Analysis and validation of a CFD-HTC-based hybrid LPTN thermal model for oil-cooled motors with hairpin windings

To address the limitation that homogenized thermal boundary assumptions cannot adequately characterize the non-uniform heat transfer behavior and inter-layer temperature gradients of oil-cooled hairpin end-windings, this paper proposes a hybrid thermal modeling method that integrates Computational Fluid Dynamics (CFD) with a three-dimensional Lumped-Parameter Thermal Network (LPTN). Based on the manufacturing characteristics of hairpin windings, this method introduces a thermal modeling rationale using a layered stacking structure. It obtains the variation patterns of the HTC for each layer with cooling flow rate via CFD simulation and integrates these into the LPTN for rapid full-domain temperature calculation. Taking a 90 kW rated power oil-cooled hairpin-winding motor for new energy vehicles as the research subject, a temperature rise experimental test platform was built to validate the accuracy of the hybrid model's calculations. The results under steady-state operating conditions show that the relative error between the model-predicted temperatures and the measured data at key end-winding locations is less than 5.2%. The model can accurately predict the inter-layer temperature gradient within the end windings, and its computational efficiency is significantly improved compared to a full CFD platform approach. This work provides a reliable and efficient modeling tool for thermal analysis and structural optimization of oil-cooled hairpin-winding motors.