Modeling, Characterization, and Control of Switched Reluctance Motors

Abstract

Switched reluctance machine (SRM) is known for many advantages such as permanent magnet-free operation, robust structure, low rotor inertia, low manufacturing cost, and excellent fault-tolerant capability. Hence, SRM has been adopted in many applications, such as electric vehicles, aerospace, and robotics. Nonlinear characteristics and pulsations in torque developed are well-known problems, rendering modelling and control of the SRM challenging. This thesis focuses on the modelling, characterization and control of switched reluctance machines. Current, torque, and speed control are all part of the scope of study. Conventionally, rotors with laminations are used in SRM. In certain applications where the shaft temperature increases very significantly, the thermal expansion of the different constituent materials in a typical laminated would be at different rates. This creates stress in the rotor assembly and could reduce the reliability of the machine. Hence, in such applications, rotors made from a single piece of magnetic material are potential candidates. Solid-rotor and, recently proposed, slitted-rotor SRMs are prospective candidates for high-temperature applications. However, research on solid- and slitted-rotor SRMs remains relatively limited. In this thesis, solid- and slitted-rotor SRMs are systematically compared through comprehensive 3D transient finite element analysis (FEA) and experimental evaluations under both static and dynamic conditions. Blocked-rotor experiments and 3D finite element analyses show that the slitted-rotor SRM has much lower core loss and higher torque density than the solid-rotor SRM. High torque density is essential for applications such as electric vehicles and aerospace systems. This thesis compares several methods to enhance the torque density of laminated-rotor SRMs through FEA-based simulations. Various magnetic structures for the SRM, including multi-toothed stators, tapered poles, non-uniform air gaps, flux barriers, and segmental rotors, are analyzed. Additionally, the performances of two different winding configurations, namely, double-layer conventional (DLC) winding and double-layer mutually coupled (DLMC) winding, are compared under unipolar and bipolar excitations, respectively. The DLMC winding concept is successfully applied to solid- and slitted-rotor SRMs to enhance torque output in this thesis. FEA-based simulations and extensive blocked-rotor experiments are conducted to demonstrate the improvement in torque characteristics due to the DLMC winding reconnection. Two new current control schemes are proposed in this research work. In the first part, an extended horizon model-based predictive current controller is proposed for SRM. An analytical equation is reported for real-time computation of the optimal duty ratio to minimize the RMS error between the future current references and predicted currents over a horizon. The proposed controller reduces the RMS error in current tracking and improves robustness to parameter variations, compared to an existing dead-beat predictive controller. Simulation results supported by experimental validation on a laboratory prototype drive are presented. Further, a fixed-frequency, model-free predictive current control is proposed for the SRM. Unlike traditional approaches, this method does not require any pre-measured characteristics of the SRM. Hence, this method eliminates the need for time-consuming characterization experiments, multi-dimensional lookup tables, and offline curve fitting to model the flux-linkage characteristics of the SRM for current control. A high-performance torque control scheme for SRMs is presented, incorporating a PI controller, feedforward compensation, high-frequency compensation, and optimized gating functions. This controller achieves a significant reduction in pulsating torque and outperforms the state-of-the-art technique across various operating conditions. Further improvement in performance is achieved through a novel PWM-based optimal predictive direct torque control scheme. In this work, a cost function, encompassing the instantaneous torque error and the RMS values of phase currents, is minimized. An analytical expression is derived for the optimal duty ratio, resulting in improved computational efficiency. This controller delivers better torque-reference tracking, higher torque per ampere, and lower sound pressure levels than the existing torque control methods, as shown by simulation and experimental results. A novel experimental method is reported for determining the combined moment of inertia and frictional torque characteristics of an SRM coupled to a load, utilizing a highperformance torque controller. The identified mechanical parameters are used in a systematic controller design procedure to achieve fast speed reference tracking with good disturbance rejection. The controller’s effectiveness is validated through simulations and experiments, demonstrating its effectiveness in improving the SRM drive performance.