Predictive and Optimization frameworks for bead geometry in Wire Arc Additive manufacturing using Machine Learning

Wire Arc Additive Manufacturing (WAAM) is being popularized as a new practical and economical approach for producing large metal components. However, the process still faces challenges of getting consistent bead geometry. Even a small variation in heat input, travel speed, wire feed rate and interlaying thermal accumulation can fluctuate the bead shape. This affects the dimensional accuracy, mechanical performance, and surface finish. Researchers have increasingly turned to Machine Learning and other data driven strategies to understand and manage these complex non-linear relationships. This study aims to combine the scattered findings on how different machine learning models have been used to predict and optimize WAAM. Commonly used data sets such as engineering particles, sensor-based monitoring methods, and hybrid frameworks that combine physics based and data driven approaches alongside summarizing existing efforts. This review will map the strengths and shortcomings of current techniques further highlighting where the work is still needed, especially in the areas of large-scale data sets, real time control, and cross-material generalization.