Generative AI for Process Modeling in the Steel Industry

This paper investigates the generation of realistic process data by utilizing adversarial networks to emulate actual steel process information. The objective is to create data resembling process data without discernible differences. Obtaining accurate process models for steel production is complex due to multiple challenges. Nevertheless, such data can prove invaluable for training operations personnel, simulating scenarios, testing software systems, and exploring novel control schemes. This research explores various generative modeling approaches to gain a deeper understanding of the constraints and practical implications of these models in steel environments, focusing on the application to a public dataset on continuous casting machine.

Authors:

Dr. Richard Marquez | ECON Tech
 

M.Sc. Esnardo Morales | ECON Tech
 

M.Sc. Alex Alvarez | ECON Tech
 

Session Chairs:

David Jiang | ArcelorMittal Dofasco
 

Bertrand Orsal | Dassault Systems
 

Michael Mayer | SMS group Inc.