CFD Electric Motor External Fan System Comparison and Validation

C3 - CFD Validation Studies

This work presents a comparison between different approaches that can be used to evaluate the external fan system of an electric motor in a Computational Fluid Dynamics (CFD) code. The entire process of comparison was supported by the experimental data, which was obtained from a special device. This device was designed and manufactured in order to consider a number of geometric simplifications aiming the validation of a CFD model. The geometry of the experimental device was transposed to commercial Computer Aided Design (CAD) software with insignificant adjusts, only when it was extremely necessary, to avoid numerical problems in CFD software. With this approach it was possible to concentrate the efforts in numerical problems, avoiding questions like, were the differences between numerical and experimental data originated by numerical error or geometrical considerations?

From numerical point of view the impact of turbulence model is a critical point in CFD simulations, so the turbulence model Shear Stress Turbulence (SST) and k-ε were confronted, both based on Reynolds Average Navier-Stokes (RANS). Especial attention was dedicated to the variation of the y+, i.e., quality of the mesh near the wall. Additionally, the problems of numerical convergence in steady-state regimes were discussed in this work. The numerical results were confronted with air velocity and fan power consumption, both experimentally obtained, and showed good agreement in external flow.

For all simulations, CAD software adopted was SolidWorks 2009 SP2.1 and the CFD software was ANSYS-CFX 13.0 SP2. The mesh dimensions exceed 5.000.000 of nodes and the processing was made in two HP Z800 Workstations with two Xeon X5690 (six-core) processors and 24 GB memory, each.