UPDATING RELIABILITY MODELS FOR EXISTING STRUCTURES USING MEASURED RESPONSES

摘要

The problem of determining the posterior safety margin in randomly parametered and driven structures when data on measured responses become available is considered. The first step in the solution consists of system parameter and (or) force identification and this is followed by the solution of an inverse reliability problem posed as a constrained nonlinear optimization problem. The solution strategy consists of the application of Bayes' theorem, a modified version of subset simulations, and numerical optimization tools. The proposed method is exemplified by considering the analysis of an 11-member truss.