Dynamic behaviour of a box girder bridge under seismic loading considering soil-structure interaction

Abstract

Dynamic response during earthquake is a very important factor for both the serviceability and safety of bridge structures. The controlling parameters that govern dynamic response of a bridge depend on different structural attributes of a particular bridge. A number of schemes for identification of dynamic parameters of bridges considering soil-structure interaction have been developed in recent years. Superstructures of bridges used in the most of the schemes are, however, very simplified elastic models. The present study is focused on system identification of a box girder bridge considering soil-structure interaction effect. In Bangladesh only the Jamuna Multipurpose Bridge is instrumented with accelerometers. 4.8 kIn long Jamuna Multipurpose Bridge over the Jamuna River connects Bhuapur in Tangail with Siraj ganj and thus provides a very important road link between the East and West of Bangladesh. The bridge is located in a seismically active region that can be subjected to moderately strong earthquakes. Special earthquake protection devices have been used in the bridge. It has been designed to resist dynamic forces due to earthquakes with peak ground acceleration as high as O.2g. The earthquakes of 26th August 2008 and 16th June 2004 had excited the bridge; the bridge vibration was recorded by the sensors placed in the bridge. Jamuna Multipurpose Bridge is supported on hollow steel batter piles. The present study is aimed at determining the support properties of the underlying batter piles of the Jamuna Multipurpose Bridge considering soil-structure interaction. A numerical tool based on the Thin Layered Element Method (TLEM) has been used along with a number of assumptions to analyze the interaction between a pile and the surrounding soil. In the present study, the soil-structure interaction has been included in a previously developed Finite Element Models of the bridge. Modal.analyses have been performed using different FE models. Time history analyses using Finite Element model of a single module of the bridge have been conducted considering and not considering the effect of soil-structure interaction. Then, simulated responses of the Finite Element analyses have been compared with the recorded responses of the bridge. Since, the depth of river bed varies along the bridge length, the free length of pile effects the dynamic properties of the bridge. Moreover, the scour depth of the river bed shows seasonal variation in a year. The effect of scour depth on the substructure dynamic stiffness has been analyzed here. The variation of modal periods and responses due to the earthquake of 2008 has also been compared for different scour depth ofriver bed. Effect of radiation and material damping of the massive steel piles have been found negligible compared to that of the stiffuess on the pile-soil interaction. In this study, it is found that the simplified FE model, where frame elements are used for modeling transverse rail girders, shows spurious modes when soil-structure interaction effect is considered. The soil-structure interaction incorporated modes with longer modal periods to the structure, which are supposed to be excited due to earthquake containing long predominant periods. Time history analyses showed that the effect of soil-structure interaction was negligible for the earthquakes of 2004 and 2008, since these earthquakes had shorter predominant periods. Increase in modal periods has been found with the increase of scour depth. In the time history analysis, reduction in peak of accelerations of the responses in longitudinal and up-down direction have been found reduced due to increase in scour depth. This study proposes a method of incorporating effect of soil structure interaction in a bridge structure, as well as, demonstrates the importance of proper modelling of super-structure and consideration of the effect of scour depth of river bed in dynamic analyses of bridges.