Oliver Grievson highlights the dynamics that will be required for the water sector to capture the full potential of digital transformation.
When we think about the digital transformation of the global water industry, we immediately start to jump to concepts such as digital twins, 3D physical models, and, of course, artificial intelligence (AI) and machine learning. All of these technologies and their adoption into the mainstream are important as we move to a modern water industry. However, underpinning all of these technologies, to a greater or lesser extent, is data.
Data transformation
For me, the fundamental start to any data transformation journey (although this has been an unpopular opinion in the past) is stakeholder engagement. From the CEO of a water operating utility to the frontline operatives and technicians, there is a need for data and situational awareness – an understanding of how everything is operating from the grand scale of the whole utility to the individual scale of a single treatment works or pumping station.
Nowadays, in a water utility environment, all of this data is pushed into a data lake, or whatever data repository you choose (lake, pond, ocean, have all been touted). I believe the distinction is the size of data you have and whether it has been structured or is unstructured. This last point of whether the data is structured or unstructured is the important one here, and was the subject of an IWA project on meta-data that was concluded earlier in 2024.
Meta-data collection
IWA’s Meta-Data Collection and Organization (MetaCO) Task Group, led by Kris Villez, aimed to describe a number of data models – i.e., structured approaches to the management and storage of meta-data – that have been deployed successfully in recent years. In addition, its scientific and technical report included:
- Guidelines aimed at avoiding duplication of efforts and databases
- Current experience with applicable standards, such as open architectures, including Open Geospatial Consortium WaterML
- The potential of recently developed technologies, including block chain and ontology-based tools.
Meta-data collection will be essential to underpin the data lakes that are currently being proposed within the industry or are actively growing in size – giving the data the structure that it needs to be used effectively in a number of different applications.
Streamlining and accessibility
The industry as a whole has been brilliant at collecting data for a single purpose, but when a single piece of data is needed for multiple purposes – and potentially in multiple different databases or models – this is when things historically became unstuck. As the industry’s collection of data is increasing significantly, the lack of meta-data becomes a significantly larger problem as we enter the realms of big data.
UK duration monitoring programme
An example of this is the event duration monitoring programme in the UK and how it ties in with different datasets. Between 2014 and 2022, around 14,000 event duration monitors were installed on combined storm overflows. However, some of these were within the wastewater network and some were on the overflows from storm tanks.
The UK is moving – from a regulatory point of view – to install monitors on overflows to storm tanks, along with monitors on emergency overflows, in addition to flow meters measuring compliance with flow to treatment conditions across the country. This is on top of the water quality monitors that are going to be installed up- and downstream of all overflows to the environment.
From a non-regulatory perspective, the water companies are also using sensing and machine learning to look at wastewater network performance and blockages, with data coming in from tens of thousands of sensors.
What is not available currently is a system to join together all of this data so that it may be operated in a logical way. This, for example, could include sewer network level monitors working with regulatory event duration monitors to give an idea of the situational awareness of network performance – something that is happening with suppliers, however. In addition, current network performance indicators could work with the front end monitoring of wastewater treatment works in a way that is compatible with the water quality monitoring that is going to be installed over the coming decade.
By bringing this data together with models of the wastewater network, treatment works and the riverine environment, we would have a very powerful tool to not only monitor the performance of wastewater systems, but also their impact on riverine environments.
Underpinning the success of this is the availability and quality of the data – a subject that was addressed in the MetaCO scientific and technical report, ‘Digital Water: The value of meta-data for water resource recovery facilities’, which adds to other IWA work undertaken on this subject.
Garbage in, garbage out
We have all heard, or even potentially used, the phrase ‘Garbage In, Garbage Out’. It was a phrase that was first used by William Mellin in the 1950s when the majority of instrumentation within the water industry was still mechanical. (Gustaf Olsson’s book, ICA and me, provides an insightful history of the development of instrumentation, control and automation [ICA] in water and wastewater.)
Mellin highlighted that if you put poor quality data into a computerised system, you will, of course, get garbage out – the computerised system will not realise what is and what is not useable data. For an industry that is increasingly using machine learning and trying to make sense of huge datasets to garner insights, poor quality data would make it impossible to see the wood for the trees.
While it may be laudable to collect data for the sake of collecting data, in reality there is a cost to gathering data, and if the value of that data is not recognised, then its collection will not be maintained. This point has been highlighted by an IWA Digital Water Programme White Paper on digital transformation and instrumentation, ‘Digital Water: The role of Instrumentation in Digital Transformation’, which proposed the concept of the instrumentation life-cycle.
Instrumentation life-cycle
The first part of the life-cycle asks the user to define the ‘instrumentation need’ – or, taking it up a step, the ‘data need’. If the need for the instrument is understood and the data that it provides has a value higher than its cost, then the data quality should be ensured.
Understanding uncertainty
The next step is to understand the uncertainty associated with the data, which was a subject that was covered in another IWA Digital Water Programme White Paper, ‘Measurement Uncertainty in Digital Transformation’, published in early 2024.
Next steps to digital transformation
As the water industry transforms digitally, ‘digital tools’ are going to help the sector address global challenges and targets – most importantly, the acceleration of the drive to achieve Sustainable Development Goal 6, access to safe water and sanitation for all.
To manage water effectively, the industry as a whole needs to adopt the concepts that digital water offers. But we need to get the fundamentals right and ensure that the data collected is accurate and in a format that can be used. For this to be achieved, we need to garner the situational awareness to which I referred, to ensure data quality and understand its limitations through our knowledge of measurement uncertainty, so that we know what the data is for and where it fits into the system as a whole. Only by doing this will the water industry apply meta-data effectively.
More information
Digital Water: The value of meta-data for water resource recovery facilities, iwa-network.org/wp-content/uploads/2021/04/IWA_2021_Meta-data_IWA.pdf
Olsson, G., ICA and me – A subjective review. Water Research (2012),46, (6), 1585-1624
See: www.sciencedirect.com/science/article/abs/pii/S0043135411008487?via%3Dihub
Digital Water: The role of Instrumentation in Digital Transformation,
iwa-network.org/wp-content/uploads/2020/12/IWA_2020_Instrumentation_WEB.pdf
Measurement Uncertainty in Digital Transformation,
iwa-network.org/publications/digital-water-measurement-uncertainty-in-digital-transformation
The author: Oliver Grievson is an Associate Director at the global engineering consultancy AtkinsRéalis and a Royal Academy of Engineering Visiting Professor of Digital Water at the University of Exeter. He is also Chair of IWA’s Digital Water Programme.
