Diagnosis Method for Partial Discharge Faults in Power Cables Based on CEEMDAN-IZOA-DBN

Power cables can be regarded as the main artery of the urban power supply system, which has an important supporting role for the operation of the power system. As an important indicator reflecting the insulation status of cables, partial discharge faults in power cables can cause damage to the insulation of cables, which may cause major power outages in serious cases. Since the traditional PRPD mapping analysis has been difficult to meet the demand for large-scale diagnosis of partial discharges in the complex power cable network underground in cities, this paper proposes a partial discharge fault diagnosis method for power cables based on CEEMDAN-IZOA-DBN. Firstly, Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) is used to decompose the collected partial discharge signals of power cables, and the signal features are extracted according to the set feature indicators. Secondly, for the problem of low adjustment efficiency and easy to fall into local optimum during the learning process of Deep Belief Network (DBN), the zebra optimization algorithm (ZOA), which is an improved hybrid strategy combining chaos theory and mutualistic symbiosis strategy, is used to optimize the parameters of the DBN network, so as to obtain the partial discharge fault diagnosis model of power cable based on IZOA-DBN. Finally, the extracted features data of power cable partial discharge signal are input into IZOA-DBN for data testing and analysis, and the test results show that the fault diagnosis accuracy of the model is as high as 94%, and the method has a high computational efficiency. The proposed method can effectively assist technicians to carry out partial discharge fault diagnosis of power cables.