R. Lee, D. Conley
RJ Lee Group, Inc,
United States
Keywords: AI Artificial Intelligence inclusion identification digital thread
Summary:
The US has become increasingly dependent on foreign-made raw steel which puts DoD weapon system supply chains in jeopardy and places our entire nation at risk. At RJ Lee Group (RJLG), we have a 38-year history of working with the US steel industry to enhance productivity. RJLG’s vision of utilizing Artificial Intelligence (AI) enhanced real-time steel contamination (defects/inclusions) identification during steel making will enhance steel purity and significantly reduce CO2 emissions by up to 30%. This vision integrates real-time AI predictions with advanced imaging techniques and Digital Thread (DT) connectivity to optimize steel production, enhance material performance, and significantly reduce CO2 emissions. Our primary objective is to expand our current IntelliSEM capabilities to develop a prototype AI-Enhanced Steel Inclusion Decisioning Tool with comprehensive capabilities designed to: 1. Utilize Real-Time AI for Inclusion Analysis during the liquid steel stage: Expand algorithms to systematically determine type and characteristics of inclusions in steel 2. Enhance Predictive Models and Enable Systematic Inclusion Remediation: Integrate a customized real-time sparse X-ray mapping into SEM analysis efforts to enhance AI predictions and improve the accuracy 3. Provide Real-Time DT Connectivity: Establish a DT that links data on demand from steel mill to analysis laboratory, facilitating optimization of the process and enabling the production of longer-life materials through advanced materials modeling The RJLG team with ArcelorMittal and Carnegie Mellon University is well-positioned to drive advancements in this sector. This team’s AI-Enhanced Steel Inclusion Decisioning Tool, when completed, would be the first end-to-end industrial implementation of any AI-driven inclusion analysis workflow for the steel industry. About 80% of total energy consumed in the steel industry is used to produce liquid steel. Final steel product quality and resulting yield is mostly determined at this liquid steel stage, where alloy additions, mixing and slag control occur. Inclusion identification, decisioning and remediation is needed quickly during this liquid steel stage to increase steel purity and reduce CO2 emissions before its solidification. The proposed tool will advance the art-of-the-possible in steel making by utilizing AI-enhanced systemic decisioning across the analysis process thereby substantially accelerating analysis tasks and minimizing contaminants. The result will be production of higher quality steel in more rapid fashion (using less energy and less CO2) and raising US domestic steel capacity. This tool would be the first end-to-end industrial implementation of any AI-driven inclusion analysis workflow for the steel industry shop floor and will be an invaluable tool for monitoring and enhancing US capabilities. RJLG’s abstract presents a forward-thinking innovative approach to steel making that leverages real-time AI decisioning, advanced SEM imaging, and DT-centric connectivity to enhance steel quality, reduce CO2 emissions, reduce energy consumption and accelerate the development of high-performance steel materials by avoiding wasteful yield losses (1-2% per published data). The total cost savings possible from a tiny reduction of 1% in yield loss for the whole U.S. steel industry is estimated at $450M/year, stemming from material, energy, labor, and other contributing items.