Welding Simulation: Relationship Between Welding Geometry and Determination of Hardening Model

Abstract

It is generally acknowledged that the material hardening model exerts a considerable effect on predicted weld residual stress fields. For this reason the choice of hardening model has attracted interest among analysts, particularly during recent validation studies. Nevertheless there is still lack of evidence for a hardening model which is generally applicable for all welding geometries. In this work we examine the predictions of nonlinear kinematic, isotropic and mixed hardening models for two different geometries: a single bead on plate weld, and a multi-bead girth weld. Hardening parameters are based on the same openly available mechanical test data. Deformation histories for the two welding geometries are presented. Predicted residual stress profiles are compared with experimental measurements. It is noted that nonlinear kinematic hardening results in good predictions for the single bead welding simulation where hardening in the weld and HAZ is dominated by a single heating and cooling cycle. Isotropic hardening results in good predictions for the 42 bead girth weld, where hardening in the weld and HAZ is heavily influenced by several heating and cooling cycles from the addition of several weld beads and where some relaxation of residual stress is possible. Mixed hardening can result in good predictions for both welding geometries. Additional strategies for development of material models based on isotropic and kinematic hardening and relevant test data are discussed with particular attention paid to intermediate weld geometries.