The rise of digital twins

Staff Writer

Nov 1, 2022

Digital image of a city layout

Part One: A space ripe for innovation & industry collaboration

For decades, civil and environmental engineers have relied on digital models to assess and strengthen designs before they enter the physical world. Now, with improved sensing and computing technology, the way engineers use and maintain models is evolving, and a concept called digital twins is taking hold.

How individuals define the term varies, but most agree that a digital twin is a model that, as much as possible, aligns with current conditions. “We don’t just use our digital model for design and forget about it,” says Mario Bergés, one of many CEE faculty working with digital twins. “We acquire information about the physical structure or system and modify the model so that it reflects what's actually happening during its life cycle.”

The second piece is that digital twins are tools for analysis and simulation. They guide real-world decisions and actions. “The data feeds from the physical to the digital, and the digital provides an environment within which we can generate scenarios,” explains CEE Department Head Burcu Akinci. “A digital twin allows us to replicate what is happening in different environments and conditions, learn from it, make predictions, and eventually control what happens in the physical world.”

A digital twin allows us to replicate what is happening in different environments and conditions, learn from it, make predictions, and eventually control what happens in the physical world.

Burcu Akinci, Department Head, Civil and Environmental Engineering

Collaborating with industry leaders while combining infrastructure, data analysis, and computing expertise, CEE faculty are working on numerous digital twin projects. Their research aims to predict and prevent vehicle and equipment failures, maintain smart habitats in space, and optimize the safety, equity, and sustainability of various infrastructure systems.

“Our department is leading important research around generating, maintaining, and utilizing digital twins to support many critical decisions proactively,” says Akinci. “Depending on which domain it originated from, digital twins get referred to as an information repository or physics-based simulations, amongst other things. We are embodying and bringing together all of these different thoughts across many research areas within CEE.”

Optimizing Equipment and Vehicle Maintenance

Among the work underway is a project with CEE faculty Mario Bergés and Katherine Flaniganto drive proactive, intelligent maintenance for a range of US Army assets and aviation equipment. They are developing a method for combining AI with data-driven and physics-based models to predict likely equipment failures and the underlying problems that need to be addressed.

“To date, while digital twins can represent physical bases, their integration with AI solutions and scalability to complex systems is not yet realized,” Flanigan explains.

Combining the two could unlock multiple benefits. While some predictions can be made by AI and machine learning alone, a digital twin could expedite AI training and enable AI to extrapolate beyond situations for which documented experiences and data exist. Integrating digital models into the AI system should also allow the platform to more clearly show the reasoning behind predictions and recommendations.

In a similar vein, CEE faculty member Pingbo Tang is exploring how digital twins could inform commercial vehicle inspection and maintenance. Partnering with vehicle safety technology providers CompuSpections and Truck-Lite, Tang is building a fleet deterioration model using historical vehicle records and real-time inspection data. The resulting model could reveal strategies for operating fleets with fewer costs, improved mobility, and less idling, not only saving companies money but also reducing emissions of greenhouse gasses and other pollutants.