2021 Nashville AISTech

Development of Quantitative Indices and Machine Learning-Based Predictive Models for SEN Clogging (Room Virtual)

01 Jul 21
8:30 AM - 9:00 AM

Tracks: Virtual Program - Metallurgy

Speaker(s): Ruibin Wang, Ph.D. Candidate, University of Toronto; Fernando Guerra, University of Toronto & Stelco Inc.; Kinnor Chattopadhyay; Chad Cathcart, Stelco Inc.; Heng Li, University of Toronto

The submerged-entry nozzle clogging as a result of alumina inclusion deposition has been a critical issue during continuous casting for aluminum-killed steel. In order to impose a strict monitoring and control over continuous casting, clogging indices were developed to precisely quantify the magnitude of clogging based on process parameters such as stopper rod position, mold level variation and argon pressure. Analysis of clogging indices per steel grade was conducted, and respective critical clogging index was investigated. A machine-learning model was also established based on the developed clogging indices to predict the clogging phenomenon and provide corrective modifications.