Finite Element-Based Multi-Parametric Analysis of Residual Stress Evolution in Thin Structural Steel T-Joints

Abstract

Optimization among welding parameters is important to minimize residual stress, distortion, and mechanical strength in the thin structural steel T-joint plates. In the present work, a finite element–based multi-parametric simulation was established to study the impact of welding current, arc voltage, and welding speed on the thermo-mechanical performance of welded T-joints. A co-developed thermo-mechanical finite element model along with a statistically rigorous Design of Experiments (DOE) using Central Composite Design (CCD) and Response Surface Methodology (RSM) was developed to systematically assess the interaction of parameters and the nonlinear effects with the least computational effort. The response variables include maximum temperature, residual von Mises stress, total deformation, and ultimate tensile strength. Quadratic regression models have been formulated and compared for 95% confidence level at ANOVA. Their predictive accuracy was impressive; all models had coefficient of determination (R²) higher than 0.98. Results indicated that welding current directly affects the residual stress and the tensile strength, and the deformation behavior depends on the welding speed. Interaction and quadratic dependence were significant implying that the welding process is coupled, as well as nonlinear. Heat input was defined as a regulating function of thermal cycles and mechanical events. The combined FEM–DOE approach is an effective and computationally efficient method for predicting and adjusting welding parameters in thin plate T-joint configurations. The approach offers solutions for data based optimization of welding processes and a strong implication for structural and fabrication systems.