High Performance Switched Reluctance Generator Controller

This proposal describes a method to reduce the acoustic noise and torque ripple of a switched-reluctance machine by solving a multi-objective optimization problem. The physical limitations of the machine, namely voltage and temperature, are used as constraints, and these are related to the objectives of noise and torque variation. Electrical, mechanical, and thermal models are used to form an analytical association, while finite element analysis forms the core of a complimentary numerical approach. Final accuracy is determined by experimental testing, and iterative improvements are used to improve the accuracy of poorly performing methods. The motor will be extensively instrumented for the experimental validation. Vibration sensors are the primary means of measuring stator displacement, with noise explicitly recorded from a high-fidelity microphone. Strain gauges will be installed to allow direct measurement of radial forces on the poles of the stator to allow for accurate comparison with analytical and numerical results. Mechanical torque is measured using a dynamometer having a high sampling rate to allow for determination of the tangential force as a function of rotor angle. With sufficiently accurate analytical models and numerical methods, the original optimization problem will be solved to yield the desired control.