulted from one-sided excavation. The BP model of artificial neural network is used for back-analyzing the mechanical parameter of the tunnel's surrounding rock. The network has a good simulating capability, and the range of the error all controls within 1%. The result shows that the BP neural network can be used for simulation network of back analysis on rock mechanic parameters, which the parameters by back-analyzed are true and believable, it is a good method for back-analysis design on tunnel surrounding rock parameters, and provides some engineering application value.