A COMPREHENSIVE MODEL FOR PREDICTING FATIGUE LIFE OF LASER WELDED LAP JOINT OF GALVANIZED HIGH STRENGTH STEEL AS A FUNCTION OF RESIDUAL STRESSES

摘要

This paper presents a three-dimensional (3D) multi-physics finite element model (FEM) to predict the fatigue life of a laser welded lap joint of dual phase (DP) 980 steel sheets based upon the level of residual stress. A FEM-based thermal analysis is first performed to numerically predict the welding-induced temperature field combined with the corresponding experimental verification. The temperature histories are then loaded into the mechanical model as thermal loading to numerically calculate the evolution curves of thermally induced stress in order to calculate the level of residual stresses after cooling to room temperature. In order to calculate the equivalent fatigue strength in the laser-welded lap joint, the resultant multi-axial stress (including the induced residual stress (RS) result) is loaded into the equivalent uni-axial stress equation via the Sine Method (SM) in order to achieve the stress curve as a function of the loading cycles. A series of fatigue tests of lap joints are also performed in order to achieve the S-N curves, from which an empirical function between the alternating stress and loading cycle is derived in order to predict the fatigue life of the DP980 lap joint. Finally, the maximum fatigue strength can be predicted numerically through the proposed FEM instead of using experimental trials. The numerical results show that a greater temperature gradient and residual stress are mainly located within the fusion zone (FZ) and close to the heat affected zone (HAZ). The residual stress plays an important role in deciding the final fatigue strength and failure of the DP980 lap joint. An X-ray diffraction technique is used to experimentally measure the residual stress distribution within the weld, for which the numerically predicted results exhibit a good agreement. Also, the numerical simulation and experimental measurements of the fatigue life versus the applied load show a good correlation of results.