Seismic fragility and resilience assessment of instrumented Cogswell concrete-faced rockfill dam using scalar and vector-valued approaches

Abstract

Assessing the seismic vulnerability of dams is crucial due to the significant risks posed to downstream populations in the event of earthquake-induced failures. Fragility analysis is an effective tool for estimating earthquake-induced damage, thereby playing a vital role in dam design and the prevention of catastrophic incidents. In this study, Cogswell dam, a Concrete Faced Rockfill Dam, is considered as a case study, leveraging its recorded seismic responses from the Whittier Narrows (1987) and Sierra Madre (1991) earthquakes, as captured by monitoring instrumentation. A series of nonlinear 2D dynamic analyses were performed using Finite Element Method (FEM) and the results were validated against the reported data by the dam's instrumentation. Then, scalar and vector fragility assessments were performed by applying 100 diverse recorded ground motions. According to the results, scalar fragility analysis highlights that Intensity Measures (IMs) like PGA, A_rms, I_c, ASI, EDA, and A95 can be considered as key optimal predictors of seismic vulnerability, with PGA demonstrating the strongest correlation and A_rms associated with higher probabilities of severe damage. Vector-valued fragility analysis reveals that IM combinations, particularly I_c-A_rms, provide more conservative estimates, capturing critical interactions between IMs. In the next step, scalar and vector resilience assessments are conducted considering various functionality loss scenarios and different recovery functions. Resilience assessment shows that earthquakes with high PGA enables faster recovery, while the ones with high A_rms significantly impacts recovery functionality. Also, it is observed that scalar-valued resilience assessment provides more conservative results compared to the vector-valued approach.