Research on Power Balance and Measurement of New Energy Power System Based on Graph Neural Network

With the surge of new technologies, such as high penetration of new energy, ultra-high voltage transmission, and intelligent digital power grids, the power system has become increasingly complex and requires stricter safety and stability standards. To address this issue, this paper introduces a transient stability analysis method using graph convolutional neural networks. This method combines short-term simulation with neural network prediction, reducing analysis time and making it suitable for various simulation scenarios. It also combines models with algorithms to quickly and robustly optimize transient stability control strategies for expected faults. This method is superior to traditional methods in terms of runtime and efficiency. The test results of IEEE-30 and IEEE-39 node systems confirm the effectiveness, efficiency, and superiority of our proposed method. In high-tech energy systems, the volatility and uncertainty of wind and photovoltaic power generation output significantly affect power balance. The increasing renewable energy production capacity in China poses a challenge to reliable electricity supply. Based on actual data analysis of daily and seasonal fluctuations in new energy output, we have summarized power balance issues at different time scales. Using a time series production simulation model, the balance problem in high-tech energy systems was studied, and the supply shortage of typical power grids was quantitatively analyzed. Solutions were proposed, with the daily peak fluctuation of new energy in the power grid reaching 79.96 million kilowatts, an increase of 41% compared to the previous year.