Abstract:

To investigate the value of soil mechanics parameters in the numerical simulation of underground engineering and the influence of pipe jacking construction on the metro structure, a case study is conducted on a pipe jacking construction spanning a metro tunnel in Xi′an, China. Base on the monitoring data of the local construction section, the designed and constructed back propagation(BP) neural network is employed to inversely calculate the physicomechanical parameters of soils, and the parameters are input into the model constructed by MIDAS GTS NX finite element software. After the project is completed, the actual value of the tunnel structure deformation measurement is obtained, the predicted value of the crown settlement and the arch waist horizontal convergence of the corresponding sections in the numerical simulation is extracted, and the accuracy of the finite element analysis and the rationality of the parameter inversion method are validated by comparing the difference between the two. The results show that the error between the predicted and the actual values is small, the deformation values are 2.6 mm and 0.5 mm respectively, and the error values are 0.24 mm and 0.07 mm respectively at the maximum crown settlement of the tunnel and the peripheral convergence section, indicating that the deformation of the tunnel structure is less affected when the pipe jacking construction is close to the metro tunnel. The inversion of the physicomechanical parameters of the soil layer based on the BP neural network is more reasonable, and the prediction of the deformation is more accurate, which can provide a reference for the construction and optimization design of the later tunnel engineering.

Key words: tunnel engineering, proximity construction, parameter inversion, back propagation neural network, numerical simulation, monitoring and measurement