Online fault diagnosis techniques for interturn short-circuits in switched reluctance motors

Abstract

Fault diagnosis is crucial to any electric drive system, ensuring their reliable operation. The significance of fault ride-through capability, which denotes the ability of an electric motor to endure faults and maintain operational integrity, is of paramount importance. For this objective, the first step involves the application of fault diagnosis methods to identify any motor fault precisely at the incipient stage. With the increasing demand for rare-earth free alternatives to traditional motors, switched reluctance motors (SRMs) have gained significant attention in different applications due to their specific features. However, to make this technology more commercially available, the different aspects, such as their designs, control strategies, and mitigation of acoustic noise and torque ripples, have been extensively researched, leaving behind the area of fault diagnosis relatively. Interturn short circuits (ITSCs), accounting for 21% of all the electrical faults inside any machine, are responsible for catastrophic failures leading to a complete winding short circuit if left unchecked due to the generation of local hotspots. The inherent problems of torque ripple, noise and vibrations associated with SRMs are also escalated when the machine is subjected to ITSCs. These faults are even difficult to diagnose as the fault features are least apparent in the electrical parameters in ITSC of fewer turns. Also, most of the diagnosis techniques for ITSCs in SRMs suffer from several issues, like lower sensitivity where the system can not detect ITSC if a lower number of turns are short-circuited. Also, there are interference of load variation on the detection reliability in which the fault index might initiate false alarm without any fault. Some of the methods are dependent on the control strategy on which the motor is operating. The thesis delves into an attempt to devise sophisticated online fault diagnosis techniques for ITSCs, considering the research gap in the existing literature. Three online diagnosis techniques have been formulated and validated experimentally on a test rig of four-phase 8/6 SRM. The first method targets low and medium-speed applications operating under chopped current control with 4% least severity detected. The second method is based on the signal injection technique applicable to the SRMs independent of control strategies. It is capable detecting ITSC of 2 turns. The third technique eliminates the additional hardware used for diagnosing the fault utilized in the second method and is also independent of the control strategies. All the techniques have been tested under different operating conditions (load/speed variations) and also under transient conditions, proving the robustness of the proposed schemes. The background, formulation and experimental results of all the schemes are discussed in detail in the subsequent chapters of the thesis.