In the keynote, Nicholas Cumins, CEO of Bentley Systems, stated that 95 % of the infrastructure intelligence needed by 2030 already exists. To what extent do you think this applies to the water sector?
Robert Mankoswki: I don’t know the actual statistics really, what I suspect is that at least 95 % − if not more − of the infrastructure we will need by 2030 already exists today. Of course, there is new water infrastructure being built − for example transmission, distribution, treatment − but there’s already a lot that exists. It also definitely varies from place to place as well. So, if we say 95 % already exists, that’s not uniformly distributed in the world − there are places which are obviously much more developed than others.
In the finalist-presentations we saw all kinds of technologies applied in the projects, and to what extent the expansion of infrastructure already influenced so many new innovations. How do those technologies help extending infrastructure projects?
Above other things mentioned during the keynote was the high demand. There is no shortage of demand for infrastructure and infrastructure projects, but there is a shortage of engineers, and there is this amazing amount of data collected.
While the percentage of data insights generated is quite good, there are still not enough actionable insights being produced. The technologies showcased in the finalist-presentations, particularly those that effectively integrate data platforms for structural and hydraulic analysis, play a crucial role in bridging the gap between our needs and current capabilities. Without these technologies, our ability to plan, design, build, and operate infrastructure would be significantly limited. Essentially, these technologies serve as enablers for engineers in their work. Without this technology, we’d be able to do much less, because essentially these technologies, they’re enablers for engineers to plan, design, build and operate the infrastructure.
By that I mean that we often talk about smart water systems, or data-driven decision making, and it all starts with the question: what are the decisions we are trying to make? Understanding what the challenges are that the water utilities or those specific communities are facing, understanding what problems they are trying to solve is the first step. And then define, how can we use the data that we’re collecting? How can we collect the right data to help make better decisions about how to build the necessary infrastructure? How to design the infrastructure more sustainably and make it more resilient. And how to do it with fewer engineers, because we just don’t have them.
I’m glad you brought that up. Could those technologies replace the missing engineers/work force?
I don’t think so. What we’re seeing is the ability of AI to enable engineers to focus their energies on the more challenging, difficult, creative sort of problems and which solutions they have for these problems. What AI can do is help in a number of ways: in the open site; or in the example we saw during the keynote, by automating some of the repetitive and mundane tasks, like placing little text labels on a drawing. Today, engineers, draft people and other professionals spend an inordinate amount of time placing text on drawings, because drawings are a great way to communicate design intent. However, they must be creative and a lot of it is very manual, tedious and really not high value work. So, if we can get the machine to automatically do that, then the engineer can spend more time on the actual design of the project, looking for more cost-effective ways to do it, looking for more environmentally alternatives, or to meet other objectives or do more projects. Because there’s a project backlog. In the end, it’s about how can we speed up the execution of any single project so that they can do more projects, and we can burn down that backlog of the projects faster.
So, it’s not about replacing the engineers, it’s about making up the gaps between the engineers we need and the engineers we have.
We often talk about smart water systems, or data-driven decision making, and it all starts with the question: what are the decisions we are trying to make?
In your opinion, what role does the digital transformation strategies we just talked about − and also IoT and data analytics − play in optimizing life, specifically in the water and wastewater systems?
If I‘m not mistaken, SCADA has been around for decades now, so water systems, I think, are pretty well. We don’t call it IoT necessarily, but it is about data acquisition and supervisory control. And that data acquisition part, that’s where IoT is today − it is about different ways of collecting data from operating assets. And this information can be used in a variety of ways throughout the whole life cycle of water infrastructure. The information can be used very clearly.
I also think that what comes to mind for most people is the operation’s phase of the infrastructure, i.e. what are the flows and pressures? What are the tank levels, the pump speeds, valve settings and so forth. All that data that’s being brought in helps to get a situational awareness and a better understanding of how the system is operating.
What we’re seeing with AI in this space is a couple of things: One is analysing the data coming in. Sensors are not perfect, as they sometimes produce incorrect readings. For example, let’s say there is a stream of data available on the pressure at a location in the system, and sometimes there’ll be a spike or a real drop in the pressure reading, but it’s just an erroneous reading from the sensor. AI has been proven to be really good at detecting those spikes and cleaning the data right, as if it would basically say “oh, this, this data doesn’t look right, it’s not the right data so let’s eliminate that”. The other thing is, sometimes the sensors go offline, or they get a stuck reading, and so AI can detect that as well and say “oh, well, let’s fill in the gap for this sensor”.
And why is that important?
Because that data can then be used in analysis. And it’s sort of a classic garbage in garbage out scenario. If you have sensor data which has bad values in it, and you feed that into analysis, then you’re going to get bad analysis results. Once you have a clean set of data, operationally, you can start doing analysis with it. And that makes it possible to detect real problems in the system.
And so that is one way in which we are seeing AI used in water systems today, just to name some real-world examples of when you have clean IoT- or SCADA data, then you can use it in analysis and help improve the operations of the system. Another thing that is important for water and wastewater is estimating the demands on the system, for example: how much water is going to be used in the next twelve hours or tomorrow? And then using that prediction to help inform your operations now. That’s another example where AI is has been proven very useful, by both predicting those near term demand values and also by taking into account the variation in demand that happens during the weekdays, during the weekend, during season changes, etc. All those things make it difficult to make a prediction, but AI has proven to be very effective at doing that.
What other possibilities are there to use AI?
We can see IoT data being used during the design process as well, and during the planning phases. If you have a better understanding of how the system is operated, you can do a better job of planning and understanding where it has its current deficiencies. Or when planning for city expansions or population growth, understanding then how to better design and plan the expansion taking into account how it’s operating now − that’s a point we often forget that we can use AI. Today, when we talk about AI, we talk a lot about large language models, like ChatGPT or Gemini.
In design process or planning, I think people forget about another type of AI, which is sort of an older one, I guess, that is genetic algorithms. And evolutionary computing is also a form of computational intelligence, and we’ve been sort of using those techniques in water applications for a long time. Artificial Intelligence can be very useful for adjusting the parameters of the model to be able to simulate “real world”-conditions, for designing systems, or for figuring out the optimal sort of trade-offs between cost and performance. It may help sorting out multi-objective optimization problems, and the scheduling of materials/activities to minimize energy usage, or to minimize carbon footprint − that kind of thing.
The water industry, obviously, is one of the key enablers of the quality of life that we come to expect. We need to invest in it, but there is still a gap to fill in technology and digital twins are one way to fill it.
How do you evaluate the current challenges in infrastructure, particularly concerning climate change and urban densification?
What we are seeing is obviously more intense and frequent storms and it’s a real challenge because how do you design or plan for resilient water infrastructure under these conditions? It is a really challenging problem, because it’s really a system of systems, so to speak, and not only the water systems. Let’s say, your city water utility relies on pumping. If those pumps are in an area which may be flooded, then how do you plan for resilience? If that pump station goes offline, how can you still supply water? Or if the electricity of a substation, for example, is flooded and the substation goes offline, do you have alternative power supply for your pumping stations? Water systems are therefore challenged “from everywhere”.
The issue of urban densification is another problem we face for many, many years: How to plan for the increased water usage and wastewater discharge of a higher urban population? How to deliver safe, clean drinking water when your population is really exploding is certainly a problem that utilities are facing, and again, this aggravated by the shortage of engineers, making it necessary to turn to technology to help fill that gap and to “make” engineers more efficient and able to do more work.
In your opinion, what measures should be taken to increase public acceptance and awareness of the importance of water infrastructure?
That’s a tough one. I think we all take water for granted too much, especially in the Global North. But I honestly don’t know exactly how to answer your question of what we should do to shine a light on the water industry. What I do know is that this is crucial. The water industry, obviously, is one of the key enablers of the quality of life that we come to expect. We need to invest in it, but there is still a gap to fill in technology and digital twins are one way to fill it.
The interview was conducted by Charlotte Quick