MARVEL team shares lessons learned through microreactor development

On June 1 at the American Nuclear Society's Annual Conference in Denver, Colo., a team from Idaho National Laboratory presented a session titled "Lessons Learned from MARVEL Reactor Fabrication." The presentation highlighted challenges that arose as they moved from design to manufacturing and assembly, with a focus on reactor part fabrication, Stirling engine implementation, and reactivity control system development.

Project scope: MARVEL, which stands for Microreactor Applications Research, Validation and Evaluation, is a 10-15-kWe (85-kWth) microreactor using sodium potassium natural circulation for the core coolant and modified TRIGA fuel elements.

It is designed to have a two-year operational life, though John Jackson, the national technical director for the Department of Energy Office's Microreactor Program, said there are ways to extend that.

As the first DOE-developed reactor in more than 50 years, he expects MARVEL to provide a platform to test the unique operational aspects and applications of microreactors and serve as a pathfinder for commercial microreactor companies.

Recent project milestones include the identification of five teams that will demonstrate test cases on MARVEL and DOE approval of its preliminary documented safety analysis. Jackson said the team is on track to achieve dry criticality in December.

He said some high-level lessons learned were that optimizing in one area can lead to penalties elsewhere, using off-the-shelf components doesn't guarantee easier integration, and that "minor" changes in the design specifications can equate to significant execution setbacks.

"Some of these seem like common sense, but I'll just tell you from experience with this with this reactor project, even things that seem like common sense catch you in the end," said Jackson. "You change one lever and 97 other levers change in unpredictable ways."

Stirling engines: MARVEL was initially conceived as a self-contained system, utilizing integrated Stirling engines to produce electricity.

"This was one of the pillars of the design, so that it could be a stand-alone power plant that you could take somewhere to deploy and just pull power off it," said Chad Ryan, a design agent at Walsh Engineering Services.

As highly developed, off-the-shelf components that have been shown to run reliably in very harsh conditions, the engines were expected to help keep the design simple.

"We were so happy with this, we thought to ourselves, maybe we won't even test the Stirling engines. They're just drop in, plug and play," said Justin Johnson, a MARVEL program manager.

The team wanted to adjust the cooling system but ran into a series of issues including flow meter failure, lead-bismuth eutectic leaks, and helium bubbling.

Then, tests showed that the mounted engines on the reactor caused significant vibrations.

"The engineer on the floor during [the engineering demonstration unit] test said he was worried that the building was going to shake apart," said Jackson.

Ultimately, the team decided to keep the Stirling engines, but as a decoupled system. They returned the engines to the box they came in, which included rubber bumpers for vibration mitigation. MARVEL's new design includes two heat exchangers between the Stirling engines and the reactor.

"Off-the-shelf does not mean 'off-the-shelf' when you don't use it the way it was designed," Johnson. "That was a stiff lesson to learn."

Materials and fabrication: Increasing design complexity and the desire to keep the reactor small had a downstream effect of ultra-tight tolerances; for example, it required fuel core components to be fabricated to within 0.0005 inches of the specified size. (For comparison, this is around 1.5 times the size of a red blood cell.)

"This haunted us all over the place on this design on MARVEL," said Ryan.

The team did figure out how to make it work but noted that the issue potentially could have been avoided with some design changes.

The MARVEL team had to find alternative methods during machining and fabrication to get past unexpected hurdles.

According to Reese Gannon, a MARVEL project manager, when the fuel core barrel began to warp during machining, their initial solution-high-temperature annealing-was found to significantly worsen the issue, making it more warped and introducing oxidized scale. The team ended up starting the barrel fabrication from scratch with an approach that used low-temperature thermalizations throughout the machining process.

Similarly, a combined TIG (tungsten inert gas) and MIG (metal inert gas) welding process was developed to produce high-quality welds without surface distortion.

Beryllium oxide was found to be difficult to procure and to machine into the precise shapes required for MARVEL's reflector design, and the team ended up needing to build allowances into the design to account for cracks and chips.

Reactor control system: During operation, MARVEL will be controlled by four control drums, which moderate neutrons through either reflection or absorption based on the drum's angle of rotation.

"There's not many control drums out there in [reactor] designs," said Anthony Crawford, a researcher at INL.

Among the designs out there, Crawford said he's only aware of one that included a scram (the SNAP 10A, which was part of the System for Nuclear Auxiliary Power program), which operated in space for 43 days, and he said there are questions about whether it went through the rigor of being a safety-related device in the way that is essential for terrestrial deployment.

"So that's [MARVEL's] novelty of balancing that robust, fast scrams and more slow, precise operations," said Crawford.

The team had to assess the critical characteristics of all parts including bearings, springs, motors, switches, and more, most of which didn't have nuclear-grade versions available off the shelf. When it came time to put all these components together, they had to develop a systematic assembly process for the multitier stack.

MARVEL target milestones: The MARVEL team submitted its dry criticality documented safety analysis in May. They will commence reactor assembly in August and expect to receive MARVEL's fuel in October, with the aim of spending October and November performing readiness activities before achieving dry criticality in December.