基于Kostiakov二参数入渗模型参数的BP预测

发布时间：2018-03-22 04:30

本文选题：耕作土壤　切入点：Kostiakov二参数入渗模型　出处：《节水灌溉》2016年10期 　论文类型：期刊论文

【摘要】：基于黄土高原区大田耕作土壤的水分入渗试验,建立了Kostiakov二参数入渗模型参数的BP神经网络预测,实现了以土壤基本理化参数为输入变量,Kostiakov二参数模型参数为输出变量的BP预测方法,并分别对二参数模型中的入渗系数k、入渗指数α以及90min累积入渗量H进行了预测值与实测值的精度比较,结果显示对入渗系数k实现BP预测的平均相对误差为6.082 3%,入渗指数α的平均相对误差为1.045 9%,90min累积入渗量H的平均相对误差为4.973 5%,三者的平均相对误差值均在7%以下,预测精度较高,预测效果较好,表明以土壤基本理化参数为输入变量的BP神经网络预测是可行的。研究结果为获取准确的入渗参数提供技术手段,进而为提高农业灌溉水管理水平和灌水效率提供支撑。
[Abstract]:Based on the experiment of water infiltration in field tillage soil in Loess Plateau, BP neural network was established to predict the parameters of Kostiakov two-parameter infiltration model. The BP prediction method with basic physical and chemical parameters of soil as input variable and Kostiakov two-parameter model parameters as output variables is realized. The accuracy of the predicted values and the measured values of the infiltration coefficient k, the infiltration index 伪 and the cumulative infiltration amount H of 90min in the two-parameter model are compared respectively. The results show that the average relative error of BP prediction for infiltration coefficient k is 6.082 3, the average relative error of infiltration index 伪 is 1.045 9 minutes, the average relative error of cumulative infiltration H is 4.973 5, and the average relative error of all three is less than 7%. The prediction accuracy is high and the prediction effect is good, which indicates that BP neural network with the basic physical and chemical parameters of soil as input variable is feasible. The research results provide a technical means for obtaining accurate infiltration parameters. It provides support for improving agricultural irrigation water management level and irrigation efficiency.
【作者单位】：太原理工大学;
【分类号】：S152.7;;TV93

【相似文献】