基于BP神经网络的渠道无干扰量水方法  

A Non-interference Flow Gauging Method for Channels Based on BP Neural Network

作　　者：闫欣[1] 彭世彰[1] 罗玉峰[1] 徐俊增[1] 熊玉江[1] 卞益龙 

机构地区：[1]河海大学水文水资源与水利工程科学国家重点实验室,南京210098 [2]高邮市水务局,江苏高邮225600

出　　处：《节水灌溉》2014年第5期79-81,共3页Water Saving Irrigation

基　　金："十二五"国家科技支撑计划课题(2012BAD08B04;2011BAD25B07);水利部公益性行业科研专项项目(201301014);江苏高校优势学科建设工程资助项目;江苏高校优秀科技创新团队

摘　　要：灌区量水是节约农业用水、灌区实行按方收费的重要手段。BP神经网络具有很强的非线性映射能力,可描述通过水工建筑的流量与上下游水位之间的函数关系。以渠道进水口上下游水位作为输入向量,以渠道流量作为输出向量,构建了3层神经网络模型。以江苏高邮灌区2012年一条斗渠的水位流量观测数据,应用Matlab神经网络工具箱构建模型,采用trainlm算法进行网络训练与检验。结果表明:模型能较好地反映影响因素与渠道流量之间的关系,计算精度较高,提出的基于BP神经网络的渠道无干扰量水方法简便可行,具有较广的应用前景。Flow rate measurement is an important means for saving agricultural water and charging according to water quantity in irri- gation districts. The complex nonlinear relationship between the channel flow rate and the influencing factors can be well reflected by a Backpropagation （BP） neural network. A neural network model with three layers is developed by using the neural network tool box of Matlab and trained by using the trainlm algorithm. In the model, the input vectors of model are the water levels at upstream and downstream while the output is the channel flow rate. The stage-discharge data of a field channel from Gaoyou irrigation district, Jiangsu province, are used to train and test the model. The results show that the proposed model can well reflect the relationships between channel flow and relevant factors.

关 键 词：无干扰量水 渠道流量 BP神经网络 

分 类 号：S275.3[农业科学—农业水土工程]