Volterra Series-based Neural Network and its Application in Tap-water Flow Forcast

The system of community tap-water is influenced by many factors, which is a typical nonlinear dynamic system. Both neural networks and Volterra series are widely used in nonlinear dynamic system. This paper discusses the relations between Volterra series and BP neural network, and proposes the Volterra series-based neural network and the solution of the hight order Volterra series kernel. In this paper, the ARMA model, BP neural network and Volterra series-based neural network are applied to short-term forecast a community tap-water flows. According to the results of the comparison, it shows that the Volterra series-based neural network is better than other methods.