Digital Twins: Paving the way for the Treatment Plant of the Future
The future of water is already here. Today, forward-thinking drinking water and wastewater operators are implementing digital solutions, many powered by digital twin technology, to drive transformation right across the water cycle – including wastewater networks, treatment plants and drinking water networks. In wastewater management specifically, harnessing the power of digital twin technology to continuously improve operational performance is critical to building a treatment process that’s efficient, cost-effective, and future-proofed.
Digital twin technology gives operators the system-level awareness, control, and guidance they need to make better, informed decisions, allowing them to identify and remove operational roadblocks, reduce risk, and increase system reliability. When applied to maximum effect, the technology not only optimizes treatment processes, but also minimizes energy consumption and empowers plant operators to meet water quality regulations at a lower cost.
As a concept, digital twin technology isn’t new – water and wastewater utilities have been using it as a tool for the best part of the last decade
Here, Brian Vu, Senior Practice Solutions Manager at Xylem, discusses how utility leaders are embracing data analytics and digital solutions to solve their most pressing challenges, and how digital twins are laying the foundation for treatment plants of the future.
Can you tell us about digital twin technology and how it has evolved?
As a concept, digital twin technology isn’t new – water and wastewater utilities have been using it as a tool for the best part of the last decade. However, traditional digital twins have been built using what are called “first principles”, and these models utilize physical, chemical, or biological equations to simulate the infrastructure, which is a long and expensive process. While efficient, decision support systems built based on first principle models are unable to learn from past experiences, unless someone calibrates their models.
However, when you introduce sophisticated tools like machine learning, the digital twin becomes ‘supercharged’ and supports a more seamless route to rapidly building a model with minimal human intervention. With artificial intelligence and computer science, the technology can utilize past data and automatically ‘calibrate’ to better represent infrastructure. As the data updates in real-time, it is immediately fed back into the system for continuous refinement and improvement.
The resulting real-time decision support system (RTDSS) is able to continuously learn and adapt, allowing operators to see what’s happening within their system at any given moment. The RTDSS can then generate multiple scenarios and provide operational recommendations, making it easier for operators to control their assets, processes, or systems. Ultimately, this advanced application of digital twin technology has the potential to deliver autonomous, optimized control.
How are utility operators applying digital twin technology in treatment plants?
Digital twins represent the single largest technological breakthrough we’ve seen in wastewater this lifetime
Now, more than ever, treatment plant managers are being asked to make proactive, informed operational decisions. Yet, many don’t have that insight or visibility into their operations previously mentioned, or the support of additional staff or funding. That’s where the digital twin can play a transformative role. The technologies and applications typically used in wastewater treatment have had little advancement over the years, but digital twins represent the single largest technological breakthrough we’ve seen in wastewater this lifetime – it’s really quite amazing.
Take Plant Real-Time Decision Support, for example. The solution leverages real-time data from a sensor network and applies this to process models, coupled with an optimizer, to arrive at operational recommendations and decision support that helps managers run their treatment plants more effectively. Having greater system insight allows them to change certain set points within the system to optimize the treatment process. That can be anything from monitoring energy and chemical usage, to adjusting influent changes and enhancing effluent quality.
Outside of driving operational efficiencies, one of the larger interests in digital twin technology is filling the knowledge gap. Globally, we’re facing a new type of transition where there’s a significant reduction in legacy operators. Ultimately, the goal of a treatment plant manager is to deliver clean water to communities, but that becomes more difficult to maintain as knowledge of operational processes diminish. Digital twin data help new operators pick up where their predecessors left off, so there’s no need to start from scratch when changes in the workforce happen.
What are some of the barriers to maximizing digital twin technology?
It really comes down to having the right system deployed. While digital twin technology can enable a whole new level of operational resiliency, it’s often viewed as a singular solution capable of solving problems on its own. That’s not the case – understanding where it fits within the digital ecosystem is incredibly important.
Many utilities have digital twins, but they don’t have the other tools combined to create true decision support systems. A lot of the challenges that plant operators come up against are the things they can’t see, and the digital twin certainly solves that. But that basic level of visibility still relies on the operator to take action. Even in slightly more advanced applications where the digital twin is capable of processing variables to predict an outcome, the operator is responsible for that manual optimization.
Many utilities have digital twins, but they don’t have the other tools combined to create true decision support systems
In slightly more advanced applications, the digital twin can generate multiple scenarios and provide operational recommendations in real time. It’s like a GPS system in your car – you plug it in, and it tells you where to go. It also tells you how long it’s going to take you to get there, right down to the minute. It’ll even tell you the fastest route based on current and historical data. Then, when combined with a RTDSS, it has the potential to deliver greater, more optimized control – almost like a self-driving car!
How do you see digital twin technology supporting the treatment plant of the future?
The journey to becoming a utility of the future starts with advanced solutions that help unlock your system’s capabilities – and that’s exactly what the digital twin does. However, we’re still lagging in terms of integrating machine learning and decision support in wastewater plant settings as a rule. Those who are deploying basic digital twins are certainly using it to their advantage, but many are one step short of reaching that level of autonomous control. If we can get to a place where we are utilizing artificial intelligence in treatment plants the same way we utilize satellite navigation systems on the road, it’ll be revolutionary.
What can we learn from the pacesetters who are deploying this technology to deliver transformative results for their communities?
While uptake has been slow, many forward-thinking treatment utilities around the world are already using treatment system optimization, powered by digital twin technology
While uptake has been slow, many forward-thinking treatment utilities around the world are already using treatment system optimization, powered by digital twin technology, to run their operations more effectively, safely and at a reduced cost. The solution provides real-time analysis of operational inefficiencies and decision-making support to help managers monitor energy and chemicals, adjust to influent changes, deliver effluent quality, and predict future conditions. Combined, these powerful outcomes also reduce response times, allowing plant managers to develop more efficient operational processes.
For example, when EWE WASSER GmbH (EWE) in Germany leveraged digital twin technology for process optimization, the plant achieved a 30% reduction in aeration energy use which equates to 1.2 million kWh saved annually. In partnership with Xylem, EWE deployed Plant Real-Time Decision Support which used machine learning to create models of the carbon, nitrogen and phosphorous elimination processes based on data from the plant’s SCADA system. Several “virtual sensors” were developed to calculate an estimate of the incoming carbon, nitrogen, and phosphorous loads of the influent.
As a result, each process receives optimal aeration and chemical inputs to match the chemical and biological oxygen demand. It’s a new level of control that has helped the plant accurately estimate influent concentration and optimize the aeration process while still meeting regulatory requirements. Not only has EWE improved overall treatment processes at the plant, but they have done so in a way that creates a positive environmental impact for the community they serve.