Mechanical Characteristics And Bayesian Analysis Of Soil-water Characteristic For Unsaturated Granite Residual Soils

Collapsing erosion is the most serious type of soil erosion in granite area in South China, particularly in subtropical regions such as Ganzhou, Jiangxi. Collapsing erosion poses substantial threats and has caused severe damages. Collapsing erosion affected by numerous factors is a complex problem. Water is the most important external cause. The rainfall infiltration will result in a transient increase of water content and a reduction of suction, as a result the shear strength of the soil is correspondingly reduced. In Gannan collapsing erosion area, there is grantie residual soil. Its physical and mechanical properties are affected significantly by water.Therefore, the unsaturated characteristics of the grantie soil are studied in paper, and which can provide an important basis for the further study of the soil collapsing erosion process and treatment. The triaxial tests under suction-controlled conditions are carried out on unsaturated residual soils to analyze strength characteristics of unsaturated granite residual soil. A pressure plate apparatus is used to determine the soil-water characteristic curve (SWCC). The SWCC is predicted by using the Bayesian updating theory. The uncertainty of SWCC fitting parameters is analyzed using the Bayesian theory, and the confidence intervals of uncertain model parameters of SWCC can be obtainet.Then,the stability of analysis of unsaturated soil slope under rainfall condition are illustrated by constucting the condence interval for the factor of safety.(1) To research the strength and deformation characteristics of unsaturated granite residual soil in collapsing erosion area, the triaxial tests under suction-controlled conditions are carried out on unsaturated residual soils with different initial dry density, net confining pressure, suction to analyze stress-strain relationship, unsaturated strength and deformation for collapsing soil. The stress-strain relationship depends on the combination of net confining pressure and matric suction. The cohesion has a linear increase with the increasing of suction.However, the suction has small effect on the friction angle, nearly negatively. The triaxial tests are performed for unsaturated lime-treated granite residual soil. The effects of lime content and gradation on the strength and deformation of unsaturated lime-treated soil are studied. The experimental program consists of 0%, 4%, 8% lime content and 1 mm, 2 mm grade sizes tests are carried out in the suction of 100kPa.The cohesion and friction angle of the lime-treated soil increases with the increasing of the lime content. The grade size has small effect on the strength, nearly negatively.(2) The soil-water characteristic curvecan be used to describe the relationship between the water content and the matrix suction in unsaturated soil and plays important roles in unsaturated soil mechanics. To research SWCC of the granite residual soils, the GEO-Experts pressure plate extractors is used to measure SWCC of different factors of dry density, grain size, drying and wetting cycles and lime content.The initial dry density has a significant impact on SWCC. With the increasing of dry density, the suction is decreased for the same water content. The larger the grain size is the greater the suction value for the same volumetric water content. Under the same suction, the volumetric water content decreases as the lime percentage increases and the water stability improves. SWCCof drying and wetting cycle demonstrates the hysteresis phenomenon. The area of hysteresis loop decreased with the increase of the dry density and drying and wetting cycle number. Hysteresis loop also became small when the soils are mixed with lime. The influence of the gradation and the vertical stress on SWCCis investigated. The experiments evaluated that the SWCC of different gradation under different vertical stresses are similar to the stress conditions in the field. The difference of the SWCCs of the different gradations is small under without the vertical stress and will become big under vertical stress. This different will increase with the increasing of the vertical stress. It is apparent that the vertical stress has a significant effect on SWCC.The Fredlund & Xing model, Van Genuchten model and Gardner model are used to fit the experimental data of SWCC. The parameters of three different functions used to fit the measured data of the SWCCtests are presented.The computed results indicates that the simulated value of the Van Genuchten model does best agreement with measured value.(3) There have the uncertainty of geotechnical parameters and calculating mode.This paper proposes the soil-water characteristic curve predicting framework based on the Bayesian updating, according characteristics of soil-water characteristic curve.Fusing the predicting method of soil-water characteristic curve and the measured information, the parameters in the soil-water characteristic curve fitting model,Fredlund & Xing model and Van Genuchten model are updated by using the Bayesian updating theory and preliminary observation data of soil-water characteristic curve.The late soil-water characteristic curve can be predicted more accurately by the updated parameter. The dynamic prediction can realized through keeping updating of the parameters. The uncertainty of geotechnical characteristics is coupled to the prediction of soil-water characteristic curve. When the standard deviation of the measurement error is reasonable, it is demonstrated that confidence in the predicted the SWCC increases rapidly with a small amount of data points by Bayesian updating.The SWCC can be predicted quickly and effectively by the combination of Bayesian updating and experimental methods.(4) There have several parameters in fitting models and have significant uncertainty of SWCC obtained by experiment. Based on the Bayesian theory, the uncertainty analysis of the SWCC has been preformed. The parameters of SWCC fitting models are regarded as the random vectors. The Fredlund & Xing model, Van Genuchten model and Gardner model are used to fit the experimental data of SWCC for the granite residual soils. The posterior distributions of model fitting parameters are obtained by the Markov Chain Monte Carlo (MCMC) simulation delayed rejection adaptive metropolis. Confidence intervals of uncertain model parameters of SWCC are obtained by proposed approach. MCMC method is an effective tool to estimate the parameters of SWCC models.(5) The SWCC plays a significant role in the analysis of seepage flow and slope stability. Combined the uncertainty analysis of SWCC, a homogeneous slope model is taked as an example to ivestigate the effect of confidence level of SWCC on the seepage field and the slope stability. The seepage filed and slope stability is affected by confidence level of soil water characeterist curve. The seepage field is affected significantly by the rainfall duration. The growth rate of pore water pressure increases with the increase of the percentage of the confidence level. The factor of safety decreases with increasing of the confidence level of soil water characeterist curve.