Fragility curves play essential roles in probability assessment. A challenge in developing these curves lies in the requirement for extensive observational or analytical data, resulting in high costs. This paper proposes a new procedure for creating fragility curves that uses Bayesian updating and Bayesian optimization. A key feature of this method is that it identifies these variables using joint distribution analysis with Bayesian updating and defines confidence intervals accordingly. Moreover, Bayesian optimization adaptively selects high sensitivity points for observation. Numerical experiments confirm the effectiveness of this approach in producing accurate fragility curves even under data constraints.
