Carryover slag weight prediction under converter tapping process by applying attention-shifting infrared thermal mapping technique

Abstract

The prediction model of actual slag weight under converter tapping process was constructed by applying attention-shifting infrared thermal mapping technique. Firstly, the linkage between programmable logic controller signal of the converter mechanical structure and the infrared thermal mapping detector image was realized to ensure the real-time monitoring of the whole tapping process. Images should be preprocessed to ensure a good distinction between the flowing liquid and background at the tapping beginning, furthermore to separated slag from liquid steel at the following process. Moreover, the use of attention-shifting block belonging to the infrared thermal detector was effect to obtain more accurate image distinction between the liquid steel and slag. Hence the gray degree threshold value was adjusted to 220–225 in the early stage, and the value was needed to be lowered to 170–176 in the end. After using the model, the slag weight was less than < 8.0kg/t when only slag-stopping awl was adopted in tapping process. The average weight of slag in converter was 7.3kg/t by using slag-stopping awl and sliding plate slag-blocking.