Friction Stir

Friction stir processing, encompassing techniques like friction stir welding and additive friction stir deposition (AFSD), aims to join or build materials through localized heating and plastic deformation without melting. Current research heavily utilizes machine learning, particularly employing algorithms like random forests, gradient boosting, and neural networks (including physics-informed neural networks), to predict material properties, optimize process parameters (e.g., tool speed, temperature), and analyze resulting microstructures from experimental data and simulations. This focus on data-driven modeling and process optimization is crucial for improving the efficiency, predictability, and quality of friction stir processes across various materials, leading to advancements in manufacturing and material science.