An adaptive methodology to predict daily consumption behaviour in water distribution systems
Water utilities can operate their systems at higher efficiencies when they are able to predict water consumption in the short and long terms. Baseform develops forward-thinking software for networked water infrastructures, including applications that use real time flow monitoring records to analyse demand patterns and forecast their behaviours. In this talk, we show how statistical methodologies can be a crucial aid in understanding human behaviour.
Several methodologies can be applied to predict the 24-hour demand pattern in district metering areas (DMA) of water distribution systems. However, in most cases, it is essential to know in advance the characteristics of each flow time-series, to choose the most suitable combination of parameters to use as the input of the method. In particular, the demand pattern derived is dynamic and may be influenced by the period of time considered. An adaptive predictive model of the pattern based on some selection criteria, which automatically choose the most appropriate parameterizations is described, in an attempt to overcome these drawbacks. This new approach is built on the concept of weighted percentiles, allowing the prediction to quickly adjust to changes in the consumption habits, as validated through the application to a large number of DMA.