Structural system identification and sensor placement

Abstract

Field test is an indispensable tool to determine the dynamic characteristics of structures in terms of their natural frequencies, modeshapes and damping. The dynamic characteristics of a suspension footbridge and a high-rise commercial building are investigated through modal identification using ambient vibration data with a limited number of sensors. This study also verifies the stationary assumption in ambient vibration data for modal identification. A conventional modal identification method based on a least-square minimization of the discrepancy between the correlation function of the measured response and modelled response is adopted. Since the quality of measured dynamic data is a critical issue not only for modal identification but also for structural model updating and health monitoring, the use of optimal sensor placement technique in enhancing the performance of system identification and structural health monitoring is investigated. A statistical approach is presented for the identification of an effective way to install a given number of sensors on a structure to extract as much information as possible, or equivalently to minimize the uncertainties associated with the results of structural model updating. The problem of optimal sensor placement is formulated as a discrete optimization problem, in which the information entropy is minimized, with sensor configurations as the minimization variables. The methodology is illustrated numerically and experimentally using shear building models. The performance of the optimal sensor placement technique is verified using the results of model updating based on measured acceleration responses of a 4-storey shear building model under laboratory conditions. In addition, a computationally efficient numerical optimization algorithm is proposed following a Genetic Algorithm to solve the discrete optimization problem for identifying a group of “optimal” sensor configurations. A bridge model subjected to an impulse and a transmission tower subjected to random excitation are adopted as examples to illustrate and verify the performance of the proposed methodology.

Date of Award	2 Oct 2009
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Siu Kui AU (Supervisor) & Heung Fai LAM (Co-supervisor)