Key points are not available for this paper at this time.
The paper presents models for a high-power (60 kW) variable-reluctance motor (VRM) and for the GTO-based inverter that excites the motor. Model development is motivated by the accurate performance predictions required to support optimised control and excitation. One component of the motor model analytically describes the nonlinear magnetic characteristics of the motor, accounting for both spatial and magnetic nonlinearities. The magnetic model is physically motivated, and its ties to motor geometry are discussed. Fitting of the general magnetic model to the experimental VRM is based on static magnetic measurements, but the model is shown to do an excellent job of predicting dynamic behaviour up to frequencies were eddy currents have significant influence. The magnetic model used for the motor is more accurate, yet easier to implement, than the piecewise linear magnetic model typically used for reluctance motors. Electrical terminal behaviour and torque are based on the analytic magnetic model. Another model component is presented for motor resistive losses and core losses. The former is treated analytically, while the latter is treated heuristically. The accuracy of the motor model is verified experimentally, demonstrating that the model has the ability to predict the drive performance in sufficient detail to allow the development of excitation strategies for the drive. The inverter losses considered here include conduction, snubbing and switching. The inverter model complements the motor model and creates an integrated description of the drive, which may be used to evaluate candidate excitation strategies. The accuracy of the models is verified experimentally, demonstrating they have the ability to predict the drive performance in sufficient detail to allow the development of optimised excitation and control strategies for the drive.
Torrey et al. (Mon,) studied this question.