Who’s afraid of the NRC part II: The Advocate
The NRC is not the enemy, said pro-nuclear advocate James Krellenstein. There’s plenty of blame to go around for why nuclear stalled out in the US…and plenty of opportunities to start working together
Welcome back to the Decouple Dispatch: This is Part 2 of our 3-part deep dive into the US Nuclear Regulatory Commission (NRC). In Part 1, we talked with Josh Payne, a self-described proud Nukebro who argued passionately that the NRC is standing in the way of nuclear development and innovation. Today, we are going to get a diametrically different view.
“I find the Nukebro view on the NRC baffling,” said James Krellenstein, a physicist, pro-nuclear advocate and part-time technical consultant in the industry.
Learning to hate on the NRC is something of a rite of passage for Nukebros. It is a part of our sacred lore of how nuclear energy went downhill in the United States. We bring out notorious examples of NRC intransigence and hold them up over and over again to justify the belief that the regulatory body is placing an unbearable burden on new nuclear construction. Vogtle and Summer delays? Blame NRC aircraft impact rules. And didn’t they change concrete strength requirements mid-build? And did you hear the one about the $30-million dollar pipe brace?
The $30 million pipe brace story has now been thoroughly debunked, including by the Breakthrough Institute, usually a stern critic of the NRC.
“The NRC is not perfect,” said Krellenstein, “but activists literally made up a situation that did not happen in order to criticize the NRC. Why?”
The Bad Old Days and how things have changed
It’s true that in the wake of the Three Mile Island accident in 1979, the NRC issued a multitude of new regulations that caused painful delays for all the nuclear power plants that were then in the midst of construction.
“You can understand how frustrating this could be to have to build something and tear it apart after it was already built to comply to a regulation that didn’t exist when it was designed,” said Krellenstein, “my dad was one of those engineers.”
Under the “Part 50” licensing process, the owners of a new nuclear plant would apply to the NRC for a construction permit. After construction is underway and nearing completion, you’ll have to apply again for an operating license. The NRC would not grant an operating license till construction on the plant was finished. But in a nuclear power plant build , 4+ years in the best-case scenario, any regulatory change that was imposed while the plant was under construction, but after it had been designed, could see a power plant caught in the middle. This was the trap that so many nuclear plant projects fell into in the wake of Three Mile Island.
Since then, “Part 52” has turned this into a combined licensing process. The NRC analyzes and approves a standard generic design (be it AP1000, ABWR, APR-1400 or NuScale). Then, after analyzing site specific characteristics for a specific new power plant, they will give a project a Combined License (COL) prior to construction.
“Under Part 52, the entire design is licensed before you pour a drop of concrete,” explains Krellenstein, “not only is the design fully licensed, the method which prove that the construction was performed in compliance with the license are also fully specified.”
Having a COL means the project is good-to-go up to fuel loading and beginning of operation, as long as the project was executed according to the original plan. This gave the engineering and design teams concrete target to hit that wasn’t going to move.
The thing is, Part 52 was finalized in 1989, well before the disastrous AP-1000 build experiences at Vogtle and Summer. Vogtle’s COL did not come down until 2012/2013. And the architecture wasn’t finalized until 2017.
“It’s easier to scapegoat the NRC and hail new shiny SMR (Small Modular Reactor) designs, but the hard truths of what went wrong at Vogtle and Summers must be confronted,” said Krellenstein, “the industry has to do the work of figuring out lessons learned and get good at building nuclear power plants.”
What actually went wrong at Vogtle
The AP1000 is a greatly-simplified Light Water Reactor. It contains 85% less cable, 45% less safety critical volume, 35% fewer safety critical pumps and 50% fewer valves than an equivalent-sized generation II Pressurize Water Reactor (PWR). In fact, you might even call it a “Large Modular Reactor.”
“The problem wasn’t with the AP1000,” said Krellenstein, “it was with the nuclear industry itself and its engineering, procurement and construction (EPC) firms. Four firms tried and failed to build Vogtle because they were not ready.”
The factory that manufactured the modules failed, meaning they had to be built on site. This not only defeated the purpose of modularization, it raised rather than lowered the cost.
Back to the “$30 million dollar pipe brace.” According to Krellenstein, what happened there was Southern and its contractors forgot to install the required plant-critical structural supports for the Automatic Depressurizing System (ADS). If it weren’t for NRC-mandated pre-critical testing, the plant would have likely gone critical without this issue being fixed. The cost of fixing this problem, if detected subsequently, would likely be at least 10 fold higher. Of course, if this problem wasn’t detected at all, the possibility of severe damage to the plant existed, which could have cost thousands of times higher. As it happened, Southern requested modification of tech specs on Jan 12. The NRC approval came down the very next day, on the 13th.
“This and other technical issues cannot be laid at the door of the NRC,” said Krellenstein, “and the industry will have the same problems delivering to their customers for the SMRs.”
Is prescriptive regulation really so bad?
Another frequent complaint thrown at the NRC is that they take a “prescriptive” approach to regulation. That is, the NRC will lay out detailed and specific pathways to how things must be done. The pro-nuclear community frequently object that this strict approach is too inflexible and not fit for purpose when it comes to advanced reactors that the NRC itself doesn’t understand fully. They want performance-based regulatory designs so that industry can find and solve its own problems. Seems to make sense?
“No. You WANT prescriptive over interpretive regulation because it spells out the exact criteria you need to meet in order to be certified,” said Krellenstein, “if you had a more interpretive set of regulations, that’s open by definition to interpretation, which leads to regulatory ambiguities.”
In Krellenstein’s book, power plants are not supposed to be complex and nifty: we should be working with designs that are well-understood. As for design innovations, there is already a mechanism in place where the NRC allows licensees to challenge individual prescriptions.
A big mistake by going small?
The US nuclear industry appears to be so scarred by Vogtle they have turned their backs on gigawatt scale reactors altogether in favor of SMRs and advanced reactors. The Biden administration is the most pro-nuclear we’ve had in a generation, but when they speak of nuclear, they speak of life extensions and SMRs only.
Ironically, it might be the best time right now to build an AP1000 in the US.
“Our workers are trained, our supply chains finally figured it out. We should start by doing more and learn by doing,” said Krellenstein, “The NRC has already granted 10 combined licenses for AP1000s at sites around the country.
“The NRC are not chock full of anti-nuclear people,” said Krellenstein, “they are engineers who got into nuclear power because they believe in it. A problem with focusing on the NRC is we are ignoring all the other policy changes we need to support the nuclear renaissance we want.”
ANGELICA’S TAKE
Nukebros — and I am absolutely including myself in this formulation — have to grow up and stop blaming the NRC for everything that goes wrong in our industry. Is there space for criticism? Absolutely. Specifically, I think they still have a lot of space to improve when it comes to SMR and advanced reactors. However, when you look back into history, the idea that they somehow tripped up Vogtle or strangled the industry because they won’t issue new licenses is just plain wrong.
In fact, I had no idea until I wrote this article that the NRC has issued 13 GWs worth of combined licenses to build large reactors. Some of those licenses would have to be reactivated, but 4.4GW of those plants have active COLs right now that could hypothetically start construction tomorrow from a regulatory point of view.
This is hard. As long as we could hold onto the belief that the Big Bad NRC is what’s standing in the way of our atomic dreams, we’ve got somebody to blame. And we’ve got the hope that if we can just change this one thing we’d be living in the future.
But reality is more complicated than that, and that’s a good thing actually.
We, as Nukebros, should absolutely be hooting and hollering about getting those 4.4GW of shovel-ready projects started. I’ve had quite ardent pro-nuclear activists straight up tell me they’re “not financeable” with a shrug as if that should be the end of the conversation. Wait…in this community we make room for molten salts, triso fuel, fast sodium, pebble beds…even fusion if we’re really feeling spicy. But somehow a high interest rate environment is too much to overcome for a proven technology in the richest country on the face of the planet?
And in the future, when it comes to the NRC, let’s resist the temptation to paint them as the bad guy in a simplistic way. First, if we do that, it takes away our power. There are in fact a lot of things we can do to promote nuclear with the NRC exactly the way it is. Secondly, it takes away our influence. The NRC is more likely to respond positively to criticism if it is targeted and fair rather than relentlessly negative. Third of all, it doesn’t fix the fundamental problem of our industry, and if we don’t do that, we are doom to repeat the mistakes of the past over and over again.
It’s not at all a simple demonization. It is just a straightforward manifestation of public choice theory. The NRC does exactly the thing they were tasked by congress to do, and they do it well, but that means we get no innovation and no ability to prototype and learn. For plants in existence, they are actually pretty decent at being a regulator. Unfortunately, there is basically zero flexibility- nobody is ever empowered to make a decision and every tiny thing becomes a bureaucratic nightmare. This isn’t just a government thing- big companies routinely have the same problems. This is how we end up worrying about making calculations accurate to 0.000001% with supercomputers when a 5% estimate with a 50% safety factor is beyond good enough and could be done with a single 80286 processor from 1988.
Unfortunately what the NRC mandate means in practice is that we get nothing but a wedge that can be infinitely used by professional scaredy cats that want to delay things forever. Of course most in the NRC actually like nuclear power- they studied it extensively in school, etc. That doesn’t mean they are in a system that allows for prudent regulation or a system of continuous improvement.
The benefit side has to be a part of any good regulation otherwise you get a one way ratchet. Even though it is not even close to the most dangerous thing in our society, it is regulated as if it were a loaded nuclear weapon pointed directly at children which is patently absurd. No other thing is regulated this way, and unfortunately the people who are regulated derive most of their revenue from the status quo so they have every incentive to keep their gravy train going even if it is worse for all of us in the long run. Incentives must be changed.
Great article Angelica!
A couple of thoughts in response to this one and some nuances that are hard to capture through an interview and article.
1) While I ascribe a large amount of blame for the current state of nuclear in the US to the NRC I do not believe that the industry is blameless or perfect. Vogtle was a colossal screwup on numerous fronts and there were many things that Westinghouse could have done differently and better. I will also say that the industry's insular and tight-lipped nature has significantly hurt it. They have absolutely failed on the publicity, advocacy, and public education fronts. They have also built up and reinforced many of the barriers to entry in an attempt to protect themselves from competition. There are few other reasons but I won't go into those here.
2) Prescriptive approaches can still be open to interpretation. There are numerous internal NRC arguments over how some rules should be interpreted and whether or not some requirements have been satisfied or not. Even if the NRC lays out a fully constrained and defined set of requirements for a design you still have to "prove" that your design meets those requirements. Since you can't build it without proving this first you are left with using models to "prove" your design. As any experienced modeler knows, you can always find a "problem" with a model. Since we haven't built anything in decades designers are left trying to convince the regulators that the models sufficiently match reality and neither the regulator nor the designers have a solid set of experience to base assumptions of what is and isn't important to model.
“Under Part 52, the entire design is licensed before you pour a drop of concrete,” explains Krellenstein, “not only is the design fully licensed, the method which prove that the construction was performed in compliance with the licence are also fully specified.”
I'm not sure how much experience James has in construction or manufacturing, but this isn't how building anything works. No engineer can produce a FOAK design that will look exactly like the final product. We have many excellent tools that let us predict and plan for most thing but there are always unknown unknowns. As anyone who as ever done a woodworking, home improvement, or really any project will know, the plan never survives first contact with the enemy. Planning and modeling can only take you so far, eventually you have to execute and learn from the mistakes made while executing.
In general I, and many others in the industry, have seen and experienced evidence that overwhelmingly indicates that the NRC is a functional, let alone good regulator. The insane pricetag of $700M+ for NuScale's design certification, the 80y life extension walk backs, the Oklo license denial, timeframes and costs for licensing extraordinarily low risk, tiny research and test reactors. The total failure of the Part 53 rulemaking, and many, many others.
A couple of good articles on this:
https://foreignpolicy.com/2023/04/08/nuclear-power-energy-net-zero-emissions-nrc-regulation-climate-change/
https://rootsofprogress.org/devanney-on-the-nuclear-flop
https://thebreakthrough.org/blog/nrc-staff-whiffs-on-nuclear-licensing-modernization
https://thebreakthrough.org/blog/blog-nrc-revises-previously-issued-subsequent-license-renewal-for-existing-nuclear-power-plants
https://atomicinsights.com/nrcs-imposition-of-aircraft-impact-rule-played-a-major-role-in-vogtle-project-delays-and-vc-summer-failure/