Search

This paper presents a unique application of spline modeling and wavelet analysis of largely experimental aerodynamic data. Wavelet analysis has more generally been associated with signal processing. Shrinking budgets for the Department of Defense and other government agencies and industry have resulted in a need for improvements in the processes for simulation aerodynamic model development and validation. A more objective and automated process for verifying and validating aerodynamic models would significantly reduce the time, and therefore cost, currently required to accomplish these vital tasks. This paper presents preliminary results obtained using multi-dimensional spline wavelet software that was developed to more quickly and consistently detect irregularities in aerodynamic data. Results of analyses of aerodynamic data with both two and three independent variables are presented. Based on spline modeling, the software is capable of analyzing aerodynamic data acquired during wind tunnel testing as well as multi-dimensional data tables that are used in aerodynamic simulation models. Two significant requirements for this particular application were that the wavelet analysis algorithm be capable of accepting unevenly spaced input data and that the resulting wavelet coefficients be produced at every data point. The current status of the software provides for the analysis of data that are a function of up to three independent variables. © 2001 The American Institute of Aeronautics and Astronautics, Inc. All rights reserved.

High fidelity aircraft simulations can dramatically reduce training and proficiency-related expenses, and can also reduce overall design and development costs by revealing problems early in the design process. However, developing, updating, and verifying simulation models is currently very expensive. Tremendous savings could result from innovative software tools that provide an objective and somewhat automated process for developing and verifying simulation aerodynamic models. The mathematical properties of wavelets appear to be a good match for the properties needed in a software tool capable of automatically developing and conducting verification and validation of simulation models. The primary reason for selecting wavelets over other methods is that wavelets are capable of performing localized analysis of a signal. Wavelets capture details in a signal at different resolutions at the same time. This paper addresses the feasibility of using wavelets to quickly and consistently detect irregularities in a large set of aerodynamic data. Two examples are presented. For demonstrative purposes, the cubic B-spline wavelet was used to smooth a noisy signal while obtaining 95% compression and retaining desired properties in the signal. Additionally, the linear wavelet was used to detect trends and irregularities in wind tunnel data consisting of seventeen separate data sets.