AI System Helps Optimize Uranium Extraction in Russia
A new decision‑support platform powered by neural networks predicts the effectiveness of repair operations on uranium‑leaching wells, improving planning across Rosatom’s mining sites.
Russian researchers have developed a decision‑support system that helps schedule maintenance work on technological wells used in underground uranium leaching. The creators—graduate student Aleksandr Berezin, associate professor Andrey Istomin, and professor Mikhail Noskov of the Seversk Technological Institute at MEPhI—integrated artificial intelligence into the platform, according to the university’s press service.
Thousands of wells operate across Rosatom’s mining division, and their performance directly influences uranium output. To avoid production slowdowns, wells must be serviced on time.
The new AI assistant generates precise forecasts for the outcomes of different maintenance operations for each well at a given site and suggests the most effective approach.
Neural Networks at Work
The neural network was trained on data from tens of thousands of completed maintenance procedures. The intelligent system is currently used at Dalur, a company developing deposits in the Trans‑Urals uranium district.
Earlier, an innovative system for direct uranium detection was launched in Buryatia. The equipment can determine uranium content in ore bodies without expensive core drilling or extensive geological and laboratory studies.
