Fuzzy Reliability Analysis Of Slope Stable Capacity Based On Fuzzy Sets

According to analyses of the meaning of slope reliability and review of the researches about reliability theory and fuzzy mathematic on slope engineering, the limitation of the conventional reliability method and the research direction on slope stability are presented. To investigate the uncertainty implicit in the slope stability analysis, a new model called Fuzzy-Random Reliability Evaluation(FRRE) is developed to analyze the stability of engineering slopes, the key to which is establish of fuzzy membership function and probability density function, and the analytic method and process are introduced to evaluating the soil slopes stability.For this purpose, Raised-Mountain-Shaped Function(RMSF) is chosen for the fuzzy membership function of slope stability. A new method, which applies fuzzy probability theory to distribute slope state into 7 fuzzy language values and considers the cases histories of slope engineering to construct fuzzy restriction condition, is proposed to confirm the coefficients of RMSF. It suggests by cases studying the proposed fuzzy membership function basically reflect the relationship between slope stability and safety factor, and the distribution of slope states suit the engineering practice and human knowledge about slope safety on the whole, which provides powerful reference with right safety factor in engineering decision-making.In this study, "Normal Fuzzy Set" is introduced for the first time to describe the fuzziness and randomness of soil parameters and "Mountain-Shaped Fuzzy Set" for fuzzy seismic intensity. Applying fuzzy set algorithm, a program validated by examination case is developed to calculate the credibility function of safety factor, on the basis of which the probability density function is defined and the fuzzy-random reliability is calculated finally.A study of FRRE for simple soil slopes is carried out, the analysis of which provides the relationship between parameter variation and fuzzy-random reliability and choosing principle of the key coefficient k in the "Normal Fuzzy Set", suggests more attention pay to parameter variation when a quite high safety factor. In addition, some significant discussion on the difference between the linear fuzzy set andnonlinear and also between the linear fuzzy membership function and nonlinear is put up. The results are applied to evaluating the state of a real complex multilayer soil slope, calculating and comparing the fuzzy-random reliability of common case and seismic case, conclusion that the conventional reliability evaluation of slope stability is always unsafe is pointed out.The study indicates: 1 .Relativing to the conventional certain methods, the FRRE can systematically, quantificationally consider the fuzziness and randomness of slope engineering, which makes for engineering design and decision-making; 2.Relativing to the conventional reliability methods, the FRRE emphasizes the importance of expert knowledge and engineering experience on the slope stability evaluation, and suites the practice of human thinking and slope safety knowledge level better.