基于小波神经网络对城市供水管网漏损的研究

摘要

The study of the pipeline leakage is an important project in the field of water supply.Taking maintenance da-ta of DN300 pipe diameter in Waterworks from 2011to 2014 as research object,wavelet neural network is used to forecaste the Water Network's leakage.Prediction model is established to compare with the BP neural network.Wavelet neural net-work is more accurate than the BP neural network in predict Water Network's leakage with the minimum error is 0.36%, and the maximum error is 13.47%.%管网漏损的研究是在供水领域一项重要的工程.论文以某水司提供的2011~2014年DN300主管的漏损作为研究对象,运用Matlab构建小波神经网络对其进行预测,并与BP神经网络模型进行比照,结果表明:小波神经网络比BP神经网络拥有更好的预测精确度,最小误差为0.36%,最大误差为13.47%.