Development of Physics-Based Tsunami Fragility Functions Considering Structural Member Failures

A probabilistic framework is presented for the development of physics and simulation-based parametrized tsunami fragility functions for structures accounting for structural member failures. The proposed framework is general and accounts for material and geometric sources of uncertainty and makes use of nonlinear finite-element structural models and the first-order second-moment (FOSM) reliability method. The application of the framework is illustrated with the development of parametrized fragility functions for an example reinforced concrete moment frame building designed to recent United States codes. Results indicate that explicit consideration of structural member failures is of paramount importance because the fragility functions based on global failure criteria that do not account for member failures tend to overpredict damage state capacities. Among the several sources of uncertainty considered, breakaway openings in the building are the dominant contributor to the uncertainty in the structural capacity. In addition, the estimation efficiency of several scalar and vector-valued intensity measures as predictors of structural damage is evaluated using the logistic regression method. The intensity measures considered consist of inundation depth, flow velocity, specific momentum flux, kinematic moment of specific momentum flux, and their interactions. The estimation efficiency of vector-valued intensity measures is found to be higher than that of scalar intensity measures. Among the scalar intensity measures analyzed, those that combine information of inundation depth and flow velocity are identified to be the most efficient predictors of structural damage, and therefore are considered to be the preferred measures to characterize the intensity of tsunami hazards for practical applications.