Priority 2030: Moscow Polytechnic University is developing an AI-based system to predict power grid failures
The project's primary objective is the early detection of problems. The AI-based power grid management system will be able to predict failures and optimize fuel consumption. TASS reports on the project, which received support through Moscow Polytechnic University's grant competition named after V.E. Fortov.
"We process a large volume of information to monitor and diagnose power systems. This enables us to accurately forecast energy resource needs, which helps us plan power grid loads more effectively," explains Valeria Kolishchak, the author of the project and a senior lecturer in the Industrial Thermal Power Engineering Department at Moscow Polytechnic University.
The system uses artificial intelligence and machine learning to analyze large volumes of power grid data. Algorithms process information about equipment operation, loads, energy consumption, and other parameters in real time. Based on this data, the system will forecast energy demand, identify anomalies in equipment operation, and warn of potential failures.
The project's primary objective is early problem detection. The system will analyze equipment behavior and identify deviations from normal parameters. This should allow for the detection of malfunctions before they lead to an emergency.
The second key function is energy system optimization. Neural networks will learn to redistribute power flows and balance loads to reduce transmission losses and fuel consumption. The system will plan capacity utilization based on consumption forecasts and the current state of equipment.
Use Prospects
The development includes several stages. Further testing is planned for real-world applications. The test site will be power grid sections supervised by the Moscow Analytical Center for Urban Management. This should allow the system's operation to be verified under real-world conditions on small sections of Moscow's energy infrastructure.
According to Valeria Kolishchak, the system should be able to operate with various types of power plants—from boiler houses and heating stations to transformer substations. The algorithms will be adapted to the specifics of each facility, taking into account its technical characteristics, operating mode, and load characteristics. This is necessary because the energy infrastructure is heterogeneous: equipment of varying ages, capacities, and operating conditions vary.
The implementation of such systems can significantly reduce the number of accidents and unplanned equipment shutdowns. Furthermore, optimizing energy systems saves fuel and reduces transmission losses, which is especially important for large urban heating and electricity systems.
If testing at Moscow facilities demonstrates the technology's effectiveness, it can be scaled up to other regions. The system is being designed to adapt to various climatic conditions and types of energy infrastructure.
Photo: https://ru.freepik.com/
Background: The grant competition named after V.E. Fortov is being held as part of the federal program "Priority 2030," in which Moscow Polytechnic University is participating with the strategic projects "Modular Platform for Compact Electric Vehicles" and "Localization of Automotive Components."
