| Rainfall-induced landslide is one of the most common natural disasters in the world,and also is the main geohazard in the Loess Plateau of China.It is of great scientific value and practical significance to study the stability evaluation method of this disaster.A large number of artificial rainfall experiments have been carried out to explore the failure variation rules of loess landslides during precipitation process,but for accurate analysis of key control factors affecting slope stability,further field and laboratory experiments are still needed.On the other hand,compared with the conventional stability analysis,the reliability method based on probability analysis can provide more information about slope safety,such as failure probability and reliability index.However,preferential infiltration and rainfall pattern also plays a significant role in failure of slope,but it is rare to study slope failure probability based on regional rainfall characteristics.Based on the above considerations,this paper analyzed the deformation inducement of rainfall-induced loess landslide through field investigation.Furthermore,author analyzed the spatial variability of soil parameters and discussed the unsaturated properties of soil by laboratory tests.Comsol Mutliphysic and Matlab were used to construct the analysis framework of the reliability and failure probability of rainfall loess slope considering preferential infiltration and spatial variability.The effects of different infiltration models on the spatial-temporal variation of slope reliability were compared.Finally,combining with local rainfall record in study area,rainfall extreme value and time-scale effect are analyzed by IDF curve and breakdown coefficients methods.Bounded random cascade model was introduced to simulate the random rainfall patterns(RRP),so to discuss slope failure probability under different rainfall patterns.Last,combining with the spatial variation of soil and preferential flow,Author took a comprehensive analysis of various factors influencing the loess slope stability in rainfall condition.In this paper,preferential infiltration hypothesis is introduced into the reliability and failure probability framework,and various uncertain factors in rainfall-induced loess landslide are analyzed comprehensively,which can provide new method to evaluate the rainfall-induced loess slope instability physical model.The research results are as follows:(1)Failure inducement: Through artificial rainfall test,the inducements of slope failure were summed.Matric suction and pore water pressure jointly affect slope failure during rainfall infiltration.after the wetting peak of the slope reaches 1.5m,the deep matric suction drops rapidly,leading to initial deformation of the slope.Then the matric suction drops below 10 Kpa,the repeated rise and fall of pore water pressure leads to the continuous deformation of the soil.Under the influence of preferential flow,continuous heavy rainfall may lead to a rapid decline in the matric suction of deep soil and also result in structural weakness surface.Soil heterogeneity also controls slope deformation and failure process and may take adverse effect on slope stability.(2)Preferential flow effect: Using the deterministic analysis merely may overestimate slope stability,while the uncertainty analysis can calculate the failure probability of slope at different depths during rainfall process,and reveal more information about slope failure.For this simulation experiment,When the rainfall intensity is low(<10mm/h),the preferential infiltration promotes drainage,while the homogeneous infiltration is more stable in the process of heavy rainfall(>20mm/h).From reliability analysis,the spatial variability of parameters is the key inducement of slope failure in homogeneous infiltration,while the failure is more likely caused by the rapid advance of wetting peak in preferential infiltration condition.Finally,based on the preferential infiltration model,hydraulic exchange between matrix domain and dominant domain is studied.It is found that the slope has greater failure probability when the hydraulic exchange is powerful,while the lower hydraulic exchange coefficient may affect the failure at bottom of the slope.(3)Rainfall pattern effect: persistent rainfall and early rainfall with late heavy rainfall are the most likely to cause loess slope failure,while the slope is stable under short duration storm rainfall and advanced rainfall pattern.Uniform rainfall model is more likely to obtain conservative results for stability evaluation.In comparison,rainfall threshold and annual failure probability calculated by random rainfall pattern(RRP)can cover the common characteristics of all classic rainfall pattern.Considering the spatial variability of soil and preferential infiltration,it is found that neglecting two factors may overestimate the slope stability seriously,and judging failure probability of slope by considering various factors can provide more comprehensive analysis for slope stability evaluation. |
PDF Full Text Request