| 引用本文: | 樊红梅,刘晓民,王文娟,等.基于多种方法的“以电折水”系数研究[J].灌溉排水学报,0,():-. |
| fanhongmei,liuxiaomin,wangwenjuan,et al.基于多种方法的“以电折水”系数研究[J].灌溉排水学报,0,():-. |
|
| 摘要: |
| 【目的】针对目前我国农业用水占比高、计量监测难度大等问题，引入“以电折水”方法计量农业用水，推进解决农业用水的计量和核算问题。【方法】以内蒙古通辽市科尔沁区为例，采用平均值预测、多元线性回归模型、BP神经网络模型3种方法对“以电折水”系数进行预测分析，并选用平均相对误差（MRE）、均方根误差（RMSE）、决定性系数（R2）对上述3种模型比较优选。【结果】结果表明，平均值预测误差平均为7.40%，预测值与实测值拟合效果较差；多元线性回归模型预测误差平均为2.40%，预测值与实测值拟合效果较好；BP神经网络模型预测误差平均值为1.65%，预测值与实测值拟合效果最好。【结论】多元线性回归模型预测精度较好，BP神经网络模型预测精度最好，其MRE、RMSE、R2分别为0.024、0.175、0.923和0.018、0.131、0.957，平均值预测模型预测精度最差。 |
| 关键词: “以电折水”系数;平均值预测;多元线性回归;BP神经网络 |
| DOI： |
| 分类号:TV93 |
| 基金项目:国家重点研发计划子课题（2018YFC0406404-3）;国家自然基金项目（51969021）;内蒙古自治区科技重大专项（2020ZD0009-4） |
|
| Research on the coefficient of “Conversion of Electricity to Water” based on various methods |
|
fanhongmei1,2, liuxiaomin1, wangwenjuan2, chenqiong1
|
|
1.Inner Mongolia Agricultural University;2.Inner Mongolia Jinhuayuan Environmental Resources Engineering Consulting Co., Ltd
|
| Abstract: |
| Abstract:【Background】Agriculture is the largest economic and social water sector in China. At present, the main problem and primary difficulty of agricultural water management is to solve the problem of agricultural water measurement and accounting. In recent years, according to the characteristics of agricultural water use and the problems existing in the process of water use, the method of "converting electricity into water" is introduced to measure agricultural water consumption.【Objective】In view of the high proportion of agricultural water and the difficulty of measurement and monitoring in China, this paper introduces the method of "converting electricity into water" to measure agricultural water, so as to promote the measurement and accounting of agricultural water. 【Method】Taking Horqin District of Tongliao City in Inner Mongolia as an example, the average value prediction, multiple linear regression model and BP neural network model are used to predict and analyze the coefficient of "converting electricity into water", and the average relative error (MRE), root mean square error (RMSE) and decisive coefficient (R2) are selected to optimize the above three models.【Result】The results show that the average prediction error of the average value is 7.40%, and the fitting effect between the predicted value and the measured value is poor; The average prediction error of multiple linear regression model is 2.40%, and the fitting effect between predicted value and measured value is good; The average prediction error of BP neural network model is 1.65%, and the fitting effect between the predicted value and the measured value is good.【Conclusion】The MRE, RMSE and R2 of multiple linear regression model are 0.024, 0.175, 0.923 and 0.018, 0.131, 0.957 respectively, and the average prediction model has the worst prediction accuracy. |
| Key words: “conversion of electricity to water”coefficient; Average value prediction; Multiple linear regression; BP neural network |
