Kristian Rönnberg Licentiate : Heat-transfer simulations applied to electrical machines

Electrification and energy efficiency are two important aspects in present scenarios describing a sustainable future. Electric motors constitute a large fraction of industry’s electricity demand today, and it is expected to remain high in the future. Electrification of the transport sector is expected in a sustainable development scenario, leading to a large increase in electric vehicles. Their propulsion systems will contain one or several motors. Development of new energy efficient motors, and generators, require high resolution methods for studying and describing heat transfer phenomena.

This since temperature level affects a motor’s efficiency and effective and efficient cooling allow for using less active material in the motor. Simulations of temperature distribution in a motor for traction applications were performed with different specifications of the loss distribution and distribution of coolant flow. Simulation results were compared to measured values. The comparison show how the simulation results differ in comparison to the measurements. It can be concluded that attention need to be paid to how the simulation is defined when comparing to measured data. In establishing high resolution simulation approaches, the heat transfer system constituting of an impinging jet on a flat plat was considered as a prototype problem. A Large-Eddy Simulation (LES) approach was employed to study the heat transfer and gather heat transfer data. Statistical analysis of sampled heat transfer data showed behavior which was previously unpublished.

The application of Proper Orthogonal Decomposition (POD), on the heat transfer field, and Extended Proper Orthogonal Decomposition (EPOD), linking heat transfer modes with fluid flow modes, regarding the impinging jet system was performed for the first time. The results show a clear correlation between structures in the heat transfer field and structures in the fluid flow field. The investigated simulation methods and approaches can be employed in studies of heat transfer in electric machines.

Contact person
Kristian Rönnberg

Email address
kriron@kth.se