Case Study: Metinvest achieves blast-furnace efficiency and fuel savings with Microsoft Azure

Metinvest achieves blast-furnace efficiency with Azure Machine learning

Metinvest, a Ukraine‑headquartered, vertically integrated steel and mining group with over 80,000 employees, set out to improve blast‑furnace fuel efficiency by controlling silicon content in cast iron—a key factor in furnace heating and coke consumption. To reduce variability without risking furnace instability, the company launched an AI pilot using Azure Data Factory and Azure Machine Learning to predict silicon levels up to nine hours ahead.

The solution uses Azure Data Factory for orchestration, Azure ML models and pipelines to generate predictions, Azure SQL to store results, and an hourly Power BI dashboard for operators; implementation also included algorithm tuning, operator training, and incentive changes. Within three months silicon variability fell from 0.16 to 0.10, delivering targeted coke savings, and Metinvest is rolling the solution out across its furnaces while expanding ML use across its operations.

Metinvest

Olga Ovchinnikova

Economics and Business System Development Director