The SIBUR Innovations Research and Development Center has taken a bold step into the future of materials science, rolling out artificial intelligence to automate experimental data analysis. 

 

The move promises to accelerate analytical testing up to seven times, speeding the creation of new materials and components critical for controlling the properties of modern synthetic materials—and bolstering the company’s technological independence.

 

Using Russian software solutions for automated image analysis, SIBUR’s AI models instantly identify objects in material structure and particle morphology images, performing quantitative analysis in seconds. 

 

Tests show these automated results match manual analysis with over 85% accuracy for key parameters. However, experts note that the system’s success hinges on the competence of operators: careful and precise marking during the AI training stage is essential for reliable outcomes.

 

“Digitalization of experimental data analysis allows us to significantly accelerate scientific research and improve its efficiency. Automated image processing reduces the time to obtain results and reduces the impact of human error, which is especially important when working with large data sets. 

 

"This allows our researchers to more quickly test hypotheses and accelerates the development of new materials and technologies,” said Dmitry Afanasyev, Director of Analytical Research and Development at the SIBUR Innovations Research Center.

 

The initiative is part of SIBUR’s broader digitalization strategy, which leverages AI not only for image analysis but also for predicting structure-property relationships, virtual catalyst screening, optimizing synthesis conditions, and reducing lab experiments. 

 

Collaborations with leading universities aim to shrink the time required to design and manage complex scientific and production challenges—strengthening the company’s technological independence while expanding its scientific foundation.

 

With AI at the core of its R&D, SIBUR is moving faster than ever in the race to innovate, proving that the future of materials science is not just in the lab—but in the code.