Leveraging AI-powered large language models to improve operational safety and efficiency in the metal and steel industry

Abstract

The metal and steel industry faces persistent challenges in enhancing operational safety, efficiency and complex decision-making processes. This article explores the transformative potential of generative artificial intelligence (AI), with a focus on multimodal large language models (LLMs), to address these challenges. Unlike traditional predictive and prescriptive AI, generative AI enables new possibilities by integrating and synthesizing unstructured data (text, images, video) with domain-specific knowledge.

Two case studies highlight practical applications: (1) A Vision AI system leveraging integrated LLMs to monitor electric arc furnace operations, identifying key operational events and potential safety hazards; (2) A generative AI approach for project scheduling optimization, achieving near-optimal performance for small- to medium-sized projects. The study emphasizes generative AI’s ability to enhance decision automation, reduce reliance on manual oversight, and drive innovation in safety and efficiency. Challenges regarding accuracy, scalability and consistency are discussed, alongside recommendations for future advancements in agent-based refinement techniques.