Focus on Second Open Call winners: FurnAIce5.0 – AI-enabled Decision Support for Smarter Aluminium Melting

WHAT ABOUT THE EXPERIMENT?

FurnAIce5.0 aims to deploy and evaluate an Edge AI system to support aluminium furnace operators in real-time during the melting and casting processes. The experiment focuses on retrofitting existing industrial furnaces with a modular, AI-based decision support layer that enhances process control, traceability, and efficiency.

Through this pilot, GHI Smart Furnaces will integrate the AI REDGIO 5.0 reference architecture into a real production environment to evaluate how its edge computing and human-machine collaboration capabilities can improve decision quality, reduce operator workload, and increase the effectiveness of aluminium recycling furnaces.

The solution will be deployed in a rotating tilting furnace, commonly used for melting aluminium scrap, where key decisions such as charging strategy, fusion time, and casting moment are often based on operator experience. By introducing Edge AI, the experiment seeks to standardize best practices, reduce variability, and improve key operational KPIs.

The solution includes the capture and processing of furnace signals on edge devices, real-time inference of process stage and AI-guided recommendations, and traceability of operator actions to evaluate system impact. The results of the experiment will provide valuable insights into the adaptability of the AI REDGIO 5.0 architecture and its potential for industrial-scale deployment in SMEs.

WHAT IS THE EXPECTED IMPACT?

● Increase in correct charging decisions from 70% to ≥95% through AI-assisted recommendations.

● Reduction in manual checks from 15/h to ≤5/h, decreasing operator overload and response time.

● Boost in OEE from 60% to 85% by reducing inefficiencies and enhancing consistency.

● 100% traceability of charging and casting events supported by AI decision logs.

● Launch of at least two AI-augmented furnace models in the GHI product portfolio, supporting market differentiation.

● Practical evaluation of AI REDGIO 5.0 reference architecture in a real industrial setting.

The pilot will serve as a replicable reference for other SMEs aiming to adopt AI-at-the-edge solutions in the metal industry