Modeling of the Vacuum Arc Remelting Process for Estimation and Control of the Liquid Pool Profile

Vacuum arc remelting (VAR) is an industrial metallurgical process widely used throughout the specialty metals industry to cast large alloy ingots. The final ingot grain structure is strongly influenced by the molten metal pool profile, which in turn depends on the temperature distribution in the ingot. A reduced-order model of the solidifying ingot was developed specifically for dynamic control and estimation of the depth of molten liquid pool atop the ingot in a VAR process. This model accounts only for the thermal aspects of the system ignoring other physical domains such as fluid flow and electromagnetic effects. Spectral methods were used to obtain a set of nonlinear dynamic equations which capture the transient characteristics of liquid pool profile variations around a quasi-steady operating condition. These nonlinear equations are then linearized and further simplified by suppressing fast modes. The resulting system was used to construct a linear-quadratic-gaussian (LQG) controller which was tested in a laboratory-scale furnace showing a good performance. A high-fidelity physics-based model is used in real-time to provide information about the solidifying ingot and potential solidification defects.