Modified Data-driven Stochastic Subspace Identification
- Paper number
IAC-07-C2.3.09
- Author
Mr. Yueyu Wang, China Academy of Space Technology (CAST), China
- Year
2007
- Abstract
Identified structural modal characteristics can be served as important basis for attitude control and real-time monitoring of large complex space structures such as space station, stratosphere airship etc. Although modal parameter identification under ambient excitation does not demand to interrupt the operation of engineering structures, it has some disadvantages such as low signal noise ratio (SNR) and great amount of response data. In the last two decades several techniques, such as data-driven stochastic subspace identification (DSSI), covariance-driven stochastic subspace identification (CSSI), and the combination of Natural Excitation Technique (NExT) and Eigensystem Realization Algorithm (ERA) are found very useful for applications to identifying structural modal parameters under ambient excitation. DSSI obviates the calculation of covariance matrices and utilizes some well-known numerical linear algebra algorithms, such as singular value decomposition (SVD), QR decomposition, and least square etc. In addition, DSSI is easy to realize the modal identification on-line. However, the identification efficiency of DSSI is limited by the efficiency of QR decomposition enormously when the amount of response data is great. To deal with the problem, this paper proposes a modified DSSI which does not require constructing high dimension block Hankel matrix (equivalent to matrix Y in this paper). In the developed identification method, simplified processing of the matrix YY T is carried out firstly. Then Cholesky decomposition to the simplified matrix YY T is performed and nonsingular triangle matrix R is gained. The upper triangle elements of matrix R are equal to that of matrix ρ obtained from the block Hankel matrix by QR decomposition. While maintaining the same identification precision, the modified DSSI raises calculation efficiency at least ten times. In this paper, the modified DSSI is presented in detail and proved to be reliable and efficient theoretically and by numerical results
- Abstract document
- Manuscript document
IAC-07-C2.3.09.pdf (🔒 authorized access only).
To get the manuscript, please contact IAF Secretariat.