PURPOSE:To predict a service water demand amount with high precision even at the time of disturbance such as irregular fluctuation. CONSTITUTION:An arithmetic part 12 obtains a survey service water demand amount from a present service water demand amount detection value by a flow meter 11 to a time before a prescribed time and the waterworks demand amounts for respective past time bands are fetched from a demand amount fluctuation time series data storage part 13. A service water demand amount prediction part 14 predicts the waterworks demand amount for a short time by a decision theoretical time series model based on non-linear dynamic system theory with respect to the time series data. A fluctuation coefficient arithmetic part 15 obtains a fluctuation coefficient J from the prediction value Nm (n) and a survey service water demand amount Nr (n-1), and fuzzy inference means 16 and 17 correct a prediction service water demand amount by fuzzy infernce from the fluctuation coefficient and the present prediction service water demand amount Nm (n).
展开▼