Girder‐to‐Girder and Abutment Chain Pounding and Collapse Mechanisms in Multi‐Frame Highway Bridges: Lessons From the 2021 Magnitude 7.4 Maduo Earthquake

ABSTRACT Two neighboring highway girder bridges exhibited significantly different damage during the 2021 Magnitude 7.4 Maduo earthquake. Yematan Bridge, a multi‐frame girder bridge (five frames each with five spans, denoted as 5 × 5 spans), suffered collapse of 70% of the total number of girders (mostly on the north side), expansion joint damage from girder‐to‐girder pounding, and severe north abutment damage due to girder‐to‐abutment pounding, whereas the single‐frame Heihe Middle Bridge with three spans (1 × 3 spans) remained almost intact. This contrast highlights the crucial roles of frame/span numbers and girder‐to‐girder/abutment pounding in bridge performance. In this regard, a comprehensive parametric analysis using finite element full bridge models is conducted to reproduce the observed damage and to explore pounding and collapse mechanisms, considering the effects of near‐fault pulse‐like ground motions (i.e., pulse period, velocity, and wave passage) as well as the frame/span numbers. A novel pounding Gantt plot is proposed to visualize the successive pounding events from girder‐to‐girder to girder‐to‐abutment, called chain pounding in this study. Results show that significant chain pounding occurs in Yematan Bridge under specific long‐period pulse‐like motions with wave‐passage effect, which is the primary cause of the severe north abutment damage and girder collapses. In contrast, such a devastating chain pounding is absent in the single‐frame Heihe Middle Bridge, where stronger abutment constraint effects on girders shift its vibration period away from unfavorable pulse periods of ground motions, thereby mitigating damage. Furthermore, decreasing the frame/span numbers reduces bearing displacements under long‐period pulse‐like ground motions, but increases them under short‐period motions. Thus, seismic design of multi‐frame and multi‐span highway bridges at near‐fault sites should carefully consider relationships between bridge length and pulse period of ground motions.