Abstract:
With the use of integrated soil-foundation-structure modeling in seismic bridge design, there is an inherent difficulty in determining the validity of the assumed stiffness and damping of the system and the contribution of various components of the structure and foundation to the overall system response. Validating these assumptions in a laboratory setting is particularly difficult due to problems associated with the appropriate representation of realistic boundary conditions, the variability of soil properties within a bridge site, and scaling of soil-structure systems for laboratory analysis. Some of these problems can be overcome by testing of full scale bridges in the field. This paper provides an overview of the forced vibration testing and computational modeling to determine the in situ dynamic properties of an in-service bridge in Auckland, New Zealand. Testing was performed over two nights with lateral excitation provided by an eccentric mass shaker in transverse axis of the bridge. The bridge-foundation system response was measured with 180 accelerometer channels on the bridge structure. Soils surrounding the piers and abutment were characterized with CPT tests to a depth of 20 m. Natural periods and mode shapes were extracted from the experimental data using a suite of system identification algorithms. Modal properties from both fixed based and a simplified integrated structurefoundation model were compared to those identified from field testing. It was determined that the simple integrated model was able to represent the in situ dynamic properties of the bridge and provided a much better prediction of seismic demand distribution than the fixed base model, which under-predicted the demand on the abutments and over-predicted the demand on the piers.