Developing an Intelligent Process-Properties Simulator for Predicting Magnetic Properties of Electrical Steels Through Synthesis of Operational Data

Big River Steel produces cold-rolled motor lamination grades for electrical steel applications that require high-resolution variability control and achieving target magnetic properties, i.e., core loss and peak permeability. As part of the advanced analytics projects for NGO, a continuous machine learning model was developed to effectively analyze, explain, and control variability of core loss and peak permeability. This modeling approach simulates process conditions across the mill and provided twofold benefits of predicting the target magnetic properties and quantifying process impact on coils. This tool currently aids BRS in process improvement, product development and achieving first-time prime product.

Authors:

John Caleb Somasundaram | Big River Steel
 

Menghan Gu | Big River Steel Works, a U. S. Steel Company
 

Codrick Martis | Big River Steel Works, a U. S. Steel Company
 

Ted Hill | Big River Steel Works, a U. S. Steel Company
 

Session Chairs:

Chirag Mahimkar | Big River Steel
 

Luis Garza Martinez | Cleveland-Cliffs Research & Innovation Center