| The response of simple isolated bridge overcrossings to a variety of multi-dimensional earthquake inputs was investigated by analysis and experiment. These studies were undertaken as part of the Protective Systems Research Program at UC Berkeley sponsored by the California Department of Transportation to establish an understanding of global and local behavior characteristics, including the effect on response of variations in isolator properties, pier flexibility, mass and strength, substructure damping, and global mass and stiffness eccentricity. A versatile 114-scale bridge model on flexible piers, enabling one- or two-span configurations, was tested on an earthquake simulator to validate the findings of the analytical studies performed. Provisions of the current AASHTO Guide Specifications for Seismic Isolation Design were also evaluated using the results of the analytical and experimental investigations. Specifically, the application of the Uniform Load Method to the design of a broad range of isolated bridge systems was studied. The effect of mass and damping assumptions, bi-directional ground motion inputs, and substructure strength on these procedures was also evaluated.; Several overriding observations were derived from these investigations. First, these studies illustrated the considerable overall ability of seismic isolation to provide an effective means of earthquake resistance in simple bridge overcrossings. The considerable durability and robustness of these systems under seismic loading was illustrated through multiple, varied simulations, including multi-dimensional inputs of far-field, near-fault, and soft-soil ground motions. Secondly, analytical evaluations undertaken were able to identify sensitivity of the nonlinear behavior of simple isolated bridge types to various ground motion and structural characteristics. Lastly, these studies established several areas in the current AASHTO Guide Specifications in need of further development. (1) Foremost, it was shown that the Uniform Load Method does not provide uniform reliability, pointing to the need for the refinement of current Performance-Based Design procedures. (2) Further, possible improvements to the Guide Specification provisions to more accurately account for bi-directional effects were suggested. (3) It was also shown that the present R-factor approach may not adequately control ductility demands in substructure components. (4) Conditions were identified for which the effects of substructure mass and damping may require more accurate consideration than presently provided for by typical approximate linearized single-degree-of-freedom design procedures. |
