Multiphase Flow and Heat Transfer in an Electric Motor

Abstract Electric vehicles are becoming increasingly common due to environmental needs. Due to this, efficiency in design process of electric motors (E-motor) is becoming critical in the industry. To assess performance capabilities for an E-motor, thermal predictions are of utmost consequence. This study describes a computational method based on unsteady Reynolds-averaged Navier-Stokes equations that resolves the gas-liquid interface to examine the unsteady multiphase flow and heat transfer in a concentrated winding E-motor. The study considers all important parts of the motor i.e., coils, bobbins, stator laminate (yolk), rotor laminate, magnets etc. The study highlights the ease of capturing complex and intricate flow paths with a robust mesh generation tool in combination with a robust high-fidelity interface capturing VOF scheme to resolve the gas-liquid interfaces. Results obtained show the dominant processes that determine the oil distribution to be the centrifugal force from rotation of the rotor, the flow rate of oil injected in the stator assembly as well as in the rotor assembly and gravity. A novel heat transfer approach (mixed time-scale coupling) is used to solve for the temperatures in the stator and rotor solids. The approach first requires achieving a quasi-steady flow solution before initiating heat transfer calculation for faster turnaround times. The approach separates the conjugate heat transfer calculation into a fluid heat simulation and a solid heat simulation while setting up a communication method to exchange the thermal boundary conditions between the two simulations. This study also considers the anisotropic nature of thermal conductivities resulting from the wound-around arrangement of the coils and the laminate nature of stator/rotor laminates in the assignment of the thermal conductivities of these solids. Results of thermal simulation show the solid temperatures to be in direct correlation with the oil distribution near those solids. This computational study was validated by comparing the computed and measured temperatures at specified locations on the coils and good agreements with experiments were found.