Equation for Predicting Yield Seismic Coefficient of Simple Soil Slopes Using Polynomial-Interaction Regression Model Based on Pseudo-Static Numerical Limit Analysis

Performance-based seismic design of a soil slope is generally determined by setting the permanent displacement of the slope as a performance criterion. Estimation of permanent displacement requires estimating the yield acceleration, which is generally obtained by either repeating the slope stability analysis or applying the yield seismic coefficient from the chart solution. However, such approaches are not appropriate for the preliminary design stage or when promptness is critical. Therefore, there is a need for a simple equation that would estimate the yield seismic coefficient directly from slope conditions. This study proposes an explicit prediction equation based on polynomial-interaction regression analysis of the mean values of the lower and upper bound yield seismic coefficients determined through mechanically rigorous pseudo-static numerical limit analysis of various slope conditions. In addition, the validity of the proposed equation was examined by comparing this solution with the results of numerical limit analysis, existing prediction equations, and existing model tests. The yield seismic coefficient in this study was thus demonstrated to be mechanically rigorous, less biased than the existing equation, and consistent with model test measurements. Therefore, the proposed equation is useful, particularly in preliminary design, owing to its ability to explicitly estimate the mechanically rigorous yield seismic coefficient from the basic information of a slope.