top of page

Feature 5 September 2022 (ComputerWeekly)

From sewage outflows to blue-green algae and hosepipe bans, water management has rarely been so much in the public eye. However, utilities are increasingly adopting cloud-based applications and platforms to make this precious resource go further.

Amazon Web Services (AWS), Accenture and Colombia’s Ecopetrol, for example, are seeking to address the full water-use cycle through a water intelligence venture announced in March 2022, which they see as contributing to net-zero initiatives and sustainability targets straddling water and energy sectors, allowing participants to share data that promotes water reuse both within and between industries.

According to Ecopetrol’s chief executive, Felipe Bayón, collaboration is needed to “truly ignite change” because companies cannot solve environmental challenges alone.

“We will use this platform to accelerate ... our goals of reducing 66% of freshwater captured and zero discharges to surface water by 2045, improving the environment for the communities where we operate,” says Bayón.

Adam Selipsky, chief executive at AWS, says the idea is to combine data from previously disparate sources across Ecopetrol’s operations, using Accenture “industry insights” and AWS cloud-based machine learning and artificial intelligence services, along with high-performance computing (HPC) and storage.

“Like other sustainability initiatives, water conservation is a big data problem,” says Selipsky.

Raymond Ma, general manager for Europe, Australia and New Zealand at Alibaba Cloud Intelligence, says this includes efficiencies around water supply and demand, via smart monitoring and automatic controls, including alerts.

Process optimisation can further reduce water use, with deep learning deployed to empower predictive maintenance, including by setting power operating intervals. Improved equipment performance and predictive capabilities that can warn of potential outages in advance can deliver even greater cost and resource savings over time.

“For example, when a water pipe bursts, the digital system can quickly find and locate the place, and provide scientific decision-making references for emergency response, which can reduce the amount of water lost in such an accident,” Ma tells Computer Weekly.

“Proven intelligence analytics technologies can predict future water consumption, while water supply and water pressure can be adjusted more accurately.”

Industry-specific cloud applications, such as Endress+Hauser’s Netilion Water Network Insights cloud-based water monitoring of flows, pressure, temperatures, water levels and so on, have been emerging in recent years, and UK utilities such as Yorkshire Water are continuing to progress along the road to digital transformation.

Cloud analytics supports ‘real opportunities’ for water industry

Gary Ashby, enterprise data architect at Yorkshire Water, says many British water firms have been collecting data from various assets above and below ground, in treatment works, pumping stations, reservoirs and the like for years, with many water utilities having existed in some form for about a century.

“Over the past 30-35 years, that has become more computerised, but historically it has been very difficult to organise, and in the past, the water industry tended to be quite segmented, with different teams and treatment works doing things in different ways, including the recording and storing of data,” says Ashby.

Today, cloud computing is supporting “real opportunities” for the water industry with analytics, through the ability to quickly process vast volumes of data. Managers and planners, meanwhile, no longer want to wait for the technical innovation to happen – they seek agility, and for Yorkshire Water, this means developing and improving predictive monitoring and maintenance capabilities.

“Cloud allows us to provide an evergreen technology environment – we’ve got a constantly managed, secured, scalable environment for our data. In the past, we had large datacentres, which we needed to create a capacity to deal with data analytics and all of the things that go with that,” says Ashby. “We’ve tried to create a data reservoir in the cloud.”

Yorkshire Water is in the process of migrating “an awful lot of our data” into that reservoir – “a sort of enterprise data catalogue in the cloud” – for analytics purposes, creating consistency and accessibility of data from multiple technologies. There are hybrid and public cloud elements.

“What the cloud offers is a really fast route to both proving the concepts and scaling the design, and changing or evolving the design,” he says.

With “close to half a million” different pieces of equipment located across thousands of miles, Yorkshire Water is now beginning to be able to pinpoint where incidents such as leaks are likely to occur and reduce the chance of those incidents happening. It’s also bringing data together to help it analyse customer sentiment.

“Water is a very multi-factorial, very weather-dependent and influenced business, and has traditionally been rather reactive,” says Ashby. “Now we can get a real understanding of how we’re performing right now and how we might perform in future.”

Leeds-based data consultancy partner Oakland Group is currently helping the firm further advance along the cloud-based path.

As “smarter water” use cases have expanded, so have ambitions around what can be achieved with a more consistent, traceable enterprise dataset, even as systems are swapped in and out, says Oakland director Andy Crossley.

Cloud considerations against an overall systems backdrop

This also means paying more attention to the governance, people, processes, operating model and so on.

“Because you can plug all the data in you want, it can be perfect quality with a brilliant dashboard that’s dead shiny and works in real time, but if nobody decides to make a decision off the back of it, it’s irrelevant,” says Crossley.

Cloud adaptability and flexibility mean Yorkshire Water need not reinvent the wheel. It can benefit from using a public cloud and a Microsoft stack, while still tackling the nuances of its own situation, including its specific legacy technologies, as well as the climate and geography that affect the way water moves through the regional ecosystem.

Iterative design takes elements that are more off the shelf and works out how to stitch those elements together in a way that reflects the need for Yorkshire Water to “own, run and evolve” the system, he says.

How does the company handle classic objections to cloud from a cyber security perspective? “All data doesn’t need to be secure and protected in the same way. You need to work out what your high-risk, high-value sort of data is,” says Crossley. “Data governance plays a really important role.”

Paul Duddy, CEO and founder of, says the cloud, blockchain and location tech provider has been working with Scottish Water. Detailed data on the delivery of water infrastructure improvements and emergency call-outs at the point of work or problem is becoming key.

“Water utilities are under immense pressure to do more with less – to reduce costs while improving or maintaining service levels and water quality,” says Duddy. “Demand is increasing, but the network itself is formed of old and new parts, making it difficult to identify problem areas, leaks and where to prioritise maintenance.”

Kieran Blackstone, co-founder and chief operating officer of consultancy Tecknuovo, agrees. Tecknuovo has helped Thames Water move from its siloed data applications and on-premise datacentre to a cloud ecosystem.

Like Yorkshire Water, the team is building a data lake on top of a single “landing zone”, incorporating data pulled from assorted applications and standardised, with machine learning models and predictive applications to be added on.

“The lake will contain all of the historical datasets but will have been quality checked and standardised, meaning there will be a single, clean copy of the data within it,” says Blackstone. “This will make the product team’s life at Thames Water much simpler and avoid rework of ...


bottom of page