Fault tolerant control of five-phase brushless direct current (BLDC) machines is gaining more importance in high-safety applications such as offshore wind generators and automotive industries. In many applications, traditional controllers (such as PI controllers) are used to control the stator currents under faulty conditions. These controllers have good performance with dc signals. However, in the case of missing one or two of the phases, appropriate reference currents of these machines have oscillatory dynamics both in phase- and synchronous-reference frames. Non-constant nature of these reference values requires the implication of fast current controllers. In this paper, model predictive deadbeat controllers are proposed to control the stator currents of five-phase BLDC machines under normal and faulty conditions. Open circuit fault is considered for both one and two stator phases, and the behaviour of proposed controlling method is evaluated. This evaluation is generally focused on first, sensitivity of proposed controlling method and second, its speed in following reference current values under transient states. Proposed method is simulated and is verified experimentally on a five-phase BLDC drive. (C) 2015 Elsevier Ltd. All rights reserved.
Interturn faults in permanent magnet synchronous motors (PMSMs) may develop fast into more severe faults such as coil-to-coil, phase-to-phase and phase-to-ground short circuits. These faults are very destructive and may irreversibly damage the PMSM. Therefore, it is highly desirable to develop suitable methods for the early detection of such faults. The effects of interturn faults are visible in both the stator currents and the zero-sequence voltage component (ZSVC) spectra. By designing appropriate fault diagnosis schemes based on the analysis of the harmonic content of such electric variables it is possible to detect short circuit faults in its early stage. However, the stator winding configuration of the PMSM deeply impacts the harmonic content of both spectra. This paper studies the effects of different stator winding configurations in both the stator currents and the ZSVC spectra of healthy and faulty machines. Results presented may help to develop fault diagnosis schemes based on the acquisition and further analysis of the stator currents and/or the ZSVC harmonic components.