Optimizing the butt-weld process of pipes based on the inherent strain approach

Abstract

Welded parts often suffer from distortion and residual stresses, leading to assembly inaccuracies and increased costs. To enhance the quality of welded structures, it is essential to predict and mitigate these issues. Traditionally, time-consuming thermo-elastoplastic (TEP) analyses are used to estimate welding stresses and distortions. In the present study, the theory of inherent strain and the elastic finite element method have been employed for accurate and rapid prediction of distortion and residual stresses in butt-welded pipes. Furthermore, to improve the overall weld quality, the genetic algorithm is used to optimize the main arc welding parameters such as welding power and the ratio of input heat to pipe thickness and the number of intermittent welding sections. The accuracy of the analyses is validated by comparing the simulation results with the FEM and experimental results of [1]. The optimization of the welding process was able to reduce the residual stress and deformation by 5 % and 9 %, respectively. The inherent strain method has been shown to reduce the computational cost of the analyses by more than 95 %, compared to the conventional TEP analysis, while preserving an acceptable accuracy in terms of the predicted residual stresses and deformations.