Prediction of the amount of water supplied in wide-area waterworks

In this paper, a fuzzy model for forecasting water supply of booster pump station, water treatment plant, and delivery well in water supply facilities of waterworks is proposed. The prediction is carried put based on pumping out water quantity of the day, and temperature and weather conditions of a few days. The structure of the premise of the model is identified by use of fuzzy clustering technique (FCM). The parameters of the model are identified by gradient descent. The consequent part of the model takes fuzzy singletones, i.e., real number. This identification technique has advantages of simplicity, flexibility, and high accuracy compared with the statistical method and conventional fuzzy modeling technique. The proposed approach is applied to a water distribution system of metropolitan areas in Korea and its effectiveness is checked