Weekly urban water demand forecasting using a hybrid wavelet–bootstrap–artificial neural network approach

Abstract

This study developed a hybrid wavelet–bootstrap–artificial neural network (WBANN) model for weekly (one week) urban water demand forecasting in situations with limited data availability. The proposed WBANN method is aimed at improving the accuracy and reliability of water demand forecasting. Daily maximum temperature, total precipitation and water demand data for almost three years were used in this study. It was concluded that the hybrid WBANN model was more accurate compared to the ANN, BANN and WANN methods, and can be applied successfully for operational water demand forecasting. The WBANN model simulated peak water demand very effectively. The better performance of the WBANN model indicated that wavelet analysis significantly improved the model’s performance, whereas the bootstrap technique improved the reliability of forecasts by producing ensemble forecasts. The WBANN model was also found to be effective in assessing the uncertainty associated with water demand forecasts in terms of confidence bands; this can be helpful in operational water demand forecasting.

Authors and Affiliations

KAZ ADAMOWSKI, JAN ADAMOWSKI, OUSMANE SEIDOU, BOGDAN OZGA-ZIELI Ń SKI

Keywords

Related Articles

Zastosowanie koagulacji i odwróconej osmozy do odzysku wody z wód wyparnych

Wody wyparne z zatężania gęstwy drożdżowej stanowią potencjalny surowiec do odzysku wody. Charakteryzują się niskim odczynem (pH 4.6–6.3), podwyższoną mętnością (3.65–13.7 NTU) i dużą zawartością ogólnego węgla organiczn...

Określenie sztywności prekonolidowanych gruntów spoistych uwzględniające historię stanu naprężenia i zakres odkształceń

Artykuł dotyczy opisu sztywności średnio spoistego gruntu silnie prekonsolidowanego. Materiał doświadczalny obejmuje badania trójosiowe na próbkach gruntu naturalnego o nienaruszonej strukturze (9 badań) i na próbkach re...

Restoration of Cnidion dubii meadows on Warsaw cross-section of the Middle Vistula Valley

The studies were carried out as the part of flooded meadows’ restoration on the Warsaw cross-section of Natura 2000 site Middle Vistula Valley. They concerned determination the possibility of restoration and improvement...

Estimating and verifying soil unit weight determined on the basis of SCPTu tests

The unit weight, as a basic physical feature of soil, is an elementary quantity, and knowledge of this parameter is necessary in each geotechnical and geo-engineering task. Estimation of this quantity can be made with bo...

Evaluation of correlation between parameters from CPTU and DMT tests and soil type behavior chart

This paper presents the correlation of soil parameters obtained from the dilatometer of the Marchetti (DMT) and the cone penetration tests (CPT). The data for the development of these correlations come from four experime...

Download PDF file
  • EP ID EP100044
  • DOI -
  • Views 158
  • Downloads 0

How To Cite

KAZ ADAMOWSKI, JAN ADAMOWSKI, OUSMANE SEIDOU, BOGDAN OZGA-ZIELI Ń SKI (2014). Weekly urban water demand forecasting using a hybrid wavelet–bootstrap–artificial neural network approach. Annals of Warsaw University of Life Sciences – SGGW. Land Reclamation, 46(3), 197-204. https://europub.co.uk./articles/-A-100044