Accuracy of computational welding mechanics methods for estimation of angular distortion and residual stresses

This study estimates the angular distortion and residual stresses due to welding using the following methodologies: thermo-elastic-plastic, inherent strain (local-global), and substructuring on two types of welded joints (T-type fillet weld and butt weld). The numerical results are compared with the experimental measurements and these methodologies are evaluated in terms of accuracy and computational time. In addition, the influence of welding sequence on distortion and transverse residual stresses has been studied numerically by implementing the thermo-elastic-plastic and inherent strain (local-global) methods on the T-type fillet weld. For the T-type fillet weld, the estimated angular distortion from these methods is much the same and in good agreement with the experimental measurements. For the butt weld, the angular distortion calculated by the inherent strain (local-global) method is largely underestimated. In order to gain a better understanding of where the underestimation of angular distortion in the inherent strain (local-global) method comes from, the study discusses the influence of block length and welding speed on angular distortion. It is found that for long weld length or slow welding speed, activating the plastic strain gradually by dividing the weld bead into an appropriate number of blocks can reduce the level of underestimation of angular distortion.