Comparative analysis of stochastic variables for seismic vulnerability assessment of concrete gravity dams
摘要
Accurately assessing the seismic vulnerability of concrete gravity dams to damage or cracking during earthquakes is crucial for evaluating their performance. This study employs a comprehensive approach that integrates both epistemic and aleatory uncertainties, addressing a significant gap in the literature regarding random variables specific to concrete gravity dams. To achieve this, the research utilizes Latin hypercube sampling (LHS) to systematically sample structural uncertainties related to material properties, modelling errors, and the inherent randomness of seismic events. A thorough sensitivity analysis is conducted, represented by a Tornado diagram, to evaluate the impact of various stochastic variables on dam behaviour. The Pine Flat Dam is selected as a case study, and its seismic response is assessed using the incremental dynamic analysis (IDA) method. The analysis reveals that variations in certain stochastic variables significantly influence the fragility of the dam. Specifically, the elastic modulus of the foundation rock and concrete, as well as the damping coefficients, are identified as critical factors affecting the maximum crest displacement and cumulative damage energy dissipation. The findings of this study contribute significantly to enhancing the robustness and reliability of seismic vulnerability assessment methodologies for concrete gravity dams by comprehensively addressing multiple sources of uncertainty and providing insights into the key variables that dictate dam behaviour under seismic loading conditions.