You probably don't know it, but if you wear a Fitbit and check the number of steps you walk on an app on your smartphone, you're using a digital twin, a concept that is gaining a lot of traction right now.
The app's dashboard is actually a digital representation of you, says Alfonso Velosa, a vice president at market research firm Gartner. "The digital twin is the dashboard that collects the data — the location, vibration, GPS — and applies analytics to highlight how many steps you've taken and if you've climbed up steps, he says.
Essentially, a digital twin is a virtual representation of a thing, person, or process, he says. Use of the concept is growing, thanks to organizations that have Internet of Things (IoT) projects underway.
With an estimated 21 billion connected sensors and endpoints expected this year, digital twins will exist for billions of things in the near future, according to Gartner. By 2022, over two-thirds of companies that have implemented IoT will have deployed at least one digital twin in production – if not sooner, Gartner estimates.
"We've been doing models in software forever, but it is different from before because [with a digital twin] there's a model populated with data about an entity,'' Velosa explains. "There's a unique one-to-one correspondence."
Eventually, widespread usage of digital twins will offer value in operational efficiency and insights into how products are used and how they can be improved, industry observers say.
Because of IoT, artificial intelligence (AI), and augmented reality (AR), "The digital twin concepts will continue to evolve, and we can expect to see a lot more use of the term in marketing materials to describe existing products and approaches whilst at the same time actual digital twin tools,'' says Ian Hughes, senior analyst of the Internet of Things at New York-based 451 Research.
"We will also hear more about the lifecycle of a digital twin, from design to operation to improvement and decommissioning, as part of what is termed the 'Digital Thread,' Hughes adds. "Digital twins are also multi-dimensional and fractal in their nature; you might have a twin of an engine, or equally, a twin of the whole vehicle, or fleet of vehicles, or of the road system and city."
Also driving momentum is the fact that we have greater computational power than ever before, coupled with advancements in AI and the ability to have almost real-time feedback on things, notes Donna Rhodes, a principal research scientist at the Massachusetts Institute of Technology.
Use of digital twins is "absolutely exploding right now across most sectors,'' Rhodes says, including in the aerospace and oil and gas industries. For example, creating a software representation of an aircraft that collects data in near-real time while it is in operation can predict when a part is about to fail, she says. "It's taking the raw data and turning it into useful information."
The most powerful use of a digital twin today is when an aircraft – or an oil platform in the middle of the ocean – is far away and someone needs to conduct diagnostics, Rhodes says. "When you want to diagnose a problem, you can do it from your desk and not have to go into the deep ocean."
Velosa points to the use of digital twins by Canadian healthcare provider Hamilton Health Sciences to prevent code blues (cardiac arrests). HHS mixed data from IoT-based sensors on patients with information from their electronic health records, and information on which staff members were on call on a particular shift, to create a digital twin that compiles all that information in one place. Previously, the data resided in multiple siloed systems, he says.
"Sometimes nurses spent a few minutes getting data and weren't seeing challenges that were going on with a patient,'' he says. By putting all the information in one place, HHS staff could assess the likelihood of whether a person would go into cardiac arrest, and reduced the number of code blues by 60% in one month, Velosa says.
If there are any drawbacks to digital twins, it is the law of numbers, he says. "It's easy to do something when it's one or 10. You can keep track of a few [software models], but what happens when you have 10,000?" he says. It becomes a lot more onerous and challenging when dealing with assets in a supply chain, Velosa adds.
Also, in most cases, someone has to custom-build a digital twin to align it to their specific operations, Velosa says.
"So the other challenge is, who maintains these living models?" They need to be updated for them to stay relevant, and someone has to pay for that; either a business unit or the IT department, he says. However, "At the moment, neither of those have the budget or responsibility for that."
Governance is another issue with a digital twin. In the HHS scenario, a hospital has to figure out whether oversight is the responsibility of IT or the cardiac unit's. Velosa suggests it is probably the latter, but "they don't have a culture that understands how to manage software objects."
Because there are multiple stakeholders in the enterprise touching a digital twin, there has to be interaction among them, Velosa says.
Rhodes says distrust of software models in general could also become a roadblock to the use of digital twins.
"Our research has shown there are many social and technical factors involved in developing trusted models, and these need to be addressed for the investment in digital twins to pay off."
Esther Shein is a freelance technology and business writer based in the Boston area.
No entries found