Inductance and active phase vector based torque control for Switched Reluctance Motor drives

Abstract

The Switched Reluctance Motor (SRM) drive technology has developed significantly over the last few years. The simplicity in both motor design and power converter requirement along with the availability of high frequency, high power semiconductor switches have made SRMs compete with conventional adjustable speed drive technologies. The subject of winding current control in switched reluctance machines has always been associated with the shaft position information. The use of inductance for direct commutation control is the central subject of this dissertation. In contrast to the conventional methods based on position commutation, new methods of control based on inductance commutation are presented. The object of a commutation algorithm is to switch the currents in the phase coils, in order to provide continuous energy conversion with maximum torque output for a given unit of input current. Since torque production in a SRM is based on the concept of variable reluctance, it makes more sense to observe the instantaneous phase inductance or reluctance instead of estimating the rotor position. The inductance sensors observe the machine parameters and provide sufficient information on the electrical characteristics of the coils. This control strategy avoids the inductance to position transformation blocks conventionally used in SRM control systems. In a typical SRM, the phase coils have a nonlinear behavior of inductance due to effects of current saturation. Also the parameters of one phase coil differ from those of the other due to manufacturing tolerances or due to bearing wear. In such cases, the algorithms written during the stage of manufacturing may not be valid after parameter changes. Optimizing torque production in the event of phase asymmetry and saturation is developed in this research. Indirect sensors connected to the active phase coil of the SRM are based on sensing the flux level in the active coil. New commutation algorithms based on flux sensing concepts are derived and commutation based on observable phase coil parameters are developed. The commutation methods are based on a composite vector of the observable parameters of the active phase coil. These methods work on a tabular approach which is ideal for implementation using digital computers.