CFD Simulations of Velocity and Temperature Distributions of the AuraGen Axial Flux Machine

Abstract: Axial flux induction machines are attractive solutions for electric vehicle applications nowadays. Thanks to their high torque density and no need for rare-earth material for permanent magnets, axial flux induction machines are the most used electric machine type with good performance and low prices. Research on the thermal characteristics of induction machines can benefit the design development of products. Generally, the finite element analysis (FEA) method is used to conduct a fast thermal simulation of machines. However, a significant disadvantage of the FEA method is that the forced convection heat transfer and the fluid motion are challenging to consider. To solve this problem, the thesis work focuses on conducting a computational fluid dynamics (CFD) model to predict the temperature distribution of the AuraGen induction machine and the velocity distribution of the airflow by accurately considering the forced convection heat transfer and the fluid motion in different operation conditions. The thesis work covers the improvement of 3D cad models of the AuraGen induction machines and airflow fields, evaluation of simulation parameters of the CFD simulation models, and the comparison of results between the CFD simulation, FEA simulation, and physically experimental measurements. Finally, the best CFD simulation model can accurately predict the temperature distribution of all components of the induction machine and the airflow in the 3000, 2000, and 1000 rpm conditions. The accuracy satisfies the desired goal which is within 4℃ of the average error and 8℃ of the maximum error. Velocity distributions of the airflow can also show characteristics of the fluid motion from inlets to the outlet. Compared with simulation results of the FEA method, the CFD simulation model has significantly more accurate results when applied for a wide range of operating speeds to predict the temperature distribution in the forced convection heat transfer condition. The good CFD simulation results can help quickly discover design problems in the early stage of the product development process without making repeated prototype constructions and physical tests. The good CFD simulation results are beneficial to reducing the number of necessary prototypes and therefore reducing development costs and time consumed.