Abstract:

Partial discharge (PD) detection is an important and effective method to evaluate the insulation status of electrical equipment. The high frequency pulse current will be generated in the ground line of transformer core when PD occurs, and the leakage current of transformer core can also reflect the multi-point ground fault of transformer core. A joint diagnosis method for detecting partial discharge fault of transformer and multi-point ground fault of iron core by using ultra-bandwidth high frequency current transformer is proposed. A partial discharge detection system is built to simulate the partial discharge of the transformer, and two signal pre-processing circuits are designed to extract the high-frequency current and leakage current information of the transformer core ground line respectively. Harmonic analysis is used to extract the fundamental amplitude of the leakage current, and BP neural network based on genetic algorithm is used to classify the fault current of the transformer partial discharge. The accuracy of the model is verified.

Key words: transformer core ground line, partial discharge, leakage current, fault diagnosis