Power system stability assessment method based on GAN and GRU‐Attention using incomplete voltage data

The social economy is growing rapidly, and the power grid load demand increasing. To maintain stability of grid, it crucial to achieve accurate rapid system assessment. In actual operation network, data loss an unavoidable situation. However, most data-driven models currently used assume that input complete, which has obvious limitations in real-world applications. This paper suggests IVS-GAN model assess using incomplete phasor measurement unit with random loss. proposed method combines super-resolution perception technology based on generative adversarial network (GAN) a time-series signal classification model. generator adopts 1D U-Net uses convolutional layers complete recover missing data. discriminator new gated recurrent unit–attention architecture here better extract voltage temporal variation features key buses. result this evaluation outperforms other algorithms high rates New England 10-machine 39-bus system.