Experimentally investigating finite element accuracy for ground vibrations induced by high-speed trains

This study compares the dynamic responses of finite element analyses and field experiments. Finite element meshes included bridge superstructures, bridge foundations, moving trains, and soils. Along the mesh boundaries, numerical schemes were also used to avoid fictitious wave reflections. The comparison indicates that the predicted numerical and measured experimental results are in good agreement, so the proposed finite element model can be used to simulate the soil vibration induced by high-speed trains moving on bridges. Theoretical equations are derived in this paper, and the parametric study using those equations shows that the bridge model is more important than the vehicle model in simulating train-bridge interaction problems. This condition was verified by the field experiments. When the continuous rail between two simply supported beams is neglected in the finite element analysis, the finite element result shows that vibrations in the bridge's longitudinal direction are twice as large as those from experiments. Satisfactory finite element results were also obtained by replacing the complicated train model by a simplified moving loads model.