Effect of ground motion filtering on the dynamic response of a seismically isolated bridge with and without fault crossing considerations

Abstract High-pass filtering not only removes the low-frequency noise from the near-fault ground motion records, but also eliminates the permanent ground displacement and reduces the dynamic ground displacement. This may considerably influence the calculated seismic response of a spatially extended engineering structure crossing a fault rupture zone. To demonstrate the importance of incorporating permanent ground displacements in the analysis and design of extended structures under specific fault crossing conditions, the dynamic response of a seismically isolated bridge located in the vicinity of a surface fault rupture (“Case A”) or crossing a fault rupture zone (“Case B”) is calculated by utilizing a near-fault ground motion record processed with and without a displacement offset. The seismically isolated bridge considered in this study is a 10-span continuous structure supported by 11 piers, resembling a typical segment of the 2.3 km long Bolu Viaduct 1 located in west-central Turkey. The Lucerne Valley record from the 1992 M w 7.2 Landers earthquake, which preserves a permanent ground displacement in the fault-parallel direction and exhibits a large velocity pulse in the fault-normal direction, is used as the basis for investigating the effect of high-pass filtering on the dynamic response of the bridge. For the seismically isolated bridge located in the vicinity of the surface fault rupture (“Case A”), the utilization of the high-pass filtered ground motion leads to underestimating the demands of pier top, pier bottom and deck displacements. However, the demands of isolation displacement, isolation permanent displacement and pier drift are almost identical for both the unfiltered and filtered versions of the ground motion record. On the other hand, for the seismically isolated bridge traversed by a fault rupture zone (“Case B”), all response quantities are significantly underestimated when the high-pass filtered ground motion is used. These results, though limited to a single bridge structure and a single ground motion input, clearly indicate the importance of permanent ground displacement on the dynamic response of spatially extended engineering structures crossing fault rupture zones.