SIBUR accelerates analytical research with artificial intelligence

26 Februaty 2026. The SIBUR Innovations R&D Centre has begun implementing artificial intelligence to automate the analysis of experimental data. With the help of AI tools, analytical testing will be accelerated up to seven times. This will bring about faster development of new materials and special components essential for flexibly managing the properties of modern synthetic materials and ensuring the industry's technological sovereignty.

For automated image analysis used in researching material structures and particle morphology, SIBUR relies on Russian software solutions. Trained models automatically identify objects in images and perform quantitative analysis within seconds, ensuring high reproducibility of results. A comparison of automated and manual analysis showed a high degree of result alignment – over 85% for key quantitative parameters. However, the success of AI application depends on the competence and professionalism of its human trainer and operator. In this case, the initial stage is crucial, requiring the specialist to be thorough and precise in data labelling, as this directly impacts the model training outcomes.

Dmitry Afanasyev, Director of Analytical Research and Development at the SIBUR Innovations R&D Centre:

"Digitalisation of experimental data analysis allows us to significantly speed up scientific research and enhance its efficiency. Automating image processing reduces the time required to obtain results and minimises the influence of human factors, which is particularly important when working with large datasets. This enables our researchers to test hypotheses more rapidly and accelerates the development of new materials and technologies."

Rolling out digital solutions in research and development is part of SIBUR's comprehensive strategy. The Company is actively testing and introducing artificial intelligence methods across its scientific processes, including for predicting structure-property relations, virtual catalyst screening, optimising synthesis conditions, and reducing the number of laboratory experiments – all of which should substantially accelerate the development of new products and materials. Research is conducted in collaboration with leading universities, and the approaches developed are designed to radically shorten the time required for the design process and management of complex research and production tasks, strengthening technological independence and expanding the Company's scientific foundation.