Results
A single pole-pair salient rotor, 15-phase synchronous reluctance motor was designed and evaluated. Full efficiency surfaces were generated for an array of different rotor geometries, and high-fidelity time series simulations were run for the optimal designs. Inverter transistor losses were investigated for the optimal design.
The results show that it is possible to build a very efficient synchronous reluctance motor without rotor flux barriers. However, low power-factor makes the transistor selection for an efficient, low-cost inverter nearly impossible.
The results show that it is possible to build a very efficient synchronous reluctance motor without rotor flux barriers. However, low power-factor makes the transistor selection for an efficient, low-cost inverter nearly impossible.
Further Work
First, I have not given up on the synchronous salient rotor (although conventional wisdom says I probably should). There is a large space of rotor geometries I have yet to explore, including increasing the number of pole-pairs. Perhaps even more significantly, there is a massive space of current profiles that were not even tested in this study. Again, there is no reason to believe that sinusoidal current control is optimal for this kind of motor, and using a high phase order design such as this gives the freedom to explore much more complex current profiles. As this is a very high-dimensional optimization, I will need to re-evaluate how I model the motor. There are some steps I can take to decrease the run-time of the finite element solver, but even if I cut it down by a factor of two or three, the problem space is far to vast for me to do on a single computer. The space increases further if I also optimize across the stator dimensions, which is necessary to truly optimize a design. So, to move forward I need to figure out how to put minimize the number of simulations to evaluate a design, and put the process on a compute cluster.
Second, if the work is to be of any use, a more complete model of the flywheel energy storage system, as well as a cost model, needs to be explored. Understanding the use statistics of a flywheel energy storage system, as well as a more complete model of the system losses, including bearing losses, bearing inverter losses (if active magnetic bearings are used) and all the quiescent losses from the electronics are important to the efficiency evaluation.
Finally, I think it would be worth exploring rotor designs that include flux-barriers. Flux barrier rotors have been shown to be superior to salient rotors in the literature, but these studies typically designing for lower rotor speeds, and do not need to include centripetal stresses as part of the optimization. While very ambitious, it would be possible to couple FEA structural analysis with the magnetic analysis to optimize flux-barrier rotors for high-speed applications, and might be worthwhile.
Second, if the work is to be of any use, a more complete model of the flywheel energy storage system, as well as a cost model, needs to be explored. Understanding the use statistics of a flywheel energy storage system, as well as a more complete model of the system losses, including bearing losses, bearing inverter losses (if active magnetic bearings are used) and all the quiescent losses from the electronics are important to the efficiency evaluation.
Finally, I think it would be worth exploring rotor designs that include flux-barriers. Flux barrier rotors have been shown to be superior to salient rotors in the literature, but these studies typically designing for lower rotor speeds, and do not need to include centripetal stresses as part of the optimization. While very ambitious, it would be possible to couple FEA structural analysis with the magnetic analysis to optimize flux-barrier rotors for high-speed applications, and might be worthwhile.