Date of Award

Spring 2006

Document Type


Degree Name

Doctor of Philosophy (PhD)


Aerospace Engineering

Committee Director

Colin P. Britcher

Committee Member

Osama A. Kandil

Committee Member

Drew Landman


The presence of nearby boundaries in a wind tunnel can lead to aerodynamic measurements on a model in the wind tunnel that differ from those that would be made when the boundaries of the moving fluid were infinitely far away. The differences, referred to as boundary interference or wall interference, can be quite large, such as when testing aircraft models developing high lift forces, or whose wingspan is a large fraction of the wind tunnel width, or high drag models whose frontal area is a large fraction of the tunnel cross section. Correction techniques for closed test section (solid walled) wind tunnels are fairly well developed, but relatively little recent work has addressed the case of open jet tunnels specifically for aeronautical applications.

A method to assess the boundary interferences for open jet test sections is introduced. The main objective is to overcome some of the limitations in the classical and currently used methods for aeronautical and automotive wind tunnels, particularly where the levels of interference are large and distortion of the jet boundary becomes significant. The starting point is to take advantage of two well-developed approaches used in closed wall test sections, namely the boundary measurement approach and adaptive wall wind tunnels. A low-order panel code is developed because it offers a relatively efficient approach from the computational point of view, within the required accuracy. It also gives the method more flexibility to deal with more complex model geometries and test section cross sections.

The method is first compared to the method of images. Several lifting and non-lifting model representations are used for both two- and three-dimensional studies. Then the method is applied to results of a test of a full-scale Wright Flyer replica inside the Langley Full Scale Tunnel. The study is extended to include the effect of model representation and the test section boundaries (closed, open and 3/4 open) on the interference. The method is also used during a test of full scale NASCAR inside the NASA Langley Research Center 14- by 22- Foot Subsonic Wind Tunnel. This part includes the effects of test section length and the inclusion of the nozzle in the solution on the predicted boundary interference. Finally, a test is conducted at the 1/15th scale Langley Full Scale Tunnel using a generic automotive model ("Davis" model) to validate the prediction of the boundary distortion and to investigate the effect of the collector.

The developed method showed reliability when compared to the classical method of images. Through the studied wind tunnel tests, the method showed enough flexibility to be applied to solve both aeronautical and automotive models and several test section configurations with a reasonable computational efficiency.