Digital Twins at Work: 8 Examples
Here's how eight organizations are using digital twins to simulate factories, water systems, entire urban areas and more.
The U.S. Apollo spacecraft may not have made it to the moon in 1970, barely avoiding tragedy when one of its oxygen tanks ruptured and exploded en route, but some count it as the first documented example of what was essentially the digital twin approach transforming industry today. Both the crew and mission control had trained using various simulators controlled by a network of digital computers. When Houston realized there was a problem, mission control teams turned to those same simulators to work out plans on the fly to get the astronauts safely home.
Fifty years later, thanks to advances in analytics, AI, and computing power, digital twin technologies – virtual replicas of systems that can model, simulate, monitor, analyze, and optimize the physical world – are being deployed across many industries. More than half (55%) of respondents who either had a digital twin program or plans to start one said digital twins are a strategic element of their digital transformation programs, according to a 2022 Capgemini Research Institute survey. And progress continues as researchers around the globe, from the Netherlands-based Digital Twin group to the Digital Twin Consortium in the United States, are partnering with industry to tame some of the complexity and further put these approaches into operation.
Let’s dive into seven examples of digital twins at work in different settings today.
Example 1: BMW’s digital twin factories support sustainability, efficiency, and digital transformation goals
BMW is all revved up. The German automaker first began developing virtual replicas of aspects of its production line in 2014. Before rolling out the drivetrains for its new electric vehicles in 2021, the German automaker had created a fully operational virtual version of them at its Regensburg, Bavaria factory where they would be made. The real-time digital twin can simulate production and scheduling at the factory, down to the work-order level.
By the end of 2022, BMW was all-in on digital twins, with plans to have virtual models of all 31 factories in its production network in operation. In its official announcement, the company said this project rolls up under its global “iFactory” strategy, noting that with digital twins as a foundation, the automaker can reduce production planning by nearly a third.
Takeaways:
For BMW, this is an efficiency play, supporting its sustainability and digital transformation goals. The iFactory initiative “focuses the BMW Group’s production expertise on three key topic areas: Lean, Green, and Digital.”
- The twinning project starts with a 3D digital scan of each factory.
- Roughly 15,000 BMW employees can currently access the data through a custom application called BMW Factory Viewer, and use it to virtually inspect specific areas, carry out precise measurements, and collaborate simultaneously across locations and time zones.
- This access will allow plant designers to create more efficient layouts and connections between different shop-floor processes.
- The company says that data science provides the foundation for real-time, data-based decision-making, as well as quick and proactive identification of root causes of issues that BMW can use to improve plant processes.
Learn more:
BMW provides information on its Web site.
Engineering Matters wrote up BMW’s initiative.
IoT World Today offers more details on the sustainability connection.
Example 2: Tata Steel turns to digital twins to support radical innovation in a centuries-old industry
One of the world’s biggest steel companies is also an ardent fan of digital twin technologies, evaluating the approach’s value for everything from remote factory management to keep production going during a pandemic; to identifying factory process failures that might affect quality before products go into production; to smart mining enablement.
One of the areas that these carbon copies may have the biggest effect is on Tata’s vision for a more sustainable steel industry. The industry is looking to redesign its production process to reduce emissions in order to meet the EU’s target of 80-95% reduction (from 1990 levels) by 2050. Tata’s European unit was picked to lead a €75 million (approximately US$79 million) project to develop new technology, and its IJmuiden, Netherlands, plant became the testing ground for an alternative to using the blast furnace for iron making. The new method, called HIsarna, processes ore almost directly into liquid iron. If it can be done, it will be more energy efficient; have a lower carbon footprint; produce lower emissions of nitrogen, sulfur oxides, and nitrogen oxides; enable more efficient use of raw materials; and lower the operating costs of steelmaking. But it’s a big if. Blast furnaces, which date back to the 14th century, are proven; HIsarna is not.
Takeaways:
Tata Steel is counting on digital simulation to help overcome a huge hurdle to introducing a more energy efficient method to steelmaking. In order to commercialize the novel HIsarna steelmaking method, the company needs to stabilize it.
- The company observed fluctuations in this new, more complex steelmaking process that were not easily explained or controlled. Researchers had some hypotheses but could not definitely nail down the cause.
- A real-time replica of the sintering method will allow the company to tinker with the process digitally to test possible performance improvements.
- The goal is to develop an adaptive digital doppelgänger that guarantees real-time, accurate representation of the physical systems. This could have applicability for any industries seeking to perform real-time updating of dynamic systems and multi-physics models (think of simulating fluid dynamics, acoustics, or electromagnetic fields, for example).
- By digitalizing the thermal process management at iron and steelmaking plants and using simulation-based digital twins and AI/ML modeling, Tata Steel can pinpoint the weak spots in the process and solve them more cost effectively.
- Tata Steel’s 2021-22 annual report indicates that its R&D efforts continue with plans to explore scaling the HIsarna process in India and develop an implementation plan over the coming decades.
Learn more:
The Digital Twin research group, helping Tata on dynamic digital model enrichment, describes the use case in more detail in “Automated Model Updating of Digital Twins” and “Use Case: HIsarna Process Stability.”
The Economic Times describes the HIsarna process beginning by injecting iron ore at the top of a reactor, where it is liquefied inside a high-temperature cyclone.
A case study offers more background on the original project to develop new steelmaking technology.
Example 3: Thames Water System plugs leaks with a replica of its water network
Aside from the fable of the little Dutch boy who placed his finger in a hole in the dike and saved the town of Haarlem, there are few stories of quick and easy methods for finding and repairing leaks in complex, crucial water systems. Digital twins could change that.
The Thames Water network provides clean water for 15 million people who live within a 13,000-square-kilometer (5,000-square-mile) region in and around London, from Gloucestershire in the west to Kent and Essex in the east. Every day, the company supplies 2.6 billion liters of water, but as it notes on its Web site: “not all of that gets to our customers.” Nearly a quarter of Thames Water’s supply is lost to leakage within the network and its customers’ pipes.
Some of those leaks are visible, rising to the surface, reported by customers, and quickly fixed. But those don’t tend to lose as much water. The vast majority of leaks – 95% of them – are insidious and invisible. The causes vary, including natural wear and tear on the infrastructure, heavy traffic that causes instability in the ground, temperature changes and extreme weather events, and sudden changes in pressure. In addition to doubling down on some more traditional efforts to stem leaks – increasing its detection, repair, and plumbing field force; analyzing meter data to better manage demand on pipes; employing no-dig technology to prevent further water loss; and enhancing approaches to balancing water pressure – Thames Water is building a digital version of the full water network.
Takeaways:
- Thames Water’s leak-finding digital replica connects an early warning system for water loss to other risk-management tools. “Reducing leakage is one of our most important priorities, and also one of our biggest challenges,” David Gable, Thames Water’s project delivery manager, shares on the company’s Web site. “Being able to capture a range of data streams and present them in a simple format adds another tool to our armoury in the fight against leakage and I hope this is a model that can be rolled out across our network.”
- Compiling data from a number of connected devices, like smart meters and acoustic loggers installed on pipes to “listen” for leaks, the water company is creating a digital dupe of the water network.
- The digital twin is one of several digital applications that feeds into the company’s system risk visualization tool which leaders and managers use to monitor network performance and identify risks.
- Thames Water is testing the digital twin – a first in the industry – in Deptford, South London, where the pilot project uncovered a number of leaks caused by high pressure and damaged valves.
- In addition to saving a million liters of water lost to leaks every day, the digital duplicate can simulate the effect of various repair methods to help identify the best solution in each case.
Learn more:
Thames Water and its partners will help create standards for digital twin use in industry, according to Smart Water magazine.
The virtual water network can help the company identify risks beyond leaks, according to a description from The Ofwat Innovation Fund for water sector innovation.
Thames Water data scientists explain how they used sensors and geospatial information system (GIS) asset data to develop the platform.
Example 4: Orlando Economic Partnership builds an immersive regional replica to guide future development
Digital twins are emerging as a cornerstone of many smart city initiatives from Singapore to Zurich. Because they allow simulation prior to implementation, they provide insights across a number of key areas of city governance, from urban planning to land-use optimization.
The Orlando Economic Partnership (OEP) recently unveiled its own immersive 3D representation of its metropolitan area, created with input from multiple stakeholders. Project leaders at the private-public coalition say theirs is the first digital twin in use by an economic development organization to map out an entire region, recreating the 800 square miles of Orange, Seminole, and Osceola counties in virtual form. Developed by Unity, first known for creating a real-time 3D engine for gaming, the first iteration was built on municipal demographics, transportation, real estate, and education data with the goal of incorporating additional data sets from public and private sources.
Takeaways:
OEP’s goal is to create an essential urban planning tool for companies, local governments, and nonprofits – a scenario-testing engine for visualizing how decisions about infrastructure, utilities, and business development could play out.
- The OEP worked backward from the digital twin it wanted to build to determine what they needed to model, and then determine what data sets they needed to create it.
- The first version of the digital representation, on display in holographic form at OEP headquarters, is being used as a marketing tool to drive further involvement in and iterations of the project.
- The OEP hopes the technology will enable greater cross-functional collaboration through the design, build, testing, deployment, and operation of complex systems, business development, and other regional projects in a more interactive and immersive way.
- Company leaders and site selectors could use the technology to explore available land and real estate along with infrastructure and connectivity, talent availability in the area, and industry-specific insights. City planners and utility companies might map out proposed infrastructure improvements and simulate their effects.
- The twin could be used to evaluate rail line expansion being voted on in an upcoming referendum, for airport master planning, or to improve hurricane recovery efforts.
Learn more:
An article in VentureBeat reports that future phases of the project will expand into VR, AR, and mobile, and support live sensor feeds for traffic and weather.
Smart Cities Dive coverage of OEP’s initial announcement notes that future versions could include an open back end that will allow a range of users to run their own data scenarios.
Example 5: Kaeser twins its way to a modern air compressor sales model
Coburg, Germany-based Kaeser entered the air compressor business in 1948, cut off from its original auto parts market in a divided Germany and sensing strong demand for compressors in the post-war economy. Kaeser changed the game in air compressor sales in 2018 when it built digital replicas of its compressors; this technology became the basis for its networked compressor products, whose operating data can thus be monitored in real time, enabling predictive maintenance ahead of any equipment failures. (For more on predictive maintenance, see “The Future of Field Management.”)
Taking its cues from other digitally disrupted spaces, Kaeser’s leaders also seized the opportunity to sell these smart compressors in a whole new way – air as a service. Rather than a large capital expense, for a monthly payment customers can get just the air with additional amounts available for a fee as needed. As the company says in its sales materials: You simply give us a few square meters of floor space, and we’ll take care of the rest.
Takeaways:
With the ability to monitor operating data in real time, Kaeser has been able to build a profitable as-a-service business. The company also uses resulting insight to improve product quality.
- The digital twins are used throughout the product lifecycle of the company’s smart air compressor stations.
- The company can monitor operating data in real time, allowing technicians to intervene before there is a problem with the product.
- Kaeser’s compressed air subscription offering saves its customers money compared to the traditional approach to equipment repairs – at regular intervals or upon request – which can result in performing maintenance too early or too late, leading to unnecessary downtime.
- Digital twin tech is part of Kaeser’s core configure, price, and quote process, allowing the company to offer bespoke product and service pairings without blowing its margins. Engineering simulations enable technical verification of customer configurations, the testing of various scenarios to balance costs against customer requirements, and increased efficiency in the sales funnel for these one-off solutions.
Learn more:
Quality Digest says charging customers for the air they use has been a profitable move for Kaeser.
The book Managing Industrial Services includes a chapter on Kaeser as an example of how the technology can underpin new service models.
Short video shows an engineer using a digital twin and virtual interface to optimize wind farm performance.
Example 6: GE Renewable Energy gets more power with digital wind farms
When it made its first spins gathering electrical power from thin air, the wind turbine was itself a radical new way of harnessing a sustainable and natural source of energy. Some 125 years on, GE Renewable Energy is innovating again by taking the whole process virtual. The company’s Digital Wind Farm – more accurately described as digital twin farms – is a project to create a replica for every farm in order to improve its design and ongoing operation.
GE, the former conglomerate, had been working on the forefront of digital twin adoption for some time, investing in the development of digital replicas of its massive, complex, and expensive industrial products that are often essential to crucial infrastructure: jet engines, locomotives, and healthcare equipment. In the mid-2010s, the company had ambitions to so fully digitalize its assets that it could shift from selling capital intensive assets to selling their capabilities as a service (much like Kaeser is doing with air compression, in example five above). That transformation ultimately stalled (and the GE Digital business unit that rose up to support it was ultimately shut down), but digital twins still play a leading role within the three companies that resulted from GE’s split, focusing on aviation, healthcare, and energy.
Which brings us back to the turbines. Wind farms are highly individual – as singular as a human fingerprint, according to one of the company’s general managers. Thus, company leaders thought if they could capture more data from the machines about their interaction with the landscape and the wind, they could improve their design and performance.
GE Renewable Energy uses the digital twin to help design the most efficient turbine for each pad, and optimize the farm's usage as conditions change. (It’s an approach that’s worked before. The company used AI to build a simulation of its gas turbines that works in the background to find optimal flame temperatures and fuel splits that minimize emissions and electrical noise.)
Takeaways:
GE Renewable Energy uses the digital twin to help design the most efficient turbine for each pad and improve the farm’s usage as conditions change. (It’s an approach that’s worked before. The company used AI to build a simulation of its gas turbines that works in the background to find best flame temperatures and fuel splits that minimize emissions and electrical noise.)
- Each three-blade, upwind, horizontal-axis 2MW wind turbine is equipped with a wind power application that can both monitor the asset’s performance and improve it as it runs.
- Running this hardware and software together can increase energy production by as much as 20% and deliver an additional $100 million in revenue over the life of the wind turbine.
- Taking things further, a cloud-based model of an entire whole wind farm enables engineers to mix and match turbine configurations to ensure the best approach based on the specifics of the wind farm location.
- Once the wind turbines are in place, the digital twin collects and analyzes data from its real-world likeness and suggests efficiency improvements.
Learn more:
These digital twins enable scenario planning what would happen to power production from the turbine if the wind blew harder, longer, or not at all, as an AWS blog post explains.
Windpower Engineering describes how digital twins are transforming the industry.
The use of digital technology in wind turbine enhances performance, reliability, and profitability, says a market report on the growth of digital wind farms from AMA Research.
Example 7: NASA deploys digital twinning at massive scale for facilities management
Returning to the OG of digital twins, NASA, provides a modern example of their deployment on a grand scale.
NASA’s oldest research site in Hampton, Virginia – the Langley Research Center where they build colossal space shuttles – encompasses more than 300 buildings housing 1,800 employees and various idiosyncratic equipment, like wind tunnels and industrial ‘shaker tables’ to simulate re-entry vibration. Dating back to the earliest days of aviation, the campus is now ground zero for digital virtualization.
The foundation for a digital double of the 764-acre campus has been developed over three decades as technologists built a GIS to better manage operations and maintenance. The exhaustive map of Langley became the core of the center’s digital twin, which is used to manage most key tasks at the research center, from safety and flood prevention to daily maintenance and sustainability. A testament to the GIS team’s attention to detail: the mapping includes intricate facets of underground utilities. “We surveyed all of our visible features, all of our utilities, all of our roads, and then we backed that up with high-resolution aerial photography,” GIS team leader Brad Ball says in a blog from GIS software maker ESRI. “We can compare the map to the aerial photo and they align exactly to where things are.”
Takeaways:
As trust in a digital twin approach grows, new applications will emerge. NASA’s mapping began with digitizing paper floor plans and employing optimization algorithms to evaluate the facility needs of each department, resulting in a first of its kind space allocation tool built in 2004. Since then, digital twin tools have taken off.
- Today, the 3D model feeds into all sorts of functionality from a locator app that guides maintenance staff to the right piece of equipment to flood impact analysis and scenario planning.
- The ReVITALization team, which manages the agency’s 20-year plan to replace legacy facilities with new state-of-the art structures, uses nearly 50 twin-based apps for space management, real property management, and planning.
- Bonus: The digital twin even helps NASA get better deals from suppliers, its facility intelligence boosting bidding for maintenance and operations.
- Digital twin adoption has its challenges even at NASA, with some paper-lovers still insisting on rolled-up blueprints.
Learn more:
The ESRI blog post covers the genesis and evolution of the Langley Research Center digital twin.
Earlier coverage includes ESRI’s 2004 peek at the first use of digital twin tech.
Example 8: Lowe's uses digital twins, augmented reality to sharpen retail operations
U.S. home improvement retailer Lowe's has begun creating digital duplicates of individual stores, combining spatial data with other information such as product location and order histories. Lowe's says these store twins give its employees "superpowers to optimize operations and localize plans."
At its official announcement in September 2022, two stores had complete replicas live. The system was built by Lowe's Innovation Labs unit, which explores new technologies and applications for the chain. "Through emerging technology, we are always imagining and testing ways to improve store operations and remove friction for our customers," said Seemantini Godbole, Lowe's executive vice president, chief digital and information officer. Lowe's operates roughly 1,700 stores nationwide.
The integration of real-time product location data, together with the use of augmented reality headsets to view the digital twin while standing anywhere in the store, gives a hint of how the utility goes well beyond a simple duplication of a physical facility.
Consider, for example, the standard retail task of resetting and restocking shelves. A store employee can use the AR headset to overlay what the shelf is supposed to look like, and adjust the products accordingly. Employees can also use the AR overlay for a complete view of a box that is partially out of sight on an upper shelf.
Takeaways:
Retailers of all sorts, as well as event spaces and multi-tenant building operators, can garner potential lessons in operational improvement from Lowe's digital twin work.
- For operational improvements, the digital twin system can display heatmaps showing foot traffic, measure the distance between items frequently purchased together, and thus suggest improved product placement or floor layouts.
- Associates can test proposed changes in the digital system before rolling them out to the store. Speaking at the National Retail Federation (NRF) show in January 2023, Mason Sheffield, senior director of creative technology for Lowe's Innovation Labs, said the company could potentially run thousands of simulations overnight, across all stores. Previously, similar testing would require weeks or months of planning and research.
- Lowe's uses Nvidia's Omniverse 3D development platform to underpin the creation of digital twins.
- In the future, the company expects to add more streaming data APIs to the platform and smart internet of things (IoT) sensors in its stores, with additional data opening up new use cases.
Learn more:
Lowe's Innovation Labs and Nvidia both offer additional details on the digital stores, including video walkthroughs of some features.
https://main--builder-prospect-prod--sapudex.hlx.page/fragments/insights/article-details