摘要
天然河道洪水期的水位~流量关系,因受洪水涨落率和回水顶托等因素的影响,一般成为非线性绳套状曲线。洪水流量过程的计算往往根据大量实测水位、流量资料,采用定性判断和实践经验手工绘制水位(或时间)与流量的时序型关系曲线.由水位过程(或时间)插补求得。本文运用人工神经网络反向传播(BP)模型,引入与流量变化密切相关的水位、涨落率、落差等因子,建立多因素相关的人工神经网络推流模型。实例应用结果表明,该模型的拟合能力强,推流计算精度能满足要求。
In general,water level and discharge relationship is non-linear ropy curve for natural river during flood season because of fluctuation rate of flood and back water pushing. Based on a large number of data of measured water lvel and discharge,through mapping water level (of time) and discharge curve by qualitative judyement and practical experience,discharge hydrograph of flood is obtained by interpoting water leve hydrograph(or time). Estimating flow models of artificial nervous network with many relative factors are established by application of back propagation(BP)of artificial nervous network and introduction of water level,fluctuation of flood and gead which is closely related to variation of discharge. Case history shows the model is fitted well. The estimated precision is satisfactory.
出处
《四川水力发电》
1997年第1期26-29,共4页
Sichuan Hydropower
关键词
水位流量
推流计算
神经网络
BP模型
洪水
water level and discharge relationship ropy curve estimating calculation of flood flow artificial nervous network BP model