An Optimized Fault Feature Extraction Method for PV Grid-Connected T-Type Three Level Inverter

Referencing to the single switch open circuit fault in T-type three level grid-connected inverter, a new fault feature extraction method based on optimal wavelet packet energy feature extraction and FFT technique is proposed in this paper. In this method, the wavelet packet is used to decompose the three-phase fault current in multiple layers, and then the proportion of the low-frequency component energy to the total energy is extracted as the basic fault characteristic value. Optimizing the basic fault characteristic value based on the three-phase current average change trend. Then execute FFT technique on the current of the fault phase, and extract the amplitude of the fundamental wave as an additional fault characteristic value. The fault feature extraction method only needs to analyze the three-phase current, and the algorithm is simple and easy to realize. It can be applied to the fault feature extraction of switch open-circuit fault in most T-type three level grid-connected inverter systems. The simulation results show that the fault feature extraction method combined with neural network can effectively realize the fault diagnosis of single switch open circuit of T-type three level grid-connected inverter.