Automatic Design of Synchronous Reluctance Motors focusing on Barrier Shape Optimization

The automated design of Synchronous Reluctance motors based on Multi-Objective, Genetic Optimization and Finite Element Analysis is considered in this paper. Three types of barrier shapes are considered, all described by an effective, limited set of input variables. The three solutions are investigated to establish which of the geometries can give the best torque output and also which one represents the best compromise between output performance and computational time. The analysis presented in this paper shows that Synchronous Reluctance motors designed automatically can give a good performance, can be designed in a reasonable time and it is also shown that not all design degrees of freedom are useful in terms of motor performance. Two prototypes of automatically designed machines have been fabricated and experimentally compared to a third prototype designed according to state-of-the-art design principles.