Comparison of explicit and implicit finite element methods with parallel computing for seismic response analysis of subway underground structure

摘要

With the 32-CPU parallel computing cluster platform based on EM64T hardware architecture, dual-path Intel Xeon processor, Linux OS, Gigabit Ethernet subsystem and 64-bit ABAQUS applications, the calculation accuracy and efficiency of the explicit finite element method, which uses the central difference method, and the implicit finite element method, which uses the Hilber-Hughes-Taylor method, are evaluated, and the accuracy with the viscous-spring artificial boundaries is also evaluated. The results show that the calculation accuracy of the explicit method and the implicit method is basically the same, but the efficiency of the explicit method is much higher than that of the implicit method. In this research, the comparisons between the explicit and implicit methods using multiple processors in the 2D and 3D nonlinear seismic response of subway station structures were achieved. The results show that for the case of the 3D large-scale model in which the DOF is 387426, the solution time of the explicit method using 8, 16 and 32 CPUs are respectively 21.89%, 23.10% and 4.32% of that of the implicit method. However, for the case of the 2D small-scale model in which the DOF is 10516, the solution time of the implicit method using 1, 2 and 4 CPUs are respectively 41.4%, 45.3% and 51.8% of that of the explicit method. So the explicit method with multiple processor parallel computing is fit for large-scale numerical computing problems while the implicit method with multiple processor parallel computing is fit for small-scale numerical computing problems.