Vessel collisions on bridge piers have become a potential threat to the safety of bridges crossing navigation waterways. Such collision will cause inevitable damage on bridge piers and hence reduce the performance of the whole structure. It is therefore critical to identify the condition of abridge pier after a vessel collision event to judge whether it can still be used or certain rehabilitation is required to recover its normal operation. This paper develops an intelligent approach based on machine learning algorithms to identify the evolution process of damage on abridge pier during collision using sensor-measured acceleration time-history data considering the effects of multi-hazards. A barge vessel is employed and atypical reinforced concrete (RC) bridge pier is considered in this study. A coupled vessel-pier collision model (CVCM) considering soil-pile interactions and material non-linearity of RC components is developed and employed to generate pseudo-experimental data to assess the accuracy of the proposed damage identification strategy. The results demonstrate the potential of the proposed strategy for intelligent damage identification of waterway-crossing bridge piers after vessel collision.
