Using AI and digital twin to improve blast furnace operations

Auteur(s) N. Aubry, M. Sukhram, J. Kim, Y. Zhang.
Presented at ECIC 2022

Successfully managing an ironmaking plant is a complex task. The need to satisfy competing interests of efficiency, quality, and cost while adhering to ever higher standards of safety and sustainability puts immense pressure on people and systems. Digital twin and AI technology meet today’s challenges of blast furnace operation from three strongly lined aspects: data integration; operational intelligence, and human-machine interaction. In this paper, a use case focusing on blast furnace thermal state prediction is presented, in which a stacked machine learning model combined with fundamental mass and energy balance calculations were applied and prototyped in Hatch digital twin platform. The model provides operators with a consistent assessment of the furnace thermal state, enabling a better hot metal temperature control strategy.