Fault Detection and Identification Using Neural Network and Fuzzy Logic

Author(s):  
K.W. Lee ◽  
H.N. Lam
2021 ◽  
Vol 2021 ◽  
pp. 1-18
Author(s):  
Mohamed Ali Zdiri ◽  
Mohsen Ben Ammar ◽  
Fatma Ben Salem ◽  
Hsan Hadj Abdallah

Due to the importance of the drive system reliability, several diagnostic methods have been investigated for the SSTPI-IM association in the literature. Based on the normalized currents and the current vector slope, this paper investigates a fuzzy diagnostic method for this association. The fuzzy logic technique is appealed in order to process the diagnosis variable symptoms and the faulty IGBT information. Indeed, the design, inputs, and rules of the fuzzy logic are distinct compared with the other existing diagnostic methods. The proposed fuzzy diagnostic method allows the best efficient detection and identification of the single and phase OCF of the SSTPI-IM association. Accordingly, after the fault detection and identification using this proposed FLC diagnostic method, a reconfiguration step of IGBT OCFs must be applied in order to compensate for these faults and ensure the drive system continuity. This reconfiguration is based on the change of the SSTPI-IM topology to the FSTPI-IM topology by activating or deactivating the used relays. Several simulation results utilizing a direct RFOC controlled SSTPI-IM drive system are investigated, showing the fuzzy diagnostic and reconfiguration methods’ performances, their robustness, and their fast fault detection during distinct operating conditions.


Sign in / Sign up

Export Citation Format

Share Document