Novel Fatigue Damage Assessment Approach for Composite Truss Railway Bridges Using a Hybrid Model Associated with Stress Influence Matrix

Advances in construction technology have rendered steel railway bridges increasingly complex, yet conventional fatigue-assessment methods still rely on global beam–column models that cannot quantify in-service damage. Therefore, building on current design codes and cumulative-damage concepts, this study proposes a quantitative fatigue-damage assessment framework that converts the qualitative guidance of existing regulations into an efficient, metrics-based methodology. Global beam–column elements alone cannot capture weld-level hot-spot stresses (HSSs); therefore, local solid sub-models are created in accordance with International Institute of Welding (IIW) recommendations. These sub-models are embedded in a hybrid finite element model (FEM) that couples global beam–column behavior with local solid behavior. To further reduce computational cost, a stress influence matrix (SIM) is introduced to reconstruct HSS time histories, thereby replacing repeated transient finite-element simulations without sacrificing accuracy. In this study, the numerical model is based on the Anxin Bridge, which is located in Taiwan. After a rigorous investigation of stress-range effects, the SIM approach achieved a Pearson correlation coefficient exceeding 0.99 in reconstructing HSSs. Relative to full finite-element time-history analyses, the computational effort was reduced by approximately 449-fold. Fatigue life was subsequently estimated using Miner’s rule in conjunction with code-based stress–number (of cycles) [Formula: see text]–[Formula: see text] curves, yielding an expected service life of about 108 years. Finally, this study proposed design guidelines for HSSs, thereby enhancing the identification of structural-member safety. This methodology provides a practical tool for quantitative fatigue verification. Additionally, it provided an important basis for structural health monitoring of existing bridges.