GBSD Using Digital Twinning at Every Stage of The Program Lifecycle

The Air Force-managed modernization of America’s ground-based nuclear missiles has emerged as a test-bed for the use of digital twins—virtual models of real weapons systems—at every stage of the program lifecycle, its chief told the Space Symposium April 7.

“I have a front row seat right now,” USAF Col. Jason E. Bartolomei, the system program manager for the Ground Based Strategic Deterrent program, told a panel titled Digital Engineering and Digital Twins. GBSD is employing digital twins at every stage of the program lifecycle from “an early conceptual design frame [at the start of a program] to currently right now in the middle of the [Engineering and Manufacturing Development, or] EMD phase [in which prototypes are built] … getting ready for first flight. I have another program going into production, and then I get to see how the Minuteman III is [using digital twinning as it is] transitioning into the sustainment arena.”

Using digital twinning in each of these phases “has its own unique challenges that really need to be taken on front and center,” said Bartolomei.  

He said that the digital tools the program used for the new Sentinel ICBM enabled it to scan and asses “six billion [potential] different system designs,” looking for the one that best balanced capabilities with cost.

As part of Space Force’s commitment to being a digital first service, “We are really focused on [using] digital engineering and digital twins in the entire ecosystem,” said Lisa Costa, the chief technology and innovation officer for the U.S. Space Force. “Not just for acquisition, but we’re really looking at how we embed digital engineering and digital twins into our training, our doctrine, our red teaming, our force design.”

Digital twinning uses software models of real components or systems to help guide designers as they develop plans for a prototype and later, as they work out how to manufacture the real thing. Once a system is in service, digital twins can also be used to work out how often parts need to be replaced, or how to minimize fuel consumption and conduct maintenance more efficiently.  But the models need to answer very different questions at each stage, panelists said.

“Digital engineering and digital twinning can mean a million things to to a million people, but it can also mean a million different things within a single program or a single program office, depending on the lifecycle, depending on the use case,” said moderator Sian Griffiths, a partner at McKinsey and Company.

She noted that Bartolomei was, “At the program pointy end of making this [digital twinning] actually work and actually deriving program value from it.”

The GBSD program had been using digital twinning for eight years, Bartolomei said, joking that was “only a few heartbeats here.” Their ambitions has expanded with each success.

Early on in the program, there was “a lot of concern” that design choices made to maximize capabilities might introduce “cost and schedule risk,” he explained.

“What the digital environment allowed us to do was to bring our multi disciplinary engineering models in with our cost models, to examine a trade space” where different capabilities and different ways to achieve them could be costed against each other, he said.

Decisions made early in the acquisition process could have huge implications downstream, and digital engineering tools made it possible to predict how choices would cost out, panelists said.

“Once you start building the wrong thing,” observed Rob Wavra, a Mckinsey partner and panelist, “recovering that is challenging.” Early choices could be helped by models that “might be lower fidelity, … but support decisions that are incredibly important at the initiation of a program to shape what it is.”

And digital twinning also opened the aperture for acquisition teams, said Bartolomei.

“Industry showed us nine booster designs. And we challenged our team to look at 1,000 booster designs. And lo and behold, our government team found many, many designs that were more affordable and better performing than the ones industry was showing us,” he said. Flush with that success, Bartolomei said, “We got greedy. And we went and looked at not just the booster design, but the total system design.” The team developed “some pretty sophisticated algorithms” that enabled it to examine cost trade offs in “a trade space of six billion different system designs.”