• Home|About JOPS|Editorial Board|Contact us|Chinese
SONG Pinggang,ZHANG Wei,CHEN Huan,LIN Jiatong,ZHOU Zhenbang.Open-circuit Fault Diagnosis of Inverter Transistor Based on Mathematical Pattern Spectrum[J].JOURNAL OF POWER SUPPLY,2018,16(5):159-166
Open-circuit Fault Diagnosis of Inverter Transistor Based on Mathematical Pattern Spectrum
Received:June 16, 2016  Revised:January 07, 2018
View Full Text  View/Add Comment  Download reader
DOI:10.13234/j.issn.2095-2805.2018.5.159
Keywords:inverter  pattern spectrum  fault diagnosis  mathematical morphology
Fund Project:
              
AuthorInstitutionEmail
SONG Pinggang School of Electrical and Electronic Engineering, East China Jiaotong University, Nanchang , China
ZHANG Wei School of Electrical and Electronic Engineering, East China Jiaotong University, Nanchang , China 735055749@qq.com
CHEN Huan School of Electrical and Electronic Engineering, East China Jiaotong University, Nanchang , China
LIN Jiatong School of Electrical and Electronic Engineering, East China Jiaotong University, Nanchang , China
ZHOU Zhenbang School of Electrical and Electronic Engineering, East China Jiaotong University, Nanchang , China
Hits: 59
Download times: 80
Abstract:
      In consideration of the open-circuit fault in an inverter transistor, a diagnosis method for the inverter's open circuit fault is proposed based on mathematical pattern spectrum. When the inverter is in a fault state, the differences in the three-phase output current waveform, such as positive half-wave missing, negative half-wave missing, positive and negative half-waves missing, and serious distortion, are utilized to obtain the pattern spectrums of normal current signal and the current signals under the above four types of fault, of which the first 11 normalized spectra are extracted as feature vectors and further input to the ELM neural network to perform classification. With the combination of simple three-phase fault diagnosis criterion, the fault transistors can be positioned precisely. Simulation results show that by using the proposed method, the fault diagnosis rate is higher than 96%. In the end, the experiment data were used to verify the effectiveness of the proposed method.
Close