Stability analysis of power grid based on generative adversarial network

Abstract:
The stability of power grid is an important guarantee to ensure the safe operation of the power system. In view of the shortcomings of the traditional analysis and calculation methods of power grid stability, such as complex processing flow and unable to judge the large-scale whole quickly, combined with the key data feature extraction method and deep learning method, a power grid stability detection and recognition method based on generative adversarial network is proposed in this paper. Compared with the previous neural network models, the generated countermeasure network perfectly balances the recognition time and recognition accuracy of the system, and can realize efficient, lossless and fast diversified anomaly detection of large-scale power grid operation state. According to the experimental results of traditional learning methods and existing models, the proposed method has good real-time performance, robustness and recognition rate, and the recognition accuracy is 95%.