CFD optimisation of the thermal design for a vented electrical machine

Optimisation algorithms hold the potential to dramatically reduce computational time whilst ensuring the optimal solution is found. Within this paper, the feasibility of using this novel approach on complex 3-D Computational Fluid Dynamics models, which are required for thermal management of electrical machines, is proven. A model of a simplified generator is parameterised with the aim of minimising the peak stator temperature by varying the axial location of a single stator vent. By generating a single parameterised case, and automating the optimisation, the simulations are run independently after initial setup, hence reducing both computational and user time. Locating a vent in the optimal position reduced the peak stator temperature by 9.4 K. A sensitivity study linking peak temperature to vent position has been carried out developing a polynomial relationship between them for the aforementioned geometry. Mass flow and pressure distribution in the vent have been analysed in detail.