Reliability analysis methods of structures considering modeling uncertainty

To fully investigate the influence of modeling uncertainty on reliability of structures, in this paper, the statistical reliability problem of structures considering two kinds of uncertainties, i. e., physical uncertainty and modeling uncertainty is deeply studied on the basis of Bayesian statistics. First, the fundamental principles of Bayesian statistics are summarized. The modeling uncertainty in structural engineering is then modeled and analyzed by Bayesian updating approach, and the various measures for structural statistical reliability are introduced. To treat with the aleatory and epistemic uncertainties, two methods, i. e., the integrated approach and the separate approach for statistical reliability analysis of structures considering both physical uncertainty and modeling uncertainty are proposed. Two examples are given, one of which is a numerical one; the other is a problem of soil liquefaction. The examples demonstrate the effects of modeling uncertainty on the reliability indices of structures.