High Impedance Fault Detection Method Based on Multi-source Information Fusion and CNN

Multi-source information fusion has been applied for fault detection in distribution network which can fuse data from different sensors according to certain rules to fully explore fault characteristics. Compared with only using a single sensor data for fault detection, it is conducive to achieving more accurate judgment and decision-making. The paper develops a high impedance fault (HIF) detection method based on multi-source information fusion and CNN. Considering the features implied by the current and voltage data are not exactly the same, the combination of current and voltage is used in the proposed method to increase fault detection accuracy. The method first uses continuous wavelet transform to extract current and voltage features, then designs neural network to realize multi-source data fusion of feature layer, and finally combines with classifier to realize fault detection, and the effectiveness of proposed method is verified by a large amount of simulation data.