Torque ripple and vibration reduction of switched reluctance machines

Recently, the limited resources of rare earth materials and growth of environmental challenges have brought increasing interest in the advancement of rare-earth-free electric machines. A switched reluctance machine (SRM) is a promising candidate due to its features of low cost, rugged structure, wide operating range, and fault tolerance capacity. However, high torque ripple and vibration are two primary disadvantages of SRMs. Torque ripple and vibration must be reduced for high performance applications, such as automobiles and home appliances. In this dissertation, comprehensive analysis and methods for reduction of the torque ripple and vibration of SRMs are presented. The origins of torque ripple and vibration of SRMs are analyzed, and design considerations of SRMs for torque ripple and vibration reduction are presented. Then a comprehensive optimization framework for SRM design is proposed by combining a method of design of experiments (DoE) and a particle swarm optimization (PSO) algorithm. Instead of optimizing machine design using performance indices computed by finite element analysis (FEA), surrogate models of machine performance indices are constructed based on DoE and incorporated with the optimization procedure to significantly reduce computational cost while ensuring high accuracy. Pareto-optimal solutions were obtained by the proposed method considering trade-offs between cost and performance. Next, a current profile optimization technique is introduced to generate control references of current-controlled SRMs to reduce torque ripple and vibration simultaneously. A modeling technique based on Fourier series is proposed to consider saturation in the models of SRMs; and the SRM models are incorporated in an optimization process to generate the optimal current profile. In addition to that, the torque ripple of SRMs is further reduced by a proposed pulse-width modulation (PWM)-based direct instantaneous torque control (DITC) algorithm, which significantly reduced torque ripple using a relatively low sampling frequency compared to a traditional hysteresis-based DITC. The researches reduced torque ripple and vibration based on comprehensive studies of both machine design and motor drive. FEA simulations and experimental tests, including measurement of machine characteristics and tests at variable-speed and on-load conditions, were performed to validate the methods described in this work