Digital twins just might be “how civil infrastructure finally enters the digital age,” said Paul Mostella, P.E., S.E., M.ASCE.
“Digital twins give us a living, breathing model of assets – bridges, buildings, and utility systems – that we can inspect, analyze, and improve using advanced surveying and sensing tools, all while minimizing field exposure and safety risks," explained Mostella, director of engineering at Quanta Infrastructure Solutions Group.
For civil engineers across disciplines, digital twins have the potential to unlock more efficient and effective construction and maintenance operations.
Opportunities abound
The term “digital twins” can have a range of definitions. They could be scans of existing structures, models of structures used during project delivery, single-structure models, or even models of wider systems, such as state highway systems, said a team of bridge engineers from the California Department of Transportation, or Caltrans.
And digital twins have applications across a wide variety of infrastructure projects.
Some of the most notable benefits are accelerated timelines and lower costs.
A 2024 Caltrans project used building information modeling to create a digital model of a roadway impacted by a landslide as part of a project to “design and build a soldier pile ground anchor wall to stabilize the landslide.”
The project was part of a pilot program to expand the use of BIM for infrastructure in the construction phase. By using the model, Caltrans was able to save $520,000.
The most immediate, widespread opportunities for digital twins are in operations and maintenance, said Houtan Jebelli, Ph.D., EIT, M.ASCE, an assistant professor in the University of Illinois’ department of civil and environmental engineering.
“A well-managed twin can reduce downtime, extend asset life, and enable predictive maintenance across infrastructure systems such as bridges, roads, tunnels, and utilities,” he said. “On active projects, digital twins are being used in 4D and 5D planning, construction sequencing, and schedule- or cost-risk mitigation.”
Mostella said digital twins “turn maintenance from a reactive chore into a proactive strategy.”
Lynn Hiel, P.E., ENV SP, M.ASCE, senior bridge engineer (specialist) in bridge design at the BIM Specialty Branch at Caltrans’ Office of Bridge Design South, also sees maintenance as a prime opportunity for the use of digital twins.
Digital twins have “the ability to identify the details and location of deterioration on a 3D model” with location-based/georeferenced defect tracking, which “could potentially be used in determining the effect on load capacity as well as providing clear guidance for maintenance crews and for maintenance projects,” she said.
A successful example is the smartBRIDGE Hamburg project, which installed a system of over 500 sensors on the Köhlbrand Bridge in Germany’s Port of Hamburg. Connected to a digital twin, the sensors provide data on the structural health of the bridge that is shown in the model.
Digital twins also hold enormous potential as part of smart cities, and municipalities are already exploring the use of large-scale models.
Launched in 2021, Columbia University’s three-year Hybrid Twins for Urban Transportation: From Intersections to Citywide Management project built a digital twin of parts of New York City to explore how sensing data and machine learning could improve traffic flows and safety.
Houston Public Works nearly a decade ago began developing a digital twin model of the city’s water distribution and transmission system, which covers 671 square miles. The model was planned to include all piping, valves, facilities, storage tanks, and wells in the system. In the future, Houston Public Works aims to create a system that uses artificial intelligence and machine learning for predictive modeling.
“We’re heading toward connected ‘ecosystems’ of twins – entire grids, water networks, or cities that sense, predict, and adapt continuously,” said Mostella.
Cities are looking into the relationship of digital twins and the Internet of Things for activities like traffic management, energy efficiency, and other infrastructure necessities.
One example is the Smart Dublin project, in which the city is working with VivaCity and Bentley Systems to develop a digital twin that would display transport data and inform efforts to expand access to safe active travel routes.
“Agencies and utilities want federated models that can support asset management and planning across multiple systems,” Jebelli said. “More twins are becoming operational tools, integrated into work orders, inspections, and budgeting processes. They are no longer just design artifacts handed off at the end of a project.
“We are also seeing models become smarter, embedding code checking, carbon tracking, and resilience scoring directly into workflows,” he continued.
Digital twins can also be used for human-centric health monitoring. One of Jebelli’s recent projects used a digital twin “that fuses physiological and behavioral data such as heart rate variability, skin conductance, and motion patterns to assess workers’ cognitive load and fatigue levels.”
His team used the technology for proactive safety management and personalized work-rest scheduling.
Jebelli has used digital twins to connect humans and robots. On one project, he used a system that “allows supervisors to visualize ergonomic load, intent prediction, and task progress in real time.
“It has been deployed during drywall installation experiments with our humanoid robot, helping identify unsafe postures and optimize robot assistance strategies,” he said.
Digital twins face challenges
Like with any new technology, there are challenges in the adoption of digital twins.
“Integrating data from design models, sensors, and decades-old records is far more complex than most realize,” Mostella said. “Then, there’s the ongoing cost of keeping those models current and secure.”
Jebelli also sees interoperability as a “core issue.”
“Connecting BIM, (geographic information systems), IoT, and legacy systems without data loss is technically and organizationally difficult,” he said. “Most twins fail when data gets trapped in silos.
“There are also challenges with data governance, including who owns the data, how it is updated, and how decisions trace back through the system,” he continued.
Hiel noted that despite the value of centralizing digital twin data, having everything in one place could make the system vulnerable to bad actors.
Dated technology presents another major hurdle.
Traditionally, Hiel said, inspection teams collect data in-person through notes and media.
“Current inspection teams do not typically possess the hardware or training background required to effectively use these technologies,” she said.
She offered drone inspection as an example: If a project requires drone use to inspect a bridge, it needs people with the skills to operate drones.
Beyond technical ability, some teams might be cautious about how much a new tool could actually help.
“Some resistance is expected, especially around perceived complexity and concerns about disrupting workflows,” Jebelli said. “Field teams in particular want to know how this helps them on Day 1, and they are right to ask. The benefits need to be tangible.”
Jebelli has seen some generational differences in attitudes toward data-driven tools, with younger generations being a bit more comfortable.
“But resistance is not strictly generational; it is more about incentives,” he said. “When leadership aligns digital twin use with project outcomes like safety, budget, and schedule, adoption happens across all age groups.”
“Civil engineers, contractors, and operators need to develop comfort with data-centric tools. Without that buy-in, adoption stalls,” Jebelli added.
Justin Alamares, M.S., P.E., senior bridge engineer for structure maintenance and investigations at Caltrans, said the best strategy for getting people on board is to ensure it is as easy as possible to use and make the benefits clear for all generations of engineers.
“There‘s nothing more discouraging to an engineer than a half-baked plan,” he said. “So, if you can’t prove the new process is an improvement of some sort, then you’re going to have a hard time with the adoption.”
“Once teams see that digital twins reduce risk and improve decisions, adoption follows quickly,” he added.
At its core, the issue with trust in digital twins is organizational, and solving it revolves around “aligning people, process, and technology around a single source of truth,” said Mostella.
Even if the technology is there and teams are supportive, there are institutional barriers to the adoption of digital twins that persist, noted Peyman Kaviani, Ph.D., P.E., M.ASCE, branch chief of Bridge Design Branch 19 at Caltrans’ Office of Bridge Design South.
Some regulations require in-person inspection. Unmanned aerial systems like drones, though useful for the creation of digital models, are often restricted in localities.
“I think we should be a little flexible on making rules and taking more risks instead of making sure everything’s perfect. Then, we can apply it nationally,” he said. “So, regulation is a big part of adopting and using digital twins.”
What’s next for digital twins?
Despite these challenges, digital twins are here to stay. And experts in the field hope that digital twins can become even more integrated with other technologies shaping the future of smart cities.
“AI is becoming the co-pilot for digital twins,” said Jebelli. “Machine learning enables anomaly detection on sensor data, surrogate physics models for rapid simulation, and even natural-language interfaces, such as asking the system to show the highest-risk assets under drought conditions.”
Digital twins can work in harmony with AI models to create “a feedback loop between data and decision,” Mostella said.
“AI is the engine that makes digital twins smarter. The twin captures the physical reality, and AI interprets it – identifying stress patterns, predicting failures, and optimizing performance before problems occur,” he explained.
The combination of engineering judgment and computation foresight is “a powerful intersection,” Mostella continued.
Where AI and machine learning could really shine is in processing and transferring data from large digital files to relevant parties, said Hiel.
“The sensor data for structural health monitoring could be processed with AI to deliver trends and summarize the data into information for the bridge maintenance engineers,” she said.
However, other technological improvements are still needed to support data storage and field access for large files.
A network of digital twins and other technologies in smart cities could also prove beneficial for disaster response, Hiel added.
Municipalities could use federated digital twins to centralize multiple domains and information from different stakeholders to assess damage and inform response in the wake of disasters. Digital twins of historic landmarks could also “help reconstruct the landmark accurately after a fire or other disaster.”
On the organizational side, stakeholder attitudes are beginning to shift from a single project-based mindset about digital twins to thinking about their full life cycle.
Continuing to challenge the “siloed mentality” can prevent the benefits and efficiencies of digital twins from being lost, Hiel said. In terms of bridges, “construction, design, and maintenance work more in lockstep.”
As digital twins gain more traction domestically and internationally, she expects enthusiasm to grow.
“Different DOTs have different angles that they’ve approached (for) digital twins,” she said.
“In Colorado, they’re really good at utilities modeling … whereas in other states like Minnesota, they’ve done a lot on their asset side, putting data together and managing an asset information system,” Hiel continued. “It’s good to look at what everybody else is doing to be inspired by different use cases and see what we can do with these models.”
