Practice of Searching a Noncircular Critical Slip Surface in a Slope with Soil Variability

Abstract

An improved approach is proposed here for searching the global noncircular critical slip surface in a slope with soil variability. To avoid the local minimum, the proposed approach combines a readily available optimization with the simulated annealing algorithm (SAA). The simplified Morgenstern-Price method was used to calculate the factor of safety (Fs) for the generated slip surface. The reliability of the searching strategy was evaluated through the effect of vertices number and the position of the initial slip surface. Subsequently, the proposed method was used to analyze two example problems involving slope profiles with different geometries and strength fields. The results showed that the location of the initial slip surface had a negligible influence on the global critical slip surface as well as the corresponding factor of safety, if the annealing time was sufficient. The location of the global critical slip surface diverged when the segment number was greater than 3. The proposed approach can seek the relative global factor of safety with a calculation accuracy of 1 × 10−4, which is suitable for engineering use. The noncircular slip surface assumption was superior to the circular slip surface assumption for both the simple multilayered soil slope and the spatially variable soil slope with a low level of strength variation. However, the circular slip surface assumption will also locate more critical slip surfaces if the variation of soil strength reaches a high level. The critical Fs value was determined from the given random strength field other than the assumption of a circular or noncircular slip surface. The noncircular slip surface searching strategy can provide a useful supplement for the limited-equilibrium method (LEM) searching for the critical slip surface in a spatially variable soil slope. Practitioners can select the smaller Fs from these two different assumptions.