
基于CEEMD-BP耦合模型的灌区地下水埋深预测
王燕鹏, 穆玉珠
基于CEEMD-BP耦合模型的灌区地下水埋深预测
Prediction of Groundwater Depth Based on CEEMD-BP Coupling Model in Irrigation Area
为了科学预测人民胜利渠灌区地下水埋深,促进灌区水资源的可持续利用,针对灌区地下水埋深具有非线性、非平稳性与预测精度低的特征,基于CEEMD具有非平稳信号平稳化的能力和BP神经网络较强的非线性和不确定性映射能力与预测效果,构建了基于CEEMD-BP的灌区地下水埋深预测耦合模型。将该模型应用于人民胜利渠灌区地下水埋深预测中,结果表明:CEEMD-BP耦合模型和其他模型相比具有较好的预测效果,平均相对误差为4.7%,纳什系数为0.96,预测精度更高。综合上可知,模型预测精度高,为灌区地下水埋深提供了一种有效的预测方法。
In order to scientifically predict groundwater depth and promote the sustainable utilization of water resources in people's victory canal irrigation district, according to the complex characteristics of nonlinearity and non-stationarity, the accuracy of prediction is usually not high, Based on CEEMD, it has the ability to smooth non-stationary signals and BP can approximate arbitrary functions, has good nonlinear mapping capabilities, and has an advantage in the prediction of uncertain factors, the CEEMD-BP coupling model is built to predict the groundwater depth in the irrigation area. The results show that the CEEMD-BP coupling model and the EEMD-BP model have better prediction results than the BP model. The average relative error is 4.7%, the Nash coefficient is 0.96, and the prediction accuracy is higher. In conclusion, the model has high prediction accuracy, provides an effective method for predicting the buried depth of groundwater in the irrigation area.
CEEMD / BP网络 / 地下水埋深预测 / 人民胜利渠灌区 {{custom_keyword}} /
CEEMD / BP network / groundwater depth prediction / People's Victory Canal Irrigation District {{custom_keyword}} /
表1 各分量与趋势项相对误差对比 (%)Tab.1 Comparison of relative error between each component and trend item |
预测项 | 相对误差最大值 | 相对误差最小值 | 相对误差平均值 |
---|---|---|---|
IMF1 | 594.74 | 4.30 | 161.55 |
IMF2 | 287.17 | 1.50 | 41.30 |
IMF3 | 9.89 | 0.12 | 2.90 |
IMF4 | 1.88 | 0.05 | 0.67 |
IMF5 | 3.37 | 0.43 | 0.77 |
趋势项 | 0.04 | 0 | 0.01 |
表2 研究区2012-2013年预测的相对误差Tab.2 The relative error of forecast in the study area from 2012 to 2013 |
年份 | 月份 | 真实值/m | 预测值/m | 绝对误差 | 相对误差/% |
---|---|---|---|---|---|
2012 | 1 | 6.02 | 6.08 | 0.06 | 0.73 |
2 | 6.22 | 6.23 | 0.01 | 0.35 | |
3 | 6.40 | 6.40 | 0 | 0.05 | |
4 | 6.50 | 6.55 | 0.05 | 0.69 | |
5 | 6.60 | 6.66 | 0.06 | 0.91 | |
6 | 6.84 | 6.78 | -0.06 | 1.04 | |
7 | 6.60 | 6.66 | 0.06 | 0.77 | |
8 | 6.52 | 6.50 | -0.02 | 0.35 | |
9 | 6.48 | 6.50 | 0.02 | 0.35 | |
10 | 6.56 | 6.60 | 0.04 | 0.47 | |
11 | 6.57 | 6.55 | -0.02 | 0.44 | |
12 | 6.59 | 6.41 | -0.18 | 3.04 | |
2013 | 1 | 5.96 | 6.03 | 0.07 | 0.96 |
2 | 5.86 | 5.92 | 0.06 | 0.88 | |
3 | 6.06 | 6.07 | 0.01 | 0.03 | |
4 | 6.30 | 6.49 | 0.19 | 2.71 | |
5 | 6.71 | 6.76 | 0.05 | 0.81 | |
6 | 6.96 | 6.90 | -0.06 | 0.67 | |
7 | 6.78 | 6.78 | 0 | 0.13 | |
8 | 6.62 | 6.61 | -0.01 | 0.04 | |
9 | 6.80 | 6.77 | -0.03 | 0.35 | |
10 | 6.80 | 6.81 | 0.01 | 0.17 | |
11 | 6.93 | 6.87 | -0.06 | 0.80 | |
12 | 6.87 | 6.82 | -0.05 | 0.77 | |
平均相对误差/% | 0.73 | ||||
纳什效率系数NSE | 0.96 |
表3 模型评价指标表Tab.3 The outcome of evaluation index |
预测模型 | MAE/m | RMSE/m | MAPE/% |
---|---|---|---|
CEEMD-BP | 0.047 | 0.242 | 0.730 |
EEMD-BP | 0.095 | 0.485 | 1.469 |
BP | 0.184 | 0.929 | 2.819 |
Elman | 0.198 | 0.997 | 3.027 |
1 |
陈笑,王发信,戚王月,等. 基于遗传算法的BP神经网络模型在地下水埋深预测中的应用:以蒙城县为例[J]. 水利水电技术,2018,49(4):1-7.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
2 |
柴成武,蒋志荣,徐先英,等. 民勤县荒漠绿洲过渡带土地荒漠化类型判定[J].西北林学院学报,2006(6):12-16.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
3 |
王根绪,杨玲媛,陈玲,等. 黑河流域土地利用变化对地下水资源的影响[J].地理学报,2005(3):456-466.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
4 |
肖彩虹,郝玉光,贾培云. 乌兰布和沙漠东北部磴口绿洲近52 a水分因子的变化[J].干旱区资源与环境,2008(6):161-165.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
5 |
谭矫,丁建丽,陈文倩,等.基于多变量时间序列模型的地下水埋深预测:以渭库绿洲为例[J].节水灌溉,2017(9):55-59,64.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
6 |
张先起,宋超,胡登奎.基于EEMD与Elman网络的灌区地下水埋深预测模型[J].节水灌溉,2018(12):86-91.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
7 |
张晓萌,王振龙,杜富慧,等.淮北平原浅埋区地下水埋深对土壤水变化的影响研究[J].节水灌溉,2019(9):6-9.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
8 |
徐存东,朱兴林,张锐,等.地下水埋深空间插值方法比较和空间变异性研究[J].节水灌溉,2019(1):49-56.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
9 |
张文鸽,侯胜玲,殷会娟.内蒙古河套灌区地下水埋深时空变化及其驱动因素[J].节水灌溉,2020(7):36-40,45.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
10 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
11 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
12 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
13 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
14 |
于海姣,温小虎,冯起,等.运用小波变换与支持向量机耦合模型(WA-SVM)预测干旱区地下水埋深[J].中国沙漠,2016,36(5):1 435-1 442.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
15 |
邵光成,章坤,王志宇,等. 基于IABC-RBF神经网络的地下水埋深预测模型[J].浙江大学学报(工学版),2019,53(7):1 323-1 330.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
16 |
曹文洁,肖长来,梁秀娟,等. RBF神经网络在地下水动态预测中的应用[J].水利水电技术,2018,49(2):43-48.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
17 |
YEH J R,
{{custom_citation.content}}
{{custom_citation.annotation}}
|
18 |
杨婷,魏晓妹,胡国杰,等. 灰色BP神经网络模型在民勤盆地地下水埋深动态预测中的应用[J].干旱地区农业研究,2011,29(2):204-208.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_ref.label}} |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
/
〈 |
|
〉 |