Detecting unaccounted sewage water

Applying data science to prevent unaccounted sewage water from entering sewage systems.

The Dutch water authority Waterschap Drents Overijsselse Delta engaged Royal HaskoningDHV Digital to help it detect where and how unaccounted sewage water was entering its sewage systems and increasing transportation and pumping costs.

Together we devised a new data science approach, fuelled by existing data sets and new data captured from sensors deployed within the sewer system. With this, the water authority can now confidently see where unaccounted water is entering the system, and take the right actions to prevent it immediately.

The challenge

Detecting foreign water in sewage systems

Sewage water costs a lot of money and energy to transport and purify. So, when unaccounted sewage water sources such as ground water, heavy precipitation, or an overflow from a local body of water enter the system, it can have a significant impact on costs.

Waterschap Drents Overijsselse Delta wanted to better understand where and when this water was entering sewage systems around Kampen and Raalte. To achieve this, it needed to go beyond simply looking at historical data and trends, and find a new way of detecting foreign flows sooner.

The solution

Machine learning and hydrological knowledge combined

Waterschap Drents Overijsselse Delta engaged Royal HaskoningDHV Digital to help it devise a new approach to unaccounted sewage water detection based on modern data science techniques.

Previously, the water authority relied on theoretical assumptions about the area to identify potential sources of unaccounted sewage water. Together, we devised a new plan that would make use of actual measurements – including readings from sensors within sewer systems, precipitation level data, and flow rates from sewer pumping stations.

The municipalities of Kampen and Raalte were selected for a pilot project that would test this approach. An algorithm was then created based on all available data for the selected region, and visualisations were added through a simple dashboard. This enables virtually anyone to instantly identify flow changes and potential sources of unaccounted sewage water in the system.

The result

Taking the right actions, faster

Now, when a deviation between pumping rates and flow rates is found, it can be compared against data from across the sewage system, and precipitation data. This provides insight into the possible causes of a deviation, for example leaks, negative overflows, faulty connections or discharges.

That insight enables the water authority to take the right actions faster, and prevent further unaccounted sewage water from entering the system before it can have a significant impact on costs and pumping throughput.

Plus, because the pilot algorithm was based on data that many water boards and municipalities already have, this model and approach can now easily be rolled out to solve the same problem across further municipalities in the Netherlands.

Digital Twins for wastewater infrastructure

Discover the vital role digital twins will play in the future of wastewater management.

Download now

Digital Twins for Wastewater Infrastructure l Royal HaskoningDHV

Anything we can help with?

Have a question, comment or business enquiry? Talk to one of our experts to see how we can help.