OJSC Magnitogorsk Iron and Steel Works ("MMK" or “the Company”) announces that it has signed an agreement with Yandex Data Factory (division of Yandex, https://yandexdatafactory.com) to develop a mathematical model for steelmaking in MMK’s oxygen-converter plant (“the OCP”), and to use this as a basis for developing related software. The project, known as “Sniper”, aims to optimise the consumption of ferroalloys and supplementary materials during steel production at the oxygen-converter plant.
IT specialists from Yandex Data Factory and MMK-Informservice (an MMK Group subsidiary) together with specialists of the OCP and research and development centre at OJSC MMK are currently analysing information covering the OCP’s operations over previous years. A Machine Learning algorithm will be incorporated into the mathematical model that is being developed based on corporate information available on the Apache Hadoop platform. The final software solution will receive data on the original composition, original weight and requirements regarding the chemical composition of the steel produced, and the mathematical model will help it indicate the optimal amount of ferroalloys and supplementary materials to be used during steelmaking. The Sniper project employs Big Data technologies – processing large sets of data in order to improve management process and optimise production.
The final stage of the project is expected to include a quality check involving a series of experimental steelmaking processes in which additive composition is based on the system’s recommendations.
“MMK is the first heavy industry client for Yandex Data Factory. Our joint project is aimed at the maximum optimisation of the production process. We expect to achieve a significant impact from raw materials savings. The implementation of this pilot project will make it possible to expand our cooperation with Yandex and extend it to the Company’s other production facilities,” says Sergey Sulimov (Deputy CEO for Finance and Economy at OJSC MMK.