基于BP神经网络的MIKE SHE模型参数率定
|
摘要
为了更精细地对水文全过程进行描述和解析,更准确地构建分布式水文模型,以丹麦Karup流域为例,对MIKE SHE模型的饱和导水率、饱和带水平水力传导系数、河床透水系数进行了参数率定,模拟流域的日径流过程。结果表明:基于BP神经网络反分析的参数率定方法比MIKE SHE模型参数自动率定计算得到的均方根误差RMSE小,模型效率系数Ens更接近1;采用BP神经网络反演率定参数后,3组测试样本的日径流模拟过程的RMSE分别为0.04,0.03,0.08 m~3/s,Ens均为0.99,且模拟结果能较好地反映径流的实际变化趋势。因此,这种基于BP神经网络反分析的参数率定方法对构建分布式水文模型具有一定的价值。
In order to describe and interpret hydrological processes in more detail, and at the same time to construct a more accurate distributed hydrological model, we took the Karup watershed in Denmark as an example and calibrated three parameters of MIKE SHE model, namely, saturated hydraulic conductivity, saturated horizontal hydraulic conductivity, and leakage coefficient of river bank, and simulated the daily runoff process in the watershed. Results demonstrate that the root mean square error(RMSE) obtained by the method of parameter calibration based on BP neural network is smaller than that by automatic parameter calibration in MIKE SHE model, with the model efficiency coefficient Ens closer to 1. Having been treated by parameter calibration by BP neural network, the values of RMSE of daily runoff of three test samples are 0.04 m~3/s, 0.03 m~3/s, and 0.08 m~3/s, respectively, and the value of Ens is 0.99. As the simulated runoff displays a trend in agreement with the real runoff, the back analysis method of parameter calibration based on BP neural network is of certain value in runoff simulation.
In order to describe and interpret hydrological processes in more detail, and at the same time to construct a more accurate distributed hydrological model, we took the Karup watershed in Denmark as an example and calibrated three parameters of MIKE SHE model, namely, saturated hydraulic conductivity, saturated horizontal hydraulic conductivity, and leakage coefficient of river bank, and simulated the daily runoff process in the watershed. Results demonstrate that the root mean square error(RMSE) obtained by the method of parameter calibration based on BP neural network is smaller than that by automatic parameter calibration in MIKE SHE model, with the model efficiency coefficient Ens closer to 1. Having been treated by parameter calibration by BP neural network, the values of RMSE of daily runoff of three test samples are 0.04 m~3/s, 0.03 m~3/s, and 0.08 m~3/s, respectively, and the value of Ens is 0.99. As the simulated runoff displays a trend in agreement with the real runoff, the back analysis method of parameter calibration based on BP neural network is of certain value in runoff simulation.
引文
[1] 李德龙,程先云,杨浩,等.人工智群算法在水文模型参数优化率定中的应用研究[J].水利学报,2013,44(增1):95-101.
[2] 万增友.MIKE SHE模型国内应用现状及其关键问题研究[J].科协论坛(下半月),2011,(5):99-101.
[3] 姜凌峰,薛联青,刘远洪,等.基于MIKE SHE模型的干旱区节水灌溉对地下水位的影响研究[J].灌溉排水学报,2016,35(2):59-65.
[4] 田开迪,沈冰,贾宪.MIKE SHE模型在灞河径流模拟中的应用研究[J].水资源与水工程学报,2016,27(1):91-95.
[5] 卢小慧,李奇龙.基于MIKE SHE模型的流域地下水水文响应[J].长江科学院院报,2015,32(1):11-15,20.
[6] 王中根,夏军,刘昌明,等.分布式水文模型的参数率定及敏感性分析探讨[J].自然资源学报,2007,22(4):649-655.
[7] 刘飞,胡斌,宋丹,等.基于BP神经网络和均匀设计的边坡敏感性分析[J].水电能源科学,2014,32(10): 113-115,165.
[8] 郑震,张静,宫辉力.MIKE SHE水文模型参数的不确定性研究[J].人民黄河,2015,37(1):23-26.
[9] MA L, HE C G, BIAN H F, et al. MIKE SHE Modeling of Ecohydrological Processes: Merits, Applications, and Challenges[J]. Ecological Engineering, 2016, 96: 137-149.
[10] 郄志红.大坝安全监测资料正反分析的智能软计算方法及其应用[D].天津:天津大学,2005.
[11] REFSGAARD J C.Parameterisation, Calibration and Validation of Distributed Hydrological Models[J]. Journal of Hydrology, 1997, 198(1/2/3/4): 69-97.
[12] SUMAN A, AKTHER F. Investigation of Water Balance at Catchment Scale using MIKE SHE[J].International Journal of Engineer and Computer Science, 2014, 3(10): 8882-8887.
[13] DHI.MIKE SHE分布式水文模型培训教程[K].Denmark: DHI Water and Environment,2012.
[14] 沈花玉,王兆霞,高成耀,等.BP神经网络隐含层单元数的确定[J].天津理工大学学报,2008,24(5):13-15.
[15] DHI. Auto Calibration Tool (User Guide)[K]. Denmark: DHI Water and Environment, 2012.
[2] 万增友.MIKE SHE模型国内应用现状及其关键问题研究[J].科协论坛(下半月),2011,(5):99-101.
[3] 姜凌峰,薛联青,刘远洪,等.基于MIKE SHE模型的干旱区节水灌溉对地下水位的影响研究[J].灌溉排水学报,2016,35(2):59-65.
[4] 田开迪,沈冰,贾宪.MIKE SHE模型在灞河径流模拟中的应用研究[J].水资源与水工程学报,2016,27(1):91-95.
[5] 卢小慧,李奇龙.基于MIKE SHE模型的流域地下水水文响应[J].长江科学院院报,2015,32(1):11-15,20.
[6] 王中根,夏军,刘昌明,等.分布式水文模型的参数率定及敏感性分析探讨[J].自然资源学报,2007,22(4):649-655.
[7] 刘飞,胡斌,宋丹,等.基于BP神经网络和均匀设计的边坡敏感性分析[J].水电能源科学,2014,32(10): 113-115,165.
[8] 郑震,张静,宫辉力.MIKE SHE水文模型参数的不确定性研究[J].人民黄河,2015,37(1):23-26.
[9] MA L, HE C G, BIAN H F, et al. MIKE SHE Modeling of Ecohydrological Processes: Merits, Applications, and Challenges[J]. Ecological Engineering, 2016, 96: 137-149.
[10] 郄志红.大坝安全监测资料正反分析的智能软计算方法及其应用[D].天津:天津大学,2005.
[11] REFSGAARD J C.Parameterisation, Calibration and Validation of Distributed Hydrological Models[J]. Journal of Hydrology, 1997, 198(1/2/3/4): 69-97.
[12] SUMAN A, AKTHER F. Investigation of Water Balance at Catchment Scale using MIKE SHE[J].International Journal of Engineer and Computer Science, 2014, 3(10): 8882-8887.
[13] DHI.MIKE SHE分布式水文模型培训教程[K].Denmark: DHI Water and Environment,2012.
[14] 沈花玉,王兆霞,高成耀,等.BP神经网络隐含层单元数的确定[J].天津理工大学学报,2008,24(5):13-15.
[15] DHI. Auto Calibration Tool (User Guide)[K]. Denmark: DHI Water and Environment, 2012.