Safety performance evaluation of a gently dipping bedding slope containing multiple weak interlayers in Southwest China
Year: 2026
Type: article
Abstract: The failure modes and safety assessment methods of gently dipping bedding slopes containing multiple weak interlayers (GBSWs) are critical issues in engineering construction. In this study, engineering geological investigations and laboratory direct shear tests on undisturbed samples were conducted to elucidate the strength degradation mechanism and the evolution of strength parameters of the weak interlayers, while finite element simulations were employed to investigate the slope failure modes. Based on the laboratory test results and numerical simulations, a back-propagation neural network (BPNN) model was developed to enable rapid assessment of the operational safety of engineering slopes. The results indicate that the shear strength of the weak interlayers decreases significantly with increasing water content, while both cohesion and the internal friction angle exhibit an exponential decay. The bedrock–overburden interface and the weak interlayers within the slope constitute unfavorable interfaces. Following excavation-induced unloading, slope deformation develops preferentially along these interfaces, resulting in an overall stepped sliding–tensile cracking failure mode. Staged excavation, together with the timely implementation of protective measures after excavation, can effectively control the factor of safety ( F s) and the maximum horizontal displacement (| U x| max ) of the slope. Comparative trial calculations showed that the 5–10–2 BPNN architecture exhibited the best performance, yielding the lowest root mean square error (RMSE) and the highest coefficient of determination (R²). For GBSWs, an analytical limit equilibrium method applicable to the engineering feasibility study and preliminary design stages, together with a predictive assessment method based on machine learning for the safe operation stage, was proposed.
Source: Environmental Earth Sciences
Language: English
Subfield: Safety, Risk, Reliability and Quality
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $2,290
OpenAlex ID: W7222788656