Search

Switched reluctance machines (SRMs) have gained more interest in the past decades due to their simple and robust structure. SRMs are classified into conventional SRMs (CSRMs) and mutually coupled SRMs (MCSRMs). CSRMs are based on single-phase excitation and torque is produced by the rate of change of self inductance. On the other hand, MCSRMs are based on multi-phase excitation and torque is produced by the rate of change of both self and mutual inductances. The drive system of CSRMs consists of the asymmetric half-bridge converter and the hysteresis current controller. That drive system is limiting the SRMs to be widely used since most applications are using AC motors where the standard voltage inverter and the vector control are used. Thus, in order to replace an AC motor with SRM, the converter and controller used need to be changed and not only the motor. That issue is solved in this thesis. This thesis presents the fundamentals and operating principles of MCSRMs. A literature review of the existing modeling and control methods of MCSRMs is introduced, followed by a performance comparison for MCSRMs with different winding configurations and different control methods. After analysing the existing control and modeling methods in literature of MCSRMs, the focus of this thesis will be on MCSRMs controlled by sinusoidal currents. I preferred sinusoidal current excitation as it enables using the standard voltage source inverter and the standard vector control with the regular modulation schemes such as sinusoidal pulse width modulation or space vector modulation. In order to test the performance of MCSRM with sinusoidal current excitation, a dynamic model is required that can predict the phase currents and electro-magnetic torque when a given voltage is applied. Hence, a new modeling method is introduced in this thesis that is based on vector representation of motor dynamics instead of instantaneous values. The proposed modeling method reduces the size of the look-up tables and the computational steps of finite element analysis (FEA) by 50% compared to other methods. It has also the minimum error compared to other methods. After having an accurate dynamic model, next is to apply the vector control on the MCSRM and observe the motor performance. It will be concluded in this thesis that the standard vector control could not create sinusoidal currents due to the effect of spatial harmonics. Those harmonics are due to the slotting effect of the stator and they are usually ignored in AC motors. However, they cannot be ignored in MCSRM due to the high saliency of stator and rotor poles. Thus, a simple and effective spatial harmonics compensation method is introduced to eliminate the spatial harmonics of phase currents in MCSRM. So far we have an accurate dynamic model and we can ensure sinusoidal current excitation. The next step is how to choose the sinusoidal currents to optimize the motor performance. In order to answer that question, a comprehensive analysis of power factor, torque ripple, and efficiency of MCSRM with sinusoidal current excitation is done. That analysis is then used to optimize the motor performance in terms of power factor, efficiency, and torque ripple. Thesis Doctor of Philosophy (PhD)