Multiscale Adaptive Fault Diagnosis Based on Signal Symmetry Reconstitution Preprocessing for Microgrid Inverter Under Changing Load Condition

In this paper, a multiscale adaptive fault diagnosis (MAFD) method based on signal symmetry reconstitution preprocessing (SSRP) is proposed to realize fault diagnosis for any switch of the microgrid inverter under changing load condition. The MAFD method is composed of SSRP, multiscale features extraction, and artificial neural network (ANN). First, the SSRP method is used to generate the input signals of multiscale features extraction, which can reduce the impact of the changing load. Then, the multiscale features extraction is realized by the means of multilevel signal decomposition and coefficients reconstruction to extract energy content of different frequency groups signal. It can represent the detailed signal change laws at different levels for three-phase current. Finally, in order to achieve data-based adaptive fault diagnosis, ANN is used to detect the type and the location of the inverter switch fault. Compared to conventional fault diagnosis methods, the proposed fault diagnosis method can accurately detect and locate fault for any switch of the microgrid inverter under changing load condition. The effectiveness of the proposed fault diagnosis method is verified through detailed simulation and experimental results.