基于NARXNN模型的降雨-水位关系研究

南京大学学报(自然科学版) ›› 2016, Vol. 52 ›› Issue (1): 133–141.

基于NARXNN模型的降雨-水位关系研究

闫中月1,2，徐静1,2*，华健1,2

出版日期:2016-01-27 发布日期:2016-01-27
作者简介:(1. 南京大学表生地球化学教育部重点实验室，南京，210023；2. 南京大学地球科学与工程学院水科学系，南京，210023）
基金资助:
基金项目:国家自然科学基金青年科学基金(41201022),国家水体污染控制与治理科技重大专项(2014ZX07204-005)
收稿日期:2015-11-23
*通讯联系人,E-mail:xujing@nju.edu.cn

Applications of rainfall-water level relationship based on NARXNN model

Yan Zhongyue1,2,Xu Jing1,2* ,Hua Jian1,2

Online:2016-01-27 Published:2016-01-27
About author:（1. Key Laboratory of Surficial Geochemistry, Ministry of Education, Nanjing, 210093, China;? 2. Department of Hydrosciences, School of Earth Sciences and Engineering, Nanjing University, Nanjing, 210093, China）

摘要/Abstract

摘要： 近年来，神经网络模型的发展为降雨-径流这一复杂的非线性过程的模拟提供了一种新的解决思路。本文基于淮河流域下游区滨海站2010～2012年的降雨及水位日资料，应用带外部输入的非线性自回归神经网络模型（Nonlinear AutoRegressive models with eXogenous input Neural Network，NARXNN），构建了以降雨为外部输入的淮河下游区降雨-水位关系模拟模型。通过设计了不同参数组合的正交模拟实验，采用相关系数，均方误差和平均绝对差评判模型的拟合优度，对模型进行验证，实验结果表明节点数对模型的拟合优度影响最大，当激励函数为logsig，节点数为7，延时阶数为4，隐含层数为9时，模型模拟效果最优。根据优化的参数组合，利用NARXNN模型对淮河下游区滨海站和长江下游区黄桥站的水位过程进行了模拟，结果表明该模型具有很强的鲁棒性。

Abstract: In recent years， rapid developments of Artificial Neural Networks (ANN) offer a novel approach for handling complex non-linear system like the rainfall-runoff relationship. This study built a rainfall-water level relationship model based on daily observation data of rainfall and water-level at Binghai station of Huai River during 2010-2012 year as well as Nonlinear AutoRegressive models with eXogenous input Neural Network (NARXNN)taking rainfall as the exogenous input. Reduced orthogonal test was designed and performed to save calculation time, since full factor experiments are highly time-consuming. The model was then calibrated with multiple goodness-of-fit criteria including correlation coefficient, mean square error and mean absolute error. The results show that model fitting is the most sensitive to node number; the model accuracy is the best when the activation function is logsig, the node number of hidden layers is 7, the input delays is 4 and the hidden layer number is 9. Then the rainfall–water level relationship was set up in the lower reaches of Huaihe River Basin and Yangtze River Basin, respectively, and the simulated results fitted the measurements well. It suggests that our model is robust in different hydro-meteorological environments.

闫中月1,2，徐静1,2*，华健1,2. 基于NARXNN模型的降雨-水位关系研究[J]. 南京大学学报(自然科学版), 2016, 52(1): 133–141.

Yan Zhongyue1,2,Xu Jing1,2* ,Hua Jian1,2. Applications of rainfall-water level relationship based on NARXNN model[J]. Journal of Nanjing University(Natural Sciences), 2016, 52(1): 133–141.

参考文献

[1] Tokar A S, Johnson P A. Rainfall-runoff modeling using artificial neural networks. Journal of Hydrologic Engineering, 1999, 4:232–239.
[2] Sudheer K P, Gosain A K, Ramasastri K S. A data-driven algorithm for constructing artificial neural network rainfall–runoff models. Hydrological processes, 2000, 16 (6): 1325–1330.
[3] Senthil Kumar A R, Sudheer K P, Jain S K, et al. Rainfall–runoff modeling using artificial neural network: comparison of networks types. Hydrological processes, 2004, 19 (6): 1277–1291.
[4] Tayfur G, Singh V P. ANN and fuzzy logic models for simulating event-based rainfall-runoff. Journal of Hydrologic Engineering, 2006, 132:1321–1330.
[5] Antar M A, Elassiouti I, Alam M N. Rainfall–runoff modeling using artificial neural networks technique: a Blue Nile catchment case study. Hydrological processes, 2006, 20 (5): 1201–1216.
[6] Nourani, Vahid, Özgür Kisi, et al. Two hybrid Artificial Intelligence approaches for modeling rainfall–runoff process. Journal of Hydrology, 2011,402 (1): 41-59.
[7] Kisi, Ozgur, Jalal Shiri, and Mustafa Tombul. Modeling rainfall-runoff process using soft computing techniques. Computers & Geosciences, 2013, 51: 108-117.
[8] Hsu K, Gupta H V, Sorooshian S. Artificial Neural Network Modeling of the Rainfall-Runoff Process[J]. Water resources research, 1995, 31(10): 2517-2530.
[9] 雷晓云，张丽霞，梁新平. 基于MATLAB工具箱的BP神经网络年径流量预测模型研究-以塔城地区乌拉斯台河为例. 水文, 2008, 28(1): 43-46.
[10] Sohail A, Watanabe K, Takeuchi S. Runoff analysis for a small watershed of Tono Area Japan by back propagation artificial neural network with seasonal data. Water resources management, 2008, 22(1): 1-22.
[11] 尚晓三，王栋. 两种不同类型的水文模型在贵州典型岩溶地区的应用. 南京大学学报(自然科学), 2009, 45(3): 409-415.
[12] 崔东文. 多隐层BP神经网络模型在径流预测中的应用.水文, 2013, 33(1): 68-73.
[13] Haykin S. Neural Networks:A Comprehensive Foundation.Newejexsey: Prentice-Hall inc,1999.
[14] 文明，张顶立，房倩等. 地铁车站施工过程中地表沉降的NARXNN时间序列预测模型. 岩土力学与工程学报, 2015,34(增1):3306-3312.
[15] 潘丽莎，程晓卿，秦勇等. 基于NARX神经网络的轮重减载率预测.铁道车辆, 2012, 50(9): 4-7.
[16] 吴启蒙，魏明，庞雷等. 基于NARX神经网络的电子电路电磁脉冲响应建模. 高压电器, 2013, 49(11): 62-68.
[17] Chai L, Qu Y, Zhang L, et al. Estimating time-series leaf area index based on recurrent nonlinear autoregressive neural networks with exogenous inputs. International journal of remote sensing, 2012, 33(18): 5712-5731.
[18] Aguilar-Lobo L M, Loo-Yau J R, Rayas-Sánchez J E, et al. Application of the NARX neural network as a digital predistortion technique for linearizing microwave power amplifiers[J]. Microwave and Optical Technology Letters, 2015, 57(9): 2137-2142.
[19] 赵良杰，夏日元，易连兴等. 基于 NARX 模型的岩溶地下河日流量预测.水电能源科学, 2015, 33(5):19-25.
[20] 周毅，徐柏龄. 神经网络中的正交设计法研究.南京大学学报(自然科学), 2001, 37(1): 72-78.
[21] 黄鹍，陈森发，亓霞等. 基于正交试验法的神经网络优化设计. 系统工程理论方法应用, 2004, 13(3): 272-275.
[22] 苑玉凤. 正交试验结果的分析. 统计与决策, 2006, 209(5): 138-139.
[23] 中华人民共和国水利部. 水文情报预报规范. GB/T 22482-2008.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed