Improved insight and control with Digital Twins at water utilities

The Hydroinformatics knowledge exchange meetings are organised as part of the Joint Research Programme (BTO) of the water utilities. An important goal is to exchange practical experience on the basis of a relevant theme.

Transparent data quality

The meeting started with a presentation by Lennart Brokx (KWR) about digital twins for ASR systems. ASR (Aquifer Storage and Recovery) systems are used to store fresh water underground using infiltration wells. They are used, for example, in greenhouse horticulture. Not all the infiltrated water can be recovered and so it is often not known exactly how much water is still available. This is where the digital twin can help. It consists of a groundwater model (Modflow6) that runs in parallel with the physical twin (the actual system), and the data from both can be displayed on an online dashboard. The groundwater model is also used to run a number of model scenarios and make an estimate of the remaining freshwater stocks in the subsurface. Lennart listed some lessons learnt from this project:

• Technically, it is possible to link data and models to create a digital twin. However, the field data must be in order before you can make a good model and dashboard.
• The complexity of the subsurface is difficult to capture in a slimmed-down groundwater model.
• A dashboard that displays field data clearly always helps to get a clearer picture of the system.

The participants from the drinking water utilities understood the added value of building the digital twin to identify the data with enough quality. The result may be that the utilities will be prepared to invest in the data.

A wide collection of applications

Other applications of digital twins were explained in four short use cases. Evides and Brabant Water use software from the market, while PWN and Vitens build their own digital twins. Joeri Legierse (Evides) began with a presentation on the digital twin WaterSight. WaterSight makes it possible to bring together a lot of fragmented data in a central location at Evides. It also produces automatic reports (Power-BI). A statistical model is used to produce a seven-day forecast. At the same time, the model receives real-time data. As soon as this produces a discrepancy, for example because of an incident such as a spontaneous leak, a message is automatically sent to the drinking water utility. The main goal of the digital twin has been achieved for Evides: incidents are recorded correctly during the pilot project and there are fewer false positive alarms than with pre-existing alarm systems in PI. The system is now being rolled out to the area served by Evides.

Suze van der Meulen (PWN) then presented the use of the Digital Twin Roadmap developed by KWR in collaboration with other drinking water utilities. PWN wants a real-time picture of the security of supply, sustainability and costs from the source of the Rhine through to and including delivery to the customer. PWN then wants to make decisions based on integrated information. The idea is therefore to have a digital twin for the whole system (from the Rhine to the customer) between now and 2050. They are doing this by making digital twins for different domains. Because these digital twins are built using the roadmap in similar ways, they can eventually be combined in a large system. A ‘central digiteam’ has been established that will embed the roadmap at PWN. Each step involves: organisation, model development, data management, and architecture and security. By using the roadmap at both the central and local levels, a complex system can be analysed in manageable steps.

Wouter Huisman (Brabant Water) then presented SOFIA and the digital twin of Brabant Water’s mains. Brabant Water’s challenges are:

• the water system is out of balance (for example because of drought);
• the retirement of experienced staff;
• the energy transition;
• limited sources;
• the digitalisation of society;
• rising costs.

The ‘company programme on the road’ was launched three years ago on the basis of these challenges. One of the tracks is a smarter approach to business operations. SOFIA consists of the data-driven control of water supplies based on data from sensors and smart water meters in the mains system. The goal of the SOFIA programme is to have a real-time hydraulic model of the mains network by 2025: a digital twin. Brabant Water wants to use this digital twin for the following:

1. real-time monitoring of the mains;
2. detecting abnormalities;
3. simulations of possible scenarios such as malfunctions;
4. displaying water-consumption patterns.

The project was launched last year and it will be scaled up and implemented this year, with more and more areas and sensors being added.

Joukje Keuning (Vitens) finished the session with a presentation about digital twins at Vitens. Vitens has launched a project to improve process control and customer satisfaction called INFRA2025. INFRA2025 includes three digital twins:

1. OPIR: a digital twin that can predict water consumption and production;
2. SLIMM: a digital twin of, among other things, water quality, production and energy consumption at production locations;
3. A digital twin for distribution.

Vitens makes the digital twins for operations and analysis & the control centre. This means the system learns faster without large amounts of area knowledge. Scenarios can also be calculated prior to an emergency.

Can water utilities still manage without digital twins?

The session ended with a discussion during which it emerged that the participants thought that a future without digital twins was unlikely. Of course, it will still be possible to keep a water utility up and running without digital twins but this is expected to require much more manpower than if they are used.

Digital twins therefore offer many benefits. In operations, they help with the rapid detection of anomalies and the optimisation of the system. Working through scenarios in advance with the digital twin makes knowledge immediately available when the scenario becomes a reality. Less detailed knowledge, such as area knowledge, is required. This is an advantage because many experienced employees, who have a lot of area knowledge, will be retiring soon. Moreover, people do not stay with the same employer for as long they once did and so they do not build up knowledge to the same extent. In addition, digital twins are generally built to be more accessible than traditional models, reducing dependence on modellers.

It is thought that digital twins also save time and costs by improving efficiency. Water utilities are constantly evolving, and digital twins can accelerate decision-making in this respect. A single platform is created that all departments can use together – in other words, a shared picture is created. This also provides more transparency (with no more separate Excel files).

Even so, there are some challenges involved in setting up and using digital twins. The quality of the data is essential. But the digital twin also shows where the anomalies are, making data quality much easier to determine and therefore helping to improve it. Another challenge is that people from different disciplines come together in the digital twin that but these people often have their own jargon and ways of working which sometimes make it difficult to establish a team. It is very important to speak the same language!

Finally, the participants said that they expected that, once you start using digital twins, it is very difficult to ever revert to a situation without them. Nevertheless, it continues to be important to be maintain the capability to process and supply water in the event of an outage of the digital twin (or the underlying digital infrastructure). Water utilities must remain prepared and able to get this job done.