AI Insight
This study presents a sequential Bayesian learning framework designed to predict thermal and material flow fields in additive friction stir deposition, a solid-state manufacturing process. The approach integrates probabilistic machine learning with physical process data to enable real-time or near-real-time prediction of complex field variables that are otherwise difficult to measure directly during deposition. The model demonstrates the ability to update predictions sequentially as new data becomes available, improving accuracy and quantifying uncertainty in the predicted fields.
Why it matters
Accurate prediction of thermal and material flow conditions in additive friction stir deposition could improve process control, reduce defects, and accelerate the adoption of this manufacturing technique for aerospace, defense, and structural applications where material integrity is critical.