Software

Digital Twins Gain Steam with Federal Agencies

NOAA and others are using virtual models to visualize and predict the behavior of physical objects and systems.

Chris Hayhurst is a contributor to the CDW family of technology magazines.

The National Oceanic and Atmospheric Administration has no shortage of data to crunch as it seeks to understand and monitor global environmental conditions. Its challenge, rather, involves tying all that together and getting disparate data pipelines to feed into a single system.

“We have so many satellites and ground-based observation sites collecting data from different sensors in different formats,” explains Sid Boukabara, former senior scientist with NOAA’s National Environmental Satellite, Data, and Information Service.

That information is highly valuable, he adds, “but it gets really complicated when we want to look at everything all at once.”

That was the problem the agency set out to solve when late last year it partnered with NVIDIA and others to build a digital replica of Earth. The Earth Observation Digital Twin will use AI technology and NVIDIA’s Omniverse computing platform to ingest and analyze NOAA’s many massive datasets and display the findings in real time.

The project is still in its early stages, but the agency plans to demonstrate the technology’s viability starting with sea-surface temperature data in September 2023.

“After that, the idea is to expand to all of the different Earth-system components,” Boukabara says. Eventually, the digital twin will integrate data from sensors on land and in the ocean, as well as NOAA and private sector assets in the cryosphere, the atmosphere, and in space.

Click the banner below to learn about the latest emerging technology by becoming an Insider.

How Digital Twins Can Help Agencies Make Better Decisions

Boukabara describes the concept of a digital twin as “an attempt to copy all the characteristics of a physical entity into the digital world.” It’s a model that, when in the right hands, can be poked and prodded and manipulated to test how physical systems will behave in real life. A digital twin can help agencies like NOAA make better decisions and accomplish their missions.

When it comes to government organizations that have deployed digital twins, NOAA remains in the minority. As of 2021, according to one survey, only 24 percent of federal agency leaders reported experimentation with virtual models.

Still, notes Brian Bothwell, director of engineering and technology assessment at the Government Accountability Office, digital twins are gaining steam in government. “I expect we’ll see more and more federal agencies using this technology over time,” he says.

LEARN MORE: What is digital twin technology and how can agencies use it?

The U.S. Army Corps of Engineers is using digital twin technology to “walk through” areas of new construction at Tyndall Air Force Base in Florida, which were badly damaged by Hurricane Michael in 2018. Meanwhile, the Navy is using digital twins to optimize infrastructure layout at its shipyards.

Similarly, the Food and Drug Administration is developing a digital twin that will help it evaluate medical devices prior to use; and the National Cancer Institute is working with scientists to create digital twins of patients with cancer.

A 2023 report on the technology from the GAO notes that digital twins are now being used for everything from supply chain optimization to agricultural management and manufacturing.

The use of digital twins isn’t new by any means. NASA, in fact, first used a form of virtual modeling as it worked to save the crippled Apollo 13.

“It’s just that now, the technologies are there for digital twins to really take off,” Bothwell says.

Digital twins are different in every scenario, but their benefits are usually clear.”

Jeren Browning Digital Engineering Researcher, Idaho National Laboratory

Digital Twin Adoption Is Driving Agency Innovation

The drivers for digital twin adoption include rapid advancements in computing power, data analytics, IoT sensors and AI and machine learning, Bothwell says.

A digital twin of traffic flow, for example, would use these technologies to model vehicle movement instantaneously as traffic data flows into the system. The approach differs from traditional computer modeling in one significant way, Bothwell says.

“Instead of just collecting data and coming back and building your model,” he says, “with a digital twin, the model and physical twin are running in parallel.”

For NOAA, Boukabara says, “the complete vision is to combine this digital twin for observing the Earth environment with a modeling component that would project the state of the environment in the future.”

Looming droughts, for example, might be identified years in advance, and heat waves and hurricanes and other extreme weather events might be forecast with far better accuracy.

Boukabara says he can even imagine a day when nontraditional sources of environmental data are used to feed NOAA’s “digital twin beast.” Many citizens, for instance, have backyard weather stations recording potentially valuable meteorological information. Cars and smartphones — and increasingly, smart cities — are doing the same with their own fleets of sensors.

“Hopefully, down the road, we’ll have the ability to really leverage the internet of things,” says Boukabara. “For now, we’re focused on learning from this proof of concept, but that’s the direction we’d like to go.”

2010

The year NASA coined the term “digital twin”

Source: IBM

The Clear Benefits of Digital Twins Across Industries

One organization where digital twin deployment is proceeding at full throttle: the Department of Energy’s Idaho National Laboratory. According to INL digital engineering researcher Jeren Browning, the laboratory is working on at least 15 different digital twin projects.

In 2022, Browning personally led a team that successfully demonstrated how a digital twin could be used to test and predict the behavior of a nuclear microreactor. The initiative entailed building a virtual model of a thermal-hydraulic device called the Microreactor Agile Non-nuclear Experimental Testbed.

Using sensor data and physics-based machine learning, the researchers subjected the microreactor to a variety of operating conditions while the digital twin made adjustments in real time to optimize the device’s performance.

EXPLORE: How the U.S. Space Force is leading technical advances within the branch.

Another INL digital twin project involves modeling energy systems to help users make better operational decisions. “For the owner-operator of an asset,” Browning explains, “you could use the digital twin to maximize revenue by incorporating information on real-time pricing in your market.”

His team is also experimenting with extended-reality tools such as Microsoft HoloLens 2, he says. “The idea is to put the human at the center so they can visualize data in more meaningful ways.”

Browning says he’s given presentations about his work on digital twins to other federal agencies, and that generally, people want to know how the technology might help them. His answer: It depends on the digital twin, and it depends on the physical asset that digital twin is modeling.

“Digital twins are different in every scenario,” he says, “but their benefits are usually clear.”

Back to the article Autoplay Full Screen Grid View Exit Full Screen