考虑土体分层不确定性的边坡系统可靠度与关键失稳区识别方法研究

Seventy percent of geo-hazards in China are attributed to landslides, resulting in hundreds of deaths and significant economic losses, around 20 billions in RMB, per year. It is a necessary pre-requisite to determine soil layer boundaries for slope safety and reliability analysis. However, uncertainties in soil layer boundaries are unavoidable due to a limited number of site observation data. It remains a challenging task to quantify uncertainties in soil layer boundaries with a limited number of site observation data using traditional statistical analysis methods. This poses further difficulties in evaluating slope reliability and identifying slope failure mass. This project will develop a Bayesian approach that makes use of multiple-source information obtained during site investigation to characterize uncertainties in soil layer boundaries. Then, probabilistic methods are proposed to incorporate uncertainties in soil layer boundaries into determination of characteristic values of soil properties and efficient system reliability analysis and risk assessment of slope stability. In addition, effects of uncertainties in soil layer boundaries on characteristic values of soil properties, slope failure modes, and slope failure risk are systematically explored, and a method is also developed to identify key failure regions that control the slope failure risk. The project will not only lead to further advance in slope reliability analysis and risk assessment theory, but also provide technical supports to prevent landslides in China, which can be widely applied in slope engineering practice.

边坡失稳导致的滑坡灾害约占我国地质灾害总数的70%，每年造成数百人死亡，经济损失高达200亿元。准确描述边坡土体分层情况是合理分析边坡安全性和可靠性的重要前提。然而，由于岩土工程勘察中获取的勘探数据非常有限，土体分层的不确定性不可避免。有限勘探数据条件下传统统计分析方法无法合理表征土体分层不确定性，导致难以准确评价边坡可靠度和识别边坡失稳区。本项目将充分利用岩土工程勘察中获取的多源信息，提出一套基于贝叶斯理论的土体分层不确定性定量表征方法。在此基础上，建立考虑土体分层不确定性的土性参数标准值确定方法、高效的边坡系统可靠度分析与滑坡风险评估方法，揭示土体分层不确定性对土性参数标准值、边坡失稳模式和滑坡风险的影响规律，提出对滑坡风险起控制作用的边坡关键失稳区识别方法。本项目的研究成果不仅丰富了边坡可靠度分析与滑坡风险评估理论，而且为我国滑坡灾害防治提供了有效的技术支撑，具有广泛的工程应用前景。

受复杂地质成因影响(如母岩性质、搬运介质、沉积环境等)，土体类别和土性参数随着空间位置的变化而不同，表现出一定的层状分布特征。根据勘察数据划分土层和确定土性参数是岩土工程勘察的重要任务。然而，由于勘探数据通常十分有限，导致土层剖面和土性参数的不确定性不可避免。定量地表征土体分层不确定性与土性参数不确定性是合理评价边坡稳定可靠度和滑坡风险的重要前提。为此，项目组首先开展土体分层和土性参数的不确定性定量表征方法研究，提出了基于静力触探试验和面波勘探数据的土体分层不确定性表征方法，探讨了影响土层边界不确定性的关键因素；建立了基于多源数据的土性参数不确定性定量表征方法，研究了土性参数不确定性的相互作用机制。在此基础上，建立了同时考虑土体分层和土性参数不确定性的边坡系统可靠度分析方法和考虑土水特征曲线不确定性的非饱和土坡鲁棒性可靠度分析方法，发展了基于随机模拟的滑坡风险定量评估和边坡关键失稳区识别方法。. 在项目支持下，发表学术论文30篇，其中SCI论文15篇，EI论文7篇；获得软件著作权2项、发明专利1项；培养博士毕业生5人，硕士毕业生2人。项目成果不仅对合理地确定边坡土体分层和土性参数、深化认识边坡失稳机理、丰富滑坡风险评估与控制理论等方面具有重要的科学价值，而且为滑坡灾害防治工作提供有效的技术支撑。