Data-driven algorithm forecasts future water demand


EDWARD (simulating Extremes in Drinking WAteR Demand) is a data-driven algorithm which, on the basis of population statistics and climate scenarios, forecasts the development of water demand in the future. Water utilities can make use of this information to organise the drinking water infrastructure accordingly.

The EDWARD algorithm was developed by KWR within its Joint Research Programme with the Dutch water utilities, and is continuously fed with tailored online data: from Statistics Netherlands (CBS) on holiday behaviour, and from the KNMI (Royal Netherlands Meteorological Institute) on climate change.

Day-peak factor

With the EDWARD model, the impact of climate change and holiday-staggering on drinking water demand can be forecast for any random supply area in the Netherlands or Flanders. Concretely, the model results offer insight into the expected increase in the day-peak factor and the average water demand under various future scenarios.

KWR heeft het zelflerende machine learning model EDWARD bij drinkwaterbedrijf Oasen getraind met grote hoeveelheden historische data.

KWR has trained the EDWARD machine-learning model at the Oasen drinking water utility with a large volume of historical data.

Always current results

The algorithm allows for the calculation of possible future scenarios in greater detail than was previously possible. Because EDWARD is continuously fed with the most recent data from the KNMI and CBS, any new developments are immediately incorporated into the model forecasts. Thanks to these model forecasts, water utilities can effectively prepare the drinking water infrastructure for the future.


Video – 01:24
Water company joint research: Modeling the water demand