MA Li-xin, WU Xing-feng, FEI Shao-shuai

(College of OpticalElectrical and Computer Engineering, University of

Shanghai for Science and Technology, Shanghai 200093, China)

 

Abstract: Aiming at the easily damaged characteristic of IGBT in the inverter, simulation, fault classification, feature extraction and fault diagnosis of IGBT open circuit fault in active power filter(APF) were researched. After the analysis of IGBT open circuit fault and fault feature, a method was presented to diagnose the fault of APF based on fast fourier transform (FFT) and neural network, the BP network was trained with fault feature vector, and the welltrained neural network was tested by IGBT open circuit fault. The experimental results indicate that the diagnosis method can monitor the condition of APF and effectively identify the fault location of IGBT.

Key words: active power filter(APF); fault diagnosis; fast fourier transform(FFT); neural network