Bridging 2D and 3D computational modeling of vacuum arc remelting: Capturing rotating arc dynamics in axisymmetric simulations

Computational modeling of the Vacuum Arc Remelting (VAR) process has been developed to provide a deeper physical and metallurgical understanding and assist in the manufacture of defect-free ingots. While fully 3D, time-resolved arc models capture arc–melt interactions with high fidelity, their high computational cost makes 2D steady arc models still the preferred option in industrial applications. In this study, a multi-physics VAR model, which accounts for magnetohydrodynamics, heat transfer, fluid dynamics, and melting/solidification, was established in ANSYS Fluent for Alloy 718 in 3D with a rotating arc. It was validated against measured melt pool morphology. A new 2D axisymmetric oscillating arc model, posed under identical boundary conditions, is introduced to assess its accuracy and efficacy. The results demonstrate that the classic 2D axisymmetric steady arc model shows limitations in representing melt pool evolution involving time-dependent arc behavior. In contrast, the 2D oscillating arc model reproduces the unsteady melt pool behaviors and flow dynamics observed in the 3D rotating arc case, with a pronounced computational efficiency. The melt pool behavior responses to arc parameter variations, including cycle time, Gaussian standard deviation coefficient, and Gaussian offset distance, in the 2D model remain consistent with those in the 3D model. This 2D oscillating arc approach offers a promising alternative for simulating arc-driven phenomena in VAR, enabling a rapid screening of process parameters and guiding when detailed 3D simulations are warranted.