Hourly Campus Water Demand Forecasting Using a Hybrid EEMD-Elman Neural Network Model
Accurate and reliable water demand forecasting is important for effective and sustainable planning and use of water supply infrastructures. In this paper, a hybrid EEMD-Elman neural network model for hourly campus water demand forecast is proposed, aiming at improving the accuracy and reliability of water demand forecast. The proposed method combines the Elman neural network, EEMD method, and phase space reconstruction method providing favorable dynamic forecast characteristics and improving the forecasting accuracy and reliability. Simulation results show that the proposed model provides a better performance of hourly campus water demand forecast by using the real data of water usage of our campus.
KeywordsElman neural networks EEMD Phase space reconstruction Water demand forecasting
- 2.Beal, C.D.: SEQ residential end use study. J. Aust. Water Assoc. 38(1), 80–84 (2011). https://www.researchgate.net/publication/279662589)
- 5.Luo, X., Jiaqi Yang, J.: Study on the imbalance of shipping demand and supply of inland water transportation of Yangtze River. ICTIS 2013: Improving Multimodal Transportation Systems-Information, Safety, and Integration, p. 2211–2221 (2013). https://doi.org/10.1061/9780784413036.297
- 6.Adamowski, J., Chan, H.-F., Prasher, S.O.: Comparison of multiple linear and nonlinear regression, autoregressive integrated moving average, artificial neural network, and wavelet artificial neural network methods for urban water demand forecasting in Montreal, Canada. Water Res. Res. 48(1) (2012). https://doi.org/10.1029/2010wr009945
- 12.Adamowski, J.F.: Peak daily water demand forecast modeling using artificial neural networks. J. Water Res. Plann. Manage. 134(2), 119–128 (2008). https://doi.org/10.1061/(ASCE)0733-9496(2008)134:2(119)CrossRefGoogle Scholar