Machine Learning Statistical Ultrasonic Non-destructive Prediction for Friction Stir Processed Metal

Puntos clave

• The approach enables accurate statistical prediction of material properties post-friction stir processing, enhancing quality assurance.
• A predictive accuracy improvement of 15% was observed using machine learning models compared to traditional methods.
• Analysis of material samples through ultrasonic testing was employed to assess properties without damaging the materials.
• These findings highlight the potential of integrating machine learning with traditional techniques for manufacturing advancements.