Efficient electric motor optimization using approximation model-based genetic algorithms

Abstract

Under the rising pressure of climate change and energy security, the solution of electric vehicles and accessory electrical infrastructure is becoming more and more prevalent in recent years. As a key enabling components for all types of electric vehicles, electric motors should satisfy a series of requirements such as robustness, high torque density and high efficiency, etc. Apart from the mainstream PM synchronous motors and induction motors which has been widely-adopted by many car manufacturers and motor suppliers, Permanent-magnet Vernier machine is also being focused by the academia and the industry, for the merits of simple mechanical structure, high efficiency, low torque ripple rate, as well as high torque density. In this dissertation, a surface-mounted vernier machine has been reproduced, analyzed and optimized with the aim of output torque power factor, and efficiency improvement. In the beginning, the background of climate change and fossil fuel depletion, recent EV and motors’ developments, and objectives of this dissertation is introduced. Then, working principles, flux modulation effect of vernier machines and various PMVM topologies are discussed. By carrying out parametric analysis to four critical design parameters of reproduced model, the scope of parameters’ optimizing space is determined. To improve the accuracy and efficiency of optimization, a mathematical RS model is constructed to approximate the relationship between input search space and output space. MOGA based on MATLAB solver is employed to complete the multi-variables, muti-objectives optimization. The results show that the target of enhancement in output torque and power factor is achieved in the optimized point compared with original design. Finally, the deficiencies of this project, potential future works and recommended research directions to vernier machine are concluded.