Abstract
In order to improve the cooling performance of a vehicle, we carried out computational fluid dynamics (CFD) optimization of the front-end cooling structure. To improve the efficiency of optimization calculation, we introduced a method using variable fidelity CFD models and proper orthogonal decomposition (POD). Firstly, in this method, the POD modes are calculated from Pareto solutions of low-fidelity CFD optimization. Secondly, an optimization calculation is performed using high-fidelity CFD models with only the dominant POD modes. As a result, we obtained solutions showing performance values equal to those of the previous method at 50% of the computational cost of it.
