摘要
根据昭平台水库综合利用特点,建立了水库调度函数的人工神经网络(ANN)模型,并与多元线性回归模型进行了对比。研究结果表明,ANN模型能够更好地反映水库月末水位与影响因素(水库月初水位、来水量、灌溉用水量)间的复杂非线性关系,而且水库实时调度结果多年均值接近多年平均最优情况。
According to the multipurpose feature of Zhaopingtai Reservoir, an artificial neural network (ANN) model of the dispatching function is established, and compared with multi-variate linear regression model. The re search result indicates that ANN model is better to reflect the complex non-lin ear relationship between the monthly end reservoir level and the influencing fac tors such as the monthly elementary reservoir level, monthly reservoir discharge , monthly irrigation requirement, and the result of the reservoir real-time dis patching is close to the average annual optimum situation.
出处
《水电能源科学》
2005年第3期20-22,34,i003,共5页
Water Resources and Power
基金
河南省自然科学基金资助项目(0411050800)
关键词
水库调度
调度函数
人工神经网络
综合利用
reservoir dispatching
dispatching function
artificial neural network
multipurpose use
