LRFD Calibration of Simple Soil-Structure Limit States Considering Method Bias and Design Parameter Variability

Abstract

A general closed-form solution for the reliability index (or probability of failure) of a simple linear limit-state design function with one load term and one resistance term is used to compute the resistance factor expressed in a load and resistance factor design (LRFD) format. The solution considers method bias, bias dependencies, and uncertainties in choice of nominal values of load and resistance determined as part of the project-specific design process. Uncertainty in the choice of nominal values for design is linked quantitatively to the concept of project level of understanding that has been recently adopted in Canadian design practice. All random variables are assumed to be lognormally distributed. Parametric analyses are carried out to show that ignoring possible correlations between random variables can lead to conservative (safe) values of resistance factor and in other cases to nonconservative (unsafe) values. Example LRFD calibrations are carried out using different load and resistance models for the pullout internal stability limit state of steel-reinforced soil walls together with matching bias data reported in the literature. The results demonstrate the practical influence of model type, method bias statistics including dependencies, and operational factor of safety on computed resistance factors.