采用基因表达式编程的沟灌入渗模型参数估计方法

发布时间：2018-06-21 03:57

本文选题：沟灌 + 入渗　；参考：《水利学报》2017年11期

【摘要】：入渗参数的准确估计是沟灌系统设计优化的关键问题。本文基于基因表达式编程(GEP)算法,通过数值试验获得了沟灌入渗模型参数γ估计GEP算法所需的有效数据集合,进行了参数γ的敏感性分析,确定了模型参数γ估计GEP算法的输入因子及其最优组合,建立了γ与最优组合因子的定量关系,进而提出了γ的估计方法。研究结果表明,γ对积水深度、初始有效含水率和沟深度最敏感;土壤水力特性参数对γ影响较大,模型参数γ估计GEP算法的最优输入因子组合为积水深度、沟深、初始有效含水率、饱和导水率和进气值参数;变积水深度条件下γ的估计值与定积水深度的γ值相近;在变积水深度条件下,与利用HYDRUS-2D模型计算得到的"精确"入渗量相比,应用基于γ估计值的沟灌入渗模型计算的累积入渗量误差小于5%,满足计算精度要求。入渗过程中边界效应对总入渗量的贡献逐渐增大,与细质土壤相比,粗质土壤中边界效应对总入渗量的贡献更大,本研究条件下黏壤土中边界效应最终贡献约为25%,砂土中约为37%。研究表明GEP算法可作为沟灌入渗模型参数γ估计的一种有效方法。
[Abstract]:Accurate estimation of infiltration parameters is a key problem in the design and optimization of gully irrigation system. Based on the genetic expression programming (GEP) algorithm, the effective data set for parameter 纬 estimation of ditch irrigation infiltration model is obtained by numerical experiments, and the sensitivity of parameter 纬 is analyzed. The input factor and its optimal combination of the model parameter 纬 estimation GEP algorithm are determined. The quantitative relationship between 纬 and the optimal combination factor is established, and the estimation method of 纬 is put forward. The results show that 纬 is the most sensitive to water depth, initial effective water content and ditch depth, and soil hydraulic characteristic parameters have great influence on 纬, and the optimal input factors of model parameter 纬 estimation GEP algorithm are water depth, ditch depth, etc. The initial effective water content, saturated water conductivity and inlet air value parameters, the estimated value of 纬 under varying water depth is close to that of constant water depth, and compared with the "accurate" infiltration calculated by using HYDRUS-2D model under variable water depth. The cumulative infiltration error calculated by using the gutter irrigation infiltration model based on 纬 -estimation value is less than 5, which meets the requirement of calculation precision. In the process of infiltration, the contribution of boundary effect to the total infiltration is gradually increasing. Compared with fine soil, the boundary effect in coarse soil contributes more to the total infiltration. In this study, the final contribution of boundary effect in clay loam is about 25%, and that in sandy soil is about 37%. The results show that GEP algorithm can be used as an effective method to estimate the parameters of gully irrigation infiltration model.
【作者单位】：中国农业大学水利与土木工程学院;中国-以色列国际农业研究培训中心;
【基金】：国家自然科学基金项目(51639009)
【分类号】：S152.7;S275.3

【相似文献】