Probabilistic seismic slope stability analysis using swarm response surfaces and rotational Newmark sliding model with primary sliding direction

The paper develops an effective and efficient swarm response surfaces to locate the slip surface with minimum critical acceleration, based on which the rotational Newmark sliding displacement is calculated. The corresponding primary sliding direction is determined using the identical critical acceleration from rotational Newmark sliding model. The proposed method is validated against the full search where the slip surface with minimum critical acceleration is located by a large number of trials through a homogeneous soil slope and a spatially variable soil slope. The comparative results have demonstrated that the proposed swarm response surfaces act as an efficient surrogate model with only 0.1% of time effort compared with full search method. Although the critical slip surfaces determined using step-wise procedure and simultaneous procedure can lead to agreeable probability density functions (PDFs) of rotational Newmark sliding displacement D compared to the proposed method, the significant discrepancies in sliding area and primary sliding directions among the three methods have been observed even for homogeneous soil slope. The variations of critical slip surfaces, the sliding area A and primary sliding direction am should be properly accounted for in seismic slope risk assessment using the swarm response surfaces.

Automated summary

The paper introduces an effective and efficient swarm response surfaces method for locating slip surfaces. Compared to the full search method, the proposed method requires much less time and effort while providing accurate results. However, significant discrepancies in slip surfaces and primary sliding directions are observed among different methods, highlighting the need to consider these variations in seismic slope risk assessment using swarm response surfaces.