Prediction of various defects and material flow behavior during dissimilar FSW of DH36 shipbuilding steel and marine grade AA5083 using FE-based CEL approach

Abstract

Abstract Dissimilar friction stir welding (FSW) of steel-Al is a very tedious job. Inappropriate welding process parameters can lead to the initiation of inevitable defects associated with dissimilar FSW processes. These can be presented as tunnel defects, void generation, excessive flash formation, and other surface irregularities. Using conventional experimental trials makes it usually challenging to identify such defects. This research adopted an Abaqus/Explicit® framework utilizing a 3D thermo-mechanical based coupled Eulerian-Lagrangian (CEL) methodology. In order to predict commonly observed defects in the FSW process, the proposed FEM uses the volume of fluid approach. By monitoring the material flow into and out of the computational/void domain, the suggested framework has made it feasible to predict surface, sub-surface, and volumetric defects. Defect formation is studied at a constant tool rotation speed of 875 rpm, welding speed of 90 mm min−1, and tilt angle of 0°. Tilt angles of 0° caused welding joints with a small tunnel defect. Thermal history, axial force variation, and material flow behavior are all strongly aligned with the principle of defect generation. An experimental trial has been conducted to validate the proposed finite element model. The previous analysis found that the average axial force closely matches the welding-related experimental findings with a percentage error of 7.85%. While a proportion error of approximately ∼0.57% was found between the compared numerical and experimental diameters of the pin end-hole defect. Furthermore, the proposed model accurately predicted the process of material flow along the thickness direction of the workpiece. It was seen that the stress generated at the root of the flashes reached a higher value ranging between 485.6 and 582.7 MPa. Finally, a good agreement between the numerical results and the experimental trial was established, showing the robustness of the developed computational FEM technique.