13 July ‘16

MMK employs Yandex Data Factory solution

Preliminary tests indicate that this new solution can deliver average savings of 5% on ferroalloys. Annual savings may potentially exceed RUB 275 mln.

Yandex Data Factory’s solution incorporates a machine learning algorithm and big data analysis. It optimises the consumption of ferroalloys and supplementary materials during steel production at the oxygen converter plant. It combines data on the original composition and mass of the burden on the required chemical composition of the steel produced with data on other production parameters. It provides operators with recommendations on the use of ferroalloys and additional materials in real-time. The project, known as Sniper, is aimed at producing steel with the target chemical composition at the lowest possible cost.

Sniper is an MMK project that uses big data technologies – processing large sets of data to improve the management process and optimise production, and has been in development since 2015. MMK-Informservice (an MMK Group subsidiary) developed the data collection system and user interface, and Yandex Data Factory developed the service providing recommendations.

“MMK is a pioneer among industrial companies in Russia in terms of digital technologies. Our joint project with Yandex Data Factory is part of a new wave in production automation involving big data. We believe that mathematical models using big data combined with the dynamic development of the Internet of Things will make it possible for industrial companies to cut costs by 5-10% over the next 3-5 years,” says MMK’s Deputy CEO for Finance and Economy Sergey Sulimov.

“Yandex Data Factory works with clients from various industries, and it was a pleasant surprise for us when a leading company in the metal sector expressed interest in adopting new technologies. Despite perceived conservative views, the metals industry, which relies heavily on technologies and engineering, in fact proved more flexible than many other sectors.”

“We appreciate MMK’s openness to adopting new technologies in this sector. It is thanks to this interest, and to the machine learning technologies themselves, that the project has been a success. We intend to further develop our cooperation.”

“The experience we gained in the course of our joint project demonstrates the importance of moving from rhetoric about the sector’s bright future to practical solutions in a timely fashion. We hope other companies and industries will follow MMK’s example and rather than wait for the application of new technologies to become clear, that they start using these technologies today,” said Chief Operating Officer of Yandex Data Factory Alexander Khaytin.

Notes for editors:

MMK is one of the world's largest steel producers and a leading Russian metals company. The company's operations in Russia include a large steel-producing complex encompassing the entire production chain, from the preparation of iron ore to downstream processing of rolled steel. MMK turns out a broad range of steel products with a predominant share of high-value-added products. In 2015, the company produced 12.2 million tonnes of crude steel and 11.2 million tonnes of commercial steel products. MMK Group had sales in 2015 of USD 5,839 million and EBITDA of USD 1,668 million.

Yandex Data Factory (YDF) is an international division of Yandex aimed at providing solutions for companies dealing with large sets of data. Yandex technologies – machine learning, image and speech recognition, deep neural networks, natural language processing – make it possible to analyse large volumes of data and identify the optimum solution. In addition to the technologies needed to process Big Data, Yandex employs the leading experts in this field.

Investor Relations Department:

Andrey Serov, Head of IR

tel.: +7 (3519) 24-52-97

E-mail: serov.ae@mmk.ru

Communications Department:

Sergei Vykhukholev

tel.: +7 (499) 238-26-13

E-mail: vykhukholev.sv@mmk.ru

Dmitry Kuchumov

tel.: +7 (499) 238-26-13

E-mail: kuchumov.do@mmk.ru

This site uses cookies as described in our Privacy Policy and Terms of Use