Jet Infosystems introduces intelligent system to save ferroalloys at Abinsky Electric Metallurgical Plant

2021: Implementation of the intelligent system "Steelmaker's Assistant"

On August 3, 2021, the company "Infosystems Jet" announced that it is introducing the intelligent system "Assistant Steelworker" to save ferroalloys at Abinskom.

An intelligent decision support system for steelworkers is being developed by machine learning IT the technology-based Jet Infosystems center. artificial intelligence At the first stage of the project, the applicability of the system in the electric steel smelting shop (ESP) of the Abinsk Electric Metallurgical Plant was proved to determine the optimal consumption of ferroalloys during the steel smelting process. The second phase of the project will result in the introduction of the system into industrial operation.

The steel smelting process is a very complex and time-consuming process, and steelworkers need to take into account many factors and requirements in order to achieve the specified quality parameters. To obtain steel of a certain grade, the steelmaker needs to add ferroalloys to the molten metal - expensive additives that give the steel the desired properties. Based on his experience, the specialist of the electric smelting shop independently predicts the content of elements in the smelting steel composition and decides on the amount of additives.

Thanks to the experience and qualifications of specialists, the plant manages to ensure high quality steel of various brands. However, the manual mode has its own drawbacks: a relatively large amount of decision time and suboptimal consumption of ferroalloys.

To automate the process, the Abinsky Electric Metallurgical Plant decided to create an intelligent system based on machine learning technologies, which predicts and recommends the exact number of ferroalloys for each smelting. The purpose of the project: preserving the quality of steel, optimizing production by reducing the consumption of ferroalloys.

{{Quote 'Jet Infosystems' team proved the applicability of the recommendation system already at the first stage and set about integrating the ML-based service into the plant's information systems, "says Anatoly Maslov, Director of IT at the management company Newtal-M. - I hope that its implementation will increase not only the quality of the steelworkers, but also the competencies of employees in working with information systems. }}

An intelligent decision support system for steelworkers is still being implemented on a limited number of metal types. She will calculate the minimum necessary additives to get into the requirements of the steel brand, based on the process charts and physicochemical parameters of the final product, and then output hints (forecasts and recommendations) to the system interface. This is a "cruise control" for a steelworker who will not replace the operator, but will become an adviser on further actions. The developed system will also be integrated with the customer's Data Lake for data processing.

The second phase of the project integrates with the customer's source systems, develops interfaces for working with the system, as a result of which the system will be deployed in the production segment. The next stage of development is the accumulation of data, improvement, adaptation, expansion of the functionality and landscape of the applicability of the service.

The recommendation system is a casual development using ML. The experience and competence of our team, as well as deep immersion in the subject area, allows you to quickly understand what the customer needs and how it is possible to solve his problem. I want to note that representatives of the plant provided us with comprehensive support in describing technological processes in production, "says Sergey Ponyaev, project manager from Jet Infosystem. - Positive results of the first stage of work allow to move to the next stage of cooperation with AEMZ.