Sensitivities of input parameters for predicting stability of soil slope

Abstract

The infinite slope stability model is the basis for predicting the slope stability; however, the fundamental theory requires various input parameters. The objective of this study is to suggest a reasonable method for assuming input parameters in difficult measurement areas. The referred values are used to perform the sensitivity analysis, which found the influence degree of each parameter on the output result. Both slope angle and soil depth are shown to be highly affected parameters in the infinite slope stability model. To gather the real values of the slope angle, soil depth, and water table height, field tests are carried out and the remaining values are estimated from laboratory test results. The geostatistical method is applied to predict the distribution of input parameters over the entire study area based on the measured data, and the safety factor in the objective area is calculated. To observe fluctuations in the safety factor, the input parameter values are changed to single values representing the average, maximum, and minimum measured values. Then the contour map, histogram, and error ratios are established using the deduced safety factors. This study suggests considerations needed to determine the input parameters and a reasonable method when assuming factors.