Toward Intelligent Process Control in Additive Manufacturing: Machine Learning Prediction of Melt-Pool Geometry in AlSi۱۰Mg

Selective laser melting (SLM) of AlSi۱۰Mg alloys offers high design flexibility but remains limited by unpredictable melt-pool behavior affecting part quality. This study integrates numerical and data-driven approaches to forecast melt-pool dimensions under varying laser power and scan speed. A finite-element thermal model was developed in Abaqus with a user-defined laser heat source and validated against experimental results. Subsequently, machine-learning regression models were trained on simulated data to predict melt-pool width and depth with high accuracy (R² > ۰.۹۵). The integrated framework enables rapid process optimization and supports adaptive control in SLM of AlSi۱۰Mg.

کلیدواژه ها:

Additive manufacturing ، AlSi۱۰Mg ، machine learning ، melt-pool prediction ، selective laser melting