Fault Detection and Remedy of Multilevel Inverter Based on BP Neural Network

• Published in: 2012 Asia-Pacific Power and Energy Engineering Conference pp. 1 - 4
• Main Authors: Jiang Wei, Wang Cong, Li Yao-pu, Wang Meng
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.03.2012
• Subjects: Biological neural networks; Circuit faults; Computer architecture; Harmonic analysis; Inverters; Pulse width modulation
• ISBN: 9781457705458; 1457705451
• ISSN: 2157-4839
• DOI: 10.1109/APPEEC.2012.6307658

An open fault detection and analysis method for 7-level Cascaded H-Bridge Inverter based on BP Neural Network is proposed in this paper. A reconfiguration method is also discussed. The output voltage is used as a diagnostic signal to detect the fault types and locations. First, the output voltage is transformed by DFT to select the main harmonic information which is then used to train the neural network. After that, the classification task is performed by a BP neural network. If there is a fault, the reconfiguration system would reconstruct the inverter to make it continually work without affecting the inverter performance. The expected and simulation results are in good agreement with each other, which represents the proposed method can perform satisfactorily to detect the fault types and locations as well as conduct reconstruction.