Critical geometric parameters of slope and their sensitivity analysis: a case study in Jilin, Northeast China

An excessively high or steep slope is one of the main causes of slope instability. This research examined the slope of the waste dump of a limestone mine in Wangqing County of northeast China’s Jilin Province. Support vector regression (SVR) and a radial basis function neural network (RBFNN) were used to train and test 76 groups of collected data, and the mean absolute error (MAE), mean absolute percentage error (MAPE), and root mean square error (RMSE) values indicated that the SVR model is better than the RBFNN model. Then, 12 × 17 combinations of angles and heights for the waste dump slope were designed, and used SVR to define β _cr and H _cr by predicting the factor of safety (FS) under different conditions. The results demonstrated that the SVR method can be used to calculate β _cr and H _cr accurately. Moreover, the effects of the parameters on FS and β _cr and H _cr were analyzed by sensitivity analysis. The sensitivity analysis results showed that the internal friction angle (φ) is the most effective parameter acting on β _cr and H _cr , followed by the unit weight (γ) and cohesion (c). Furthermore, H _cr is more sensitive to γ, c, and φ than β _cr . These results could be useful and convenient for engineers involved in slope design and management.