Research On Prediction Technique Of Water Requirement About The Small Town Based On Chaos Theory

Water is an important factor which influences the economy development, and plays the key role for the exist and development of the town. In order to assure the reliability of the town’s water supply, we need to have the accurate town water short-term prediction by compiling the overall plan and project plan of the town water supply to guide the construction of the town water supply infrastructure, making the water supply to match it demanded. The town water requirement is associated with many factors, and the all factors have the high nonlinearity. At present, there are two most common methods on water supply of short-term prediction:one is the explanatory prediction, it needs to find all the factors which influences the prediction and build the regression analysis model; the other is the time series analysis method, it bases on the former predicted observation data, then by the sequential analysis to get its order-changing rule.The town water requirement is influenced by many factors such as the economy, life standards, population and all day’s weather, So it is fairly difficult to find the system internal changing rule and the influence over the all factors. Moreover, the factors affect the water requirement are different and the regression analysis model is not universal, making the regression prediction method limited for the application in the water requirement prediction process. The time series analysis method which formed by the data observed or recorded.it simplifies the complicated external influencing factors’function, it only concerns the internal rule of the historical observed data and its data-changing mode along with the time. Thus, it does the relative description and explanation to the whole system, and predicts the changing system in future. The method fits the town water requirement changing character along the time better, so it has the wide application in terms of predicting the water supply.As a nonlinearity dynamical system, the neural network has an evident advantage in handling the nonlinear problems. This paper combines the chaos theory and BP neural network, and sets up the BP neural network water requirement’s prediction model based on the chaos time sequence. It does the correct short-term predication for the town water requirement effectively and with the good matching effect and high prediction accuracy. Furthermore, the BP neural network itself is a very easy distinguishing model. We can confirm the number of the input nodes easily no need to establish the prediction mode base on the actual system math mode, avoiding the steps for building the system model before predicting. However, theory of artificial neural network is introduced into prediction of urban water consumption in this paper. Since BP neural networks owns disadvantage of slow convergence and local minimization, the proposed prediction model of urban water consumption is based on chaos optimal method and BP algorithm. Typical examples prove that t he model is a very effective and accurate water consumption forecast model and easy to handle with. This method is expected to be widely used in practice.