Integrated computational materials engineering (ICME) for predicting tensile properties of additively manufactured defect-free single-phase high-entropy alloy

Abstract

In metal-based additive manufacturing (AM), understanding the process-structure-property (PSP) relationship – where the microstructure is tailored to control performance – is crucial for part customisation. Establishing this PSP linkage often requires extensive experimentation and computational analysis. Our study proposes an integrated computational approach for predicting mechanical properties in the laser powder bed fusion (L-PBF) process. This approach combines microstructure evolution simulation, micromechanical property calculations, and machine learning to model the structure-property relationship. We also introduce a novel sampling method based on two-dimensional (2D) cross-sections with grain-level and slice-level predictive designs. Our findings indicate that microstructure evolution during solidification, modelled via the cellular automata (CA) method and finite volume method (FVM) thermal field, aligns well with experimental results. The proposed model achieves high accuracy in predicting mechanical properties, demonstrating the potential of our approach to advance metal-based AM by streamlining the PSP linkage.

Keywords

• Laser powder bed fusion
• structure-property relationship
• cellular automata
• machine learning
• integrated computation