Severstal: Scientific Footprint Search and Analysis System

Main article: Data mining Data mining

2023: Developing an algorithm to find relevant patent information

Severstal is building its own search and analysis system. The company's experts have developed an industry-specific algorithm for finding relevant information on patents, scientific articles and the news stream called the Science Trail. The solution can replace foreign systems that have left the Russian market. Severstal announced this on December 25, 2023.

Employees of the Directorate for Technical Development and Quality of the Severstal Russian Steel Division "" created their own search and analytical system two years ago. A scientific trace is a set of tools for quickly searching for knowledge and constructing hypotheses. Such a system is in demand in the implementation of high-tech and knowledge-intensive projects requiring patent analytics, applicable in the development of new types of products and solving complex production problems. Its users can be employees R&D of subdivisions and technologists of enterprises, research institutes and centers.

Over two years, we have confirmed the demand and effectiveness of such a service among Severstal employees. More than three thousand requests are received monthly. We saw a real financial effect on reducing the time for the development and launch of new steels on the market. In addition, the "Scientific Trace" allows you to save time on the study of the technical task - up to 35%, on the selection and involvement of an external expert on the topic - up to 40%. We associate the further development potential of this service with the entry into the domestic market of search and analytical systems for use in other industries industries of Russia, "said Petr Mishnev, Director for Technical Development and Quality of Severstal Russian Steel.

The system is able to recognize the context of published articles and existing patents. Due to the applied technologies, where the search engine offers 100 articles, Severstal's proprietary classifier will reduce the selection of relevant publications to 20.

"Scientific Trail" offers analytics of the development of the investigated technological topic, allows you to find external experts with the required competencies based on patent activity, selects patents and publications in the field of materials science on the chemical composition of the product, taking into account the minimum and maximum content of elements. In comparison with competitors, the solution covers a complete list of user tasks in knowledge-intensive projects.

As of December 2023, the system combines 40 million articles and patents on metallurgical topics. In 2024, the project team plans to expand knowledge bases to other areas and increase the volume to 40 million patents and 80 million scientific publications.