城市供水管网水量预测的小波神经网络方法

Wavelet Neutral Network Forecasting Method for Water Consumption in Municipal Supply Water Networks

为提高城市供水优化调度的可靠性和实用性,对城市管网水量预测的方法进行了研究.提出了利用小波分解与人工神经网络相结合的小波神经网络管网水量预测模型,该模型以非线性小波基为神经元变换函数,通过伸缩因子和平移因子计算小波基函数合成的小波网络,并从理论上给出了严密的算法;同时通过逐步检验算法,科学地确定了网络结构,克服了普通人工神经网络难以确定网络结构、存在局部极小点等缺点.仿真结果表明,该模型比普通人工神经网络预测模型的预测精度高,并具有很强的适应能力.

In order to improve the reliability and the practicability of optimal operation of water supply system, a forecasting method for water consumption in municipal supply water networks is put forward. This method is based on wavelet neutral network in which the nonlinear wavelet basis function is used as the transform function of neurons instead of sigmoid function. The wavelet network can be got by calculating the flexing and the expansion factors. Moreover, a rigorous arithmetic is proposed in theory. And also the network configuration can be scientifical confirmed through checking the arithmetic step by step. Compared with the ANN method, this (method) is easy in configuration decision and can overcome the local minimal problem. Simulation results show that this wavelet neutral network forecasting method has the advantages of high forecasting accuracy and strong (applicability.)