| Abstract Scope |
We present our digital twin platform for EAF that fuses high-fidelity metallurgical models, sensor data, machine learning and AI to simulate, predict and optimize furnace operation and chemical makeup.
The solution helps reconstruct the complete chemical composition of incoming metallic charges based on per heat data.
By creating a full chemical composition table, the platform clarifies the long-standing “statistical dead heat” condition in steelmaking – where operators see no statistically meaningful difference between heats thus lack precise, time-specific information.
Based on continuously collected data, the platform enables control of the melting process not blindly, but through a dynamic mass-energy model that reflects the real thermodynamic state of the furnace. It also tracks key parameters such as electrical efficiency, slag practice, thermal performance, and flux balance.
This approach paves the way for real-time, model-based optimization of EAF operations, enabling smarter decision-making and more efficient, transparent, and sustainable steel production |