At the National Scientific Conference "Control, Automation and Artificial Intelligence in Smart Grids", Assoc. Prof. Dr. Dinh Van Chau - Rector of Electric Power University - raised the key question: How to combine control, automation and AI so that the power grid has the ability to observe more deeply, forecast earlier, operate more flexibly but still ensure stability, safety, network security and accountability in the decision-making process?
Also according to Assoc. Prof. Dr. Dinh Van Chau, a system with many sensors and software does not necessarily mean a smart grid. The level of intelligence must be reflected in operating results: Safer, more reliable, more economical, more flexible, more efficient use of resources and better resistance and recovery when incidents or unusual conditions occur.
One of the most important values of AI in building smart grids is the ability to analyze big data to make forecasts and support decision-making. AI has the ability to analyze data from measuring systems, sensors, monitoring devices, weather data, operating history to forecast load demand, renewable energy output, overload risk or abnormal signs.
In the field of renewable energy, this is a particularly important tool. Solar and wind power depend heavily on weather conditions. AI can help forecast power generation capacity, thereby supporting dispatchers to proactively balance supply and demand, reducing the risk of system instability.
Through analyzing data on temperature, vibration, current, and voltage of the device, AI can identify unusual signs, warn of the risk of damage before an incident occurs. This helps reduce power interruption time, reduce repair costs and improve system reliability.
Another application direction is to combine AI with digital clones (Digital Twin). This is a technology that creates digital models simulating physical systems, allowing testing of operating scenarios, risk assessment and optimization before deployment in practice.
According to experts, AI and Digital Twin need to be developed in the direction of combining physical models and data, capable of quantifying uncertainty, detecting abnormal data, saving traces and ensuring human intervention capabilities.
At the conference, scientists introduced research on "Continuous decentralized learning for testing equipment on power transmission lines using UAVs" and solutions "Applying UAVs and AI problems in monitoring smart power system operation".
This is proof of the trend of combining many new technologies to build an electrical system with deeper monitoring and faster response capabilities. According to Assoc. Prof. Dr. Dinh Van Chau, AI cannot replace physical principles, operating procedures and human responsibilities. The application of AI in the electricity industry needs to be accompanied by a tight management system, in which technology plays a supporting role, while humans are still the ultimate responsible actors.
The application of AI in building smart grids is no longer a story of the distant future but is becoming an urgent requirement in the process of modernizing the electricity industry. This is consistent with the orientation in Resolution 57-NQ/TW on developing science, technology, innovation and digital transformation; Resolution 70-NQ/TW on developing synchronous and smart energy infrastructure; as well as Adjusting Power Plan VIII with the goal of building smart grids capable of integrating large-scale renewable energy sources.
