System Reliability Analysis And Risk Assessment Of Slopes Considering Spatial Variability

Spatial variability of geotechnical properties is a major source of geotechnical uncertainties and has remarkable influence on slope reliability and risk assessment.Considering the spatial variability of geotechnical properties,there exists a large number of potential failure modes in slopes,which have a strong systematic effect on slope reliability and risk assessment.To address this problem,this thesis investigates a multiple response surfaces and representative slip surfaces based method for slope reliability analysis and risk assessment considering the spatial variability of geotechical properties.A proper form of response surface for a single slip surface considering the spatial variability is derived,and the approximate linearity between the accuracy of reliability analysis and the accuracy of random field discretization is analyzed.They effectively reduce the computational cost of constucting the multiple response surfaces and significantly increase the computational efficiency of slope system reliability analysis.The application of representative slip surfaces on slope risk assessment considering spatial variability has been further developed and the effect of choice of representative slip surfaces on complex slope system risk assessment is systematically explored.Finally,three simple slope examples and an engineered slope project are used to verify the effectiveness of the investigated method.Several conclusions can be drawn:(1)The multiple response surfaces method employs multiple low-order response surfaces to replace the single high-order response surface.Even though the former constructs more number of response surfaces than the latter,it is more efficient in fact.In addition,the former can provide reliability estimates for every single slip surface as well as the system reliability.It not only facilitates the identification of representative slip surfaces,but also provides the global reliability level of the slope.(2)The risk assessment considering the representative slip surfaces not only assess the system failure probability and risk exactly,but also requires very limited failure consequence assessments and siginificangt improves the computational efficiency.With consideration of the spatial variability,more effective representative slip surfaces are needed to represent the whole slope system as the systematic effect of slope strengthens.In addition,the identified representative slip surfaces according to the contribution to slope system failure probability can represent the slope system failure risk to a certain degree.Even for a complex slope system with strong systematic effect considering spatial variability,the slope system risk assessment based on representative slip surfaces can also be deemed as a feasible way.