Distributed photovoltaic station data enhancement method based on spatial-temporal feature clustering

The invention discloses a distributed photovoltaic station data enhancement method based on spatial-temporal feature clustering. The method comprises the following steps: step 1, collecting existing historical measurement data of a distributed photovoltaic station; step 2, building an auto-encoder model to mine time sequence characteristics of each distributed photovoltaic station, and designing an encoder and a decoder capable of realizing the auto-encoder model; 3, introducing a graph convolutional network to improve an auto-encoder, and mining spatial features between stations; 4, taking data of other stations belonging to the same class as the station to be subjected to data enhancement as input data of the step 6; 5, establishing a data enhancement model based on the conditional generative adversarial network, and taking the generated data as input data of the model in the step 6; and step 6, taking data obtained by the trained XGBoost as enhanced data. According to the method, the problem of insufficien