Efficient and Accurate Computational Theory for Long-Span High-Speed Railway Bridge Systems under Multiple Environmental Excitations

This study constructs a set of efficient and accurate modeling and solution methods for long-span high-speed railway bridge systems under multiple environmental excitations. First, an efficient partitioned non-iterative computational method is developed to enhance the computational efficiency and accuracy of vehicle-railway-bridge systems. The method decomposes complex dynamic systems into subsystems with homogeneous physical properties, enabling independent dynamic analyses with distinct temporal resolutions and spatial discretization while maintaining energy conservation characteristics across both global and partitioned systems. The proposed method integrates both Newmark scheme and generalized-α time integration schemes. Subsequently, an innovative 3D vehicle-track-bridge partitioned model is established, eliminating repetitive assembly of time-variant system matrices through strategic integration of wheel-rail contact mechanics into the vehicle subsystem. By incorporating a nonlinear Hertzian spring model into vehicle subsystem considering wheel-rail dynamic contact, the original time-variant coupled system is decomposed into three time-invariant subsystems (vehicle, railway, and bridge subsystems) connected through time-dependent interfaces. This formulation effectively decouples the complex wheel-rail and track-bridge interactions while preserving system nonlinearity. Furthermore, a Non-Iterative Partitioned Method with Sliding Window (NP-SW) is proposed to optimize computational efficiency. This approach confines rail computations to a moving window synchronized with vehicle position, dramatically reducing the track subsystem’s degrees of freedom without compromising solution accuracy. Finally, through comprehensive comparisons with conventional models and approaches, the energy stability, numerical accuracy, and computational efficiency of the proposed methodology are rigorously validated, demonstrating significant advantages over existing methods.