Russian Nuclear Engineer Trains AI to Monitor Power Plant Valve
At Novovoronezh Nuclear Power Plant, a young engineer has created an AI system that can predict valve positions even when hardware fails—boosting safety and reliability in nuclear energy.
Ivan Makarov, an engineer at the Novovoronezh Nuclear Power Plant, has developed an artificial neural network that precisely determines the position of a control valve, even in cases of equipment malfunction. His project, which won first place in an internal competition at the plant, addresses the critical issue of data inaccuracy during hardware failures—an especially important factor for safe and stable electricity generation. The innovation highlights how Russia’s nuclear industry is advancing not only through new construction but also through digital transformation of existing infrastructure.
Makarov’s solution, developed in the plant’s thermal automation and instrumentation shop, accounts for real operating conditions such as mechanical wear and vibration. Unlike the current mathematical model, which works only under ideal circumstances, the neural network analyzes multiple factors to more accurately predict the valve’s true position. This substantially improves data reliability and overall equipment performance.
The Russian government has been actively supporting innovative projects from young specialists in the nuclear sector. This year, 12 projects were submitted to the competition jury, with six advancing to the final round.
