Tackling Water Loss with Data-Driven Solutions

Is there anything we take for granted more than the clean, clear water we get straight from our kitchen taps? That water has likely travelled hundreds of kilometers across all sorts of terrains and conditions ­­­­– a journey only possible thanks to the people working behind the scenes to combat water scarcity and prevent water loss.

To find out more about reducing water loss, The Urbanista Podcast spoke to Maximilian Oberndoerfer, the Product Manager at Diehl Metering. They have extensive experience with water utilities, specifically in the water monitoring area.

 

Can you tell us more about water scarcity?

The issue of water availability, but also the scarcity of water, has increasingly become a topic that concerns us as companies, utilities, customers, and individuals. At the end of the day, we are all relying on water to be available when we need it.

Can you give us an example of water scarcity?
In France particularly it's been an issue over the last two years. They have imposed national regulations where certain regions have limits on how much water they can supply. Despite that, certain towns in France have had situations where there was no more water to be sourced from the wells, and they had to drive it in with trucks. It is honestly very scary relying on trucks like this.

How important is it to act quickly to resolve this problem?
We must act quickly. Firstly, by driving awareness and secondly, by providing solutions in collaboration with the utilities. We, as a company, see it as our responsibility to really contribute positively to that space and tackle this issue both domestically and worldwide.

Diehl Metering uses a 10-Minute Analysis. What is this?
The core concept of 10-minute analytics comes from utilities having fewer and fewer staff and less and less time. So, the idea is to take just 10 minutes a day to focus on the losses in your distribution system.
It might be difficult for an individual to digest the data and come to any meaningful results that will determine what course of action should be taken. Our user software analyses the data for you and presents the results so that you can easily interpret what's happening in your network and determines what you need to do.

How does the 10-minute analysis work in practice?
The operator joins their team in the morning to have a look at the system. They can see the network split into different zones, usually district metered areas (DMAs). Then it's very simple, like traffic lights: the red zones mean ‘I have water losses and I need to act now’, the orange or yellow zones mean ‘okay, there might be something I should do if I have the time’ and the green zones basically tell me ‘I'm fine, I don't need to do anything’.
The underlying magic is the way the different data sources are combined. We have SCADA (supervisory control and data acquisition) information from distributed sensors, information that tells where my pipes are, how old they are, what they are made of, the network pump information, and more. We take this data, and we combine it in a way that we can display this traffic light system on the DMA level.

What KPIs (Key Performance Indicators) are tracked, and how are they processed?
For the traffic light system that I mentioned, we look at a daily water balance. That’s a very simple comparison of what goes in and what comes out of the water network. That gives you a good indication how many physical losses you have in your network, especially if you know what is happening from the consumer side.
In addition to that, we build layers on top, for example, the minimum night flow, high numbers which we can presume are not due to real consumption, but instead losses in the network.
Finally, we ask, does my data make sense? Is it of a quality that enables me to make good decisions? For example, if I only get information regarding 10 out of 100 meters of a section, it's probably not very accurate.
The quality of the data is especially important, and we try to help our customers improve it over time with recommendations. For example, if there are irregular measurements from the district meter, maybe it's time to replace the data logger.

Is it difficult to assess water balance when there is varying flow and pressure?
Firstly, we look over longer periods of time. Within a day you can see if there is a peak at a certain time, for example during a football game, or more specifically during the 15-minute break.
With those kinds of events the total volume per day will be larger, so we also look at the consumer side. That could mean getting data from the households, or from the stadium in this case. We're not interested in the individuals but in how the overall population consumes water at a certain point in time to determine, is that a physical loss or is that just somebody using the toilet flush?

Are there any other benefits of this analysis for network managers?
Oftentimes we come into a scenario where people are not aware of the losses that are happening, and when you provide this level of transparency, these customers and utilities decide to act. Then we can easily prioritize areas to see immediate impact.
I can give you an example. The town of Brønderslev in the north of Denmark, which supplies water to around 20,000 inhabitants, faced very high water-losses. On average about 14% of the water supplied was being lost. In Denmark there is very strict regulation. If you have losses that are larger than 10%, you have pay hefty fines so there was a real urgency to act. The utility company decided to install DMAs, smart meters, etc. to get the data necessary. But they quickly realized that the data is only a start and utilizing this data is the key. We were lucky enough to work together with them on a solution using our 10-minute analytics and water management dashboard. Their water loss is now at 4.5% percent on average. That's quite a significant reduction.

How important are digitalization and AI in the water industry?
Everyone is thinking about AI. I try to only talk about AI when it brings benefits. When I think about broader digitalization trends, I see the trend for the Internet of Things (IOT), which suddenly provides an abundance of data. Before the IOT, we had SCADA stations, and we knew what was happening at very distinct points in our water distribution network. Now we know what's happening in the entire network. This can be good and bad because it becomes hard to manage. This is somewhere that AI can help. It’s not always necessary, but it’s a good thing to have up your sleeve.
The other useful area for AI is when we talk about hydraulic modelling, i.e., the static hydraulic models that are often used in planning and setting up new districts. Here, AI can play a crucial role bringing live data constantly into an application and help recalibrate them regularly to help understanding.

Have you seen resistance to digitalization or concerns about internet connectivity?
I would say that reluctance and pushback to digitalization is usually due to the regulation, with recent EU regulations being one of the most often mentioned. For the most part, I wouldn't say that utilities are scared, but it's not always easy to navigate the regulatory boundaries. For us, it's important to work together with these customers and often it turns out that together, with very transparent documentation, we can show that it is still beneficial. Digitalization and regulation may seem a hard road to start down, but it's safer in the long term, and it's not impossible.