Intelligent Decision Technologies pp 171-180
Dealing with Seasonality While Forecasting Urban Water Demand
- Cite this paper as:
- Froelich W. (2015) Dealing with Seasonality While Forecasting Urban Water Demand. In: Neves-Silva R., Jain L., Howlett R. (eds) Intelligent Decision Technologies. Smart Innovation, Systems and Technologies, vol 39. Springer, Cham
Forecasting water demand is required for the efficient controlling of pressure within water distribution systems, leading to the reduction of water leakages. For the purpose of this research, it is assumed that the control of water pressure is performed by pressure reduction valves (PRVs) working in the open loop mode. This means that water pressure is controlled on the basis of the daily water demand profile, a 24-step ahead forecasting of hourly time series. A key issue in such time series that affects the effectiveness of its forecasting is seasonality. Three different techniques to deal with seasonality are investigated in this paper: auto-regressive, differentiation, and the application of dummy variables. This paper details a comparative study of these three techniques with respect to water demand time series and different predictive models. We show that an approach based on dummy variables and linear regression outperforms the other methods.