<<
[an error occurred while processing this directive] | [an error occurred while processing this directive] >>
Back Analysis of the Incremental Displacement of a Long-Span Bored Metro Station
1 Key Laboratory and Engineering and Technological Research Center of Ministry of Education, Southwest Jiaotong University 2 School of Civil Engineering, Shouthwest Jiaotong University 3 China Railway Shanghai Design Institute Group Co., Ltd
Abstract In this paper, a uniform design method is adopted to design a numerical simulation test with 16 levels of 5 factors（surrounding rock density, elasticity modulus, Poisson ratio, cohesive force and internal friction angle）and a 3D calculation is conducted for the construction of a long-span bored metro station by the finite difference method to obtain the values of surface subsidence for different experimental conditions and different procedures. The incremental displacements of 11 observation points induced by the demolition of the temporary support and corresponding rock-soil parameters are used as the input set, and the output set of a BP neural network helps to complete the training and learning. Additionally, the measured displacements induced by the demolition of the temporary support are taken as input parameters and the corresponding rock-soil parameters were inversed. The displacements of the rock-soil mass were obtained by putting all the inversion parameters into FLAC3D for a 3D-simulated calculation, and the rationality of those parameters are verified by measured increments of surface subsidence at 7 measuring points. The results show that the equivalent physical and mechanical parameters of the surrounding rock of a long-span bored metro station can be effectively obtained by a uniform design method and BP neural network in light of the surface subsidence increments induced by the temporary support demolition.
Abstract：
In this paper, a uniform design method is adopted to design a numerical simulation test with 16 levels of 5 factors（surrounding rock density, elasticity modulus, Poisson ratio, cohesive force and internal friction angle）and a 3D calculation is conducted for the construction of a long-span bored metro station by the finite difference method to obtain the values of surface subsidence for different experimental conditions and different procedures. The incremental displacements of 11 observation points induced by the demolition of the temporary support and corresponding rock-soil parameters are used as the input set, and the output set of a BP neural network helps to complete the training and learning. Additionally, the measured displacements induced by the demolition of the temporary support are taken as input parameters and the corresponding rock-soil parameters were inversed. The displacements of the rock-soil mass were obtained by putting all the inversion parameters into FLAC3D for a 3D-simulated calculation, and the rationality of those parameters are verified by measured increments of surface subsidence at 7 measuring points. The results show that the equivalent physical and mechanical parameters of the surrounding rock of a long-span bored metro station can be effectively obtained by a uniform design method and BP neural network in light of the surface subsidence increments induced by the temporary support demolition.
2015, Vol. 52 