Severstal deploys AI at Cherepovets steel mill

Russian metals firm Severstal has started using a machine learning model to control speed optimisation at the Cherepovets steel mill’s continuous pickling line.
Severstal deploys AI at Cherepovets steel mill Severstal deploys AI at Cherepovets steel mill Severstal deploys AI at Cherepovets steel mill Severstal deploys AI at Cherepovets steel mill Severstal deploys AI at Cherepovets steel mill

The Cherepovets steel mill

The new system was created by integrating "Adelina", a digital model already in use at the NTA-3 pickling line since November 2019, with "Ruban", a new AI agent.

Ruban uses a "deep - reinforcement learning algorithm" —a relatively new technique in which neural networks learn by trial and error.

Adelina controls the speed of the unit, while Ruban adjusts the speed to achieve optimal results.

"The Adelina model had already met our expectations, demonstrating an initial increase in productivity of NTA-3 by more than 5%. In March 2020, we produced a record volume of pickled metal at this unit - more than 130,000t", said Evgeny Vinogradov, chief executive of Severstal Russian Steel Division.

"After introducing the Ruban agent, we recorded a further 1.5% increase in productivity, and we estimate that using the two technologies in parallel could provide more than 80,000t of additional metal each year. This is a remarkable increase for one of the most significant units in the production of flat rolled products."

The company noted that Ruban differs from classic machine learning models in that it learns not just from historical data, but also exploring the digital twin of NTA-3.

The operating speed at the unit largely depends on the parameters of the passing steel strip - the length, width and thickness of the roll, its steel grade and temperature, among other factors.

"Ruban learns from combinations of different parameters, specifically created for it by a generative adversarial network, which uses two neural networks to generate new data. It also sets a production plan and creates unique situations for training purposes," said the company.

For effective learning, the agent was assigned a training system based on rewards and penalties; Ruban experiments to find a solution where the reward amount surpasses the penalties as far as possible.

Boris Voskresenskii, chief digital officer of Severstal, added: "The use of reinforcement learning to control production units is not widespread, particularly in metallurgy. We believe the use of Artificial Intelligence at NTA-3 to be the first such case in Russian practice. The performance improvement recorded on NTA-3 following the introduction of digital tools proves that a data.