Probabilistic Modeling of Bridge Deck Unseating during Hurricane Events

Abstract

Although coastal bridges have exhibited significant susceptibility to damage during hurricane-induced wave and surge events, probabilistic models are lacking to quantify the vulnerability of bridges under a range of structural or hazard parameters and support hurricane risk assessment and mitigation activities. This paper introduces a computationally efficient methodology to assess the fragility, or conditional failure probability, of bridges, targeting the deck shifting or unseating failure mode, which is a predominant severe mode of failure for vast inventories of simply supported span bridges. The method propagates uncertainties in parameters affecting the structural capacity and demand, such as mass density, connection strength, material properties, workmanship, and wave parameters. Fragility surfaces are derived in which failure probability is presented over a range of relative surge elevation and wave heights. The application of this technique is shown through a regional fragility assessment of 136 bridges in the greater Houston area. The results reveal that a failure zone emerges in the fragility surfaces of the bridges studied, with a dramatic transition in probability of failure. Additionally, bridges with similar trends in fragility surfaces can be readily classified based on estimates of their mean mass per span length. Finally, application of the new vulnerability models in a case-study regional risk assessment for Hurricane Ike reveals consistency with observed damage and also offers an opportunity for future studies to investigate alternative scenarios for risk-mitigation planning or extensions of the methodology to consider other regions or failure modes.