Thermal Modelling and Validation of a Direct Rotor Cooled Permanent Magnet Electric Machine

Abstract

A novel direct rotor cooling method for electric machines is proposed and the performance is evaluated based on a thermal zonal model that is validated with measurements on a rotor heater setup. The trend of increasing power density of electric machines urges the need for high performance rotor cooling methods. A thermal model is needed to assess the performance of these methods on a certain geometry and to optimize the design. A previously developed thermal zonal model of an interior permanent magnet synchronous machine is extended with the novel direct rotor cooling method and validated with measurements on a rotor heater setup for different oil inlet temperatures (40 to 80°C), flow rates (0.5 to 10 l/min) and heating powers (261 to 873W). With a maximum absolute and relative deviation between the measurements and simulations of respectively 6.3°C and 13.3% on the maximum rotor core temperatures for flow rates higher than 1 l/min, it is concluded that the model is sufficiently accurate when taking into account the assumptions and uncertainties on the input parameters. A model based comparison with the validated model shows the great potential of the novel direct rotor cooling method to push the power density to higher levels, since the relative magnet temperatures can be decreased by 71% compared to a hollow shaft liquid cooled machine.

Enhancing thermal modelling of electric machines: steady-state verification of a novel methodology

“…All other internal and external surfaces are assumed to be adiabatic. The direction specific thermal conductivities in the radial plane (rad) and axial direction (ax) are 0.583/181.1 W/mK (rad/ax) for the windings, 0.14 W/mK for the liner and wedge, 21.9/1.8 W/mK (rad/ax) for the cores, 10.2 W/mK for the magnets, 0.21 W/mK for the magnet slot potting, 210 W/mK for the housing and flange, and 46.6 W/mK for the shaft [8]. The thermal contact resistances between housing and stator core (249 mm²K/W at 7 MPa), rotor core and shaft (187 mm²K/W at 10MPa) and housing and flange (47 mm²K/W at 1MPa) are calculated using the correlation of Cousineau et al [9].…”

“…For other contact regions, where no relevant correlations are available, measured values from scientific literature are applied. These include the contact resistance between the stator core and liner (608 mm²K/W), winding and liner (1799 mm²K/W), magnet and rotor core (857 mm²K/W) and at the bearing (10264 mm²K/W with air at 120 °C) [8].…”