A Short-Term Water Demand Forecasting Model Using a Moving Window on Previously Observed Data

Abstract

Abstract:In this article, a model for forecasting water demands over a 24-h time window using solely a pair of coefficients whose value is updated at every forecasting step is presented. The first coefficient expresses the ratio between the average water demand over the 24 h that follow the time the forecast is made and the average water demand over the 24 h that precede it. The second coefficient expresses the relationship between the average water demand in a generic hour falling within the 24-h forecasting period and the average water demand over that period. These coefficients are estimated using the information available in the weeks prior to the time of forecasting and, therefore, the model does not require any actual calibration process. The length of the time series necessary to implement the model is thus just a few weeks (3-4 weeks) and the model can be parameterized and used without there being any need to collect hourly water demand data for long periods. The application of the model to a real-life case and a comparison with results provided by another model already proposed in the scientific literature show it to be effective, robust, and easy to use.

Forecasting Hourly Water Demands With Seasonal Autoregressive Models for Real‐Time Application

“…Wong et al () arranged the factors impacting water demands as a hierarchy of base, seasonal, and calendric variables and applied Multivariate Linear Regression (MLR) to predict demands. Pacchin et al () forecasted hourly demands based on simple ratios of recent 24 h total demands and specific time of day demands. Apart from the statistical approaches of MLR and time series analysis, other approaches such as Artificial Neural Networks (ANNs; Bougadis et al, ; Jacobsen & Kamojjala, ; Jain & Ormsbee, ), ensemble ANNs (i.e., generating ANNs for individual hours within a day; Rangel et al, ; Romano & Kapelan, ), wavelet‐bootstrap‐neural networks (Tiwari & Adamowski, ), Relevance Vector Regression with wavelet transforms (Bai et al, ), Support Vector Regression (SVR; Bai et al, ), and SVR with a Fourier series representation of the predicted deviations (Brentan et al, ) have also been used for demand representation and/or forecasting.…”

Overview, Comparative Assessment and Recommendations of Forecasting Models for Short-Term Water Demand Prediction

“…The stochastic nature of water demand during the day and week is influenced by several factors; namely, climatic and geographic conditions, commercial and social conditions of people, population growth, industrialisation, technical innovation, cost of supply, and condition of WDS [1][2][3][4]. Therefore, water utilities need accurate and appropriate short-term water demand (STWD) forecasts in order to continually satisfy consumers with quality water in adequate volumes, and at reasonable pressures [5][6][7].…”

Smart Water Demand Forecasting: Learning from the Data

“…Predictions of urban water consumption are essential in water economies, especially under the threat of unprecedented water shortages [1]. Short-term water demand forecasting provides estimates of demand over the next hours or weeks to make informed operational, tactical, and strategic decisions that will improve the performance of the network [4,10].…”

“…However, predicting demand is a challenging task, due to its dynamic nature and inherent randomness, as well as the underlying relationships between consumption and multiple other household, socio-economic, and climatological factors that are not yet fully understood. The majority of short-term forecasting models use past consumption over the past week, month, or year as the main predictor, although some studies have considered the effect of climatic variables such as temperature, humidity, and precipitation [10,12].…”

“…A variety of studies have explored the application of machine learning methods such as Artificial Neural Networks, Support Vector Machines, and fuzzy logic in short-term water demand forecasting [6,7,8,12,18], or even hybrid models that combine one or more methods, but Random Forests have rarely been used in the water demand literature. The few studies that attempted this [1,2,10], although limited by data availability, demonstrated the potential of this method to accurately capture the dynamic nature of water consumption.…”