Time-Efficient Heat Transfer Prediction for Hairpin Windings with Flooded Stator Cooling Based on Reduced-Order Computational Fluid Dynamics

Flooded endwinding cooling is regarded as an effective direct cooling approach for high-performance electric motors, although its application for hairpin windings remains relatively limited. The challenge in the scalability of flooded cooling design makes it difficult to translate heat transfer data within parametric motor geometries, leading to excessive computational resources needed for large scale optimisation. In this paper, a reduced-order modelling methodology for investigating the performance of flooded cooling is proposed. Using computational fluid dynamics, the cooling of hairpin conductors is characterised against a variation in conductor height, width, layer number, as well as spacing between adjacent conductors and coolant velocity. The results are summarised in the form of a Nusselt number correlation, providing an equation that can be directly used for predicting heat transfer coefficients of the flooded endwinding region. A customised test rig with a multi-layer hairpin winding sample packaged within a flooded cooling environment is designed and manufactured, capable of providing a variation in the coolant flow rate, inlet temperature and current density. Test results with current density ranging from 20-35A/mm² show a strong match to the simulation, validating both the correlation and modelling methodology for their application for flooded cooling of hairpin windings.