Data enhancement partial discharge detection method based on generalized S transformation

The invention relates to a data enhancement partial discharge detection method based on generalized S transformation, and belongs to the field of partial discharge detection. The method comprises the following steps: carrying out generalized S transformation of an initialization parameter on a sampling signal to obtain a time-frequency distribution grey-scale map of a partial discharge signal; inputting the obtained time-frequency distribution grey-scale map into a lightweight neural network, and calculating the average loss value and the accuracy rate of the model after each round of training; constructing a multi-target nonlinear optimization problem with the minimum average loss value and the maximum average accuracy; the optimal solution for solving the optimization problem is used for updating parameters of generalized S transformation and used for model training of the next round; deploying the trained lightweight model to micro MUC hardware for online identification of partial discharge of electrical e