Strip Break Classification in Tandem Cold Rolling of Steel using Operational AI
Project Lead: Falkonry
Partners: ArcelorMittal (AMNS Calvert)
Member % Cost Share: 50%
CESMII % Cost Share: 50%
Duration: 6 Months
Problem Statement
Strip breaks in continuous tandem cold rolling steel mills causes line stoppages and equipment damage and are responsible for up to 15 days of lost productivity annually (~$3,500,000 loss). It is very difficult to diagnose and classify the cause of strip break so that corrective action can be taken.
Project Goal
The project objective is to demonstrate an automated strip break classification system that is capable of predicting the reason for a ‘strip break’ event.
Technical Approach
Develop the digital twins to detect and classify strip breaks using the Falkonry Time Series AI platform by analyzing historical and real time data. Tune the digital twins using semi-supervised reinforcement learning and validate the solution to monitor and classify in near real-time condition.
Deliverables/Outcomes/SM Marketplace
- Delivered Classification Solution Performance Report.
- Published Use Case:
Strip Break Classification in a Cold Rolling Steel Mill Using Machine Learning. - Metal Strip / Weld Break SM Profile.
Potential Impact
In the AM/NS Calvert tandem cold rolling mill, the annual cost attributable to strip breaks and weld breaks in the last year is $3.5 M which includes lost production time and yield loss. Economic impacts from the solution are to be realized in two dimensions: Reduction in labor by automating strip break classification and reduction in the economic impact of strip breaks due to reduced frequency of occurrence.
Benefits
SM Profile of a tandem cold rolling mill available for the CESMII SM Marketplace. A solution architecture and methodology for the use of digital twins for real-time strip break classification and explanation. An industry use case for predicted classification of strip breaks in tandem cold rolling mills.
