Design of experiments based optimization of synchronous and switched reluctance machines

During the electrical machine design process, here specifically for synchronous reluctance and switched reluctance machines, a large amount of design parameters interact with each other influencing the resulting geometry. A methodology to handle multi-parameter calculations for electrical machine design optimization is presented. With statistical methods the multi-parametric problem can be fitted and optimized by an analytic expression, containing all relevant interactions. The advantages of this approach is shown by finite element method results for a reference design of a 3kW synchronous reluctance and a 1kW switched reluctance machine. The benefit of this approach is shown for geometry optimization with different, partly contradicting, design criteria such as lowest torque ripple and highest torque density. Design of experiment is highly recommended to understand the interaction of parameters and especially in regard to design time reduction.