Granite residual soil (GRS) slopes are widely distributed in the humid and rainy regions around the world. Due to their unique mineral and pore composition, these soils exhibit a loose structure and are prone to softening and disintegration upon wetting. Such slopes often suffer from collapses and landslides along highway in the southeast coast of China, posing a serious threat to transportation safety. Therefore, it is necessary to evaluate the stability of GRS slopes. This study investigates slopes along three expressways in Guangdong Province, China. A systematic analysis of instability mechanisms was first conducted. Based on this analysis, an evaluation system was developed. It includes four primary indicators: geomorphology and geology, geotechnical character, environmental factors, and engineering activities. These are further divided into 13 secondary indicators. Subsequently, the study proposes an improved combined weighting approach to address uncertainties. Subjective fuzziness is quantified by enhancing the order relation method (G1 method) with Monte Carlo simulation. Meanwhile, objective weight distortion is mitigated by optimizing the entropy weight method (EWM) with a sigmoid function, and the linear weighting method is adopted to achieve a scientific balance between subjective and objective weights. Subsequently, a slope stability evaluation model is constructed based on the cloud model theory. A case study of K21 slope of the Yunmao Expressway is conducted. The evaluation result from the model is consistent with the field situation. Twenty typical slopes from the three expressways were furtherly selected for model verification. Results indicate that the proposed model increases accuracy rates by 10% and 5% higher than the analytic hierarchy process (AHP) and the EWM, respectively. The method proposed in this paper can effectively and accurately evaluate the stability of GRS slopes. The method uniquely combines subjective and objective factors and integrates qualitative and quantitative analyses for stability evaluation.
Jiang et al. (Thu,) studied this question.