State-of-art review on the process-structure-properties-performance linkage in wire arc additive manufacturing

Abstract

Wire Arc Additive Manufacturing (WAAM) can well offer improved design flexibility and manufacturing versatility for the integrated molding of large components. However, it is challenging to achieve high productivity in arc additive metal part applications, as it requires consistent manufacturing, reliable quality, and predictable performance. The service performance of arc additively manufactured components is often influenced by microstructure, widely distributed defects, deep residual stresses, and complex surface roughness. To this regard, investigating the Process-Structure–Property-Performance (PSPP) relationships via both experimentation and simulation is a proven strategy for furthering the capabilities of additive manufacturing. Nowadays, Machine Learning (ML) can also be a powerful tool for modelling these complex, nonlinear relationships. This paper begins with a brief overview of WAAM process classification, and a generic description of process control. It then proceeds to a comprehensive review and discussion of how component microstructure, internal defects, surface roughness, and residual stress, all impact mechanical and fatigue properties of WAAM components. Additionally, it includes a detailed exploration of the latest advancements in using ML to predict these effects, focusing on PSPP modelling. Finally, the paper discusses the current limitations of ML approaches in PSPP modelling, and outlines future trends and technological prospects.

Keywords

• Wire arc additive manufacturing
• microstructure and mechanical properties
• internal defect
• residual stress
• fatigue performance assessment
• machine learning