Smart sensor placement
Better detection of contaminations in distribution networks
More and more sensors are coming onto the market to help water companies monitor the quality of the water in their distribution networks. But setting up such a sensor network is not an easy matter. To begin with, you have to decide what you want to achieve. Do you want to catch every single contamination?
Do you want to ensure that the network goes on functioning properly if a sensor fails? Thanks to a new computer model, water companies can now optimise their sensor networks to achieve their goals at minimum cost.
The Netherlands has more than 123,000 km of drinking water distribution pipes. Water companies want to keep the quality of drinking water in the network as high as possible. To do so, they can make use of a growing variety of water-quality sensors. But these are not cheap. You have to make smart use of the sensors by optimising their placement. Within the water sector’s joint research programme (BTO), KWR has developed a computer model which allows for the optimal placement of sensors in a distribution network. The model integrates hydraulic models, contamination scenarios and a numeric optimisation method.
You have to make smart use of the sensors by optimising their placement
Noordbergum Innovation Playground
The model is tested in 2014 on the network model in the Leeuwarden-Bergum subsector of Vitens’ Noordbergum Innovation Playground. A determination of the best placement of the available sensors is made in accordance with the different objectives. The model performs better than the uniform distribution of the sensors, or their placement according to the experience of the company’s staff. “What is always fundamental,” says KWR researcher Peter van Thienen, “is what the water company wants to achieve with its sensor network placement. A network that is optimised to practically never miss a contamination – maximizing “detection likelihood” – might actually be less suitable to enable the rapid identification and elimination of the contamination source.”
With this model, KWR provides the water companies with a tool with which they can optimally set up their new sensor networks depending on their selected objective. They can therefore strike a good balance between costs and protection levels.
© 2017 KWR Watercycle Research Institute
TagsSee all tags
Peter van Thienen
Senior scientific researcher