A Generative AI Copilot for Supporting Plant Maintenance Operations in a Steel Plant

This paper describes an innovative solution based on generative AI technology aimed at driving a transformative change in plant maintenance operations with the help of artificial intelligence. The virtual assistant, built on a fine-tuned Large Language Model (LLM) specific to the steel industry, utilizes Retrieval Augmented Generation with data from sources like CMMS databases and maintenance manuals. Available both on cloud and on-premise, it supports text and voice inputs in multiple languages. Results from the implementation in ABS steel plant are presented, with optimization in the efficiency of maintenance operations, reduced intervention times and simplified onboarding for new personnel.

Authors:

Luca Cestari | Danieli Automation S.p.A.
 

Giorgio Felline | ABS – Acciaierie Bertoli Safau S.p.A.
 

Carlo Rossi | ABS – Acciaierie Bertoli Safau S.p.A.
 

Michele Cancedda | ABS – Acciaierie Bertoli Safau S.p.A.
 

Daniele Guastamacchia | ABS – Acciaierie Bertoli Safau S.p.A.
 

Cinzia Canderan | BeanTech Srl
 

Marco Antonelli | BeanTech Srl
 

Session Chairs:

John Accurso | Quaker Houghton
 

Cory Langhoff | The Timken Company
 

Chaitanya Sonarikar | ANDRITZ Metals USA Inc.