摘要

Bidirectional shake-table tests were performed on two types of relatively large two-span continuous elevated-girder bridge models. One was supported on thin-walled hollow steel (HT) piers, and the other was supported on concrete-filled tubular (CFT) piers. The CFT piers were made by filling the hollow columns of the HT piers with normal concrete. One objective of the shake-table tests was to acquire basic data to establish an advanced seismic design method of the HT and CFT piers under bidirectional horizontal seismic accelerations, considering the interaction between the piers, rubber bearings, and a superstructure. The other objective was to experimentally confirm the effectiveness of upgrading the elevated-girder bridges with the HT piers by using CFT piers. It was shown by the present tests that the ultimate state of the respective HT piers in the bridge model was predicted reasonably well by the interaction equation expressed in terms of the multiaxial force and moment components acting at the top of the columns. As soon as all the piers reached their ultimate states, the entire physical bridge model with the HT piers started to collapse in the transverse (TR) direction. Because of the load redistribution between the HT piers in the bridge model during the shake-table test, the energy dissipation capacity of the piers increased considerably compared with that evaluated by the conventional shake-table test on single cantilever column models with a mass fixed to their top. The physical bridge model with the CFT piers exhibited an excellently upgraded seismic performance (i.e., small local buckling deformations of the piers and negligibly small residual sway displacements of the bridge system) despite the increased input acceleration wave with an amplitude 1.5 times larger than that used for the physical bridge model with the HT piers. This upgraded performance of the CFT piers was due to their enhanced energy dissipation capacity and unique local buckling restraining and/or restoring behavior.