NLMK and Jet Infosystems increased gas consumption efficiency using artificial intelligence

2023: Development of a service to improve gas consumption efficiency

To improve energy efficiency, IT specialist enterprises power engineering specialists NLMK and together with specialists "" Jet Infosystems have developed a service that issues recommendations to the TPP boiler driver: when, in which boilers and in what proportions to burn various types. gas This was reported to Jet Infosystems on June 29, 2023.

At NLMK, by-products of iron and coke production - blast furnace and coke gas - are captured, cleaned and used to generate their own electricity. In order for the caloric content of the resulting mixture to be stable, that is, to give a sufficient amount of heat for the production of energy in the form of steam, natural gas is added to it.

This digital solution allows you to more accurately determine the proportions for mixing different types of fuel. The machine learning model analyzes historical data for several years, calculates the ratio of energy resources and gives recommendations for its optimization.

Earlier, the decision on the volumes and proportions of the use of various types of gas for the generation of thermal energy was made by the operator of the CHPP boiler unit based on the readings of boiler equipment devices, among which there were no indicators characterizing the efficiency of the unit.

Thanks to our joint work, it was possible to tie together the main parameters characterizing the efficiency of boiler equipment and take another step to optimize the operation of the main equipment of the thermal power plant in metallurgical production, "said Ivan Morev, chief specialist in digital projects of the Management Directorate of the Energy Complex of NLMK PJSC.

The service for optimizing energy consumption is operated by NLMK and provides savings of 1-3 million rubles per month compared to the corresponding period in 2022, when the CHPP operated without using AI. Annual savings due to the efficient use of energy resources are estimated at tens of millions of rubles.

The creation of a service for saving energy resources at a thermal power plant is an example of a project in which significant savings are achieved using machine learning, "said Kirill Minaev, a specialist at the Jet Infosystems Machine Learning Center. - Its development required a complete immersion in the heat and energy processes of the equipment under study and a deep analysis to identify dependencies between technological parameters.