Working with sensor data: still a lot of decisions to make

Knowledge exchange meeting for BTO companies

Drinking water companies use sensors for automated and regular measurements. The sensors generate large amounts of data. Decisions and solutions are required to determine how the companies collect the data, store them, manage quality control and transform the data into information products. Moreover, sensor data also provide specific opportunities for creating joint information products. On 17 February, the participants at a BTO knowledge exchange meeting concluded that a lot of water companies still lack a vision for sensor data even though they do have a vision in the area of digitalisation and data-driven working.

During the first knowledge exchange meeting of the BTO Hydroinformatics research theme in 2023, representatives of drinking water companies swapped their experiences in the field of sensor data.

Sensor data management in BTO and EU projects

Mollie Torello and Siddharth Seshan of KWR began their session with a presentation of sensor data management on a BTO and an EU project. They presented a roadmap that was developed to implement a digital twin for a drinking water treatment plant. Robust data management and connections with a range of disciplines are very important in this regard. The Waterverse EU project is developing a water data management ecosystem (WDME) to manage data from, among other sources, sensors in affordable, secure, fair and straightforward ways.

They discussed a case study conducted in collaboration with PWN in which a WDME is being developed to support decisions about the IJsselmeer lake. It is important to get stakeholders involved here.

The EU project Fiware4Water, in collaboration with Waternet, developed a real-time automated tool based on AI to clean up and validate crude sensor data from water treatment plants.

Sensor platform for water quality data

Jasper Hol of Aquon gave the second presentation. Aquon conducts water quality research for nine water authorities and it therefore works with a large network of sensors. An extensive architecture is needed to facilitate the flow, processing and validation of those data. The sensors, which belong to different companies, measure water quality factors. The data are sent automatically to the various platforms of these companies. The sensor data are then retrieved through a ‘pipeline’, processed and stored on a single platform. This process is known as ETL (extract transform load). In addition, Aquon uses a Delta

A Lakehouse is an open architecture that combines the properties of a Data Lake (a repository for crude data in various formats) and a Data Warehouse (a database for centrally storing and processing structured data from a range of underlying databases). A Lakehouse does this by implementing the data structures and data management properties of a Data Warehouse on a Data Lake. See also Datawarehouse | Explanation and definition – and Een Lakehouse: Data Warehouse en Data Lake één – Motion10 | Maak(t) uw digitale innovatie succesvol.

where data are stored, cleaned up and validated. The data are then sent to an SQL server and a dashboard. You can look at measurements with this dashboard and compare measurements from monitoring stations.

Internet of Water

Pieter Jan Haest then explained how sensor data are being used at de Watergroep as part of the Internet of Water project, a large project with a wide range of stakeholders. De Watergroep has developed a Data Space. This is an infrastructure that makes it possible to share data reliably. A data provider shares data and a data consumer receives them. The Water Agentschap in Flanders is building a digital water Data Space. The drawback of a Data Space for the data user is that the data are located externally and are therefore not managed in house. With a Data Dump, the consumer does have the data in house. The drawback here is that the data are not displayed in real time like they are in the Data Space. De Watergroep is looking at both options in the project.

Figure 1: Data users and how they are linked (©de Watergroep) alsdo see

Figure 2. Sharing data in a Data Space (©de Watergroep)

Familiar experiences

The presentations were followed by a discussion. The participants arrived at the conclusion that a lot of water companies still lack a vision for sensor data even though they do have a vision in the area of digitalisation and data-driven working. In addition, there are major differences between the companies in the extent to which they are already working on products for collecting and analysing sensor data.