Matching a real-world asset with a simulation does not mean there is a digital twin.
There is no shortage of online articles comparing simulations and digital twins. In basic terms, they all come to the same conclusion: a simulation is a digital model or representation of a process, phenomena, system or object. Meanwhile, a digital twin is born once that simulation has a corresponding real-world equivalent—aka its physical twin. Though many explanations end there, the truth is that this should be the start of the conversation.
Think about it. By the definition above, every simulation of a product on the market is a digital twin—even if it sits unused since the design was locked in. Is a simulation a digital twin if it doesn’t benefit the real-world asset? By extension, does a real-world asset need to service a simulation to become a physical twin?
At Future.Industry 2022, Altair CEO James Scapa paraphrased famed simulation expert and director of University of Texas at Austin’s Oden Institute, Karen Willcox, by defining a digital twin as “a personalized, dynamically evolving model of a physical system … grounded in physics, assimilating data from inspections, sensors, etc.”
So, not only does a simulation need a physical system to become a digital twin, but it also needs to be bound by physics, able to ingest information directly from the physical asset and offer value throughout the system’s lifecycle. In other words, the relationship between a physical and digital twin is symbiotic in nature—both benefit from the existence of each other. That relationship, itself, deserves further study.
For instance, Cynthia Dean, Director of Model Based Systems Engineering (MBSE) Training and Solution Development at Belcan said, at Future.Industry 2022, that “physical twins are more likely to blow up, and we [can] learn from that.” But it’s a heavy cost. Dean argues that as an amalgamation of numerous other models and simulations, digital twins represent a way to safely interrogate a system from various perspectives without risking the physical asset.
“We can ask twins questions about each other,” she said. And by asking these questions, the relationship between the twins gives engineers a better idea of how the system truly works.
The Symbiotic Nature of Physical and Digital Twins
Scapa notes that an example showcasing the symbiotic nature between physical and digital twins is well known to the public thanks to NASA’s Apollo 13 mission and its subsequent film. Scapa said, “[the concept of digital twins] is not this new idea that came last year. T. K. Mattingly, the very famous astronaut from Apollo 13 … had to stay back and was replaced by a different astronaut. But he played a huge role. He was described as a physical twin of the astronauts when he used a mock-up of the lunar module.”
To clarify, Apollo 13 had technical issues in space that risked the lives of its crew. Mattingly and others performed tests, back on terra firma, to help get the crew home safely. These tests were performed within the same simulators of the Lunar and Command Modules that trained the astronauts on Apollo 13. By using this strategy, NASA was able to determine a workflow that could be used to conserve enough energy to get the crew home while also ensuring all systems would run safely during reentry.
In this specific example, it could be argued that the simulator wasn’t a true digital twin—as it, too, was a physical mockup. But since the responses and actions of the simulator were run by NASA’s computers, formulas and programs, an argument can also be made that this scenario created one of the first digital and physical twin relationships with humans-in-the-loop.
Either way, the twin on Earth was used to simulate various workflows that could save the lives of those in space. If the Earth-bound system detected a failure during a workflow, there would be no risk to the astronauts. Meanwhile, data from the twin traveling through space could be plugged into the simulator to help NASA know what to expect.
“We can learn so much more from digital models than static documentation,” said Dean. “The models provide valuable feedback loops, promote higher quality decision making and increase cross functional communication that creates stronger trust networks.”
Royston Jones, Executive VP of European Operations and Global CTO of Altair, said it differently in his presentation. Basically, he states that physical twins have the advantage of being here with us. Thus, they can be used to monitor what is going on.
The advantages of digital twins are that they can be developed, tested, modified, optimized and reorganized quickly to assess various scenarios safely and affordably. Digital twins also make it easier to reverse engineer what went wrong, as engineers can dig into the substantial numerical results from the simulations. “They [digital twins] are also time travelers,” joked Jones. For digital twins, time can be sped up, reversed or jumped ahead to give engineers a more in-depth image of what could go on.
How Simulations Become Digital Twins
By deeply defining the relationship between physical and digital twins, it begs the question: how can simulations become digital twins? For that matter, how can real-world assets become physical twins?
Dean suggests that model-based systems engineering (MBSE) can act as a good way to glue simulations together. She said, “the best MBSE methodology provides a framework for organizing, managing and improving decision making throughout complex engineering lifecycles.”
Engineers can add many analytical tools to the digital twin, through the MBSE model, including 1D simulations, 3D simulations, analytical models, reduced order models, surrogate models and more. Additionally, this multi-modeling, MBSE approach can lead to the simplification of multidimensional optimizations and analyses.
“Modern systems engineers use digital systems engineering models to optimize collaborative, integrative, scalable and responsible outcomes. But that also reduces costs,” said Dean. “A cross-functional, model-based approach provides more control for selecting optimized systems and performance.”
As for how the real-world asset becomes a physical twin? These days, much of that functionality is implemented using IoT sensors, communication and control systems. The sensors are used to detect what the asset is experiencing, and then IoT systems communicate that information to the digital twin.
By gathering data from the real world via IoT sensors, physical twins can inform the digital twin about the environments and scenarios the asset faces in the real-world. This data can then be used by the digital twin to help suggest improvements for future products or even inform the control systems in the real-world model how to best deal with its current situation—much like in the Apollo 13 example.
In short, the relationship between digital and physical twins represents the ultimate manifestation of continuous validation, optimization, efficiency and cost savings.
