This paper undertakes an investigation into thermal modelling of Totally Enclosed Fan-Cooled (TEFC) motors used in traction applications. In the process, a 3D Computational Fluid Dynamics (CFD)-based Conjugate Heat Transfer (CHT) model is utilized, including the realistic stator geometry with cooling channels and considering the rotor rotation through the multiple reference frame approach. Extended air flow measurement data are used to determine the modelled air speed through the cooling channels and account for the impact of partial duct blockage. Loss distributions obtained via a transient electromagnetic Finite Element (FE) model are used in the process of heat source definition. Additionally, a simplified net radiation model is integrated in the boundary definitions along with natural convection. The main purpose of this paper is to highlight how different modelling approaches affect the obtained temperature distribution. The particular focus of this work is on the stator side and can be extended to various rotor structures. The methodology developed is applied for the thermal characterization of a prototype Interior Permanent Magnet (IPM) motor.

Impinging jets are being used in a wide range of applications, e.g., food processing, turbine blade cooling, and automobile windscreen defrosting. While early studies are about half a century old, impinging jets are still being explored and new features are revealed constantly through experimental and numerical investigations.

The heat transfer arising from an impinging jet at a Reynolds number of 500 0 is studied through LargeEddy Simulation (LES), with special attention on the heat transfer dynamics. The obtained heat transfer and flow fields are decomposed and studied using proper orthogonal decomposition (POD) and extended proper orthogonal decomposition (EPOD). The heat transfer appear to be distributed according to a gamma distribution, in time, with location-dependent shape and scale parameters. The results obtained show that, over time, many locations on the impingement plate experience large over- and undershoots compared to the time-averaged Nusselt number distribution. The POD analysis show that the low order heat transfer modes, while having low relative intensity, are associated with distinct flow features. The flow features are identified by application of EPOD. The two dominant modes are associated with ring-like vortex structures organized concentrically around the impingement point. Reconstruction of the heat transfer field using the three first modes and the mean field show radially outward moving structures with a phase velocity of 0.23 U-b.

A-priori defined thermal networks usually show reasonable agreement between measured and simulated temperature results. Discrepancies are, however, encountered on a regular basis depending on how well the model is replicating reality, and as a result, the accuracy of the thermal network may be reduced. This paper shows an evaluation of machine learning methods for adjusting the thermal network of an electric machine based on temperature inputs. The thermal resistances constituting the network are thus updated with regards to the operating conditions. Three machine learning methods (linear, nearest neighbour, and perceptron regressors) have been assessed to decide the best one for this application and machine specifications. Investigations performed revealed good predictive capability of the machine learning model with the multi-layer perceptron regressor. The machine learning-based adjustment of the thermal network exhibited promising results for the potential development of self-tuned thermal models.

This paper describes end-winding heat transfer under forced external cooling, a cooling arrangement which is not commonly studied in literature. The average end-winding heat transfer under different rotational speeds and external coolant supply is studied through the use of computational fluid dynamics. As a result of the analysis, an analytical expression for the combined effect of rotation and external forced cooling is proposed. Finally, the established expression is used in an lumped-parameter thermal-network analysis and resulting temperatures are compared with corresponding measurements.