• Framework couples thermal welding with XFEM for OSD fatigue assessment. • Tensile residual stresses reduce OSD fatigue life by 14.2%. • Welding stresses suppress crack closure, maintaining Mode I opening states. • Results indicate non-conservative gaps in standard-compliant fatigue life. • Stress reduction strategies significantly influence crack-tip driving forces. Orthotropic steel decks (OSDs) are key structural systems due to their high efficiency and low self-weight. However, welded joints in OSDs remain vulnerable to fatigue cracking under repeated traffic loading. Welding residual stress (WRS), combined with cyclic loads, significantly influences crack propagation, necessitating accurate fatigue assessment. This paper presents a validated three-dimensional numerical framework to investigate the impact of WRS on fatigue crack growth in OSDs. A three-dimensional finite element (FE) model of the U-rib-to-deck plate connection was developed in ABAQUS, utilising geometric symmetry to optimise computational efficiency. The gas metal arc welding (GMAW) process was simulated through transient thermo-mechanical analysis using Goldak’s double ellipsoid heat source model with temperature-dependent material properties to generate realistic WRS fields. The resulting WRS fields were validated against laboratory measurements and incorporated into an extended finite element method (XFEM) framework to simulate crack propagation under loading specified by Australian bridge design standards. Comparative simulations with and without WRS show that tensile residual stresses increase local stress intensity, energy release rates, and the cyclic J -integral (ΔJ), accelerating crack growth and reducing fatigue life by approximately 14.2%. These findings underscore the need to incorporate WRS into fatigue models for welded bridges. The developed framework provides engineers with an efficient tool to quantify the impact of welding procedures and residual stresses on the fatigue performance and inspection planning of OSDs under realistic traffic conditions.
Karimi et al. (Fri,) studied this question.