This February, when the National Academy of Engineering named the 114 Americans it was elevating to engineering’s highest honor, it cited Dennis Henneke, of GE Vernova’s nuclear business, GE Hitachi Nuclear Energy, for an imposing achievement: “applying probabilistic risk assessment [PRA] to enhance nuclear reactor safety.” The achievement, a kind of engineering grand slam — arcane complexity and vital real-world importance — was not, Henneke says, the result of a lifelong, noble quest.

“I’ve just always loved the mathematics and pure science of nuclear engineering,” says Henneke, a self-described “geek” and “average engineer” whose passions, as GE Hitachi Nuclear Energy’s consulting engineer in risk and reliability, have played a crucial role in a developing, unpredictable world.

A nonprofit organization within the National Academies of Sciences, Engineering, and Medicine, the NAE advises the federal government, supports education and research, and pools resources from across the commercial and academic spectrum, with inductees including a professor at Switzerland’s Laboratory of Photonics and Quantum Measurements, Apple’s vice president of hardware engineering, and the CEO of Moderna. Henneke joins them as he approaches the end of an illustrious career that in some ways began with a University of Florida master’s thesis — in which he executed a mathematical modeling of the university’s training reactor core using what was then a breakthrough technology, the personal computer — and evolved with the developing technology of nuclear power itself.

A Gift for Envisioning Answers

In other ways, Henneke’s journey began much earlier, largely in private. He recalls an almost ethereal joy when, growing up on Florida’s Atlantic coast, he’d witness the swirl of complex calculations in his head, and sometimes in his sleep. “I realize I’m a little bit of an anomaly in this,” he says, describing high school math competitions where he was an especially ferocious competitor in the speed round. “In every single one, I always answered first out of 40 or 50 people,” he says. “Of course, I was only right about half the time,” he adds with a laugh. “But I had this strange confidence that after raising my hand, the answer would just appear, because it often did.”

At the University of Florida, he once got frustrated with a complicated wave calculation using the Schrödinger equation to model nuclear pellet fusion, “until I could suddenly see it in my head,” he says. “I had the same thing with statistics, where you look at a piece of data with a probability distribution curve, take another piece of data with another probability distribution curve, and multiply or add them together to imagine what that final distribution will look like.” The resulting curve would materialize in his mind as if unbidden.

While today such calculations rely on computer code, Henneke retained an intuitive sense of the larger shapes within the data. “It was always just fascinating to me to think about how it all worked,” he says. “If you take millions of basic events for a PRA and combine them in a mathematical model,” he continues, “how would it work, how would it change over time, what might affect it?” — and, significantly, “what am I not seeing?”

In his first job, at North Carolina’s Duke Power (now Duke Energy), he was part of the first major utility PRA team for a nuclear power plant. While engineers use deterministic calculations to design and operate a plant, a PRA team provides what Henneke calls “a side check. We’d use mathematical, data-driven models to predict the likelihood and consequence of an accident at a nuclear plant, and if we see things that might be more likely or might have a higher consequence than we’d like, we recommend design or operational changes.”

While conceptually the Duke job was not so different from his responsibilities at GE Hitachi Nuclear Energy, the nature of risk calculation at the company — and the sophistication of today’s technology — have made the work exponentially more complex. “The typical nuclear plant has some 100,000 components,” Henneke says. “We have to assess the probability of failure in any one and put a probability and a frequency on each.”

Predicting What Has Never Been Seen

This fall, Henneke turned 65 and began a transition to what he expects will be a half-time role with GE Hitachi Nuclear Energy. As he steps back from his full-time duties with the company, he’ll continue to serve as standard committee chair on the Joint Committee on Nuclear Risk Management and help in the development of U.S. standards through his volunteering with the American Nuclear Society (ANS) standards development committees. He also plans to spend more time supporting the nuclear industry through his work with the International Atomic Energy Association (IAEA) in developing industry standards and technical reports that help in building the next generation of advanced nuclear reactors.

For the past several years, Henneke has mentored the next generation of GE Hitachi Nuclear Energy PRA engineers and he’s confident that he’s leaving the efforts to evaluate advanced reactor risk assessment in great hands. This work has included overseeing a team of 40 people who will carry on developing models for the company’s two advanced reactors: the 300-megawatt BWRX-300 small modular reactor (SMR) and Wyoming’s Natrium, a 345-megawatt sodium fast reactor for which GE Hitachi Nuclear Energy is supporting Bill Gates’ company TerraPower, and whose gigawatt-scale storage system will support grids where variable power output is a concern. SMRs represent one of the fastest-growing markets in nuclear energy and play a key role in decarbonization strategies for a growing number of countries.

Part of the success of the particularly low-risk BWRX-300 — the first of which is scheduled to begin operation in Ontario in 2028 — reflects its novel approach to PRA: deploying experts in the field at the beginning of the core design rather than after. “In fact, one of the core design members was a PRA person,” Henneke reports. “Since we helped lower the risk in design from the very beginning, we call this a ‘risk-informed design.’”

While Henneke feels that today’s models and computing power should inspire confidence in nuclear power, he recognizes that the stakes are high enough that math alone won’t do it. “The biggest risk is for something you don’t model,” he says. He’s particularly haunted by the 2011 nuclear disaster at Japan’s Fukushima Daiichi plant, triggered by an earthquake and tsunami. “They’d done some PRA work, but they hadn’t done a full seismic evaluation,” Henneke says. “If they had, they’d have seen that the risk of nuclear accident was quite high.”

Discovering this was no cool clinical exercise for him. “When I realized this, you know, I cried,” says Henneke. “I mean, I can’t personally fix every reactor around the world, but when I go to conferences or international meetings and talk about these things, I do try to raise the flag. I say that the basic point Fukushima taught me is that it’s not just a matter of doing the math. It’s the scope. Have you looked at all possible events? An aircraft crash, a meteorite — we should look at everything.” 

Henneke’s models — and his imagination — have benefited from the nearly 60-year history of nuclear power. “Right now, there’s some 450 operating nuclear plants, which means there’s a lot of data out there,” he says. “We’ve seen sites flooded from overflowing rivers. We’ve seen the effects of electromagnetic fields (EMF), such as sunspots, causing communications issues.” Then there are things that risk assessors haven’t seen but can predict. “Like drones,” he says, “either in warfare attacks or collision during an inspection of the containment area.” He says he has always asked the question “‘Can you think of anything else bad that can happen that we haven’t thought about?’ We can’t predict absolutely everything, but we can predict most things. You just have to go down that pathway and predict it.”