盐碱土壤Kostiakov入渗模型参数的BP预报模型

发布时间：2018-05-12 20:31

本文选题：盐碱土壤 + Kostiakov二参数模型参数　；参考：《中国农村水利水电》2017年07期

【摘要】：为改良和改善盐碱地提供土壤入渗参数技术支撑,基于在山西省北部盐碱地进行的野外系列入渗试验,获取了累积入渗量与入渗历时的数据样本,并计算回归了kostiakov二参数入渗模型的入渗系数k与入渗指数α,建立了盐碱土壤基本理化参数与入渗参数之间的数据样本,利用BP神经网络的方法,建立了以土壤含水率、容重、质地、有机质、全盐量以及p H为输入变量,kostiakov入渗参数为输出变量的预报模型。结果表明:盐碱地土壤条件下,以土壤基本理化参数为输入变量,kostiakov入渗模型参数为输出变量的BP预报是可行的,入渗系数k的相对平均误差为0.29%、入渗指数α的相对平均误差为1.28%,以及根据两个入渗参数计算得到90 min累积入渗量的相对平均误差为2.37%,对所建立的模型进行检验时,以上三个参数检验误差的平均值均能控制在3%以下,确定所建立的BP预测模型能获得较好的效果。
[Abstract]:In order to improve and improve the technical support of soil infiltration parameters in saline-alkali land, the data samples of cumulative infiltration amount and infiltration duration were obtained on the basis of a series of field infiltration tests in the north of Shanxi Province. The infiltration coefficient k and infiltration index 伪 of kostiakov two-parameter infiltration model were calculated and the data samples between basic physical and chemical parameters and infiltration parameters of saline-alkali soil were established. The method of BP neural network was used to establish soil moisture content and bulk density. The prediction model of texture, organic matter, total salt content and pH as the input variable and the parameters of kostiakov infiltration as output variables. The results show that BP prediction with basic soil physical and chemical parameters as input variable and Kostiakov infiltration model parameters as output variables is feasible. The relative average error of the infiltration coefficient k is 0.29, the relative average error of the infiltration index 伪 is 1.28, and the relative average error of the cumulative infiltration amount of 90 min calculated according to the two infiltration parameters is 2.37. The average error of the above three parameters can be controlled below 3%, and the established BP prediction model can obtain better results.
【作者单位】：太原理工大学水利科学与工程学院;
【基金】：国家自然科学基金资助项目(40671081) 山西省科技攻关项目(2007031070)
【分类号】：S152.7;S156.4

【相似文献】