The present study has attempted to apply the Bayesian updating to the LRFD, or Load and Resistance Factor Design method. The LRFD method takes into account the statistical distribution of the material resistance and those of the applied loads. The LRFD method can reflect the degrees of different uncertainties of the resistances of the materials and the loads. Thus, the LRFD method can attain the optimal design which can keep up an adequate reliability level of the components designed, whereas the conventional allowable stress design (ASD) method cannot. The LFRD method, however, requires vast amount of statistical data for the material resistances and the applied loadings of different kinds. The present study proposes the Bayesian updating scheme which requires only a small amount of statistical data for the material resistance and the various load item distributions to calculate the values of the partial design factors used in the LRFD method. It is revealed that the median of the updated distributions of the estimated standard deviations can give adequate reliability index values higher than the target reliability index value corresponding to a fracture probability of 0.01% even for a small number of the statistical data, say, less than 20. This paper also compares and discusses the LRFD method with the updating scheme and the conventional ASD method, showing that the updated LRFD method can maintain the reliability index value higher than the target index value whereas the ASD method cannot.

This content is only available via PDF.
You do not currently have access to this content.