CFD Validation of the Imperial Front Wing

Objective

Develop a computationally efficient RANS simulation that captures both qualitatively and quantitatively the flow structures and Cl/Cd of the published benchmark, providing a validated baseline for further external aerodynamic studies.

Methodology

Simulation

The turbulence model compared was k-Omega SST. The flow structure was compared against 5 PIV planes and the total pressure coefficient isosurface (Cpt=0). The table below contains the boundary conditions used for the simulation.

Mesh strategy

Refinement groups with values expressed as % of base size (12 mm):

Prism layer settings:

The front wing was run with a low y+ approach to resolve the viscous sublayer, allowing more accurate modelling of near-wall gradients and accurately resolving the interaction between the wing elements. This was considered important since the near-wall flow strongly affects drag and downforce predictions.

Mesh

Results

Cl and Cd monitors converged after 3000 iterations. A further 500 iterations were run and averaged to extract Cl and Cd.

y+ distribution:

Downstream flow validation (PIV planes)

Comparison between the 5 PIV planes from literature (PBS) and CFD.

LiteratureCFD

PIV Plane 1 (x = −294 mm)

PIV Plane 2 (x = −250 mm)

PIV Plane 3 (x = −150 mm)

PIV Plane 4 (x = −24 mm)

PIV Plane 5 (x = 150 mm)

The scale was kept the same. A slightly different colour bar was used due to the lack of information in the analysed paper.

Vortex isosurface validation (Cpt=0)

LiteratureCFD

The major vortex structures were captured well by the 30 million cell model. However, further refinement would be needed to match the accuracy of the literature PIV planes and Cpt=0, which were obtained with a finer mesh (120 million cells).

Cl and Cd comparison

The reference surface area used was 0.175 m² (c = 0.25, W(s) = 0.7).

The lift and drag coefficients with the 30 million cell mesh were −5.6406 and 0.597. Relative to the reference values, they differ by 0.0006 (0.011%) and 3.25% respectively, showing strong agreement in lift and a slight offset in drag. Given the similar mesh count, y+, refinement sizes and boundary conditions, the drag difference suggests a greater sensitivity of Cd to numerical and modelling assumptions compared to Cl.

Engineering outcome

The developed 30 million cell mesh captured the major vortex structures and Cl/Cd in good agreement with the benchmark. Cl and Cd errors of 0.011% and 3.25% respectively were achieved, while reduced accuracy was observed for the PIV planes and the Cpt=0 isosurface. The model will therefore be used as an efficient baseline for further external aerodynamics studies.