Skip to main content

[QTalks Ep.3]

The Digital Twin Journey: From Pioneers to Mass Adoption

Digital Twins are the talk of the town — but how are they evolving? From the early adopters to the potential for mass adoption, how can the latest solutions help water companies bridge the digital chasm?

By offering a virtual mirror of the physical world, digital twins are enabling water professionals to test the impact of changes before they are made.

Join this latest QTalks with environmental journalist Tom Freyberg to hear from digital twin experts as they cut through the hype with examples of digital twin adoption, lessons learned, and insights into how far this digital aspiration can be pushed. 

Our host Tom Freyberg was joined by three thought leaders:

The historic pioneers of digital twins

Tom began by asking Biju to reflect on how the digital twin journey started at DC Water. Biju said that the true pioneers of the digital twin were those who developed the physical model for the water distribution and water collection systems, flooding models, and treatment models as this enabled them to plan and design the facilities. The use of those models combined with real-time data then helped them make operational decisions. 

Biju also said that digital twins have been implemented throughout the whole water treatment cycle, from the drinking water network right through to wastewater treatment plants. 

Tom then commented on how Valencia’s digital twin is an oft-cited case study of the typical digital twin journey and asked Pilar to recap the path they took. She said that Valencia’s digital twin journey has been long and is the result of the company’s digitalization process that began 15 years ago, and that they began the journey so early that they had to develop the hydraulic model from scratch. 

In terms of lessons learned, Pilar said that ensuring the quality of data is crucial for anyone hoping to achieve the realization of a reliable digital twin.

Operational cost reductions with digital twins

Tom then asked Gigi about who else she considers digital twin pioneers — aside from DC Water and Valencia’s digital twin. Gigi said that cloud computing in and of itself is a true game changer in terms of digital twins. She said that more and more utilities are realizing the benefits of cloud computing in terms of cost-efficiency and only paying for what they use. 

Reflecting on why this is such a game changer, Gigi said this changes business models and allows vendors to not only sell things that are capital-heavy that only larger utilities can afford but to sell software and data as a service on a monthly subscription which is highly cost-effective for smaller utilities. This, she said, allows utilities to view it as an operational expenditure rather than a large, upfront capital cost. 

Gigi also commented on how cloud computing allows utilities to expand the use of CPU resources which allows them to run more complex simulations that involve a lot more resources than they would have on on-premise servers. 

Lessons learned

Tom went on to ask Biju how the availability of cloud computing and software as a service would have changed the way things were done at DC Water. 

Biju said this availability would have meant more and better-facilitated choices in the creation of services for utilities. He said, for example, a utility with one water source needs to be alert to any pollutants in the water, and so a live flow model that is accessible to all that use the water source would have helped them make quicker decisions. 

Posing Pilar the same question, she said that the availability would have shortened processes as a whole. She also said that their digital twin is evolving all the time.

The cost of digital twins

The panel received a question about how expensive it is for a digital twin to achieve the requirement of a calibrated model.

Gigi began by referencing a recent experiment with Lakewood, California which had issues with calibrating its model. She said that they provided the existing EPAnet model of their network and two months of SCADA data to Qatium, and within a few weeks they were able to bring up the model on the Qatium platform. Gigi said that this is a great example of how quickly you can calibrate a model using inexpensive tools like Qatium. 

Biju went on to say that what’s making a difference from a digital twin perspective is real-time sensors. He said that the level of accuracy they provide lowers costs since your model becomes a live model as opposed to that something that’s only used once in a while. Therefore, real-time calibrations become part of everyday operations.  

He also said that he doesn’t predict that costs will increase since models are becoming better, and real-time systems allow for real-time decisions and quicker responses to issues.

How machine learning is changing how digital twins are created

The panel was asked another question about the impact of machine learning on how digital twins are created. 

Pilar began by saying that machine learning is an important part of a digital twin and that we have to take advantage of machine learning algorithms. She said that combining the information provided by AI algorithms with the potential of hydraulic models enables them to extract much more complete information. She said that machine learning enables them to develop patterns, optimal operations, and data cleaning. 

Biju said that machine learning is the tool that is enabling the need for fewer sensors since you can train your systems to fill in the gaps. He also said that machine learning is enabling the modeling of situations where there has not been much intervention previously. 

Rounding off the discussion, Tom asked the panel whether the traditional path of achieving a digital twin needs disrupting. Pilar said that if there are new ways to shorten the process then these should be considered, while Gigi raised the point that we are already in a phase of disruption in terms of AI and cloud computing.

Ready to discover more QTalks content?

Visit Qatium’s YouTube Channel to watch this QTalks episode and all previous ones.