Physics-driven modeling of electron beam welding of Al-Cu alloys from molten pool flow, microstructure to mechanical properties

Abstract

The process-structure-property relationship has always been the research focus in welding, where computational prediction is valuable. However, some important physical phenomena are often simplified when modeling the melting and solidification processes. In this work, an integrated modeling framework is adopted to comprehensively predict the welding quality, from a computational fluid dynamics (CFD) model for molten pool dynamics, to a cellular automata (CA) model for dendrite growth, and finally to analytical and finite element method (FEM) models for mechanical properties. Using Al-Cu alloy as a model material, the complex melting and solidification processes in electron beam welding is reproduced, and various strengthening mechanisms are considered to predict the mechanical response. Notably, the material property inhomogeneities due to inhomogeneous microstructure in the welds are considered when predicting the overall mechanical properties. Based on the simulation results, the second phase strengthening proportion reaches almost 50%. The maximum relative errors of predicted Vickers hardness and tensile strength are 13.62% and 13.30%, respectively. The simulation results show quantitative agreement with experimental data, demonstrating the appealing potential of this modeling framework. This study could be helpful for optimizing the actual welding process to tailor the microstructures and mechanical properties. • A physics-driven process-microstructure-property modeling framework is adopted. • Various strengthening mechanisms based on the microstructure modeling are considered to predict the mechanical response. • The material property inhomogeneities are considered when predicting the overall mechanical properties. • The simulation results show quantitative agreement with experimental data.