Reliability Applied To Stability Analysis Of Loess Slope Based On The Point Estimating Method In Lanzhou-Huining Area

The Lanzhou-Huining area is located in the western of the loess plateau，ravines criss-cross，complex geological tectonic，new tectonic movement strongly，is the area of landslide disaster inabnormal development in our country． The landslide disaster has been brought a greatinconvenience of the region’s economic construction and national production and life，soreliability applied to stability analysis of loess slope is particularly important for this region．In this paper,the geological model is established by the investigation and measurement of theLanzhou-Huining region and the statistical analysis of the slope height and slope angle of thelimit slope at three different confidence level（95%，90%，85%）. The probability model isdetermined by the statistical analysis of the Lanzhou-Huining area’s loess physical andmechanical parameters which is obtained by sampling,doing direct shear test and triaxialcompression test and collecting in geotechnical investigation report in recent years.Making useof five point estimate methods(Rosenblueth method,Harr method,Modified Harr method,Limethod,Duncan method) to evaluate the loess slope stability and calculate the stability factor andthe probability of failure of the loess slope at three different confidence level inLanzhou-Huining area. The calculation and analysis show that:1. The slope height is from20meters to260meters in the study area, no matter steep slopeor gentle slope. The maximum difference of the failure possibility which is calculated by fivepoint estimates mothods (Rosenblueth method, Harr method, Modified Harr method, Li method,Duncan method) is less than5%.2. When using the safety factor for slope stability evaluation, the impacts of the correlationcoefficient, the skewness coefficient and the kurtosis coefficient on the stability factor can benegligible. But when using the target failure probability for slope stability evaluation, the impacts ofthe correlation coefficient, the skewness coefficient and the kurtosis coefficient on the failureprobability need to consider the slope height and target probability of failure, when the targetfailure probability is small, the impact is more significant, can not be ignored.3. Loess slope in the study area, when the stability factor reach to1.33, can ensure that the slopefailure probability is less than10%. According to the slope designed in95%confidence level can meet therequirements of the10%target failure probability. When the stability coefficience reach to1.40, slope failureprobability will be less than5%.