Forecast model of water consumption for Naples |
| |
Authors: | B Molino G Rasulo L Taglialatela |
| |
Affiliation: | (1) Institute of Hydraulics Constructions, University of Basilicata, Via della Technica 3, 85100 Potenza, Italy;(2) Department of Hydraulics, University of Naples Federico II, Via Claudio 21, 80125 Naples, Italy |
| |
Abstract: | The data refer to the monthly water consumption in the Neapolitan area over more than a 30 year period. The model proposed makes it possible to separate the trend in the water consumption time series from the seasonal fluctuation characterized by monthly peak coefficients with residual component. An ARMA (1,1) model has been used to fit the residual component process. Furthermore, the availability of daily water consumption data for a three-year period allows the calculation of the daily peak coefficients for each month, and makes it possible to determine future water demand on the day of peak water consumption.Notation
j
numerical order of the month in the year
-
i
numerical order of the year in the time series
-
t
numerical order of the month in the time series
-
h
numerical order of the month in the sequence of measured and predicted consumption values after the final stage t of the observation period
-
Z
ji
effective monthly water consumption in the month j in the year i (expressed as m3/day)
-
T
ji
predicted monthly water consumption in the month j in the year i minus the seasonal and stochastic component (expressed as m3/day)
-
C
ji
monthly peak coefficient
-
E
ji
stochastic component of the monthly water consumption in the month of j in the year i
-
Z
i
water consumption in the year i (expressed as m3/year)
-
Z
j
(t)
water consumption in the month j during the observation period (expressed as m3/day)
-
evaluation of the correlation coefficient
-
Z
j
(t)
water consumption in the month j during the observation period minus the trend
-
Y
t
transformed stochastic component from E
t
: Y
t
=ln Et
-
Y
t+h
measured value of stochastic component for t+h period after the final stage t of the observation period
-
Y
t
(h)
predicted value of stochastic component for t+h period after the final stage t of the observation period
- j
transformation coefficients from the ARMA process (m, n) to the MA () process |
| |
Keywords: | Forecast model monthly peak coefficients daily peak coefficients |
本文献已被 SpringerLink 等数据库收录! |
|