Quantifying model uncertainty of the simplified method for evaluating the soil liquefaction potential

The uncertainty in the model and parameters should be considered for the failure probability of engineering using the empirical model. However, there are few studies about the model uncertainty at present. In this paper, a new method for quantifying the model uncertainty c₁ is presented by the repeated and random sampling of case histories. The model uncertainty can be defined by the distribution of the variable c₁. The evaluation model with c₁ < 1 is non-conservative; otherwise, the evaluation model with c₁ > 1 is conservative. The model uncertainty of the Seed simplified method (Seed method) and the Robertson and Wride simplified method (RW method) for liquefaction evaluation are also illustrated. The analysis results show that the proposed method is simple and feasible, and has some degree of accuracy. The uncertainty of the Seed method may be characterized with an expectation value of 1.06 and a coefficient of variation of 0.06. Similarly, the uncertainty of the RW method may be depicted with an expectation value of 1.16 and a coefficient of variation of 0.12. Both two simplified methods are conservative models.