Axial flux machines with super high torque density or super high efficiency: Design Optimization of an Axial Flux Permanent Magnet Machine Using Genetic Optimization

Abstract

An analytical model of an axial flux permanent magnet machine has been made. The model were based on a 2-D magnetic circuit calculated at average length assuming a constant magnetic and electric loading. The model were optimized using the built-in features in MATLAB.Genetic Algorithms (GA), Gradient Based Interior-Point optimization and a hybrid combination of the two were used to optimize the design. The model were objective functions used were total material cost and total lifetime cost.The lowest average material cost were found by using hybrid GA. The material cost of the cheapest design were 58.86 \% lower than the original design. The efficiency of this solution was 13.7 \% and the power factor was 0.3.Constraints were put on the power factor to improve the optimization result. A machine design was then found with a 0.85 power factor, a 33.3 \% efficiency and a material cost 40.94\% lower than the original design.Optimizing the design with regards to total lifetime cost increased the efficiency of the machine to 58.5 \%. The material cost of the machine increased 4.3 times compared to the cheapest design, but the total lifetime energy cost of the machine was decreased 82.58 \%.A graphical user interface has been made for the optimizations by using GUIDE, MATLABs Graphical user interface designer. Through the user interface, constraints can be defined and the objection function can be changed. The most important results are displayed directly in the GUI, and additional results are displayed in the command window in MATLAB.