基于TCN-Attention-GRU模型的枣树需水量预测

doi:10.3969/j.issn.1004-1524.20240058

浙江农业学报 ›› 2024, Vol. 36 ›› Issue (12): 2823-2831.DOI: 10.3969/j.issn.1004-1524.20240058

基于TCN-Attention-GRU模型的枣树需水量预测

李侨¹^,²(), 张华东¹^,², 孙三民¹^,²^,^*(), 殷彩云³

1.塔里木大学水利与建筑工程学院,新疆阿拉尔 843300
2.塔里木大学现代农业工程重点试验室,新疆阿拉尔 843300
3.新疆生产建设兵团第一师农业技术推广站,新疆阿拉尔 843300

收稿日期:2024-01-10 出版日期:2024-12-25 发布日期:2024-12-27
作者简介:李侨(2000—),女,四川眉山人,硕士,研究方向为节水灌溉。E-mail:2774110317@qq.com
通讯作者: *孙三民,E-mail:ssmaqx@126.com
基金资助:
新疆生产建设兵团科技项目(2021CB021);第一师阿拉尔市科技计划项目(2022XX01);新疆棉花产业技术体系(XJARS-03)

Water demand prediction of jujube tree based on TCN-Attention-GRU model

LI Qiao¹^,²(), ZHANG Huadong¹^,², SUN Sanmin¹^,²^,^*(), YIN Caiyun³

1. College of Water Resources and Architecture Engineering, Tarim University, Alar 843300, Xinjiang, China
2. Key Laboratory of Modern Agriculture Engineering, Tarim University, Alar 843300, Xinjiang, China
3. Agricultural Technology Extension Station of the First Divition of the Xinjiang Production and Construction Corps, Alar 843300, Xinjiang, China

Received:2024-01-10 Online:2024-12-25 Published:2024-12-27

摘要/Abstract

摘要：

为提高新疆南疆农业用水效率,针对南疆农业灌溉制度特点,构建了基于TCN-Attention-GRU的枣树需水量预测模型。模型以枣树为研究对象,将气象数据作为模型的输入参数,输出量为枣树需水量。首先采用注意力机制(Attention)将数据进行重要性特征提取,再将处理后的数据送入时间卷积网络(TCN)抓取时序特征融合成一个新的特征向量,最后利用门控单元(GRU)进行预测,根据组合模型的特点与多步预测的验证,所提出的TCN-Attention-GRU组合模型预测良好。经测试,基于TCN-Attention-GRU的枣树需水量预测模型决定系数(R²)达到94.4%,平均绝对百分比误差(MAPE)、均方误差(MSE)分别为7.9%、28.8%,实际值与预测值相对平均误差为12.23%,与其他模型相比该模型具有更高的预测精度,能有效提高水资源利用率。所提出的预测模型可为农业节水稳产提供一定参考。

关键词: 时间卷积网络, 组合模型, 作物需水量预测, 枣树

Abstract:

In order to improve the efficiency of agricultural water use in southern Xinjiang, a TCN-Attention-GRU-based water demand prediction model for jujube trees was constructed to characterize the agricultural irrigation system in southern Xinjiang. The model took jujube trees as the research object, took meteorological data as the input parameters of the model, and the output quantity was the water demand of jujube trees. Firstly, the attention mechanism (Attention) was used to extract the importance features from the data, and then the processed data were sent into the temporal convolutional network (TCN) to grab the temporal features to fuse them into a new feature vector, the final prediction was made using a gating unit (GRU), according to the characteristics of the combined model with the validation of the multi-step prediction, the proposed TCN-Attention-GRU combination model predicted well. The tested TCN-Attention-GRU based jujube trees water demand prediction model determination coefficient (R²) reached 94.4%, mean absolute percentage error (MAPE) and mean square error (MSE) were 7.9% and 28.8%, respectively, and the relative average error between the actual value and the predicted value was 12.23%, which had higher prediction accuracy compared with other models, and it could effectively improve the rate of water resources, and the proposed prediction model provided some references to the agricultural water conservation and stabilization. The proposed prediction model provides a certain reference for agricultural water conservation and stable production.

Key words: time convolutional network, combination model, crop water demand prediction, jujube

中图分类号:

S274.3

李侨, 张华东, 孙三民, 殷彩云. 基于TCN-Attention-GRU模型的枣树需水量预测[J]. 浙江农业学报, 2024, 36(12): 2823-2831.

LI Qiao, ZHANG Huadong, SUN Sanmin, YIN Caiyun. Water demand prediction of jujube tree based on TCN-Attention-GRU model[J]. Acta Agriculturae Zhejiangensis, 2024, 36(12): 2823-2831.

图/表 9

参考文献 24

[1]	邓铭江. 南疆未来发展的思考:塔里木河流域水问题与水战略研究[J]. 干旱区地理, 2016, 39(1):1-11.
	DENG M J. Prospecting development of South Xinjiang:water strategy and problem of Tarim River Basin[J]. Arid Land Geography, 2016, 39(1):1-11. (in Chinese)
[2]	尹立河, 张俊, 姜军, 等. 南疆地区水资源问题与对策建议[J]. 中国地质, 2023, 50(1):1-12.
	YIN L H, ZHANG J, JIANG J, et al. Issues and countermeasures on water resources in the Southern Xinjiang[J]. Geology in China, 2023, 50(1):1-12. (in Chinese with English abstract)
[3]	连晓晗. 基于PSO-GRU青椒生长期需水预测算法的智慧灌溉系统研究[D]. 邯郸: 河北工程大学, 2022.
	LIAN X H. Research on intelligent irrigation system based on PSO-GRU water demand prediction algorithm for green pepper growth period[D]. Handan: Hebei University of Engineering, 2022. (in Chinese with English abstract)
[4]	李少伟. 基于人工神经网络的智能节水灌溉系统的研究[D]. 合肥: 安徽大学, 2021.
	LI S W. Research on intelligent water-saving irrigation system based on artificial neural network[D]. Hefei: Anhui University, 2021. (in Chinese with English abstract)
[5]	李真真. 基于作物需水量预测的农田智能灌溉系统的研究[D]. 郑州: 华北水利水电大学, 2021.
	LI Z Z. Research on intelligent farmland irrigation system based on crop water demand prediction[D]. Zhengzhou: North China University of Water Resources and Electric Power, 2021. (in Chinese with English abstract)
[6]	李学军, 程红. 基于LSTM算法的智能节水灌溉预测模型研究[J]. 农机化研究, 2022, 44(3):22-27, 32.
	LI X J, CHENG H. Based on the Internet of Things precision farmland irrigation system key technology research[J]. Journal of Agricultural Mechanization Research, 2022, 44(3):22-27, 32. (in Chinese with English abstract)
[7]	孙博瑞, 孙三民, 蒋敏, 等. 基于LSTM神经网络的智能灌溉系统设计与试验[J]. 中国农机化学报, 2022, 43(4):116-123.
	SUN B R, SUN S M, JIANG M, et al. Design and test of intelligent irrigation system based on LSTM neural network[J]. Journal of Chinese Agricultural Mechanization, 2022, 43(4):116-123. (in Chinese with English abstract)
[8]	李莉, 李伟, 耿磊, 等. 基于RF-GRU的温室番茄结果前期蒸腾量预测方法[J]. 农业机械学报, 2022, 53(3):368-376.
	LI L, LI W, GENG L, et al. Prediction method of greenhouse tomato transpiration in early fruiting stage based on RF-GRU[J]. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(3):368-376. (in Chinese with English abstract)
[9]	李志磊, 周建平, 魏正英, 等. ET₀预测的卡尔曼滤波修正ANFIS模型研究[J]. 干旱地区农业研究, 2017, 35(3):114-119.
	LI Z L, ZHOU J P, WEI Z Y, et al. A study on the modified ANFIS model by the Calman filter for ET₀ prediction[J]. Agricultural Research in the Arid Areas, 2017, 35(3):114-119. (in Chinese with English abstract)
[10]	刘玉甫, 曹伟. 基于支持向量机优化粒子群算法的作物生育期ET₀预测[J]. 现代农业科技, 2014(2):219-220, 228.
	LIU Y F, CAO W. Prediction of ET₀ based on particle swarm optimization and support vector regression[J]. Modern Agricultural Science and Technology, 2014(2):219-220, 228. (in Chinese with English abstract)
[11]	张丽, 吴金亮, 杨国范. 基于BP神经网络的东港灌区作物需水量预测研究[J]. 水土保持研究, 2012, 19(6):207-210, 216.
	ZHANG L, WU J L, YANG G F. Prediction research for crop water requirement based on BP neural network in Donggang irrigation district[J]. Research of Soil and Water Conservation, 2012, 19(6):207-210, 216. (in Chinese with English abstract)
[12]	吴家葆, 曾国辉, 张振华, 等. 基于K-means分层聚类的TCN-GRU和LSTM动态组合光伏短期功率预测[J]. 可再生能源, 2023, 41(8):1015-1022.
	WU J B, ZENG G H, ZHANG Z H, et al. Dynamic combination of TCN-GRU and LSTM photovoltaic short-term power prediction based on K-means hierarchical clustering[J]. Renewable Energy Resources, 2023, 41(8):1015-1022. (in Chinese with English abstract)
[13]	SIDHU R K, KUMAR R, RANA P S. Machine learning based crop water demand forecasting using minimum climatological data[J]. Multimedia Tools and Applications, 2020, 79(19):13109-13124.
[14]	TONG Z Y, ZHANG S R, YU J X, et al. A hybrid prediction model for CatBoost tomato transpiration rate based on feature extraction[J]. Agronomy, 2023, 13(9):2371.
[15]	MA Y Z, LV B, WANG Y F, et al. Crop water requirement prediction method based on EEMD-attention-LSTM model[J]. Journal of Physics:Conference Series, 2023, 2637(1):012028.
[16]	李发崇, 李鹏, 高莲, 等. 基于多尺度模型融合和VMD-TCN-RF混合网络的短期电力负荷预测方法[J]. 电子器件, 2023, 46(4):1035-1042.
	LI F C, LI P, GAO L, et al. Short-term power load forecasting method based on multiscale model fusion and VMD-TCN-RF hybrid network[J]. Chinese Journal of Electron Devices, 2023, 46(4):1035-1042. (in Chinese with English abstract)
[17]	吴珺玥, 赵二刚, 郭增良, 等. 基于Spearman系数和TCN的光伏出力超短期多步预测[J]. 太阳能学报, 2023, 44(9):180-186.
	WU J Y, ZHAO E G, GUO Z L, et al. Ultra-short-term photovoltaic power multi-step prediction based on spearman coefficient and tcn[J]. Acta Energiae Solaris Sinica, 2023, 44(9):180-186. (in Chinese)
[18]	张泉. 作物需水量及灌溉需水量趋势性分析及方法[J]. 河南水利与南水北调, 2021, 50(6):81-83.
	ZHANG Q. Analysis and method of trend of crop water demand and irrigation water demand[J]. Henan Water Resources and South-to-North Water Diversion, 2021, 50(6):81-83. (in Chinese with English abstract)
[19]	苏向敬, 邓超, 栗风永, 等. 基于MGAT-TCN模型的可解释电网虚假数据注入攻击检测[J]. 电力系统自动化, 2024, 48(2):118-127.
	SU X J, DENG C, LI F Y, et al. Interpretable detection of false data injection attacks in power grid based on multi-head graph attention network and time convolution network model[J]. Automation of Electric Power Systems, 2024, 48(2):118-127. (in Chinese)
[20]	刘晓晓, 孔云峰. 气象监测数据的时空特征分析与建模[J]. 地理空间信息, 2009, 7(4):104-107.
	LIU X X, KONG Y F. Spatiotemporal analysis and modeling of the meteorological data[J]. Geospatial Information, 2009, 7(4):104-107. (in Chinese with English abstract)
[21]	陈天涯, 陈盛, 郑阳, 等. 基于VMD-CIMFs-TCN的水电机组振动预测[J]. 水电能源科学, 2023, 41(9):159-163.
	CHEN T Y, CHEN S, ZHENG Y, et al. Prediction of vibration trend of hydroelectric unit based on VMD-CIMFs-TCN[J]. Water Resources and Power, 2023, 41(9):159-163. (in Chinese with English abstract)
[22]	张上要, 罗军刚, 石国栋, 等. 基于VMD-TCN模型的渭河流域月径流量预测研究[J]. 人民黄河, 2023, 45(10):25-29.
	ZHANG S Y, LUO J G, SHI G D, et al. Monthly runoff prediction in Weihe River Basin based on VMD-TCN model[J]. Yellow River, 2023, 45(10):25-29. (in Chinese with English abstract)
[23]	安元超, 张岳君, 林文文, 等. 基于BO-GRU神经网络的锂离子电池剩余使用寿命预测[J]. 机械制造, 2023, 61(12):50-55.
	AN Y C, ZHANG Y J, LIN W W, et al. Prediction of remaining service life of lithium-ion battery based on BO-GRU neural network[J]. Machinery, 2023, 61(12):50-55. (in Chinese with English abstract)
[24]	王则玉, 谢香文, 刘国宏, 等. 干旱区绿洲滴灌成龄枣树耗水规律及作物系数[J]. 新疆农业科学, 2015, 52 (4):675-680.
	WANG Z Y, XIE X W, LIU G H, et al. Jujube drip irrigation water consumption and its crop coefficient in oasis of arid areas[J]. Xinjiang Agricultural Sciences, 2015, 52 (4):675-680. (in Chinese with English abstract)

编辑推荐

Metrics

阅读次数

全文

276

HTML			PDF

最新录用	在线预览	正式出版	最新录用	在线预览	正式出版
0	0	19	0	0	257

来源	本网站	其他网站

次数	233	43
比例	84%	16%

摘要

134

最新录用	在线预览	正式出版

0	0	134

	来源	本网站

	次数	134
	比例	100%

可调参数Adjustable parameter	数值Value
卷积核大小Convolutional kernel size	3
残差块数量Residual fast number	1
扩张率列表List of expansion rates	1,2,4
舍弃率Dropout	0.1
残差块激活函数Residual block activation funtion	ReLu
学习率Learning rate	0.01

可调参数Adjustable parameter	数值Value
卷积核大小Convolutional kernel size	3
残差块数量Residual fast number	1
扩张率列表List of expansion rates	1,2,4
舍弃率Dropout	0.1
残差块激活函数Residual block activation funtion	ReLu
学习率Learning rate	0.01

模型Model	MAPE	MSE	R²
CNN	14.7	70.3	80.1
TCN	12.9	58.4	84.1
TCN-Attention	10.1	51.9	90.0
TCN-GRU	11.6	57.9	86.9
TCN-Attention-GRU	7.9	28.8	94.4

模型Model	MAPE	MSE	R²
CNN	14.7	70.3	80.1
TCN	12.9	58.4	84.1
TCN-Attention	10.1	51.9	90.0
TCN-GRU	11.6	57.9	86.9
TCN-Attention-GRU	7.9	28.8	94.4

日期 Date	预测需水量 Predicted water demand/ (mm·d^-1)	实际需水量 Actual water demand/ (mm·d^-1)	相对误差 Relative error/%
2023-04-16	3.62	3.73	2.82
2023-04-17	3.71	3.99	7.02
2023-04-18	3.83	4.36	12.18
2023-04-19	3.95	3.87	2.05
2023-04-20	4.09	3.90	4.78
2023-04-21	4.21	4.11	2.58
2023-04-22	4.33	5.69	23.89
……
2023-07-24	7.29	5.82	20.05
2023-07-25	7.31	5.28	27.73
2023-07-26	7.21	6.24	13.35
2023-07-27	9.41	10.11	7.41
2023-07-28	7.89	8.48	7.49
2023-07-29	8.42	9.48	12.53
2023-07-30	10.16	10.07	0.90
相对平均误差Relative mean error			12.23

基于TCN-Attention-GRU模型的枣树需水量预测

Water demand prediction of jujube tree based on TCN-Attention-GRU model

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 24

相关文章 2

编辑推荐

Metrics

本文评价

[1]	张露荷, 王多锋, 张德, 张广忠, 赵通, 吕斌燕, 张洋军, 李毅. 枣树novel-miR16靶基因ZjTCP4鉴定及生物信息学分析[J]. 浙江农业学报, 2024, 36(3): 534-543.
[2]	卢牧原, 刘源, 刘桂建. 基于组合模型的黑土区土壤有机质含量预测分析[J]. 浙江农业学报, 2020, 32(8): 1427-1436.