Research On Unsteady Aerodynamic Modeling Method Of Aircraft

In this thesis, we proposed new principles of modeling method based on support vector machine to solve unsteady aerodynamics problem. I found that this method has a big advantage on precision compared with traditional modeling methods.It get the sample data from wind tunnel test on large amplitude motion. This experiment is in the NH-2 low speed wind tunnel of University of Nanjing Aeronautics and Astronautics. In this experiment, the model has three motions, single degree of freedom roll movement, single degree of freedom to yaw movement and yaw roll coupling motion. It get the unsteady aerodynamics of model from dynamic testing system for large amplitude motion. Then use fast Fourier transform do with experimental data to filter. It convert the data into frequency domain, find the noise signal frequency range and remove them, then converted back to time domain, to implement the filtering of the data.Support vector machine modeling is the main content in this thesis. It introduces the principle and method of support vector machine modeling and writes the original programs and get the unsteady aerodynamic model. The research shows that support vector machine modeling is a reliable modeling method. Compared with fuzzy logic modeling, it has a better result. In addressing some problems of small sample size, it has the advantage of quick and accurate.Finally, I did some application on support vector machine modeling. I computed the longitudinal damping derivative of model based on dynamic derivatives simulation principle and pitching moment coefficient. The result shows that the longitudinal damping derivative reflects the model of longitudinal damping characteristics. In the same time I can get more accurate dynamic derivatives from adjust the simulation calculation of vibration frequency.