Russian steelmaker Magnitogorsk Iron and Steel Works (MMK) will save more than £3 million in steelmaking costs by using a bespoke machine learning and big data analytics service from Yandex Data Factory.
MMK produced 12.2Mt of crude steel and 11.2Mt of commercial steel products in 2015, meaning that even small efficiencies in raw material costs would yield enormous savings. Yandex Data Factory created a machine learning service that uses seven years of MMK’s granular steelmaking records, with more being added, to predict the optimal combination of ferroalloys needed to produce specific steel grades with international standard chemical compositions, and at the lowest cost for each specific smelting. These insights and practical recommendations enabled MMK to achieve an average decrease of 5% in ferro alloy use, equating to annual savings of more than £3 million, while maintaining the high quality of steel produced, claims MMK.
The Yandex Data Factory service receives data on the composition of the working mixture, compares it to the required chemical composition of the output steel, and then uses the data on MMK’s historical smeltings to make real-time recommendations on the amount of ferro alloy and supplementary materials needed to produce the necessary end products at lowest cost.
Sergey Sulimov, deputy CEO for finance and economy at MMK said that the Russian steelmaker was a pioneer among the industrial companies of Russia in the application of digital technologies. "Our work with Yandex Data Factory is a new wave in industrial automation with the use of big data analytics,” he said . “We believe that the possibilities of mathematical models using big data analytics, as well as the rapid development of IoT technologies, will reduce the costs of industrial companies by 5-10% over the next 3-5 years.”
Alexander Khaytin, chief operating officer of Yandex Data Factory, said that MMK’s forward thinking in the use of machine learning is a core reason for the successes it has delivered. "We are looking forward to continuing our work together,” he said, adding that the commercial value of the project shows the importance of moving from talking about the better, but abstract, tomorrow of the industry to practical problem solving. "We are confident that others will see this as an example of leading the way in an industry with the application of new technologies," he concluded.