A frequency‐domain noniterative algorithm for structural parameter identification of shear buildings subjected to frequent earthquakes. (19th September 2019)

• Record Type: Journal Article
• Authors: Zhang, Jingyao; Aoki, Takayoshi
• Other Names: Beck James L. guestEditor.; Bursi Oreste S. guestEditor.; Kurata Masahiro guestEditor.
• Is Part Of: Computer-aided civil and infrastructure engineering. Volume 35:Number 6(2020:Jun.)
• Page Start: 615
• Page End: 627
• Publication Date: 2019-09-19
• Subjects: Civil engineering -- Data processing -- Periodicals; Computer-aided engineering -- Periodicals; 624.0285
• DOI: 10.1111/mice.12502
• Languages: English
• ISSNs: 1093-9687

Abstract

Abstract: System identification is the key technique for damage detection in application of structural health monitoring. In contrast to modal parameters, changes in structural parameters (stiffness and damping) are more sensitive and straightforward for damage detection of a building under severe environments such as earthquakes. In this study, we first present the fundamental theory for direct identification of structural parameters by using the frequency‐domain responses of a shear building in frequent earthquakes. Shear buildings are widely adopted for structural analysis of low‐ and middle‐rise buildings in practice. Modal information, in terms of spectrum ratios, is implicitly used in the proposed noniterative algorithm to greatly improve the estimation accuracy as well as to avoid any human intervention. The fundamental theory is validated by the numerical and physical examples. The numerical examples are further used to verify the high efficiency, accuracy, and robustness of the proposed algorithm against noised responses. The proposed algorithm is highly efficient because no iterative computation is necessary, while the necessary Fourier transform of the dynamic responses is not very time consuming. Furthermore, the proposed algorithm is highly accurate and robust because (a) the fundamental theory behind the algorithm is straightforward: the identification values should have the same value irrespective of circular frequencies, according to the theory; (b) error in modal parameter identification is completely avoided because it is unnecessary to identify the exact values of the frequencies as in many existing methods.