An Intelligent Recognition Method of Partial Discharge of Power Cables Based on Data Augmentation

The oscillating wave test to assess the insulation status of power cables and identify faults is widely used. In this paper, based on the periodic type of partial discharge (PD) data, an intelligent identification method of partial discharge of power cables based on data augmentation is proposed. The algorithm uses the empirical mode decomposition (EMD) to filter the original data, and then proposes a translation expansion method to augment the original data, which solves the overfitting problem caused by the insufficient amount of original data. The expanded data are fed into a neural network with convolutional neural network as the main structure for recognition, and finally a weighted composite judgment strategy is proposed. In order to evaluate the feasibility of the algorithm, the corresponding experiments are set up, which show that the recognition accuracy of the algorithm can reach 97%, which is much better than other conventional algorithms.