Russian Students Develop AI Platform to Predict Concrete Properties
Researchers at Lomonosov Moscow State University have developed an AI platform that predicts the strength and workability of concrete for up to 90 days after mixing based on the composition of the concrete mix.
Students and postgraduate researchers at Lomonosov Moscow State University have developed the AI-powered MaterialsGen platform to create new concrete formulations. The system analyzes the composition of a concrete mix and determines within seconds when it will reach the required performance characteristics, replacing laboratory testing that traditionally takes days or weeks.
Concrete mixtures are traditionally tested over a 28-day period before engineers measure their final compressive strength. The industry faces a shortage of specialists with the necessary expertise, while many laboratories still operate without digital tools.
According to Smirnov, manually developing a concrete formulation typically takes about a month, while prediction accuracy does not exceed 64%.
The platform can also generate new concrete formulations tailored to specified performance requirements. According to the university, the technology could deliver economic benefits of up to 41.5 million rubles (approximately $530,000) per year. The project was recognized among the leading entries in the Shanghai Cooperation Organization innovation competition and received support from the Intellekt Foundation.
