Self-Sensing-Oriented Optimization of Synchronous Reluctance Machine Design

This paper deals with self-sensing-oriented optimization of synchronous reluctance machines. This kind of machine is among the most challenging to control without the position sensor at low speed. In fact, typical position estimations adopt high-frequency voltage injection which heavily relies on the intrinsic machine saliency. However, both at low and high currents, such a saliency is not guaranteed due to the presence of the iron ribs and to the saturation of the iron material, respectively. Furthermore, the estimation algorithm could also become unstable due to the absence of convergence points. The aim of the paper is to tackle this issue, embedding proper sensorless-control capability into the design through multi-objective optimizations.