Transient Frequency Stability Assessment Method for Power Systems Based on the Integration of GAN and CNN

With renewable energy accounting for an increasing portion of power systems, system operating modes and physical characteristics have become increasingly complex, thereby imposing more stringent requirements on transient frequency stability. In this context, this study proposes a transient frequency stability assessment approach combining data augmentation with deep learning techniques. First, a generative adversarial network is used to expand the original dataset, effectively expanding the training samples available for deep learning models. Subsequently, a frequency stability assessment a model integrating a Convolutional Neural Network with a Long Short‐Term Memory network is constructed and trained. Finally, the effectiveness of the proposed method is evaluated using the IEEE 39 system. The experimental findings demonstrate that the proposed approach achieves accurate transient frequency stability prediction while exhibiting strong robustness.