Reliability updating in the presence of distribution parameter uncertainty

Abstract

Reliability updating (RU) is often used to re-evaluate the failure probability of structural system when new observations are obtained. However, in engineering practice, the specific distribution parameters of input variables are generally unavailable due to the lack of the data. To this end, two reliability updating models with distribution parameter uncertainty (DPU) are firstly constructed in this paper based on the theory of nested reliability approach (NRA) and augmented reliability approach (ARA). The constructed reliability updating models can accurately measure the effect of the DPU on the posterior failure probability. Then, two reliability updating algorithms based on subset simulation (SS), named as SS-NRA-RU and SS-ARA-RU, are developed to solve the corresponding RU model, which can significantly improve the computational efficiency of RU of rare event. Finally, three examples prove the efficiency and accuracy of the proposed reliability updating algorithms.