Digital Twin Empowered Operational Status Monitoring of High-Power Induction Furnaces: A PINN-Based Approach

In the realm of industrial automation and smart manufacturing, improving the efficiency of high-power induction furnaces (IFs) has become an essential goal. Precise monitoring of the operational status is indispensable for achieving targeted optimization goals. The challenges encountered in this endeavor are twofold in nature. IF, as a highly complex thermoelectric system, spans multiple disciplinary fields, complicating the achievement of precise modeling. Furthermore, its electrical dynamic parameters continuously change throughout the melting process, eluding direct measurement methods. In this paper, we introduce a novel equivalent circuit framework that simplifies the analysis of the IF operational status. Secondly, we establish a PINN-based digital twin (DT) for IF to simulate the dynamics throughout the entire melting process, thereby supporting the status monitoring. State-space constraints are introduced to overcome the deficiencies of the traditional PINNs in handling multidimensional systems. Finally, we assess the feasibility of the approach using actual operational data spanning the complete working cycle of IF.