基于小波神经网络对城市供水管网压力的预测  

Prediction of urban water supply network pressure by wavelet neural network

作　　者：汪健[1] 王煜[1] 秦正飞 高磊[1] 

机构地区：[1]昆明理工大学,云南昆明650092

出　　处：《供水技术》2015年第4期23-26,共4页Water Technology

摘　　要：为了对城市供水管网的压力进行更准确的预测,提出了运用小波神经网络的方法,对供水管网压力进行预测。该方法以BP神经网络为基础,用小波基函数取代了BP神经网络的隐含层函数,并对其进行了优化,并用该方法进行模拟仿真。结果表明,小波神经网络比BP神经网络更能准确预测水压值,最大误差为4.39%,最小误差为0.31%。该方法具有更强的学习性能、精确度及容错能力,在水压预测中有更实用的价值。In order to predict the urban water supply network pressure exactly, the method of wavelet neural network was put forward, and the water supply network pressure was forecasted. This method based on BP neural network, and hidden layer function of BP neural network was replaced by the wavelet basis function. The optimization was conducted, and a simulation was carried out in this method. The results indicated that wavelet neural network was more accurate than the BP neural network to predict water pressure value, which the maximum error was 4.39%, and the minimum error was 0.31%. This method had a stronger learning performance, accuracy and fault-tolerance performance, and had more practical value in the prediction of water pressure.

关 键 词：供水管网 小波神经网络 BP神经网络 水压值 

分 类 号：TU991.33[建筑科学—市政工程]