In-Slot Oil Cooling Arrangement of Hairpin Windings for High Power Density EV Traction Motors

Liam Portanier Mifsud,Peter H. Connor,Adam Walker,Tianjie Zou,Hailin Huang,Xiang Ren,George Batho,Oliver Tweedy,Christopher Gerada,Christian Egger

DOI: https://doi.org/10.1109/vppc63154.2024.10755199

2024-01-01

Abstract:Ambitious electrification roadmaps for the automotive industry have prompted a revolution in winding technology for traction motors in electric vehicles (EV). Hairpin windings with reduced power losses and improved coolability are overtaking the market compared with stranded windings. Currently, most hairpin winding solutions use state-of-the-art end winding cooling strategies. This, however, still cannot bridge the gap for high performance requirements, so supplementary in-slot cooling is needed to make the next step change in achieving high power density. This paper will investigate direct in-slot cooling of hairpin windings with a focus on cooling channel arrangement. A novel analytical parametric Lumped Parameter Thermal Network modelling methodology is developed to choose the thermally optimal in-slot cooling channel radial position for a 150kW high speed traction motor. Two operating points (corner and peak speed) on the peak torque-speed curve are studied to investigate how machine losses influence the choice of the optimal cooling channel position. Changing loss distributions showed an optimal middle channel position for the corner speed and shift towards the airgap at peak speed. A typical slot-opening channel position was shown to reduce hotspot temperatures by 23 K compared with a no-in-slot cooling benchmark, for the most thermally challenging operating point. A further 8 K reduction is achieved, compared to the benchmark, for the most thermally challenging operating condition, when a central cooling channel position is chosen. The novelty of this paper is the multiphysics electromagnetic and thermal modelling methodology integration which the trade-off between losses and temperatures, as impacted by a changing channel position, capture. An FEA model was built to validate the LPTN results for the benchmark case; the temperature rise between the hotspot slot peak and the oil inlet temperature, was predicted within 5%.