Study On Stability Evaluation Of Landslides In The Red Beds Of Yunnan

In the quantitative evaluation system of landslide stability,the parameters of rock and soil greatly influenced the results of evaluation when the method of evaluation is reasonable and the geological structure is accurately.Usually,most of engineering rock and soil is spatial and variable,today the same type of rock and soil are treated as a homogeneous body when evaluating the stability of landslide,so,the rationality of the method needs to be study.This paper takes the landslide of Liushutang in Yunnan Province as an example,which mainly study the effect of the soil parameters determined by different methods on the stability of landslide.Some parameters are derived from the field and indoor tests,which laid a reliable data foundation for the later research.For results of the test results,this paper is treated in two ways:one of them is calculating the test results by the standard in the "Specification of soil test",another one is controlled the plane coordinates of each parameter by the existing test data,then the contour map of each parameter would be established after the interpolation of the parameters in the study area,finally the contours are partitioned.The physical mechanics parameter of rock and soil would be obtained,after the area weighted average of the similar parameters in each partition;the former is named the homogenization parameters,and the latter is called the partition parameters.Combined with the large of the exploration data,the geological structure is established,with the 3D and 2D models for the parameters determined by different methods is established too.In the 3D and 2D homogenization model,the homogenization parameters are given to the same rock mass;and the partition parameters are given to the different partition in the 3D and 2D partition models.For the 3D and 2D model,the stability of the homogenization model and the partition model are calculated by the finite difference method under the natural and rainfall conditions,and the safety factor of them are solved by the strength reduction method.On the other hand,the stability factors of 2D models are calculated by using the finite difference methods respectively.The results of finite difference method shows that the maximum shear strain increment,the range of state and the displacement of the partition model are smaller than those of the homogenization model under the natural and rainfall conditions,which means that the results of the homogenization model are generally conservative.Under the rainfall conditions,the displacement of Y and Z directions of each monitoring sites is 2.2-12.4 times and 4.1-8.4 times of the actual detection point Y and Z direction respectively in 3D homogenization model,while it is 1.2?2.1 times and 0.8?1.3 times in 3D partition model,which means that the results of the 3D partition model analysis is closer to the reality.And the correlation between the displacement of the monitoring point and the actual detection point is not strong.It could be known from safety and stability factor calculated by the strength reduction method and the limit equilibrium method are under the same conditions,while the safety and stability factors of the partition models are more than 10%of the homogenization models in 3D and 2D models,which means that the accuracy of the evaluation results of partition parameters are better than the homogenization parameters.