Camera- and Sensor Data-Based Process Observation In Liquid Steel Production

In the upcoming years, the steel industry faces major challenges to meet increasing production costs and CO2 reduction goals. The latter necessitates a shift from the blast furnace to the electric arc furnace process rendering vacuum treatments increasingly important. Subsequently, the Ruhrstahl-Heraeus (RH) degassing plant, which is essential for producing ultra-clean steel, will become a bottle neck vessel in the steelmaking process. In order to increase the time efficiency while maintaining consistent quality, process steps in the RH plant, which mainly follow strict recipes, need to be optimized. Within the scope of this paper, the focus lies on improving the image-based surveillance of the melt behavior and on enhancing the utilization of sensor data obtained during the RH process. We introduce a camera-based monitoring and image analysis solution, which in combination with enhanced sensor data evaluation (e.g. chamber pressure or offgas composition) determines the status of the process and detects malfunctions of the plant. The proposed research helps to deeper understand the RH process in steelmaking and provides the basis for a future on-line decision support system for operation.