Study On Stability And Rainfall Threshold Of Typical Sandy Soil Slope In G318 Bomi-Nyingchi Section

Precipitation is one of the main factors inducing geological disasters,among which continuous rainfall and short-time heavy rainfall are the main factors affecting slope instability,and precipitation infiltration makes the internal soil and water state of slope change,thus leading to slope instability.The geological conditions between Linzhi and Bomi of National Highway G318 are complex,with abundant rainfall and geological hazards,and the sandy slopes along the route form landslides under the action of rainfall,which affects the safety and smoothness of the highway.In this paper,the relevant soil parameters of potential landslide soil samples are obtained through field investigation,indoor and outdoor parameter tests,and the mechanism of landslide occurrence,especially the mechanism of slope instability under different rainfall conditions,is analyzed through the investigation of landslide hazards in the study area and the review of relevant literature.Then the theoretical calculation is carried out by using the simple cloth strip division method and the numerical simulation is carried out by using Geo-studio finite element software to calculate and verify each other the stability coefficients of the study slopes.Combined with the existing rainfall data,the relationship between different rainfall,rain intensity,duration and rain type and slope stability is analyzed by using logistic unit and multiple regression to obtain the relationship between its slope stability and rainfall threshold.The main research contents and results are as follows:(1)Through field survey sampling,field parameter testing and indoor parameter testing,parameters such as cohesion C,internal friction angle φ parameter,water content,soil grain gradation and permeability of each sandy slope soil were obtained.(2)The safety coefficients of the slopes along National Highway G318 under different rainfall intensity,rainfall durations and rainfall amounts were obtained through two methods of analysis: theoretical calculation and numerical simulation by Geo-studio finite element software,and the safety coefficients calculated by both of them were within 5%,which mutually verified the accuracy and reliability of the stability of the slopes.(3)The sandy soil slopes in the study area are stable in their natural state,but as the rainfall intensity increases and the rainfall time continues,the stability of the slopes gradually decreases until they become unstable.The greater the rainfall intensity,the faster the slope safety factor drops,the shorter the time required for slope destabilization,and the easier the slope slides on the surface;the smaller the rainfall intensity,the slope safety factor drops more slowly,the longer the time required for slope destabilization,and the slope does not slide easily.(4)The By comparing the logistic unit regression model with the SPSS linear regression model to analyse the relationship between rainfall and slope stability,it was found that the logistic unit regression model resulted in a better model fit and more accurate results.The regression calculation results show that the sandy slopes are very likely to be unstable when the rainfall amounts reach 219.52 mm、118.60 mm、82.62 mm 、 42.61 mm under the cases of light rain,medium rain,heavy rain and heavy rain respectively,and these rainfall amounts can be used as the macroscopic critical rainfall thresholds for rainfall-type landslide warning in the study area.(5)A logistic multiple regression model was constructed to analyze the three influencing factors of rain intensity,rain duration and rain volume on the stability of slopes,and the rainfall threshold model of rain duration and rain volume under four rain intensity conditions of light rain,medium rain,heavy rain and heavy rain was obtained.Then the stability of nine known slopes under different rainfall factors is tested by random simulation,and it is found that the best fitting effect of the logistic multiple regression model can reach 88.89%,indicating that the model has certain practical value.