Induction furnaces often rely on manual or spreadsheet-based charge calculations, leading to variability in chemistry control, inefficient alloy use, and costly remelts. This study presents Intelomelt, a smart charge optimisation platform that integrates linear programming (LP) and AI-driven metallurgical assistance to improve melting efficiency and consistency.

The system optimises charge composition using available scrap, returns, and alloy inventories while meeting multi-element chemistry constraints and grade-specific requirements. A spectro-based trim correction module calculates the most cost-effective alloy additions to achieve target chemistry, reducing operator dependency and decision variability. Additionally, an embedded AI assistant provides real-time guidance for heat analysis, optimisation strategies, and process decisions.

By digitising the furnace workflow and enabling data-driven decision-making, the platform improves repeatability and process control. Industrial applications indicate measurable reductions in off-chemistry heat, alloy consumption, and overall melting costs, demonstrating the potential of digital technologies to enhance productivity and sustainability in modern foundries.