Transient voltage stability advanced discrimination method based on conditional generative adversarial network and transfer learning

The invention discloses a transient voltage stability advanced discrimination method based on a conditional generative adversarial network and transfer learning, and the method comprises the steps: S1, taking a transient voltage stability and instability data set obtained through system measurement as an original sample to train a CGAN network, and completing the directional amplification of a transient voltage sample set; s2, constructing a transient voltage stability prediction model based on transfer learning: inheriting pre-training parameters of a generator in the CGAN by utilizing the transfer learning, and constructing a prediction model; s3, transient voltage prediction based on a prediction model: taking the amplified sample set as a prediction model training set, predicting data within later ten seconds by using data of the transient voltage in the previous second, and performing prediction effect evaluation; and S4, transient voltage discrimination based on the voltage prediction data: according to