Study on a new approach of municipal hourly water demand prediction

Author

Xu, Hongze ; Zhang, Wenjing

Author_Institution

Dept. of Autom., Beijing Jiaotong Univ., China

Volume

3

fYear

2004

Firstpage

2639

Abstract

Various factors that affected the municipal water demand and characteristics of municipal hourly water demand series were analyzed. Based on these, wavelet neural networks´ prediction models with periods of 24 and 168 hours were presented for working days and weekends respectively. In addition, the models were trained with gradient method. Finally, hourly water demand series of one city were predicted and compared with corresponding real values. Simulation results show the effectiveness of the proposed wavelet neural networks prediction models. At the same time, this type of hourly water demand prediction method can meet the demand of modeling and optimal dispatch of water supplying system.

Keywords

demand forecasting; gradient methods; neural nets; water supply; wavelet transforms; 168 hour; 24 hour; gradient method; hourly water demand prediction method; municipal hourly water demand series; optimal dispatch; water supplying system; wavelet neural networks prediction models; Automation; Cities and towns; Gradient methods; Neural networks; Prediction methods; Predictive models; Water;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Control and Automation, 2004. WCICA 2004. Fifth World Congress on

Print_ISBN

0-7803-8273-0

Type

conf

DOI

10.1109/WCICA.2004.1342075

Filename

1342075