Study On The Variables And Models Of Soil Water Characteristic Curve

With the rapid development of social economy and the rapid expansion of the scope of human engineering activities,the problem of unsaturated soil in practical engineering is becoming more and more prominent.As an important component of describing unsaturated soil behavior,soil-water characteristic curve（SWCC）has become a factor that must be considered in the study of engineering properties of unsaturated soil.Aiming at the variables and models of soil-water characteristic curve,this thesis took four layers of undisturbed loess L1,L6,L7and L13 in Luochuan standard section as the research object,by means of matrix suction measurement,theoretical analysis and modeling,studied the influence of pore structure of undisturbed loess on its soil-water characteristic curve,put forward the method of determining the variables of unimodal and bimodal SWCC based on VG model,established a soil-water characteristic curve model based on two-parameter Weibull cumulative distribution function and discussed the fitting accuracy of Weibull model.The following results have been achieved:（1）the basic physical parameters of four layers of undisturbed loess in Luochuan were tested,and the soil-water characteristic curves were measured by filter paper method and fitted by VG model.According to the data of mercury intrusion porosimetry,the pore distribution curves were obtained.The charts of pore distribution curve,soil-water characteristic curve and its slope were obtained using Young-Laplace equation.The results show that the pore structure has a direct influence on the SWCC,and there is a good corresponding relationship between the pore distribution curve and SWCC:the maximum slope point of the soil water characteristic curve has a good one-to-one correspondence with the peak point of dominant pores in the pore distribution curve;L1 is a bimodal pore structure soil,and its SWCC has two significant steep descending segments,while L6,L7 and L13 are unimodal pore structure soils,and their SWCC has only one significant steep descent,which corresponds to its dominant pore range respectively.（2）based on VG model,the methods for determining the variables of unimodal and bimodal SWCC have been proposed,and the feasibility of this method was verified by using the SWCC data points of four layers of undisturbed loess in Luochuan.The water evaporation test in the natural state was carried out to determine the evaporation residual saturation S_rzf.According to the relative errors between S_rzf and S_r1,S_r2(S_r1 and S_r2 were acquired from the two methods for determining the residual state,respectively),the two methods to determining the residual state were compared.The results show that the proposed methods can obtain reasonable variables of SWCC and determine its residual state effectively for both unimodal and bimodal SWCC.（3）A SWCC model（i.e.,Weibull model）based on two-parameter Weibull distribution was proposed.The Weibull model was used to fit 24 groups of SWCC measured data of 12 kinds of soils from UNSODA 2.0 database.The statistical results show that the determination coefficient R² is 0.999 and the root mean square error RMSE is 0.010,so the fitting accuracy of Weibull model for all types of soil is high.According to comprehensive evaluation system based on root mean square error,adjusted determination coefficient and the corrected Akaike Information Criterion,it is found that the form of Weibull model is simple and the fitting accuracy is high.As a result,its fitting performance is slightly better than that of VG model by comparison of the fitting capacity with other models.