Seismic Monitoring of a Bridge: Assessing Dynamic Characteristics From Both Weak and Strong Ground Excitations

Abstract

SUMMARYAn array of 24 strong-motion accelerometers produced records for the New-Lian River Bridge, a five-span continuous bridge, during 25 February 1995 earthquake (weak motion) and 25 June 1995 earthquake (strong motion). This paper describes the application of linear discrete-time system identification methodology to the array of strong-motion measurements, in order to assess seismic response characteristics of the bridge. The structural system identification will concentrate not only on the global identification but also on the local structural system identification. Results of this application show that: (1) weak and strong ground excitation will induce significant differences on the dynamic response of the bridge; (2) linear models provide an excellent fit to the measured motions of the bridge from the records of these two seismic events; (3) the rigid-body rocking of the bridge pier during strong shaking is significant and cannot be ignored during identification; (4) the transverse motion at mid-span of the bridge is controlled by the quasi-static response from the boundary system and this phenomenon is quite significant during strong ground excitation. Also, systematic estimates of modal damping ratio and equivalent assessments of pier stiffness developed in the bridge during earthquake are discussed.

Variation of modal parameters of a highway bridge extracted from six earthquake records

“…An increase in the period of the fundamental vibration mode is found from one earthquake to another. Later, both weak and strong ground motions are used to assess seismic response characteristics of a five-span continuous bridge [10]. That study shows that weak and strong ground excitation induce significant differences on the dynamic response of the bridge.…”

“…In that paper, the first few vibration modes of the bridge that participate significantly in response are identified using the bridge's response to the 1979 Imperial Valley earthquake. Later, both weak and strong ground motions are used to assess seismic response characteristics of a five-span continuous bridge [10]. Later, the performance of the two finite element models (a linear stick model and a detailed nonlinear model) of the Painter Street Overpass is compared [8].…”

Effects of liquefiable soil and bridge modelling parameters on the seismic reliability of critical structural components

“…Uncertainty in dynamic soil properties due to the innate variability of soils and the inherent and induced anisotropy may necessitate soil sampling to derive more statistically significant databases. Furthermore, many bridge monitoring paradigms target interrogation of structural response in order to assess global dynamic characteristics and damping ratios (a significant parameter in this study) via various system identification techniques (Ghanem and Shinozuka 1995, Farrar and James 1997, Loh and Lee 1997, Alvin et al 2003. Given the significance of such modelling parameters as bearing coefficients of friction that may change in time with debris or corrosion products, localised monitoring and parameter estimation (Kim 2003, Catbas et al 2008) as well as deterioration monitoring, such as corrosion rate sensing (Brinkerhoff 1993), may also warrant increased attention.…”

Advanced signal processing for structural identification: experimental studies

“…To overcome these shortcomings, a sub-structure strategy is adopted to perform local structural identification. Loh and Lee [12] wrote substructure motion equations in the state form and used a Kalman filter to estimate sub-structure parameters. Yun et al [3,13] proposed sub-structural identification methods based on a sequential prediction error method and neural networks for local damage assessments.…”