System Reliability Analysis Of Soil Slopes Under Rainfall Condition

As one of the most important natural disasters, landslide has been an important subject of scientific research. As a mountainous country, China's annual economic losses caused by landslide disasters were up to billions Yuan according to statistics. Moreover, economic losses and casualties during water conservancy and hydropower construction process, which caused by the dam and reservoir landslide accident are regrettable. Domestic and foreign experts made great efforts to the safety and reliability of soil slopes with various attitudes and a variety of methods, achieved abundant conclusions. Along with the continuous development of technology, and new research methods come out, old methods for slope safety and reliability analysis continuous replaced by new technology which developed by geotechnical researchers all over the world.Large numbers of statistics show that most of the landslide occurred during or after the course of rainfall. Rainfall is the main reason for the failure of soil slope, which has become the consensus of the geotechnical academic circles. There are many articles about the failure of soil slope caused by rainfall around the world, but the research on the effect of rainfall on the system reliability of soil slopes is very little, and there are seldom unified opinions in these researches. One main reason for this difference is that the characteristics and permeability of the soil are different in different regions, so the influence of the rainfall to soil slopes will be different.Through the study of other researchers'works, considering the particularity structure of the soil slopes, the spatial variability of characteristics of soil slopes has great influence for its shear strength. Ignoring the spatial variability of soil parameters will overestimate the safety of soil slopes. System reliability of soil slopes considering the spatial variability of parameters and rainfall condition is seldom being studied.Through study spatial variability and machine learning theory, the saturated-unsaturated seepage theory, and the limit equilibrium theory, combined with my previous research results, using the optimized artificial neural network to establish the response surface, which can analyze the system reliability of soil slopes. The main research work in this paper are as follows:1. Take spatial variability theory into the system reliability analysis of soil slopes under rainfall condition, which make the analysis more closely to reality.2. Written the MATLAB program called Spatial_V_midpoint, which can make the 2-dimensional space discretization of parameters of shear strength of soil.3. Optimized extreme learning machine by artificial bee colony, and made a response surface model ABC-ELM to predict the safety reliability of soil slops. Compared with the results of Monte Carlo method, the results of ABC-ELM model are accurate, and the computation efficiency is improved by several hundred times.4. Analyzed the effect of rainfall on slope reliability by case studies.