Accelerating Product Chemistry Refinement Using Explainable Machine Learning
Wednesday, 22 July • 10–11 a.m. EDT
Jena Kreuzer, Gerdau
Bryan Williams, Gerdau
Berk Birand, Fero Labs
Berk Birand is the co-founder and chief executive officer of Fero Labs, a machine learning (ML) software company focusing on industrial process optimization. He has been working closely with steel manufacturers since 2015 to bring cutting-edge explainable ML models to the factory floor. Berk is a seasoned speaker, having given keynotes at various industry conferences around artificial intelligence (AI)/ML, as well as being featured as a TEDx speaker. He holds a Ph.D. in electrical engineering and computer science from Columbia University, where his academic research focused on Internet of Things systems for enterprise-grade, resilient networks.
Jena Kreuzer is a process engineer in Gerdau’s technology group. She joined Gerdau after college and completed their management associate program and has worked in melt shop process improvements for the Knoxville and Cartersville steel mills. Currently, she focuses horizontally across all Gerdau mills with systems integration to ensure the best cost decisions from a process and technology view. She holds a B.S. degree in mechanical engineering from South Dakota School of Mines and Technology and is working on completing her M.B.A. at Ashford University.
Bryan Williams is a metallurgist at Gerdau Long Steel North America Cartersville Mill. He joined Gerdau in 2014. Prior to Gerdau, he was the quality manager and plant metallurgist at a commercial heat treatment facility in Rockford, Ill., USA, primarily servicing the automotive and aerospace industry. He is currently responsible for providing quality support for rolling operations within Cartersville. He holds a B.S. degree in metallurgical technology from Arkansas State University as well as a certified lead nuclear auditor.
There is significant hype around artificial intelligence (AI)/machine learning (ML) and its implications for manufacturers. Unfortunately, this hype has not translated into replicable return on investment, impeding widespread integration of this technology into the daily workflows of plant engineers. Many companies are left with failed pilots and digital transformation initiatives that do not deliver results.
This webinar describes how Gerdau has successfully navigated beyond the conceptual phase of digital transformation and achieved significant cost-savings results, leading them to deploy Fero Labs to many additional plants. They use machine learning to not only accelerate product chemistry refinement but also to continuously track realized cost improvements. Gerdau’s plants have been leveraging ML to automatically model production data to better define the relationships between process parameters (e.g., chemistry, rolling mill temperatures, product properties) and mechanical results. The use of explainable ML software has accelerated the prediction of mechanical KPIs, enabled accurate simulations of the effects of process changes, and led to the optimization of alloy content of specific grades.
In this webinar, Gerdau and Fero Labs will walk the audience through the metallurgical design process using ML software and highlight the new workflow of plant engineers and operators. Part of the webinar will be devoted to the description of explainable ML and how it complements the existing process optimization toolkit. The discussion will also cover the steps needed to integrate the new technology into the existing IT infrastructure for near-real-time alloy optimization. The session will conclude with a summary of the year-to-date improvements Gerdau has achieved and the resulting reduction in alloy costs of the new product grades.
Ramesh Khajjayam obtained his bachelor of technology in electrical and electronics engineering from Nagarjuna University India in 2003 and a master’s degree in electrical engineering from West Virginia University in 2006. He worked for Siemens from 2006 to 2015, and has since been working for Primetals Technologies since 2015. Khajjayam is a licensed Professional Engineer in electrical engineering. He is an active volunteer of IEEE Industry Application Society’s Atlanta chapter. He served as papers chair for AIST’s Electrical Applications Technology Committee in 2017–2018. Currently he serves as chair and DT Forum Core Team member. Khajjayam has presented technical papers at both IEEE conferences and AISTech. His current interest is in digital applications in the steel industry with a focus on electrical systems.
AIST’s Electrical Applications/Sensors Subcommittee and Digitalization Applications Technology Committees.