Date of Award


Publication Type


Degree Name



Mechanical, Automotive, and Materials Engineering


CFD;Corrections;Open Jet;Parametric Model;Vehicle Model;Wind Tunnel


Jeffrey Defoe



Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.


The simulation of automotive wind tunnel flow around a vehicle typically relies on the use of simplified models due to the complex, and often proprietary geometries of real test configurations. Current simplified models achieve a low computational cost in computational fluid dynamics (CFD) but fall short at qualitatively capturing a modern vehicle’s complex geometry due to being overly abstract; other models are more accurate but offer only discrete dimensional configurations. The first part of this thesis describes new parametric models for both vehicles and open-jet wind tunnels which aim to represent widely varying geometries through continuous dimensional variability, while remaining simple enough to be usable in modern CFD codes. The models are built in SALOME, an open-source CAD package. The parametric vehicle model is defined by nine independent parameters, which are readily available for most production vehicles, and can accommodate vehicle shapes ranging from sedans to minivans. The parametric tunnel model is defined by twelve independent parameters. Both models are shown to closely represent the geometry of real vehicles and industrial facilities. The vehicle model is demonstrated to be able to predict the drag coefficient of its representative detailed model with near 100% accuracy. The second part of this thesis employs the new models in a computational parametric study of vehicle and wind tunnel geometries with the goal of introducing a CFD driven wind tunnel correction methodology to improve upon the current state-of-the-art wind tunnel corrections. An empirical correction function which describes the expected difference between open road and wind tunnel drag coefficient measurements was built from the results of the parametric study. This new function, when applied to an experimental data set of 5 vehicles in 2 wind tunnels, showed a reduction in the standard deviation of the corrected drag coefficient measured for the same vehicle in different wind tunnels from 0.0137 with current methods to 0.0054 – a 59% improvement.