In pursuit of autonomous nuclear plant operations
It turns out the research for the master’s was just the tip of the iceberg. There was still work left to be done.
For the future viability of nuclear power, small plants, located in rural areas, are a distinct possibility. Thus far, operations conducted in legacy plants have involved intensive manual operations, which are viable because these facilities operate at 100 percent capacity and the power delivered can justify the costs of maintaining a large staff. But microreactors distributed at scale and in remote areas can’t afford a large bench of human talent. It’s where supervised and thoroughly vetted autonomous operations will help.
A primary question was: “How do we transition to autonomous operations in nuclear power plants?” Fortier wanted one integrated approach, a central supervisory control system instead of many interlinked parts. While a legitimate requirement, the challenge here was in modifying operations that were primarily meant for humans to now also include machines.
Fortier realized that no matter how thorough the framework for a supervisory control system, its scope would be severely limited if its rigidity couldn’t accommodate both humans and machines. “Because everything is human-centric, it doesn’t allow you to choose the best way to do a procedure,” Fortier observes. What if, instead, humans and computers could tag-team and do what each does best, with strategic human intervention only delivered as necessary?
MIT-facilitated collaborations
The goal was elegant, but its scope extended far beyond a master’s thesis. A doctorate seemed like a natural progression, so Fortier has continued her research toward a PhD after completing her master’s in 2025.
It was in pursuit of autonomous operations that Fortier discovered the power of collaborations at MIT. Her research advisor, Sacit Cetiner, has a joint appointment with MIT NSE and the Idaho National Laboratory (INL). To address the challenge of devising an autonomous supervisory control system with an effective and easily adopted human-machine interface, Fortier worked with Katya Le Blanc, a senior human factors scientist at INL. A collaboration with the Human System Simulation Laboratory at INL helped Fortier develop a better understanding of designing cyber-physical systems.
“I’m very much an engineer and don’t have a lot of experience in human behavior, so the collaboration with INL was a huge benefit for me. I got better insights into many aspects, including what you want to see when a human has to take over for a machine when it’s no longer working,” Fortier says.
She also availed of collaborations with Westinghouse, a leading design organization and vendor for current- and next-generation nuclear power plants — Fortier completed a summer internship there in 2025 — to test drive ideas she had about autonomous operations solutions.
At MIT, Fortier learned control theory from one of her co-advisors, Anuradha Annaswamy, a founder and director of the Active-Adaptive Control Laboratory in the Department of Mechanical Engineering. “She’s a control systems expert, which really benefits me because while I can explain what to do with a nuclear power plant, she can help me understand better how to go about operations from a control systems perspective,” Fortier says.
Annaswamy advises Fortier on supervisory control system framework and execution. Fortier has also been taking classes related to control systems related classes to build a foundation in the discipline. Curtis Smith, the former director for INL’s Nuclear Safety and Regulatory Research Division, and now KEPCO Professor of the Practice of Nuclear Science and Engineering at MIT NSE, is Fortier’s other co-advisor.
A step-by-step progression toward autonomy
Fortier clarifies that the operations systems she’s designing will incorporate a gradual and systematic move toward autonomy so as to build trust with users. “When we introduce an automated procedure that walks you step by step through what you would be doing anyway, it is reassuring and builds trust,” Fortier points out. Her doctoral work focuses on incorporating objective-oriented operations where a control system can create the sequence of events needed to reach the objective, rather than follow a predetermined operating procedure.
Another point of reassurance is that Fortier is developing automation based on a process called finite state automata, which, unlike AI, is very transparent in its execution. The advantages of finite state automata, which she also studied extensively during an internship at INL in summer 2024, is that it’s a discrete event system. This means that every move in the automation framework is event-driven — if this happens, do that — so it adjusts for current conditions in the plant and transitions between various states or events very clearly. It’s addressing a complex problem through conventional automation, not AI-driven automation. “We’re not using a data-driven statistical approach like machine learning because we do not yet have the tools to validate the operation of such systems,” Fortier says.
Future impact
So promising is Fortier’s work on developing automation for nuclear plants that she was one of the winners of the 2025 edition of the Innovations in Nuclear Energy Research and Development Student Competition from the Department of Energy’s Nuclear Energy University Program.
Using the nuclear plant automation program on next-generation equipment will deliver necessary traction in developing and deploying commercial microreactors.
But first will come the work of scaling the supervisory control system, borrowing from insights gained from work on a small aspect of control. Fortier is excited about the road and the possibilities that await. “The collaborations with other people, and the relationships we have established with stakeholders, have really helped make an impact and supported the relevancy of the work,” she says. “Sometimes when you’re stuck in your own bubble, that outside perspective is really useful.”
