Fault Diagnosis of Converter Based on Wavelet Decomposition and BP Neural Network

摘要

The modular multi-level converter (MMC) plays an important role in high voltage direct current (HVDC) transmission. Troubleshooting plays a large role in the safe operation of the converter. Therefore, this paper proposes a model based on wavelet multiscale analysis and the BP. After wavelet multiscale analysis of the original fault data, the energy of different levels and different information contained in different frequency bands are extracted. Then the feature information is further input into the BP for adaptive classification. Finally, the result of the fault diagnosis is obtained. The proposed model solves the problems of long training time and poor adaptability in fault diagnosis. Moreover, the proposed model is applied in the fault diagnosis of the MMC of HVDC transmission system, which proves its effectiveness and high accuracy.