Experimental and numerical assessment of welding residual stress effects in a full-scale plate–truss composite steel girder joint

Welding residual stresses (WRS) are fabrication-induced initial stresses in welded steel joints and may influence local yielding and nonlinear stress redistribution. This study presents an experimental and numerical assessment of the effects of WRS in a full-scale integral lower-chord joint of a plate–truss composite steel girder used in a long-span suspension bridge. The specimen was fabricated using site-consistent welding and post-weld heat treatment procedures. Near-surface WRS at critical welded regions were measured using the blind-hole method, with calibrated strain-release coefficients and plasticity correction applied in high-residual-stress zones. The measured WRS were generally limited to 0.3–0.4 times the material yield strength, reflecting the post-treated fabrication state. Based on the measured distributions, equivalent self-equilibrated WRS patterns were established and introduced into a shell-element finite element model as initial stresses. The elastic loading response of the full-scale joint was used to evaluate the numerical model, and the effects of WRS on stiffness, yielding initiation and plastic zone development were then assessed. The results show that WRS have negligible influence on the initial elastic stiffness of the joint, but promote earlier local yielding and faster plastic zone development. As the external load increases, the influence of WRS becomes less pronounced, and the response is increasingly governed by the global structural stress state. Sensitivity analyses indicate that WRS magnitude mainly affects the onset and evolution rate of nonlinear behaviour, without changing the fundamental load-transfer mechanism. The findings provide experimental data and practical modelling guidance for assessing fabrication-induced residual stress effects in complex welded bridge joints and similar large-scale steel structures.