Metinvest, headquartered in Ukraine and employing over 80,000 employees, is a leader in the Ukrainian mining and metallurgical industry, a significant player in the global metallurgical market. Aiming to improve the efficiency of blast furnaces by reducing fuel consumption through monitoring the silicon content in pig iron, the company began a pilot project using Azure Data Factory and Azure Machine Learning. Seeing efficiency gains over industry leaders, the company plans to introduce solutions for all of its ovens.
Metinvest is an international vertically integrated mining and metallurgical group of companies. The structure of the group includes mining and metallurgical enterprises in Ukraine, Europe and the USA, as well as a sales network in all major regions of the world. Metinvest controls the entire production chain – from ore and coal mining to the production of semi-finished and finished products.
Improving the efficiency of blast furnaces
At the end of 2018, Metinvest launched a large-scale program aimed at improving the efficiency of its enterprises. “One of the main aspects was to improve fuel consumption in blast furnaces with an expected value added of more than $ 100 million,” explains Kirill Makarov, Director of Continuous Improvement at Metinvest Holding. Indeed, the fuel consumption in a blast furnace depends on several factors, one of which is the silicon content in the cast iron, with an indirect dependence on heating. The higher the silicon content in cast iron, the higher the heating and fuel consumption. However, reducing the silicon content requires an accurate approach, since with a sharp decrease in the silicon content, there is a risk of a temperature drop and a shutdown of the blast furnace. The accuracy of the assessment of the future thermal state of the blast furnace and the corresponding control actions are the key to the stability of pig iron smelting. “Reducing the silicon content by 0.1% can save up to ten kilograms of coke. Thus, we needed to stabilize the blast furnace process and reduce the variability of the silicon content in pig iron, “says Kirill Makarov.
To achieve this goal, Metinvest launched a pilot project using artificial intelligence and machine learning technologies to predict content of silicon in cast iron in the time horizon up to nine o’clock.
Successful deployment of the pilot program
“We use Azure Data Factory as the main tool for organizing the data integration process. Data is loaded regularly, and the model, the one we developed at Azure Machine Learning predicts the silicon content in cast iron and then runs Azure Machine Learning pipelines to help retrieve data and run Python scripts, including those responsible for preparing the data. Machine learning models use this data to prediction of silicon content Prediction results are uploaded to Azure SQL Database “, – explains Alexander Perkhun, Curry Inspector of the Data Management Department of Metinvest Digital. “We use a Power BI dashboard to visualize data, which is updated every hour. Our production specialists have access to the panel and, by adapting key indicators, control the processes in the furnaces in accordance with the forecasts. This helps maintain the silicon level in the desired range, ”explains Vladimir Kravchenko, business transformation expert at Metinvest Holding.
The implementation of the solution was accompanied by comprehensive measures, starting with the adaptation of algorithms for controlling the thermal conditions of the blast furnace to training operators and viewing the goals and motivational component of the shop workers. It took three months to stabilize the thermal conditions and reduce the silicon content in the cast iron. As a result, by the end of the year, the variability of the silicon content decreased from the historical 0.16% to 0.1%, which made it possible to stabilize the silicon content in pig iron and obtain the necessary coke savings. In the near future, Metinvest plans to quickly deploy a turnkey solution for all furnaces at the group’s enterprises.
Accelerating data-driven process management
Metinvest is currently considering machine learning technologies for implementation in most of its manufacturing plants. “For us, this is just the beginning of a digital journey,” says Andrey Trach, director of business transformation at Metinvest Holding. “We intend to embed digital solutions throughout the value chain to enable significant operational improvements at a new level of data-driven process management.”
“Big data is a new asset that creates added value, we perceive it as one of our key business drivers. The complexity of big data analytics requires a strategic cloud map that includes not only infrastructure development and the creation of a set of tools for efficient data processing, but also