Methodology Development of a Seam Weld Fatigue Analysis

This thesis presents a comprehensive study on the fatigue behavior of thin-sheet steel structures joined by seam welds, combining experimental testing and finite element (FE) simulations based on the meshing and fatigue evaluation methodology proposed by Fayard et al. The primary objective was to establish a reliable and practical workflow for the fatigue assessment of welded joints using conventional FE software—Abaqus and OptiStruct—without requiring specialized fatigue analysis packages. Experimental tests were performed on lap-jointed specimens made of 2 mm thick, 50 ksi steel plates welded using the GMAW-P process. The resulting force–cycle (F–N) data were converted into stress–cycle (S–N) curves using both maximum principal stress and an equivalent stress parameter derived from the Dang Van multiaxial criterion. These curves were then used to predict the fatigue life of T-joint specimens subjected to multiple loading conditions (Fx, Fy, and Fz). A strong correlation was observed between experimental and simulated fatigue lives, particularly when using the maximum principal stress, confirming the robustness of Fayard’s meshing approach and fatigue model for thin-sheet welded structures. To improve modeling efficiency, two Python-based plugins were developed for Abaqus: one automates mesh generation and weld element creation following Fayard’s rules, while the other performs post-processing to calculate equivalent stresses and estimated fatigue lives. These tools significantly reduce manual effort and increase consistency in fatigue evaluations. Overall, this study demonstrates that accurate and efficient fatigue assessment of seam-welded structures can be achieved using standard FE tools enhanced by automated modeling and analysis routines.