About Designing an Intelligent System for Slag Detection in Oxygen Converter Steelmaking

In this paper, proposed are the schematic of the intelligent system for slag detection in oxygen converter steelmaking and the algorithm for slag detection based on artificial neural networks and evaluation of informative features. Automatic non-contact slag recognition is carried out based on a multi-structured approach. In the developed system, digital images are obtained from thermal imaging and thermographic video cameras. Digital images are presented in the RGB color model. A conventional video camera is used to detect flames and smoke, as infrared cameras can be faulty in the presence of such interference. The proposed system for real-time slag recognition expands the functionality of the existing automated converter control system and improves the quality of steel. The computational modelling of the slag recognition process was carried out using in situ video frames of the basic oxygen steelmaking process.