本研究的目的在建立以循環類神經網路(Artificial Neural Networks)做為空氣動力荷載(Aerodynamic Loads)計算工具(的發展)，並將之應用於顫振流場(Flutter Flow)的分析當中。本研究的物理模型為二自由度典型葉片(Typical Section)，對其在不可壓縮流(Incompressible Flow)中因陣風(Gust)所導致的任意運動(Arbitrary Motion)做顫振分析(Flutter Analysis)。藉由有限的訓練數據來訓練循環類神經網路，必須避免過度配適(Overfitting)的情況發生，使網路可以做適當地廣義化(Generalization)，以擴大網路應用的範圍。本研究目前的工作內容包括循環神經網路的訓練與氣彈時間積分(Aeroelastic Time-Marching)程式的驗證。本研究在循環神經網路的訓練上已有良好成效，有助於後續研究發展。

The objective of the study is to develop the artificial neural networks as a calculate tool of aerodynamic loads, and apply on flutter flow analysis. The physical model is using the 2-D typical section to proceed the flutter analysis which is the arbitrary motion caused by the gust in incompressible flow. Training the artificial neural networks by limited training statistics must avoid the overfitting problem occurred, enable the networks generalization appropriately and broaden the application coverage of the networks. The research up to present contains the training of the Artificial Neural Networks and validation of the aeroelastic time-marching program. It shows a good effect upon the artificial neural networks and it is expected to benefit the future study.

[1] Dowell, E.H., “A Modern Course in Aeroelasticity,” Sijthoff and Noordoff, The Netherlands, 1995.
[2] Fung, Y.C., “An Introduction to the Theory of Aeroelasticity,” Dover, New York, 1993
[3] 葉怡成, “類神經網路模式應用與實作,” 儒林圖書有限公司, 2000年4月.
[4] Haykin, S., Neural Networks, A Comprehensive Foundation, Second Edition, Prentice Hall International, Inc. 1999.
[5] 徐自珍, “航空發動機人工神經網路故障診斷法,” 國立成功大學航空太空工程研究所碩士論文，1997年.
[6] Chen, S.-K., “Acoustic Flutter Suppression of Cascade in Inviscid and Viscous Transonic Flow,” Ph.D. Thesis, IAA NCKU, Taiwan, R.O.C. 1999.
[7] Faller, W.E., Schreck, S.J. and Luttges, M.W., ”Real-Time Prediction and Control of Three-Dimensional Unsteady Separated Flow Fields Using Neural Networks,” AIAA Paper 94-0532, January 10-13, Reno, NV, 1994.
[8] Faller, W.E. and Schreck, S. J., “Neural Networks: Applications an Opporunities in Aeronautics, ” Progress of Aerospace Science, Vol.32, 1996, pp.433-456.
[9] Edwards, J.W, Bennett, R.M., Whitlow, W., Jr., and Seidel, D.A., “Time-Marching Transonic Flutter Solutions Including Angle-of-Attack Effects,” AIAA Paper 82-0685, 1982.
[10] 陸鵬舉, “以類神經網路法預估及控制二維非定常流場,” 結案報告, 計劃編號NSC-88-2212-E-006-079, 1999年.
[11] Bennett, R.M. and Desmarais, R. N., “Curve Fitting of Aeroelastic Transient Response Data with Exponential Function,” Flutter Testing Techniques, NASA SP-415, May 1975, pp.43-58.

[12] Huang, L.J., “Optimal Flutter Control of Airfoils Using Active Acoustic Excitation,”Ph.D. Thesis, IAA NCKU, Taiwan, R.O.C. 1991.
[13] F.D.Marques and J.Anderson,”Identification and Prediction of Unsteady Transonic Aerodynamic Loads by Multi-Layer Functions,”Journal of Fluids and Structures, Vol.15, 2001, pp.83-106.
[14] 羅華強, “類神經網路─MATLAB的應用,” 清蔚科技股份有限公司, 2001年9月.