Study On Inverse Method Of Soil Hydraulic Parameters Under Two-dimensional Negative And Positive Pressure Infiltration

Estimation of soil hydraulic parameters is a key factor affecting the accuracy of numerical calculation of unsaturated soil-water movement.The present research adopted numerical simulation,theoretical analysis,and indoor-experimental comparison methods to study parameter identification of soil-water movement under two-dimensional soil negative pressure infiltration and ponding infiltration conditions,on the basis of soil hydrodynamic,numerical simulation,numerical inversion,multi-objective optimization,agent model,and the integration and advanced testing techniques of various computer languages.Some important conclusions are as follows:A new method defined as“two-stage-method”was proposed for hydraulic parameter inversion.In the first step,the function?(?_final)that is the final soil moisture of negative or positive pressure infiltration was selected as the objective function;based on which,the saturated soil moisture was inversed through GA.In the second step,the cumulative infiltration??Q?and infiltration rate??v?of negative or positive pressure infiltration were then selected as the objective functions,based on which,the hydraulic parameters?,n,and K_s were inversed through the mixed algorithm combined with VEGA and PSO.Results showed the proposed method could effectively solve the issues with weight coefficients of different objective functions.This method was of higher accuracy and stronger robustness,relative to the traditional multi-objective inversion methods.Based on the proposed“two-stage-method”,the hydraulic parameters in van Genuchten model were inversed under negative pressure infiltration and ponding infiltration conditions.Results showed the estimated values of the obtained hydraulic parameters were consistent with the measured values of the typical soils,further justifying the proposed method.Besides,the soil-water retention curve?SWRC?and soil-water conductivity curve?SWCC?were plotted on the basis of the obtained hydraulic parameters,which were consistent with the measured ones.This indicated that the obtained inversed parameters could be adopted for SWRC and SWCC estimations.Furthermore,the SWRC and SWCC estimated from inversions with measurement errors considering soil moisture and accumulative were quantitatively compared with the measured curves,and this indicated that the proposed inversion method was of strong robustness due to small difference and satisfying estimation accuracy.Moreover,the hydraulic parameters under ponding infiltration condition were inversed,and this indicated that the proposed inversion range had wide range of application,due to the consistent measured and actual SWRC and SWCC.And this has potential to be applied in practice.A Kriging surrogate model hydraulic parameter inversion model was proposed.The hydraulic parameters were inversed and estimated on the basis of accumulative infiltration and final soil moisture,and this showed a high consistency between the inversed and numerical,and measured values.Besides,the SWRC and SWCC estimated from inversions with measurement errors considering soil moisture and accumulative were quantitatively compared with the measured curves,and this indicated that the proposed agent model was of strong robustness and practicality due to small difference.Furthermore,the effects of the number of sampling points on inversion accuracy of agent model was analyzed,and this indicated that the distribution pattern of sampling points could affect inversion results;and specifically,a 3×3 distribution pattern?sampling from the vertical and radial distances of 0,5,and 10 cm from the center point of the infiltration surface?was recommended under a low workload condition.The soils from Yulin,Ansai,Luochuan,and Yangling were regarded as the representative soils,and their SWRC and SWCC were measured,which were adopted to verify the feasibility of hydraulic parameters inversed from the established agent model.Results showed the SWRC and SWCC obtained agent model were consistent with the measured curves,further justifying the parameters inversed from the established agent model.The present study mainly comparative analyzed the results obtained from numerical and indoor experiments of typical soils and further indicated that the selected methods is of fast and accuracy;however,the comparative experiments of the actual soils were ignored due to the limitations of time and conditions,which still needs further analysis.Besides,the following issues meanwhile need future studies:firstly,the effects of soil type,lamination,and variability and inversion of field soil hydraulic parameter,which also needs future studies.Secondly,the feasibility verification of advanced multi-objective optimization algorithm and adaptive agent model algorithm in the studies of hydraulic parameter inversion.