AIAA 2002-2794 Uncertainty of Videogrammetric Techniques used for Aerodynamic Testing

The uncertainty of videogrammetric techniques used for the measurement of static aeroelastic wind tunnel model deformation and wind tunnel model pitch angle is discussed. Sensitivity analyses and geometrical considerations of uncertainty are augmented by analyses of experimental data in which videogrammetric angle measurements were taken simultaneously with precision servo accelerometers corrected for dynamics. An analysis of variance (ANOVA) to examine error dependence on angle of attack, sensor used (inertial or optical), and on tunnel state variables such as Mach number is presented. Experimental comparisons with a high-accuracy indexing table are presented. Small roll angles are found to introduce a zero-shift in the measured angles. It is shown experimentally that, provided the proper constraints necessary for a solution are met, a singlecamera solution can be comparable to a 2-camera intersection result. The relative immunity of optical techniques to dynamics is illustrated. INTRODUCTION As demand and usage increases for a particular test technique, issues related to the uncertainty of the measurement technique become more important. For instance, it is not uncommon for better accuracy to be desired as a given technique is used more and more. This has certainly been the case for pitch angle measurements in wind tunnels where the desired accuracy has slowly changed through the years down to less than 0.01° for some applications. Possible new uses of the data from the technique may also arise, once the data is available on a nearly routine basis, that may either require better accuracy or the use of the measurement technique in a different mode. Thus not only is the accuracy of a given technique of ________________________ * Senior Research Scientist, Senior Member † Research Scientist, Member ‡ Senior Research Scientist, Member Copyright © 2002 by the American Institute of Aeronautics and Astronautics, Inc. No copyright is asserted in the United States under Title 17, U.S. Code. The U.S. Government has a royalty-free license to exercise all rights under the copyright claimed herein for Governmental purposes. All other rights are reserved by the copyright owner. fundamental importance, but, in addition, the sensitivity of the technique to variations in setup geometry, etc. are of interest to assess the possible loss in accuracy that might be necessary to accommodate unique requirements. The importance of supplementing uncertainty analyses with experimental error assessments and the use of modern design of experiments methods (MDOE) has been emphasized in reference 1. Videogrammetric techniques combine photogrammetry, solid-state area-array cameras and image processing to yield rapid spatial measurements. Videogrammetric techniques have been used at a number of NASA facilities, primarily for the measurement of static aeroelastic model deformation. Variations of videogrammetric techniques include single-camera and multiple-camera implementations, depending on the nature of the application. For example, a singlecamera, single-view implementation at the Langley National Transonic Facility (NTF) has been used for the last ten customer tests. A typical image used for measurements at the NTF is presented in figure 1. For most of these tests model static aeroelastic data were obtained for nearly every data point acquired throughout the test, resulting in many thousands of data points of static aeroelastic data per test such as the typical example in figure 2. The interest and demand by industry for these aeroelastic measurements has increased dramatically as the technique has been improved without compromising facility productivity. A more complete and thorough uncertainty analysis will improve the value of this deformation data supplied to industry. More recent investigations with videogrammetric techniques include the measurement of aerodynamic loads, developments for use with micro air vehicles (both fixedand flapping-wing), for ultra-light and inflatable large space structures during and after deployment, and for in-flight aeroelastic measurements. Techniques are also under study for laboratory and wind tunnel dynamic measurements at frequencies up to 1000 Hz. The second aerodynamic measurement to be considered is pitch angle measurement, which is a fundamental measurement requirement of all wind tunnels. Note