How 'Digital Twins' Are Transforming Manufacturing, Medicine and More (time.com) 59
Time reports on virtual doppelgangers — also known as "digital twins". (Alternate URL here.)
Created by feeding video, images, blueprints or other data into advanced 3-D mapping software, digital twins are being used in medicine to replicate and study internal organs. They've propelled engineers to devise car and plane prototypes — including Air Force fighter jets — more quickly. They allow architects and urban planners to envision and then build skyscrapers and city blocks with clarity and precision. And this year, digital twins began to break into the mainstream of manufacturing and research. In April, chipmaker Nvidia launched a version of its Omniverse 3-D simulation engine that allows businesses to build 3-D renderings of their own — including digital twins. Amazon Web Services announced a competing service, the IoT TwinMaker, in November...
The need was always there. In the 1960s, NASA created physical replicas of spaceships and connected them to simulators so that if a crisis ensued on the actual vehicle hundreds of thousands of miles away, a team could workshop solutions on the ground. Dave Rhodes, the senior vice president of digital twins at Unity Technologies, a video-game and 3-D-platform company, says that digital-twin technology is only now being widely released because of several confluent factors, including the increased computing power of cloud-based systems, the spread of 5G networks, improvements in 3-D rendering and the remote work demands of COVID-19....
Digital-twin technology is being trialed across the medical landscape, for planning surgical procedures and exploring the heart risks of various drugs... Digital twins are also being used in other complex and potentially dangerous machines, from nuclear reactors in Idaho to wind turbines in Paris.... Digital twin humans are coming too: the NFL and Amazon Web Services have created a "digital athlete" that will run infinite scenarios to better understand and treat football injuries.... BMW could soon implement digital twins at all facilities.
Frank Bachmann, the plant director in Regensburg, says that the advantages of digital twins will only be fully realized when every factory is digitized in a standard way. "We need these processes of digital twins everywhere," he says.
The technology "raises questions about privacy and cybersecurity," Time warns. "Many of these digital twins are made possible by a multitude of sensors that track real-world data and movement.
"Workers at factories with digital twins may find their every movement followed; the hacker of a digital twin could gain frighteningly precise knowledge about a complex proprietary system."
The need was always there. In the 1960s, NASA created physical replicas of spaceships and connected them to simulators so that if a crisis ensued on the actual vehicle hundreds of thousands of miles away, a team could workshop solutions on the ground. Dave Rhodes, the senior vice president of digital twins at Unity Technologies, a video-game and 3-D-platform company, says that digital-twin technology is only now being widely released because of several confluent factors, including the increased computing power of cloud-based systems, the spread of 5G networks, improvements in 3-D rendering and the remote work demands of COVID-19....
Digital-twin technology is being trialed across the medical landscape, for planning surgical procedures and exploring the heart risks of various drugs... Digital twins are also being used in other complex and potentially dangerous machines, from nuclear reactors in Idaho to wind turbines in Paris.... Digital twin humans are coming too: the NFL and Amazon Web Services have created a "digital athlete" that will run infinite scenarios to better understand and treat football injuries.... BMW could soon implement digital twins at all facilities.
Frank Bachmann, the plant director in Regensburg, says that the advantages of digital twins will only be fully realized when every factory is digitized in a standard way. "We need these processes of digital twins everywhere," he says.
The technology "raises questions about privacy and cybersecurity," Time warns. "Many of these digital twins are made possible by a multitude of sensors that track real-world data and movement.
"Workers at factories with digital twins may find their every movement followed; the hacker of a digital twin could gain frighteningly precise knowledge about a complex proprietary system."
How is this new? (Score:5, Insightful)
We already have a perfectly good word for that: simulation.
Not cool enough these days?
Re:How is this new? (Score:5, Insightful)
"You've been pwned!" (Re:How is this new?) (Score:2)
They'll get hacked just the same, and probably the hackers will invent their own special word that is even cooler.
Isn't "pwned" good enough? We need something else?
Can someone clarify the use of "pwn" for me? If someone getting hacked is done by a hacker then what do we call those that have pwned someone? A pwner? What do we call the target of the pwn? Pwnee?
Re: How is this new? (Score:2)
And we will further destablize an already mentally unstable population.
"Close the mental hospitals. The community will take care of them."
Yeah, because of this, the clerk at the local Kwik-E-Mart has to take on the role of a mental ward orderlie as well as selling Slushees, wothout additional training or pay. There are plenty of videos on Youtube showing this in action if you really need proof.
Rant aside, and stupid buzzword aside, this is an awesome technology, and something you would see i
Re: (Score:2)
some people just like to see a digital twin in the metaverse
TFA: chipmaker Nvidia launched a version of its Omniverse 3-D simulation engine
So which is better? A metaverse or an omniverse?
:-/ confused
On the one hand omni includes the whole ~verse, but on the other, meta is above or beyond the ~verse everything.
Re: (Score:2)
We already have a perfectly good word for that: simulation.
Not cool enough these days?
Apparently. Note the fake profoundness implied by the use of the word "twin". Probably a metaverse reference or something equally stupid.
Re: (Score:2)
My take on it is that it's a crude translation of virtual doppelganger.
Re:How is this new? (Score:5, Interesting)
They're complementary, but not identical. A simulation aims to give you a realistic source of synthetic data that mimics the thing under simulation to the best of its abilities. In most cases this is an approximation under ideal conditions, and the physics models remain very much a best-effort. In the case of a digital twin, rather than requiring the simulator to come up with the readings and predict how something is behaving, real data is read in and used to update state instead.
This can be great for a lot of different things, like validating the parameters of the simulation model, exposing a general API that can support concurrent read-side access that real instrumentation can't keep up with or that you don't want to directly expose, allowing the development of AI/ML models using synthetic data while using twinned data for training some additional convolutional layers in CNN/DNN models, etc.
Overall it's a great technology where you want to maintain a layer of separation between the thing/device under monitoring and the various data-driven applications you want to build around it.
Re: (Score:2)
That is not to say that using real world da
Re: (Score:2)
Re: (Score:2)
I understand it as monitoring to maintain replica model. While simulation is imitation, not necessary synced to the reality object.
Good explanation of "Digital Twin" (Score:2)
What is Digital Twin Technology and How Does it Work? [twi-global.com]
Quoting:
"The developers who create digital twins ensure that the virtual computer model can receive feedback from sensors that gather data from the real world version. This lets the digital version mimic and simulate what is happening with the original version in real time, creating opportunities to gather insights into performance and any potential problems."
Re: (Score:2)
I simply can not believe that updating a simulation with real-world data is new or novel enough to warrant a new term.
Can someone with experience making or using simulations chime in?
Re: (Score:3)
That's not a twin. That's monitoring. And a digital representation. It's a model, not a copy.
What a fucking stupid name.
FAR more complete monitoring (Score:2)
Eventually we will be able to have a digital understanding of everything that happens in our bodies. For example, why did someone eat what they ate? How did eating that affect the person's brain and body?
"Digital Twin" is an attempt to communicate the completeness of the computer model. The name is not perfect, and, in the short time I looked, I wasn't able to find a better explanation.
Re: (Score:2)
At the point where we know the whole thing, sure. We're not anywhere near there. We've found whole systems in the body in the last decade!
Re: (Score:2)
Eventually we will be able to have a digital understanding of everything that happens in our bodies.
What would make you think that? It seems impossible to me, given that we don't live in a deterministic universe.
Re: (Score:2)
Eventually we will be able to have a digital understanding of everything that happens in our bodies.
It's still a fucking simulation.
Eventually we will be able to time travel in a DeLorean,
powered by ground-up leprechauns.
Re: (Score:2)
Also, make sure it's in your resume. Don't want to miss out on that keyword!
The problem is verifying the simulation accuracy (Score:5, Insightful)
Re: (Score:2)
Indeed. Simulation is a valuable tool for people that really understand the simulation and the real thing being simulated very well. For all others it makes the situation worse, because they will mistake the simulation for the real thing and make critical errors. This is decidedly an "experts only" thing. Of course it has been and will be misused to get jobs done by cheaper engineers (or even non-engineers), with the predictable outcomes.
Re: (Score:3, Informative)
Simulations in general, and digital twins in particular are great tools, but verifying that the simulation is accurate in all of the conditions under which you are using it can bey very tricky.
I think you misunderstood what digital twins are
Simulations create data that is supposed to mimic real data. Yes, their inaccuracy is a problem, which makes the data from them less like reality.
Digital twins don't create data. It takes real life data to use instead, which by definition is 100% accurate.
A simulation of road traffic conditions would attempt to show how vehicles react given a pre-programmed set of rules those virtual drivers follow.
A digital twin of road traffic conditions would use real wor
Re: (Score:2)
Digital twin model will be aware of a glaze on the surface of the road, or will detect/predict presence of the fog, display condition warning before bad stuff even happens. Traffic monitoring and control is already a given.
Re: (Score:2)
> Digital twins don't create data. It takes real life data to use instead, which by definition is 100% accurate.
This is absolutely not the case. Not just because it's not part of some definition.
No measurement is ever 100% accurate, due to physics, engineering, economy and all kinds of errors. And even if it were very accurate, measurements become outdated. Frequent or continuous measurements are often costlier, so there may be creep over time. Digital value representations are also subject to precision
Re: (Score:3)
You're not only forgetting that the monitored sensors are a less-than-perfect representation of the real-world conditions, but that the "digital twin" model is only an approximation of the real-world behav
Re: (Score:2)
It then uses that logged data to give a 100% accurate dataset
for what non-virtual drivers actually did.
Only if the underlying model is 100% accurate.
Which is impossible, since no complex system is
100% understood.
Re: (Score:2)
That's the whole point of the digital twin. If you can produce an accurate simulation based on first principles then everything would be fine anyway. Digital Twins exist to feed real world data to the simulation. It's a verification of a simulation solution.
It also works really well. We used a high performance simulator to design a new injection system at a plant based on first principles and got a nice theoreitcally perfect solution.
Threw that in to the digital twin and ... fail. Turns out the real world d
Re: (Score:2)
Re: (Score:2)
Not sure how it helps if you are only able to verify it for data you already have with real hardware.
That's the whole point of a digital twin. You take all you knowledge of your real hardware and use it to verify your simulation. If you don't have real hardware already with actual recorded operating data then you're not using a digital twin but likely a simulation on theoretical principles. In your aircraft example you're going outside of recorded operating parameters, the digital twin there can only help if you can extrapolate existing data to your scenario.
Example: Suppose we have a vessel with a valve c
Re: (Score:2)
Re: (Score:2)
Yes that is the basis for it, but the goal is ultimately to use this regression analysis make a high resolution simulation of your plant. These things are largely used to identify how changes we make can affect things. As per my original post, we have simulators, but they rely on some ideal formulas and documentation (theory of how the plant works). Our project passed the simulation but failed to work on the digital twin because the data feeding the digital twin model correctly identified that it was not po
Tell this to GE Predix (Score:1)
Article text. (Score:1)
How Digital Twins Are Transforming Manufacturing, Medicine and More
Illustration by Shout for TIME
BY ANDREW R. CHOW
DECEMBER 30, 2021, 4:29 AM EST
There are two versions of a BMW factory in the medieval town of Regensburg, Germany. One is a physical plant that cranks out hundreds of thousands of cars a year. The other is a virtual 3-D replica, accessed by screen or VR headset, in which every surface and every bit of machinery looks exactly the same as in real life. Soon, whatever is happening in the physical factory will be reflected inside the virtual one in real time: frames being dipped in paint; doors being sealed onto hinges; avatars of workers carrying machinery to its next destination.
The latter factory is an example of a “digital twin”: an exact digital re-creation of an object or environment. The concept might at first seem like sci-fi babble or even a frivolous experiment: Why would you spend time and resources to create a digital version of something that already exists in the real world?
But digital twins are now proving invaluable across multiple industries, especially those that involve costly or scarce physical objects. Created by feeding video, images, blueprints or other data into advanced 3-D mapping software, digital twins are being used in medicine to replicate and study internal organs. They’ve propelled engineers to devise car and plane prototypes—including Air Force fighter jets—more quickly. They allow architects and urban planners to envision and then build skyscrapers and city blocks with clarity and precision.
And this year, digital twins began to break into the mainstream of manufacturing and research. In April, chipmaker Nvidia launched a version of its Omniverse 3-D simulation engine that allows businesses to build 3-D renderings of their own—including digital twins. Amazon Web Services announced a competing service, the IoT TwinMaker, in November. The digital-twin market already generated sales of more than $3 billion in 2020, according to the research firm Research and Markets, and tech executives leading digital-twin efforts say we’re still at the dawn of this technology.
Digital twins could have huge implications for training workers, for formulating complicated technical plans without having to waste physical resources—even for improving infrastructure and combatting climate change. “Health care, music, education, taking city kids on safari: it’s hard to imagine where digital twins won’t have an impact,” says Richard Kerris, Nvidia’s vice president of Omniverse development.
An image from the digital twin of BMW's factory in Regensburg, Bavaria, created in NVIDIA's Omniverse BMW
The need was always there. In the 1960s, NASA created physical replicas of spaceships and connected them to simulators so that if a crisis ensued on the actual vehicle hundreds of thousands of miles away, a team could workshop solutions on the ground. Dave Rhodes, the senior vice president of digital twins at Unity Technologies, a videogame and 3-D-platform company, says that digital-twin technology is only now being widely released because of several confluent factors, including the increased computing power of cloud-based systems, the spread of 5G networks, improvements in 3-D rendering and the remote work demands of COVID-19.
Digital twins can replicate real-world objects ranging in size from millimeters to miles. In Poland, a team of doctors and technologists is starting with one of the smallest objects imaginable: the human fetal heart. About 1 in 100 newborns has a congenital heart disease, which can be fatal if not treated. But studies have shown that more than half of those diseases go undetected. Sonography simulators are expensive and bulky, and most medical schools don’t include hands-on training. “Most likely you will encounter a congenital heart defect for the first time when you are already in your clinic,” Marcin Wiechec, an ob-gyn and associate professor at Jagiellonian University in Krakow, says.
So Wiechec and his team created Fetal Heart VR, which allows doctors to guide a probe across a belly-like dome in order to study normal and abnormal beating fetal hearts—re-created identically from real-life scans—through a VR headset. Wiechec has been using the app to train students in Krakow. Jill Beithon, a retired sonographer and educator based in Fergus Falls, Minn., says the app could have enormous benefits for medical workers in areas with less access to resources or cutting-edge education centers. “The fetal heart is very intimidating—and the size of a quarter at 20 weeks, beating at 100 beats per minute. It takes additional training, and it’s not training you can easily find,” she says. “With the VR, you don’t have to go to expensive courses or try to find a mentor. This is going to replace hands-on experience.”
A promo image for Fetal Heart VR. Courtesy Marcin Wiechec
Digital-twin technology is being trialed across the medical landscape, for planning surgical procedures and exploring the heart risks of various drugs. In November, seven medical researchers from around the U.S., writing in the journal Nature Medicine, called for an increase in clinical studies of “cancer patient digital twins” to precisely track a patient’s physical state and adjust treatment accordingly. “[Digital twins] are poised to revolutionize how cancer and a host of other complex diseases are treated and managed,” the researchers wrote.
The automotive industry is also being transformed. Back in Regensburg, BMW can now test or tweak parts of the assembly line without having to move around heavy machinery; the company estimates the technology will cut the time it takes to plan out factory operations by at least 25%. A few months ago, factory managers created their first piece of new equipment inside the digital twin: a machine that puts door seals on a car frame. “Under the old-fashioned way, we would have had to draw it and build cardboard simulations, which is very time consuming. With COVID-19, we were very restricted in getting people on site,” says Frank Bachmann, the plant manager. “So with the digital twin we were able to work virtually, test it and have variations of the plan for more or less no money.” The machine was installed over the holidays—but before the break, Bachmann was able to show employees how to navigate their new workstations via the digital twin.
In Pittsburgh, Ding Zhao, an assistant professor of mechanical engineering at Carnegie Mellon, has been working with carmakers to use digital twins to improve the safety of self-driving vehicles. In his lab, he leverages vast quantities of data collected from real tests of self-driving cars to build complex digital-twin simulators. The simulations, he says, help predict how a car’s AI will react in dicey situations that could be dangerous and difficult to re-create IRL: when merging onto a dark snowy highway, for instance, or when jammed in between two trucks.
Crucially, digital twins also allow researchers to run crash-test simulations countless times without having to destroy cars or endanger real people. That means digital-twin technology is becoming essential to the development of self-driving cars. “Real-world testing is too expensive and sometimes not even effective,” Zhao says. Digital twins are also being used in other complex and potentially dangerous machines, from nuclear reactors in Idaho to wind turbines in Paris.
Others are deploying the technology at an even larger scale, to create digital twins of entire cities or even countries. This year, the Orlando Economic Partnership, a nonprofit community-development organization, announced it was partnering with Unity to build a digital twin of 40 sq. mi. of the Florida city. CEO Tim Giuliani hopes that the twin will eventually be used as a public resource and “backbone infrastructure,” allowing transportation experts to see how a rail system might impact the region, for utility companies to map out 5G networks and for ecologists to study the potential impacts of climate change. He estimates the project will cost $1 million to $2 million.
Of course, creating digital replicas at increasingly large scale raises questions about privacy and cybersecurity. Many of these digital twins are made possible by a multitude of sensors that track real-world data and movement. Workers at factories with digital twins may find their every movement followed; the hacker of a digital twin could gain frighteningly precise knowledge about a complex proprietary system. Zhao, at Carnegie Mellon, stresses the need for regulations. “Legislators and companies need to work together on this. You cannot just say, ‘I am a good company, I will never do evil things, just give me your data,’” he says.
Before the regulations arrive, technology companies are rolling full steam ahead. Many of them believe digital twins will gain importance with the rise of the metaverse, a collection of connected virtual worlds that increasingly impact—or even replace—what happens in the real world. Digital twin humans are coming too: the NFL and Amazon Web Services have created a “digital athlete” that will run infinite scenarios to better understand and treat football injuries.
BMW plans to take its digital-twin factory model to the world. The company is in the process of building a new plant in Hungary that is modeled completely in Nvidia’s Omniverse. But BMW could soon implement digital twins at all facilities. Bachmann, the plant director in Regensburg, says that the advantages of digital twins will only be fully realized when every factory is digitized in a standard way. “We need these processes of digital twins everywhere,” he says.
Marketing to the rescue! (Score:2)
Computer model, 3d model, virtual model, etc. have all lost their pizzazz. Time for a new name for the same old thing!
That is called "simulation" and it is old (Score:3)
Giving a fancy new name implying fake meaning does not make something "new" or "revolutionary". It just makes the ones doing it look stupid.
Re: (Score:3)
Giving a fancy new name implying fake meaning does not make something "new" or "revolutionary". It just makes the ones doing it look stupid.
There's a lot of that going around in tech nowadays.
Re: (Score:2)
Giving a fancy new name implying fake meaning does not make something "new" or "revolutionary". It just makes the ones doing it look stupid.
There's a lot of that going around in tech nowadays.
Indeed, there is. People trying to create the impression of "groundbreaking" for rather small incremental changes.
Re: (Score:2)
I work in academia (staff, not faculty), and I have to tell you... I sometimes roll my eyes at the number of times some iterative research gets publicized as "disruptive".
It's often still very cool research, don't get me wrong. But it ain't disruptive by any stretch of the imagination.
"Steak" is called "food" (Score:2)
Just like not all food is a steak not all simulations are "digital twins". A digital twin is a simulation specifically setup and based on the analysis of real world data rather than physical models.
We use different words for reasons. Mainly so people like you can understand what is being done very much is new and revolutionary, despite your very low IQ take of the world around you. The only one making themselves look stupid is you.
Re: (Score:2)
A digital twin is a simulation specifically setup and based
on the analysis of real world data rather than physical models.
Then what do you feed that "real world data" into?
That's right, a simulation based on a physical model!
Re: (Score:2)
Then what do you feed that "real world data" into?
That's right, a simulation based on a physical model!
No you don't feed the real world data into a simulation based on a physical model. You feed the real world data into an analysis tool which *builds the simulation model from your data*. The digital twin is a form of simulator, one specifically *not* based on idealised models.
Alarming (Score:3)
This section of the article was alarming. It seems like there is already the precursor of this at Amazon warehouses and probably elsewhere.
The technology "raises questions about privacy and cybersecurity," Time warns. "Many of these digital twins are made possible by a multitude of sensors that track real-world data and movement.
"Workers at factories with digital twins may find their every movement followed; the hacker of a digital twin could gain frighteningly precise knowledge about a complex proprietary system."
Re: (Score:2)
More like "alarmist", but not informed by any actual understanding of the subject. Anybody hacking a warehouse system can find out how it works and what is in there. Anybody that can monitor employees can find out what they are doing. Not in any way surprising.
Re: (Score:2)
The article is clearly describing a degree of digital modeling that goes beyond what's in the warehouse and who is working there.
Re: (Score:2)
Anybody compromising a warehouse system or an employee monitoring system has information way beyond "what's in the warehouse and who is working there" already. Your point?
Re: (Score:2)
Already made my point, which is not what you described.
Re: (Score:2)
To have twin model, you will stuff much more of sensors - which is also applicable to the human beings. Like those assigned by court for wrongdoings, but possibly more of them - to know where are your hands and feet and head any given moment. Or image recognition would achieve the same. This heads into unprecedented digital processing of everything under monitoring system. Digital you becomes not only completely accounted now, but also modeled for the possibilities of the reality. It is different position f
Re: (Score:2)
Depending on the level of detail in the digital twin, proprietary information could be lost to hackers.
Re: (Score:2)
Depending on the level of detail in the digital twin, proprietary information could be lost to hackers.
As well as to your, and not only, government.
Re: (Score:2)
Secure military systems don't have a firewall that can be hacked, they have an air gap that can't be crossed.
There are some systems that aren't completely air gapped. There's allowances for one way communications. An example is a remote pilot system for an unmanned aircraft. There's an in-only policy on files to update the maps for the navigation computer. A big deal was made about a virus that was able to piggyback on one of these updates and infect the system. Because there was no internet connection
Re: (Score:2)
Nothing alarm. Workers are tracked at factories even without digital twins. It's basic safety precautions to know where your employees are in a complex and potentially dangerous environment.
Re: (Score:2)
"the hacker of a digital twin could gain frighteningly precise knowledge about a complex proprietary system."
Re: (Score:2)
A hacker of a digital twin is already so deep in your network you're already fucked.
Re: (Score:2)
This is up another level. Being able to make a digital model of your entire process so as to replicate it is a lot more valuable than just stealing your data, encrypting it and getting a ransom.
Re: (Score:2)
This is up another level. Being able to make a digital model of your entire process so as to replicate it is a lot more valuable than just stealing your data, encrypting it and getting a ransom.
You're making assumptions. Our digital twins are close enough to the process network to be fed up to date data in real time which is precisely what makes them so useful. If a hacker got at it, the digital twin isn't the problem, they'll already have access to our historian, our IDM tools, and being several layers deep in the corporate network already I have zero reason to believe they wouldn't also be able to access the process control network itself.
All bets are off if a hacker has breached any level of yo
Re: (Score:2)
Good points, but why assume that a hacker will gain access to every aspect of your corporate infrastructure including your identity management just because they "breached any level" of it? There are various forms of zero-trust architecture that assume the network is always breached and can wall everything off individually. Not impossible to hack but it isn't a given.
re:: (Score:1)