| As an important part of hydropower engineering project,the slope plays a decisive role,thus the estimation of slope stability is imperative.Traditional method uses a single safety factor to measure the safe level of slope.However,shear strength parameters have statistical uncertainty,which means the single safety factor to evaluate the slope stability is not reasonable,failure probability could reflect the actual safety level of slope,so reliability analysis should be done for the slope stability problems.The shear strength parameters are the base of slope system reliability analysis,and it is recognized that the shear strength parameters are dependent,which may lead to huge impact to the result of slope system reliability analysis.At present,there are two main methods for slpoe system relability analysis,the limit equilibrium method(LEM)and the finite element method(FEM).The LEM has high calculation efficiency but requires a lot of assumption.The FEM fits the actual fail pattern with high calculation accuracy but low efficiency.It will be significant to combine the advantages of two methods.The slope system reliability analysis mainly focuses on the failure probability.Nevertheless,the shear strength parameters have statistical uncertainty,which may lead to the statistical uncertainty of failure probability.How to consider such statistical uncertainty is also a problem worth studying.What`s more,the efficiency of slope system reliability analysis is very low,how to improve the efficiency is worth thinking about.This thesis studies on the aforementioned issues.Three different slope system reliability interval estimation methods are developed.The theory of slope system reliability interval estimation methods is built.In the end,the proposed method is applied to engineering practice.Major conclusions of this thesis are summarized as follows:(1)This thesis first introduces the background and significance of slope system reliability analysis.The current research of slope system reliability analysis is detailed from three aspects,such as the dependence of rock and soil parameters,the methods of slope stability analysis and the methods of slope system reliability analysis.After that,the shortcomings of the current research are pointed out,and the basic research frame of this thesis is conducted.(2)The copula functions are used to construct the two-dimensional joint distribution of shear strength parameters,which could overcome the shortage of traditional method on characterizing the relevant structure.What`s more,the slope system reliability analysis is carried out.The subset simulation method is used to improve the efficiency of slope system reliability estimation,and the copula-based subset simulation method is developed.The statistical uncertainty of shear strength parameters is characterized by bootstrap method,moreover,the interval estimation of slope system reliability is given by the aforementioned method.(3)The limit equilibrium method based slope system reliability interval estimation method is developed.The calculation efficiency is promoted with the method of representative slips choosing.The point estimation and interval estimation are given further.The performance of aforesaid method is illustrated by a slope case.(4)The finite element method based slope system reliability interval estimation method is developed.The surrogate model of input index and output index is built by the polynomial chaos expansion.The point estimation and interval estimation are given further.The performance of aforesaid method is illustrated by the same case,and the advantages and disadvantages of finite element method and limit equilibrium method are compared.(5)The auxiliary analysis method based slope system reliability interval estimation method is developed.This method utilizes the advantages of limit equilibrium method and finite element method,and the key problems are studied.The performance of aforesaid method is illustrated by the same case,what`s more,the advantages and disadvantages of three methods are compared.(6)The engineering application of the proposed methods is presented.The impact of sample size change and critical safety factor change on the interval estimation results of slope system reliability is studied. |