Multi-Objective Optimization of MAG Welding Parameters for Minimizing Distortion and Residual Stress in Industrial-Scale Heavy Machinery Assemblies

Abstract Residual stress, thermal deformation and warpage are the most important engineering problems for the welding structures. This problem is more significant for large industrial-scale structures, where higher heat input is demanded and temperature gradient is more significant. Residual stress and thermal deformation after welding could lead to inconsistency of tolerances, and in cases lead to reduced mechanical strength and ultimately failure. Although it is impossible to reduce thermal distortion to zero, finding optimum welding parameters could highly reduce the thermal deformation during welding. However, due to the large scale of the structure, thermal deformation and stress prediction and parameter optimization are extremely time consuming to be implemented in real practice. In this paper, we have performed a Multi-Objective Optimization (MOO) on the welding parameters on a case study provided by Ponsse Oyj company in order to reduce thermal deformation and thermal warpage due to heat delivered to the structure during Metal Active Gas (MAG) welding process within an acceptable computational time. A thermomechanical elasto-plastic model is created using Finite Element Method (FEM) to determine thermal deformation and residual stress. Furthermore, L-ANN-GWO MOO algorithm, which adopts adaptive surrogate modeling strategy to reduce the number of FEA evaluations, is used to find the optimum welding parameters (starting angle, travel speed, voltage, and current) subject to the preliminary Welding Procedure Specification (pWPS). Based on MOO, three process parameter combinations were suggested which could lead to up to more than 30–40% reduction in thermal distortion and 20–30% reduction in residual stress. Our studies show the importance of selecting proper welding starting point. Performed modelling-optimization approach used in this paper could be used for other large structures where there is excessive delivered heat.