Abstract:
This report describes a study which calculates the structural properties of a base isolated bridge from acceleration records obtained from· dynamic tests. This study provides information for engineers to evaluate the base isolation design philosophy, and gives guidance to how system identification can be proceeded with a nonlinear behaviouring structure. The work is based on the dynamic test records obtained from the Mangatewai-iti bridge which was tested in 1985 by a team of engineers jointly sponsored by the National Roads Board of New Zealand and the National Science Foundation in the United States. A band pass filtering technique is used to isolate individual modal responses of the slightly nonlinear experimental records, in order to obtain a better understanding of the dynamic behaviour of the bridge. The Fourier integral, with its upper and lower frequencies changed accordingly, is used as band pass filter in this study. The modal properties of the bridge can be obtained from these filtered results. However, this filtering technique can not be applied to the nonlinear structural response. For nonlinear structural response, the filtering technique does not give accurate results due to severe modal interferences. For the nonlinear behaviour of the bridge, a system identification method has been developed to identify the in-situ structural properties of the bridge, by comparing the power spectra of the experimental records to the power spectra of the acceleration records of the simulated bridge test. The comparison of these power spectra is carried out in the frequency bands which envelop the significant modal responses only. With the bridge structural properties identified, earthquake simulation is performed to study the dynamic behaviour of the Mangatewai-iti bridge. Based on: the results of the earthquake simulation of the bridge, recommendations are made for the analysis and design of base isolated bridges. From the results of the earthquake simulation, the localized vibration of piers are found to have a significant contribution to the maximum pier moments. A simplified procedure is developed for the estimation of the contribution of the localized pier vibration to the maximum pier moment. The use of sliding bearings is found to be effective in reducing the dynamic response of the bridge model in these simulations. As a result of this study, a base line error correction method has been developed to minimize the baseline distortion of the experimental acceleration records. The records are divided into a number of stages, and the baseline of each stage is approximated by a straight line. This correction method is based on the assumption that the first modal response dominates the vibration. This correction method provides a new direction for the research of baseline correction, but there are still some difficulties to be overcome before it can be accepted as a standard baseline correction method.