Second-order reliability methods: a review and comparative study

Abstract Second-order reliability methods are commonly used for the computation of reliability, defined as probability satisfying an intended function in presence uncertainties. These can achieve highly accurate predictions owing to a second-order approximation limit-state around Most Probable Point failure. Although numerous formulations have been developed, lack full-scale comparative studies has led dubiety regarding selection suitable method specific analysis problem. In this study, performance is assessed based on problem scale, curvatures at failure, first-order index, and contour. The assessment three metrics: capability, accuracy, robustness. capability measure ability compute feasible probabilities, i.e., probabilities between 0 1. accuracy robustness quantified mean standard deviation relative errors with respect exact reliabilities, respectively. This study not only provides review classical novel methods, but also gives insight appropriate given engineering application.