Geometry Optimization of an Interior Permanent Magnet Machine for Electric Vehicles - Life Cycle Cost Minimization for City, Rural and Highway Driving

Abstract: This thesis investigates the benefits of customizing an Interior Permanent magnet Machine (IPM) for a specific combination of city, rural and highway driving. Focus lies on the potential to reduce the Life Cycle Cost (LCC) of the IPM and power electronic converter, for an example Battery Electric Vehicle (BEV), by custom-making the IPM. The IPM customization is done with an optimization work-flow where both the size and shape of the IPM geometry is varied to minimize the LCC for a given driving scenario. The driving is represented with a distribution of energy and time over the entire torque versus speed plane of the IPM, which allows a large number of drive cycles to be aggregated and included without affecting the computational time. The optimization work-flow includes scaling of electromagnetic finite element results, to automatically select the best feasible version of each IPM geometry, considering performance requirements as well as thermal and mechanical constraints. Customized IPM designs are created for city, rural and highway driving and compared with an IPM optimized for mixed driving. The city, rural and highway driving types are represented both with single drive cycles and with large amounts of measured driving. Customizing the IPM geometry was found to give significant LCC reduction in cases with pronounced city or highway driving. The LCC was found to be 7% lower for a city driving cycle and 1% lower for a highway driving cycle, for the customized IPM compared to the IPM optimized for mixed driving. For the city driving cycle, the optimization results in an IPM design with large slots, high amount of copper and relatively shallow flux barriers which are well filled with magnets. Comparatively, the IPM custom-made for the highway cycle has smaller slots and lower copper and magnet amounts. Still, the IPM optimized for mixed driving has a large and well-placed operating region with high efficiency and has a low LCC for most of the included driving scenarios. None of the included driving distributions based on large amounts of measured driving were specialized enough to benefit significantly from customization of the IPM. Overall, the results indicate that the usage that is currently typical for passenger cars can be well met with an IPM optimized for general driving and the potential for further improvement by custom-making the IPM is rather limited. However, the benefit of custom-making the IPM design is likely higher for BEV applications with more specialized driving type and larger amounts of driving.