Singular Value Decomposition for the Application of Modal Estimation

in International Symposium (oral presentation paper), 國際研討會(全文口頭發表)
標題Singular Value Decomposition for the Application of Modal Estimation
AuthorsLin, 林章生
出版日期Nov 14 2016 12:0

Previous studies show that when a system has repeated modes, the frequency-domain method of using multiple-input and multiple-output method can effectively identify the modal parameters. The topic of this study is to investigate the effectiveness of time-domain modal-identification method for systems with repeated eignvalues. When employing the eigensystem realization algorithm method (ERA) for modal identification from response data in consideration of the measurement noise, numerical simulations show that the identification result in the damping ratio is usually poor. Therefore, by introducing the concept of the correlation function in modal identification, the influence of noise on identification result is reduced. Furthermore, the correlation matrix is directly applied to extract modal parameters, and the process of constructing the generalized Hankel matrix is therefore simplified so that it provides a more efficient method of modal identification when comparing the conventional ERA and ERA/DC.

校址:912 屏東縣內埔鄉學府路1號 總機:886-8-7703202 傳真:886-8-7740165 系統開發統維護單位:國立屏東科技大學 電算中心 版權所有