Study On The Uncertainty Of Stability Reliability Of Subgrade Slope

In recent years, along with the development of high-speed railway construction and the promotion of railway technology, the traditional single safety factor design approach of subgrade slope is gradually being replaced by the limit state method based on reliability theory. There are many factors which have greater uncertainty (physical uncertainty, statistical uncertainty, model uncertainty) that affect the slope engineering safety, leading to the subgrade slope reliability calculations appear large deviation. Study on the uncertainty of slope reliability analysis can provide technical support and data base for the preparation of limit state law design specification for the embankment slope engineering. The main work is as follows: There are six railway lines’geotechnical survey data what were collected in order to study the parameter uncertainty of shear strength index. The probability distribution characteristic values of shear strength index were calculated, and the variability of c and φ were analyzed. The correlation coefficients of c and φ were calculated. The probability distribution models of two parameters were tested. The relationships of shear strength index’s variability and other physical and mechanical indicators were evaluated. The reliability results of the different slip surface search strategies and the different safety factors methods were analyzed when the Monte-Carlo was used, and the reliability index values of the Fellenius method and the simplified Bishop method with same safety were calculated.Through the study on the uncertainty of slope, the main conclusions are as follows:(1) Shear strength indices c and φ have a large coefficient of variation and because of various soils with a wide trading range, such as the coefficient of variation of cohesion in the range of 0.26 to 0.78, the coefficient of variation of the angle of internal friction in the range of 0.15 to 0.73, It is unreasonable to take definite variability of design parameters, fractionation of variability solve design parameters to meet the needs of different projects. The probability distribution model of c and φ indicators can accept normal and lognormal distribution what can be used as Monte-Carlo method and J-C law probability model. Liquidity index IL and the coefficient of variation of c and φ associated with the greatest degree, it is reasonable to refer to the physical index to estimate the level of variability of shear strength indexes.(2) Geotechnical parameters variability impact on embankment slope failure probability significantly, the failure probability of slope sharp rises with the improvement of the level of variability of parameters. In the condition of FS= 1.25, The engineering failure probability maybe more than 10% with the geotechnical parameters of variability reaching middle variation or above the level. The correction principle of safety factor should be established to consider the geotechnical parameters variability, to selection corresponding slope safety factor Fs by controlling the target reliability index.(3) The reliability results of different critical slip surface search strategies resulting were not significant at the homogeneous embankment slope, but large deviation occurs when foundation soil strength is less than the embankment. The probability density function curve shapes of Monte-Carlo safety factors analyzed by the Fellenius method and the simplified Bishop method were Basically the same, but the reliability indexes of different methods were significant difference. The target reliability index should be chosen based on stability analysis method, in order to ensure embankment slope have the same degree of physical security.