A computationally efficient multiphysics model for friction stir welding with polygonal pin profiles

Abstract

This study aims to model temperature distribution in friction stir welding (FSW) using various backing plates and polygonal pin profiles since temperature significantly modifies the microstructure and texture of the weld zone affecting the weld quality such as weld strength, hardness, etc. The experimental results depict the importance of temperature on the grain size and tensile strength of the materials. However, determining the temperature at each point of the weld is difficult and expensive in the case of experiments. Therefore, in order to accomplish the objective, it is necessary to perform simulations. This paper presents a 3-D transient multiphysics model developed for FSW combining multiple physical phenomena such as heat transfer and structural mechanics in a unified framework, COMSOL. A viscoplasticity model is chosen as behavior for the AA1100 aluminum material using Anand viscoplasticity. It is computationally efficient and accurate. The model fidelity to the twin FSW process is achieved by considering temperature-dependent yield strength. Modeling results show the polygonal pin profile edges to be influencing the temperature. Increasing the number of faces on the pin sides leads to a higher temperature. Specifically, transitioning from an octagonal to a decagonal profile results in a minimal increase in total heat generation. As the pin shape approaches cylindrical, there is a gradual convergence in heat generation with that of a cylindrical pin. Experiments are also carried out that validate simulation results. Overall, the model is sufficient to twin the process for predicting weld quality and is Industry 4.0-compliant.