Particle swarm optimization approach to optimal design of an AFPM traction machine for different driving conditions

Abstract

Axial flux permanent magnet (AFPM) machines can be employed as the traction motor of electric vehiclesdue to their high torque capability, high efficiency, modular and compact construction, and capability of integrationwith other mechanical components in integrated systems. Besides, the system efficiency can be further improved byoptimal design of the selected electric machine. In this paper, an AFPM machine is optimized against two well-knowndriving cycles called the New European Drive Cycle (NEDC) and US06 and the influence of the driving cycle on theobtained machine parameters is evaluated. US06 is the more demanding driving cycle and thus the machine designed forthis driving cycle demands more electrical loading compared to the machine designed for NEDC. Therefore, the copperloss minimization becomes more important for the US06-optimized machine compared to the NEDC-optimized machine.Consequently, the machine design parameters optimized for different driving cycles would be quite different. Comparedto the NEDC-optimized machine, the US06-optimized machine has a lower number of coil turns, lower height of teeth,and lower diameter ratio to limit the copper losses. Furthermore, fewer magnets are needed for the motor optimized forthe NEDC compared to the motor optimized for US06. A quasi-3D approach and particle swarm optimization algorithmare used in the semianalytical design optimization process. Additionally, computationally efficient 3D finite-elementanalysis and measurements made for the prototype AFPM machine are carried out to validate the accuracy of thequasi-3D approach.