Experimental and CFD Comparative Case Studies of Aerodynamics of Race Car Wings, Underbodies with Wheels, and Motorcycle Flows 2008-01-2997

The validity and usefulness of low-complexity “fast-turnaround CFD” for motorsports design is investigated using results from three different combined experimental and CFD analyses of racing or high-speed vehicles. Analyses using both wind tunnel experiments and CFD simulations (with commercial software and moderate computing resources) found good agreement in some aspects of interest over a variety of applied situations. Key results were the ability for relatively simple CFD models to consistently predict C_L in complex flows within 15-25% of experimental findings, predict the effect of design changes on flow, and accurately show qualitative flow phenomenon. However, C_D values were not accurately predicted with the low-complexity simulations. Simulations were run using the commercial Fluent^© 6.3 application. Experimental results were performed in the Cornell University 4 by 4 foot wind-tunnel.

The first investigation explored the development of front and rear wing packages for a formula-style race vehicle. After validation of Fluent^© 6.3, basic 2D wing designs were optimized using CFD and tested in the wind-tunnel and on the track.

The second investigation explored the aerodynamics of underbody-equipped bluff bodies in ground-effect flow. Focus was on the influence of multiple surrounding elements (wheels) on underbody diffuser performance. Flow phenomena, such as trailing-edge diffuser separation and vortex development, and how they are affected by wheels and their wakes are discussed.

The third investigation explored flow structure predictions of motorcycle engines surrounded by interfering objects (wheel, forks, frame). CFD simulations isolated key factors influencing flow phenomenon and were useful in predicting the effects of modifications to the design.

“Fast-turnaround CFD” was found to yield robust and sufficiently accurate results for varied aerodynamic configurations useful for applied vehicle design. It accurately calculated C_L in multiple configurations (less than 25% error) as well as successfully modeled flow field dynamics for single and multiple body situations. This was done in 2D with low y+, and in 3D with high y+ value models which were relatively easy to mesh and run. Limitations with the approach, such as inability to accurately predict C_D, are discussed.