Methods For System Reliability Analysis And Risk Assessment Of Soil Slopes With Engineered Applications

Landslides(i.e.,slope failures)are well acknowledged in geotechnical engineering in China and therefore it is of great importance to assess the landslide probability and the corresponding risk in a quantitative manner.Aiming at the inherent spatial variability in soils and the subsquent diversity and uncertainties in the slope failure modes and failure volumes,great efforts must be made to identify the failure modes and failure volumes in a systematical way and to evaluate the probability and risk of slope failures.Based on the theoretical analysis,numeircal modeling and re-visiting of the existing large-scale slope model test,the paper is structured as follows:(1)Following the comprehensive understanding of correlations among potential slip surfaces,a static group based strategy and dynamic strategy based on mutually exclusive and collectively exhaustive principles for identify representative slip surfaces are presented respectively.Slope reliability methods combining representative slip surfaces distinguished via the static and dynamic strategies and Monte Carlo simulations are propsoed.The proposed methodolgoies are validated against three typical slopes and the recommended one is given in accordance with the effectiveness and efficiency.It is concluded that the contribution of deterministic slip surface to the response of slope system decreases as the complexity of slope system increases and that the identification of representative slip surfaces is of crucial to the determination of the responses of slope system.(2)A second order polynomial-based response surface function is calibrated for each of representative slip surfaces discriminated in the preceding section thereby leading to multiple response surface functions.The validated multiple response surface functions are sustituted for the limit equilibrium methods to calculated the factors of safety for representative slip surfaces in slope system reliability analysis.The propsoed method is illustrated through typical slopes and the comparative studies have demonstrated that multiple resposne surface functions can simulate the limit state function of slope system with promising computational efficiency.(3)The concept of approaching index for two slip surfaces is originated and then an equation for quantifying the correlation soefficients between factors of safety for two slip surfaces is developed.The representativeslip surfaces for slope systems considering two dimensional spatial variabilities are selected based on the propsoed equation and they are used in subset simulation algorithm to evaluate the failure probability of slope system in two dimensional spatial variabilities.The effectiveness of the presented equation is validated through the case studies.It is found that if either the vertical fluctuation scale or the horizontal fluctuation scale is lower than the respective slope height or the slope width,the effect of spatial variability on the slope reliability should be considered.Otherwise,conservative slope reliability result may be obtained.The slope failure probability under noncircular failure mechanism can be two orders than that under circular failure mechanism.(4)An assumption is presented following the revisit on the large scale slope model test to quantify the sliding area using the slip surface with maximum sliding area and factor of safety equal to one.The assumption is verified through the numeirical simulation of the slope failure process from slope initiation,propogation,and termination.The propsoed method is combined with Monte Carlo simulation method to assess the landslide risk for a typical slope.It is noticed that the distribution range of the sliding areas obtained by the improved method tends to be widen than that by the traditional method where the deterministic slip surface is used to calculate the sliding area.The non-zero intervals form the former method is 20 whereas it is about 10 for the latter method.(5)The proposed methodology such as the system reliability methods based on deterministic slip surface and the representative slip surfaces and the method following updated sliding area for slope risk assessment is applied to the stability analysis and risk evaluation of Sanshandao Xinli tailings dam.The comprehensive analysis of the computed results provides much insight into the safety evaluation and risk mitigation of the Xinli tailings dam.It is recommended that besides the slip surface with minimum FS,proper attention should be paid on the slip surfaces with relatively high occurrence or those with relatively larger sliding area.