Geotechnical Spatial Variability Characterization And Collaborative Risk Assessment Of Slopes

The risk assessment and management of landslide hazard caused by slope instability is of great significance to the social and economic development.The landslide risk is affected by various uncertainties,among which the inherent spatial variability of geomaterials is one of the most important uncertainties and has great impacts on the failure mode,reliability and risk of slopes.The spatial variability preserves strongly 3-D anisotropy and needs to be identified through probabilistic site characterization.Due to the limited geotechnical data,it is challenging to characterize the 3-D spatial variability properly.Secondly,the 2-D probabilistic slope stability analysis cannot satisfy the requirement of complex engineered slopes.How to incorporate the 3-D spatial variability into the quantitative risk assessment of 3-D slopes is another challenging issue.Thirdly,most achievements of slope reliability analysis remain in the stage of theoretical research,still far from the practical applications.It is necessary to develop user-friendly software for practitioners to apply these achievements to real problems.Focusing on the abovementioned three key issues,the thesis quantifies the uncertainty of slope materials based on site investigation,and evaluates the reliability and risk of slope stability within a probabilistic framework.The main work and conclusions are as follows:(1)Regarding the quantitative characterization of stratigraphic uncertainty,a combination stratigraphic model is proposed to integrate the advantages of both boundary-based and category-based stratigraphic models.The model has the ability not only to generate almost arbitrary geotechnical strata but also to take into account the material spatial distribution trend and engineering judgment.Besides,a three-level probabilistic framework is proposed for geotechnical stratification modeling that allows the stratigraphic uncertainty to be incorporated properly into geotechnical reliability analysis and reliability-based design.(2)With respect to the 3-D spatial variability characterization,a matrix decomposition technique is proposed to address the computational difficulty of maximum likelihood estimation for high-dimensional issue and to facilitate the development of a maximum likelihood estimation based 3-D spatial variability characterization approach.The influences of site investigation plan on the probabilistic site characterization are explored.The proposed approach provides the technical support for spatial variability characterization in the presence of limited geotechnical data.After the characterization,a stepwise covariance matrix decomposition approach is further proposed for 3-D spatial variability modeling,with a rigorous proof of the equivalence to the general method.It achieves the efficient simulation of multivariate and large-scale 3-D spatial variability,and links the probabilistic site characterization to the risk assessment.(3)As for the 2-D slope risk assessment,the thesis develops a subset simulation enhanced efficient random finite element method to address the difficulty in slope risk assessment with high-dimensional uncertainty,multiple failure modes,and low failure probability.This lays the foundation for slope risk assessment considering the spatial variability of geomaterials.Furthermore,a collaborative reliability analysis approach is proposed based on both limit equilibrium analysis and finite element analysis.The advantages of the two analyses are integrated to achieve an accurate and efficient slope reliability analysis and to create conditions for applying the random finite element method to risk assessment of engineered slopes.(4)With regard to the 3-D slope risk assessment,a collaborative risk assessment(CRA)approach is proposed,with collaborative utilization of finite element models in different scales.The unified equations are derived for calculations of failure probability and risk in the preliminary and target analyses of CRA.It is of great importance to incorporate the 3-D spatial variability of geomaterials into 3-D slope stability analysis.The 3-D spatial variability has significant impacts on the failure mode,reliability and risk of slopes.The CRA enriches the theory on quantitative risk assessment of slopes and provides an effective analysis tool for risk assessment of complex engineered slopes.(5)In the end,the thesis explores the practical applications of slope reliability and risk assessment from two aspects,namely the reliability-based design of slopes and the development of probabilistic analysis software.A full-probabilistic reliability-based design approach is proposed with the consideration of spatial variability of geomaterials,which lays the theoretical foundation for the formulation of slope reliability-based design standard.In addition,a user-friendly software NIGPA(non-intrusive geotechnical probabilistic analysis)is developed,which promotes the practical application of slope reliability and risk assessment theory.