| Abstract Scope |
Vacuum Induction Melting of metal feed stock is a complex manufacturing process characterized by strong coupling between electromagnetism, heat transfer, and fluid flow phenomena. The process has many crucial applications (casting, AM powder production, etc.) relying on potential advantages (fast program change, precise temperature controls, controlled environment, etc.). Accurate modelling of these interactions is necessary for understanding process dynamics, optimizing performance, and enabling control in advanced manufacturing settings. In this work, we propose a Multi-physics model to capture the electromagnetically driven melting, convective transport, and free surface deformation of the liquid metal (Aluminum in our case) under experimental operating conditions. Verification and preliminary validation are carried out using feasible sensor data acquired from an in-house vacuum induction furnace. Further, we build emulators for the numerical model through Gaussian Process based surrogate modelling methods for fast predictions, laying the foundation for a Digital Twin framework fostering autonomous control of the system. |