本文探討資料型隨機子空間識別法(Data-driven Stochastic Subspace Identification,SSI-DATA)於定常環境振動模態參數識別之應用。隨機子空間識別法可分為協方差型的隨機子空間識別法(Covariance-driven Stochastic Subspace Identification,SSI-COV)及SSI-DATA，SSI-COV是由輸出資料計算相關函數，再進行模態參數識別；SSI-DATA則是藉由計算未來輸出向量投射至過去輸出向量空間的投影矩陣，再進行模態參數識別。SSI-COV在計算上需進行相關函數運算且相關函數須滿足無限長條件，但因資料無限長條件未能滿足之情況，故SSI-DATA在推導過程較為完備。

而SSI-DATA在計算的過程，容易有資料過大，計算效率不彰的情況，故吾人也提出幾點改良以提高其計算效率。吾人將SSI-COV求取系統矩陣的作法引入SSI-DATA，並且探討可觀測矩陣的選取，進而去推導投影矩陣的改良，爾後對改良之投影矩陣進行奇異值分解(Singular Value Decomposition, SVD)時，才不會有矩陣維度過大使計算難度增加的情況發生。也提出另一點是由同ㄧ筆預估狀態矩陣去提取出兩個有遞迴關係的預估狀態矩陣，因此可省略掉重新計算下一時刻預估狀態矩陣的步驟，進而提高計算效率，且也可配合改良之投影矩陣的作法，進而使計算效率達到最大化。

The purpose of this thesis discusses the application of Data-driven Stochastic Subspace Identification (SSI-DATA) in Modal-Parameter Identification of stationary ambient vibration. SSI can be divided into Covariance-driven Stochastic Subspace Identification (SSI-COV) and SSI-DATA. SSI-COV is about calculating the correlation functions by output data, and then proceeding the Modal-Parameter Identification. However, SSI-DATA calculates future outputs, projects to the projection of past outputs, then, proceeds the Modal-Parameter Identification. It’s necessary to count the correlation functions which must reach the unlimited condition when using SSI-COV. But the unlimited condition can’t be satisfied. Therefore, using SSI-DATA in deductive process is more complete.
This thesis also proposes several improvements to enhance the computational efficiency, as a result of the poor computational efficiency, and the excessive quantity of data in the calculating process when using SSI-DATA. After introducing the method that SSI-COV quests for the system matrix into SSI-DATA, discussing the selection of observability matrix, and then deriving improvements of projection. The situation that dimensions of the matrix is too many which raises the calculation difficulty won’t happen when using Singular Value Decomposition (SVD) in the improved projection. Another point is to acquire two estimate matrix which have recursion relation bases on an estimate matrix. It’s not only can neglect the step to recalculate the next moment estimate matrix to enhance the computational efficiency, but also can coordinate the manner of the improved projection to reach the maximum computational efficiency.

參考文獻
[1] Ewins, D. J., “Modal Testing: Theory and Practice,” Research Studies Press, 1984.
[2] Ibrahim, S. R. and Mikulcik, E. C., “A Method for the Direst Identification of Vibration Parameters from Free Response,” Shock and Vibration Bulletin, Vol. 47, Part. 4, pp. 183-198, Sept. 1977.
[3] Ibrahim, S. R. and Mikulcik, E. C., “The experimental Determination of Vibration Parameters from Time Responses,” Shock and Vibration Bulletin, Vol. 46, Pt. 5, pp. 183-198, Aug. 1976.
[4] Pappa, R. S. and Ibrahim, S. R., “A Parametric Study of Ibrahim Time Domain Modal Analysis,” Shock and Vibration Bulletin, Vol.51, Part. 3, pp. 43-72, 1981.
[5] 林宇哲, 亞伯拉罕時域法於具模態干涉之系統模態參數識別, 碩士論文, 國立成功大學航空太空工程學研究所, 2016.
[6] Juang, J. N. and Pappa, R. S., “An Eigensystem Realization Algorithm for　Modal Parameter Identification and Modal Reduction, ”Journal of　Gudance　and Control Dynamics,AIAA,” Vol.8, No. 5, pp.620-627, 1985.
[7] Juang, J. N. and Pappa, R. S., “Effects of Noise on Modal Parameter　Identification by the Eigensystem Realization Algorithm, ”Journal of　Gudance and Control Dynamics,” AIAA, Vol. 9, No. 3, pp.294-303, 1986.
[8] Juang, J. N., Cooper, J. E. and Wright J. R., “An Eigensystem Realization Algorithm using Data Correlations(ERA/DC) for Modal Parameter Identification,” Control-Theory and Advanced Technology, Vol. 4, No. 1, pp. 5-14, March, 1988.
[9] 陳忠義, 相近模態在時域之模態參數識別, 碩士論文, 國立成功大學航空太空工程學研究所, 2015
[10] James, G.H., Carne.T.G.and Lauffer, J.P., “The Natural Excitation Technique for modal Parameter Extraction from Operating Wind Yurbines,” SAND92-1666.UC-261, Sandia National Aboratories, 1993.
[11] Carne, T.G., Lauffer, J.P., Gomez,A.J.and Benjannet, H., “Modal Testing an Immense Flexible Structure Using Narural and Artificial Excitation,” The International Joural of Analytical and Experimental Modal Analysis,The Society of Experimental Mechanics, Oct. 1998.
[12] 林章生, 相關函數法於非定常環境振動之模態參數識別研究, 碩士論文, 國立成功大學航空太空工程學研究所, 2004.
[13] 曹松華, 隨機遞減法於非定常環境振動模態參數識別之應用, 碩士論文, 國立成功大學航空太空工程學研究所, 2004.
[14] 蘇芳禾, 特徵系統實現法於定常環境振動之模態參數識別研究, 碩士論文, 國立成功大學航空太空工程學研究所, 2006.
[15] Van Overschee, P., De Moor B., “Subspace Algorithm for the Stochastic　Identification Problem, ” In Proceedings of the 30th IEEE Conference on Decision and Control, pp. 1321-1326, 1991.
[16] Peeters, B. and Roeck, G.D., “Reference-based stochastic subspace identification for output-only modal analysis,” Mech. Syst. Signal Process. 13 (6), 855–878. 1999.
[17] Van Overschee, P., and De Moor B., “Subspace Algorithm for the Stochastic Identification Problem,” Automatic, Vol. 29, No. 3, pp. 649-660, 1993.
[18] G. H. Golub and C. Reinsch., “Singular Value Decomposition and Least Squares Solutions,” Numerische Mathematik, 14:403-420, 1970.
[19] Strang G. and Borre K., “Linear algebra, geodesy, and GPS,” Wellesley-Cambridge Press,”1997.
[20] Erwin Kreyszig., “Advanced Engineering Mathematics,” 10th Edition, 2010.
[21] James Hu, S. L., Yang, W. L., Liu, F.S. and Li, H. J., “Fundamental comparison of time-domain exoerimental modal analysis methods based on high-and first-order matrix models”, Journal of sound and Vibration, Vol.333, No.25, pp.6869-6884, 2014.
[22] K. A. Petsounis and S. D. Fassois, “Parametric Time-Domain Methods For The Identification Of Vibrating Structure-A Critical Comparison And Assessment”, Mechanical Systems and Signal Processing, Vol.15, No.6, pp.1031-1060, 2001.
[23] Chang-sheng Lin, “Ambient modal identification using non-stationary correlation technique”, Archive of Applied Mechanics, 2016.