Study On Water Demand Forecast In Water Supply Planning Of TEDA

Water demand forecasting is the main guide for the planning and construction of municipal water supply and saving. The select of forecasting method is also important for scientific and accurate water demand forecasting. There have been many forecasting methods all around the world, most of which are practically limited and have less precision. Commonly-used forecasting models are all based on the analysis of factors such as: population, gross national product (GNP), industrial output value, historical growth of Water supply and selling quantity. From the precision of previous forecasting, we can see that many long-term forecasting are greater than the actual value, for the limitation of these forecasting methods. The water supply planning, authorized by Tianjin development area in 1994, simply introduced the index forecasting method, the outcome of which exceeds the actual value by 78.5%. Based on the water-using structure of Tianjin development area and the characteristic of water demand growth, this paper aims at improving the precision of medium-term or long-term Water demand forecasting. To ensure its accuracy, advanced theory and forecasting methods, as well as rational water using indexes are adopted in this paperThis paper worked over the data analysis and the built-up of the mathematical model of the water demand forecasting of Tianjin development area. Firstly, it analyzed the methods and outcome of former forecasting in water supply planning (95). After a contrast of actual water supply quantity of recent years, the principal problems existent in the forecasting were found out. Secondly, the paper collected the correlative data of Tianjin development area, such as: the population, GNP, industrial output value, historical growth of Water supply and selling quantity and so on. Thirdly, analysis and comparison of the historical data of water supply system and water use index of cities of the same level were carried out. Then, a classified water use index was confirmed and adopted in this paper as main forecasting methods, which proved practical for Tianjin development area. After validating the accuracy of different forecasting methods, optimal medium-term and long-term forecasting methods were obtained in this paper.The result of this paper indicated that each forecasting method has its merits and demerits. The paper also analyzed the result of several forecasting models, such as: classification index, growth curve, regression analysis, neural network. The analysis indicated that classification index is mainly used for the medium-term and long-term water requirement forecasting of Tianjin development area. And the outcomes of other forecasting models were used as reference to ensure the accuracy and reliability of forecasting of classification index model.Growth curve is suitable for medium-term and long-term forecasting. The outcome of this model fits better with historical data, and its forecasting accuracy and reliability conform better to the growth trend of water requirement. The water demand forecasting serve as guidance for the short-term construction of water supply system. For the reason of little change in water-using structure and the obvious trend in historical data, different forecasting outcomes are close to each other. Compared with other forecasting models, neural network has better accuracy and reliability in short-term than long-term forecasting.