Smooth Torque Operation And Sensorless Control Of Switched Reluctance Motors

Nowadays the global major countries promote vigorously the technology ofenergy-saving and emission-reduction. Against this background, the energy efficiencystandard of electric machines is raised up continuously, and the electric vehicleindustry develops fast. Due to the advantages of high efficiency, robust and lowmanufacturing cost, the switched reluctance motor (SRM) is getting a gooddevelopment opportunity. However, the notable torque ripple and dependence of theposition sensors of SRM restrict its applications in more fields. In this dissertation, thekey problems including electromagnetic characteristic modeling, sensorless controland smooth-torque control for the SRM are studied, and the main works can besummarized as follows.An accurate model is essential for the research on control strategies of the SRM.A fast measurement method of fluxlinkage characteristics for SRM is given, and thenthe nonlinear model of the SRM is constructed. Based on the symmetrical geometrystructure of the SRM, without extra mechanical devices, the rotor can be fixed atsome particular positions when single phase or multiple phases in series is excited.According to this, the fluxlinkage curves can be measured. However, the fluxlinkagecurves, which are measured when the multiple phases in series is excited, can not beused directly for modeling because of the flux mutual coupling between phases.Therefore, the magnetic equivalent circuit method is applied to eliminate the effect ofmutual coupling, and then the equivalent fluxlinkage curves with single phaseexcitation are acquired. Based on the five fluxlinkage curves, the analytical models offluxlinkage and torque characteristics are built. The Hardware-In-Loop technology isadopted to test the precision of the model under the dynamic operation. The precisionof the model and the feasibility of the proposed method are verified by theexperimental results.The torque ripple of the SRM is one of the main reasons that limit its applicationin the servo system. The smooth-torque generated by the SRM system is constrainedby the supply voltage and the maximum permissible current, and the maximumsmooth-torque can be represented as a function of the speed. This function is definedas “smooth-torque speed characteristic”, which is determined by a proposed numeraliteration algorithm. The smooth-torque speed characteristic and average-torque speedcharacteristics represent the capabilities of smooth-torque generation and average torque generation respectively. Smooth-torque speed characteristic can be applied toevaluate the torque utilization by comparing with torque speed characteristic.Furthermore, the optimal Torque Sharing Function (TSF) calculated by the proposedalgorithm can be applied to smooth-torque control, which produces a highersmooth-torque over a wider speed range in contrast to four conventional TSFs. Thevalidity of the proposed method is verified by the simulation and experimental results.The Direct Instantaneous Torque Control (DITC) is more practical comparedwith TSF method, since it does not need the predefined reference phase current orfluxlinkage. DITC mainly consists of the torque estimator and torque hysteresiscontroller. The realtime total torque is obtained based on the lookup-table method, andthe torque hysteresis controller is focused on in this study. During the commutation,the torque hysteresis controllers of two adjacent phases have a group of thresholdsrespectively, and the relationships among the thresholds determine the voltage levelsof two phases. Based on the research conclusion of the smooth-torque speedcharacteristic, the suitable thresholds are set to form a novel DITC which can generatethe maximum range of the smooth-torque. Due to the fact that the novel DITC doesnot demand the extra optimization algorithm to determinate the turn-off angle, it issimpler than conventional DITC.Many existing sensorless methods rely on the motor model, which weakens itsuniversality and practicability. In this dissertation, the sensorless control methodsbased on the detection of the three particular positions, which arbitrary SRM with thenormal structure has in the phase inductance profile, are proposed. The first particularposition is the starting point of the minimum inductance interval, the second is theending point of the minimum inductance interval, and the third is the starting point ofthe maximum inductance interval. The phase current has extreme points at these threepositions, thus the corresponding position can be determined through detecting theextreme point of the phase current. The second particular position has been utilized bythe current gradient sensorless method, while the first and third particular positionsare applied by the proposed methods to realize the sensorless control. The simulationand experimental results show that the proposed methods can implement thesensorless control without the motor model. Furthermore, the zero-phaseshift filter isused to analyze the position detection precision of the three methods. The comparisonof the operational speed range and the other performances of the three methods aregiven.