Torque Performance of Optimally Designed Multi-Phase Reluctance DC Machines
Thesis (PhD (Electric and Electronic Engineering))--University of Stellenbosch, 2006.
The focus of this thesis is on the critical evaluation of the torque performance of the optimum designed reluctance DC machine (RDCM). The thesis focuses on multi-phase RDCM drives with normal laminated salient pole rotors allowing for high power and high-speed applications. An RDCM is a normal reluctance synchronous machine (RSM) but with direct control of flux and torque just as in brush DC machines. Flux and torque of the RDCM are controlled directly by the use of special phase current waveforms. Specific attention is given in the thesis for the selection of the best current waveform for the RDCM allowing for smooth rotating airgap MMF and less ripple torque. The absolute optimum designed RDCM can best be obtained by the use of the finite element (FE) method in the design optimisation process. In this thesis a multi-dimensional FE based design optimisation method for the optimum design of the current controlled RDCM is implemented. To compare the torque performance of the RDCM with other RSMs the torque performances of optimum designed 3-phase, 5-phase and 5-phase with the injection of third harmonic current RSMs are performed under the same copper losses and stack volume. The torque performances of RSMs are done with both salient pole rotor and the round rotor with internal flux barriers. The armature reaction effect of 6-phase RDCMs is also investigated in detail by considering three different rotor structures. These rotor structures are the standard salient pole rotor, the salient pole rotor with slitted poles and the salient pole rotor with chamfered poles. It was shown that the RDCM with the salient pole rotor has a severe armature reaction effect, which can be reduced by slitted or chamfered salient pole rotors. A per-phase equivalent circuit model of the 6-phase RDCM is also proposed in this thesis. The torque of the machine is calculated based on the per-phase equivalent model and compared with the torque calculated by the FE Maxwell stress tensor method. There is a good agreement between these calculated torques. This thesis shows that the implemented FE based optimisation method can be applied with success to optimally design current controlled RDCMs. It was found, amongst other things, that the torque performance of the optimum designed 6-phase RDCM is slightly higher than that of the optimum designed 5-phase RSM with the injection of 3rd harmonic currents and with the same copper losses and stack volume. The analytical and FE calculated results are confirmed by measured results on a 35 kW 6-phase RDCM drive.