The digital economy is transforming water utilities and industrial players, and bringing potentially profound changes for desalination and reuse specialists. The common thread is data.

Whether it is asset-level data collected from individual membranes, pumps, and water tanks, or network-wide data that covers an entire asset portfolio and combines with other data sources, new digital capabilities are transforming water management.

The most immediate impact of digital technology in the sector is at asset level. Pump manufacturer Xylem claims to be the first to market with a digitally enabled wastewater pump, Flygt Concertor, which has “integrated intelligence,” the company says. The product uses pressure sensors, monitors torque and current, and automatically starts a cleaning cycle if it becomes clogged. It also adjusts running speed to use minimal energy.

“If you have traditional on-off pump, you will use level regulators. The water will enter the pump and at that level, you start to pump. It will pump down to a level, and then it will wait,” says Xylem vice president and director of product development, Christian Wiklund. “This new system automatically runs the next pump cycle at a lower speed, and will continue to do so until the energy demand increases, so then it finds the optimum.”

The pump has the potential to produce “tremendous savings” of up to 70 per cent on energy bills, adds Wiklund.

Grundfos, another pump-maker, is eyeing the opportunity for digital products. Fredrik Ostbye will take up the newly created role of group vice president and head of digital transformation at Grundfos on 1 May 2017. Ostbye has a track record as an entrepreneur, most recently as “a kind of entrepreneur inside Telenor,” the largest telecoms operator in Scandinavia, he says. Ostbye sold Telenor an engineering business in 2013, which he had built up from scratch, developing digital solutions for utilities in Scandinavia as early as 2005.


“Many of the big manufacturing companies now have a challenging decision to take: either they develop and use digital technologies that are coming out, to change their role and offer; or they stay as they are today, a product company developing and producing physical products,” says Ostbye.

“It’s very tricky for a customer to do a lifecycle cost analysis. If they buy as a service, it’s much easier.” Fredrik Ostbye, group vice president and head of digital transformation, Grundfos

He envisages a future in which suppliers no longer sell a product as a one-off transaction, and instead offer product-as-a-service. In this new model, the supplier potentially monitors product performance and provides maintenance services, as well as timely replenishment of products, for a recurring fee.

“It’s very tricky for a customer to do a lifecycle cost analysis. If they buy as a service, it’s much easier. There are benefits for customers and manufacturers in transforming,” he says.

GE Water is among the desalination and reuse specialists that are already using data to help make asset maintenance more cost-effective. Its InSight platform, which began life in 2010, enables customers to view field data together with machine and sensor data, and other data such as laboratory data, in a single dashboard.

“The first step is to take those silos of data and to connect them in one platform. You can look at them, graph and trend them together, report on them all together and, most importantly, it gives the operations team visibility on one dashboard,” explains GE Water leader for business development digital water, Steve Davis.

“If you constantly have a leak on the same membrane, one out of 100, let’s replace that one instead of going after 20.” Steve Davis, leader for business development digital water, GE Water

Among the top three analytics packages that GE Water has developed on InSight is a predictive maintenance application for membrane cleaning. “The types of data we look at for membrane cleaning are things like pressure, conductivity, temperature, and total suspended solids, taken from sensors that are running live, and along with that we look at data from offline testing, such as on organics,” says Davis. The data analytics software uses these multiple datasets to predict where build up will occur on the membrane, and to alert customers.

Comparative data

Further, GE Water is repackaging a tool specifically for large membrane installations. “Let’s say they have 100 membranes, they look at those 100 membranes as one, typically, in customer sites. Whereas in reality those are 100 different membranes that each has its own asset life, serviceability, and quality. By looking at them individually we can make the whole much better. If you constantly have a leak on the same membrane, one out of 100, let’s replace that one instead of going after 20, potentially. This gives you much better efficiency within the membranes,” says Davis.

The analytics software can compare the efficiency of different technologies and membranes to see which one works best, whether in particular water conditions or in a specific installation. Additionally, GE Water anticipates that in future it will be able to benchmark performance using anonymised, aggregated data, whether by region, customer type, or industry.

Another popular application on InSight relates to water tanks. “We use some very simple analytics to predict ‘How many days to run out?’, and ‘When do we need to schedule a truck?’, and ‘What’s your rate of use?’ Then if we get higher or lower rates we can notify customers,” says Davis. The app helps customers to get the best out of their chemicals inventory. “It’s an example that everyone can relate to, and we’re growing traction very easily,” says Davis.

Needle in a haystack

A further point on the horizon of predictive maintenance is the ability to combine asset data with other data sources, particularly weather information. This could mean anticipating storm water surges, flooding, or at the other extreme, drought. “It could be, if you have a treatment plant, you want to have an even flow into it. You could use pump stations as a buffer, and make the flow from the pump stations as even as possible. So you can have intelligence on that level. You know that there will be a storm coming in five days, and that by emptying all your pump stations you will create a buffer for the excess water,” Wiklund says.

However, although combining multiple data sources seems potentially powerful, the sheer volume of data that’s sloshing about presents a challenge in terms of identifying what’s most relevant and presenting it in useful ways.

“We see a smart water network as a whole network of assets that need to be operating at optimum level to deliver service. You need to know what assets you have, where they are, how they are interconnected, and how they’re performing,” says Adrian Newcombe, business areas director for Trimble subsidiary, Telog, which has a suite of software applications that support network operations and asset performance management.

“It’s about being able to show your network assets against a map and to understand why assets are behaving a they are.” Adrian Newcombe, business areas director for Trimble subsidiary, Telog

“Visualisation and spatial views using global positioning systems (GPS), in terms of field engineers being able to absorb information, are particularly important. It’s about being able to show your network assets against a map and to understand why assets are behaving a they are,” says Newcombe.

“You want to present actionable, relevant information to an engineer in the field or a control room so that they can make good, timely, and informed decisions. They have the expertise, and they need the proper information enables that them to make better decisions. They don’t want to have to extract that needle out the haystack of all the data that’s collected,” he says.

Small data

It’s a view echoed by Water Planet chief executive Eric Hoek. “That big data approach is fine, the problem with it from my perspective is that you’re monitoring a lot of noise and you’re expecting these statistical algorithms to be able to detect signal amidst this massive amount of noise. It’s kind of like looking for a needle in a haystack. That’s why they call it big data. Our approach we call small data, the only signal that matters is the membrane,” he explains.

Water Planet’s IntelliFlux application, a software patch that helps to optimise the cleaning regime for ultrafiltration and microfiltration installations, is a different proposition from other data-driven analytics offers, in that it uses artificial intelligence to build a digital memory of what happens at individual sites.
Hoek describes it as like GPS company Waze, which was acquired by Google for $1.1 billion in 2013.

Waze differs from traditional GPS navigation software because it constantly checks for updates to the best available route based on real-time information on driving conditions. “Conventional controls, you plug in the address to GPS, it picks a course and then you stay on that course until you get there. What IntelliFlux represents is an autonomous vehicle. You get in, shut the door, programme the address, and it just takes you. So you’re free to do other things,” says Hoek.

“All the time, Intelliflux is saying, ‘How are we doing? How healthy is the membrane?’” Eric Hoek, chief executive, Water Planet

The software effectively steps through a ladder of options, starting with the least wasteful. “The first rung is permeate relaxation, you pinch the permeate valve closed and the water scours across the surface of the membrane on the feed side, and it washes off the membrane a little bit. Then a really short, really mild backwash. Then we do a harder one, a longer one, more frequently. You’re stepping up the ladder the whole time. Sometimes you may hit it with hot water. You may use some chemicals. All the way to shut it down and do a CIP (clean-in-place). All the time, Intelliflux is saying, ‘How are we doing? How healthy is the membrane?’ It’s extracting the data from how the membrane is performing. And it’s always assessing, ‘Do we need to change? Do we need to change? Do we need to change?’”

When a change occurs, such as a shift in the water’s characteristics, the software remembers which step was most effective in the last similar such situation, and goes there first, then steps up and down the ladder as needed.

“The cool thing about IntelliFlux, we can put it on anybody’s filtration system just with the standard sensors, flow meters, pressure gauges. We don’t need to change any hardware, just the software batch. When we install IntelliFlux, and it has been in operation for a while, and we have a baseline of operating data, we see up to 20 per cent lower operating costs, just by turning the software on,” he adds.

Investment decisions

Arguably the biggest potential impact of digitisation is at the network level. Improved visibility of network performance as a whole can support strategic decision-making by influencing thinking on capital investment, and, in jurisdictions such as the UK, have a role in returns to the regulator.

“In the UK, the utility has to track how it responded to an event, and explain that to the regulator: the activities, the timeline. An end-to-end system enables you to do that, to identify an event, push out work orders to the field, and then track what’s happened in the field,” says Newcombe.

A network-level view of asset failures and maintenance can further be used to help prioritise capital investments and to support other investment decision-making. “If they have good data, it helps that process of securing investment and funding,” adds Newcombe.

GE’s Davis believes good quality performance data will be instrumental in securing investment in the growing water reuse market. “It is going to be really critical to getting that market going, and going fast. As you look at indirect potable reuse (IDPR) and direct potable reuse (DPR) and energy neutrality, these opportunities are going to need funding, and the people that are funding these projects are going to want clear visibility to the operations, the savings and performance that we see. That’s a great emerging market that we see data helping,” he says.

The implication is that desal and reuse projects could in the future be contracted in a way that links digital performance data to returns.

Arguably bigger still, the network-level digital view of water will feed into utilities’ and potentially industrial players’ wider decisions about water resource management. “Our systems are distribution and collection networks,” explains Newcombe. “We have a role around water conservation. Utilities are building desal plants because source water is scarce. Or where you have reuse, you have to understand what you’re reusing. We can measure levels in aquifers and storage tanks. Pressure management, transient monitoring, leak detection. To us, water conservation is one of the big challenges that utilities are facing.”

The opportunity for desalination and reuse is to ensure that their value is understood in the context of this bigger conversation. Says Xylem’s Wiklund: “If you produce a certain amount of clean water from your treatment plant, and supply that to the consumers, to some extent there’s a correlation about how much clean water you produce and how much that will be sent backwards in the wastewater side. I don’t know exactly what kind of conclusions you will make, but it’s likely that there will be a correlation that you can utilise for something.”