Embankments are widespread throughout the world and their safety under seismic conditions is a primary concern in the geotechnical engineering community since the failure events may lead to disastrous consequences. This study proposes an efficient seismic slope stability analysis approach by introducing advanced gradient boosting algorithms, namely Categorical Boosting (CatBoost), Light Gradient Boosting Machine (LightGBM), and Extreme Gradient Boosting (XGBoost). A database consisting of 600 datasets is prepared for model calibration and evaluation, where the factor of safety (FS) is regarded as the output and four influential factors are selected as the inputs. For each dataset, the FS corresponding to the four inputs is evaluated using the commercial geotechnical software of Slide2. As an illustration, the proposed approach is applied to the seismic stability analysis of a hypothetical embankment example subjected to water level changes. For comparison, the predictive performance of CatBoost, LightGBM, and XGBoost is investigated. Moreover, the Shapley additive explanations (SHAP) method is used in this study to explore the relative importance of the four features. Results show that all the three gradient boosting algorithms (i.e., CatBoost, LightGBM, and XGBoost) perform well in the prediction of FS for both the training dataset and testing dataset. Among the four influencing factors, the friction angle φ is the most important feature variable, followed by horizontal seismic coefficient K_h, cohesion c, and saturated permeability k_s.

Probabilistic risk assessment of slope failure evaluates the slope safety in a quantitative manner, which considers the failure probability and failure consequence simultaneously. However, risk assessment of unsaturated slope accounting for spatially variable soil-water characteristic curve (SWCC) model parameter and saturated hydraulic conductivity has been rarely reported. A probabilistic risk assessment approach is proposed in current study for rationally quantifying the unsaturated slope failure risk with the aid of Monte Carlo (MC) simulation. The SEEP/W and SLOPE/W modules contained in Geostudio software are applied to carry out deterministic analysis, where factor of safety (FS) of the unsaturated slope is calculated by Morgenstern–Price method. The spatially variable hydraulic parameters are characterized by their respective random fields that are transferred from the random void ratio field in this study, rather than generating them separately. The proposed approach is subsequently employed to an unsaturated slope example for exploring the influences of spatially variable void ratio. Results show that the unsaturated slope failure risk is considerably affected by the spatially variable void ratio, and the single exponential autocorrelation function (ACF) popularized in geotechnical engineering tends to underestimate the failure risk in the unsaturated slope risk assessment.

Soil-water characteristic curve (SWCC) is a significant prerequisite for slope stability analysis involving unsaturated soils. However, it is difficult to measure an entire SWCC over a wide suction range using in-situ or laboratory tests. As an alternative, the Arya and Paris (AP) model provides a feasible way to predict SWCC from the routinely available particle-size distribution (PSD) data by introducing a scaling parameter. The accuracy of AP model is generally dependent on the calibrated database which contains test data collected from other sites. How to use the available test data to determine the scaling parameter and to predict the SWCC remains an unresolved problem. This paper develops a Bayesian approach to predict SWCC from PSD. The proposed approach not only determines the scaling parameter, but also identifies fitting parameters of the parametric SWCC model. Finally, the proposed approach is illustrated using real data in Unsaturated Soil Database (UNSODA). Results show that the proposed approach provides a proper prediction of SWCC by making use of the available test data. Additionally, the proposed approach is capable of predicting SWCC in the high suction range, allowing engineers to obtain a complete SWCC in practice with reasonable accuracy.