Skip to Main Content

Digital twins are the future. Whether they know it or not, most Americans have already seen digital twins in science fiction and superhero movies. Through the magic of CGI, digital twins are often projected into the air as floating maps and diagrams that can be manipulated like a real object, effortlessly combining and displaying real-time information from an array of hidden sensors.

A digital twin is a digital representation of an object that exists in the physical world. They can come in sizes as varied as the objects they represent, from spark plugs, to sky scrapers, to human beings themselves.

Ideally, a digital twin can be an accurate representation of the object as it exists in real time, or a “living” digital twin.

This is incredibly difficult to create during the construction of an object as complex and dynamic as a large structure, yet that’s exactly what Burcu Akinci and her Ph.D. student Yujie Wei have done. They’ve worked with construction and maintenance personnel at Carnegie Mellon’s Mill 19 to create a living digital twin of the university’s state-of-the-art facility.

Akinci, a professor of civil and environmental engineering, and her team first used laser scanning that casts numerous laser beams in every direction. As each beam meets a surface it reflects back to the scanner, creating an array of dots called a point cloud outlining the structure in space. Researchers took 120 scans between the beginning of construction and completion, capturing important features that would eventually be concealed within the finished walls like mechanical, electrical, and plumbing infrastructure.

Mill 19 will be a home for digital twinning research within manufacturing environments, and so the building having its own digital twin is quite appropriate!

Gary Fedder, Faculty Director, Manufacturing Futures Initiative

Though accurate to a few millimeters, these scans must be performed often to include features like pipes and vents that may later be hidden. This may not always be practical in the dynamic environment of an active construction site. The team’s solution was to fuse this method of digital twin creation with another, using images collected for various purposes during construction.

Images from multiple angles can be layered over the structure outlined by the point-cloud data, acting like a scaffold. Advances in computer vision and image analysis help the researchers match objects within the image to their corresponding data point coordinates within the point cloud. This means that this living digital twin can be updated in a streamlined manner as the building it represents changes throughout its life cycle.

A living digital twin with the flexibility and ease of use to grow with the building it mirrors is a valuable asset for infrastructure managers. Wei provides an example of a common scenario, in which a maintenance manager might have to identify the source of a water leak. Without the peek behind the drywall provided by a living digital twin, the manager might need to spend extra time to identify whether the leak source was from ducting, plumbing, an external source, or something else.

An up-to-date model combining the precise location of every asset and object integral to the building may save decision makers time, labor, guesswork, and money. From a general standpoint, the ubiquity of digital twin technology across numerous industries will only continue to increase. In computing and electronics, a fully realized, living digital twin of a device might be able to sense and locate internal problems almost instantly. In medicine, a patient’s digital twin could hold the secret to diagnosing the disease or injury ailing them. And Akinci and her team will continue to redefine the accuracy and value that a living digital twin can provide for construction and infrastructure managers.