Outstanding as always, thank you. One minor quibble though. The Regulated Asset Base model was introduced by the previous Conservative government a couple of years after Wylfa fell through.
You are of course right that the new Labour government has adopted it without demur and it is the funding model for the second major reactor being built at Sizewell C.
The first one, Hinkley Point C, was funded by EDF and the government. Both are eyewateringly expensive at c. $14m per MW, despite Kepco apparently being around $2.5m per MW.
Rolls Royce are the preferred bidder for a fleet of SMRs, coming in at a chunky 470MW, but they're FOAK so it will be at least 10 years to first completion.
The existing fleet (c. 5GW)are due to be closed by 2030, with numerous extensions already, with one exception at Sizewell B. But the graph is pointing down not up at the moment, despite all the recent hype.
When you get in the weeds of everything the ONR has done, it all sounds superficially reasonable. It's only when your step back and look at the whole package that you can see how ridiculous the multiple layers are.
Anyway, more power to your elbow and thanks for all your insights.
Thanks correction. RAB in its various forms has been around for a long time. If you think about it, the regualted monopoly is the only system that can work in a situation where the regulator's rules are unclear or the regualtor can change the rules as he sees fit. ALARA or ALARP is both. No rational investor can allow himself to be exposed to that risk. RAB allows him to trasnfer that risk to the ratepayer, who has no say in the matter.
The problem is now the limit on how far up the regulator can push nuke cost is the cost of alternative sources of electricity. If no other dispatchable source of electricity is legal as is becoming the case in the UK and elsewhere, that limit is very high indeed, impoverishingly high, civilization destroyingly high.
EDF made public claims that they had been required to make 7,000 changes from their pre-existing design which had been approved in France. They also argued that they had been required to use 25% extra concrete and 35% extra steel. In addition to the ventilation changes you mention, there was also a requirement to install a backup hardwired control system because the main system duplicated the software and therefore there could be a double failure. They also raised the sea wall and insisted on a second floor for the diesel engine buildings.
The ONR have pushed back quite firmly on some of these claims and argued that the overwhelming majority of these changes were approved during the GDA process and that most of the extra concrete etc was site specific due to local geography. They also argued that the sea wall changes and diesel engine elevations were post-Fukushima (the GDA was awarded in 2012) and that the hardwiring was internationally agreed as best practice and doubled down on the "hacking" risk. They basically denied imposing any real changes post GDA ('have remained largely unchanged since 2012'), instead blaming the Environment Agency and the Health & Safety Executive.
Less acceptable was the ONR response to the radiation filter which was, frankly, concerning: https://www.onr.org.uk/news/all-news/2025/07/our-assessment-was-not-bananas. The flippant nature of the press release suggests that the absolute catastrophe of GEH withdrawal from Wylfa has not been internalised by UK regulators. I accept that the primary reason was financing, but that does not absolve the ONR from some blame for the heavy handed regulatory environment which must have formed part of GEH decision making matrix.
What should really be happening is that the UK government should be in urgent negotiations with GEH to come back to Wylfa, pointing out that the very RAB model they were requesting is now in place. Of course, GEH may turn round and say that this kind of chopping and changing, plus some evidence of regulatory bone headedness creates a continuing unstable political and regulatory environment, and it would be hard to argue with that. Nevertheless, eating humble pie is something politicians and regulators need to do sometimes.
You are absolutely right that the deadline set by the Climate Change Act 2008 (as amended by Regulation in 2019) of Net Zero by 2050 creates a further looming problem for dispatchable power, particularly as the UK approaches grid level limits on non-alternative current sources of power in wind and solar etc. There is zero sense of urgency, despite government projections of peak demand increasing from 48GW to 72GW by 2030 and to ~100GW by 2035. Domestic non-intermittent supply is <40GW at the moment so, well, you can do the maths (no, not math, maths!).
ALARP is a pernicious ball and chain and has to be reformed. LNT is a equally ridiculous. But the good news is that there are straws in the wind that people in senior positions may be starting to question some of the most insane shibboleths. A new Nuclear Regulatory Taskforce recently issued an interim report. They even mention LNT - while they pay official lip service to its validity, if you read between the lines you can see they are at least aware of the pushback against LNT dogma. They are inviting evidence submissions by 8 September. I would encourage readers of this blog to submit evidence to that committee. I plan to do so.
I understand your cynicism about RAB, but when an industry is forced to invent a term for global project cost excluding financing costs, it is just not possible to continue with upfront models that are hobbled by the time value of money. The UK has a centralised regulatory model, so with a due nod to your concerns, RAB seems to be the least worst option to reduce the harm caused by the magic of compounding.
There are other potential options for improving the situation but that must await a paper I am currently drafting that I hope will be ready reasonably soon.
It's not cynicism. It's an indisputable fact. No sane investor is going to expose his capital to this nonsense, The solution is not to make the poor ratepayer bear all the costs of misdirected regulation. The solution is to change teh regualtory system to one that balances nuclear benefits and risks in a manner that is consistent with overall societal welfare. UCert is such a system. Pls read
Apologies, and withdrawn - even though I think deep cynicism would be an entirely appropriate reaction to the regulation of civil nuclear power in general and the RAB model in particular (whether in the UK or USA). I fully accept that your position is one of stating plain facts based on ample evidence, rather than cynicism.
Thanks for the link. I have read huge chunks of your blog, as well as the book you linked to, which happens to be sitting on my desk. I pretty much agree with your core theses, especially on LNT and ALARA, but also the insane DEGB and sequelae, your comparison with private shipyards and in general the over-regulation of the civil nuclear sector. I don't think the world is remotely ready to confront the second Big Lie though. That doesn't mean you are wrong of course, or that I disagree with you, but we are where we are.
I remain particularly struck by your account of the creative tension between regulators and constructors in the early years, and how nuclear competed successfully with coal when the latter was as cheap as it has ever been. I was also unaware until I read your work that the global limit for radiation exposure was 2mSv per day, and that this level is roughly 10% of the observable harm levels in the literature. I am also impressed by the 2km zone you suggest, for the reasons you give.
I have read your UCert suggestion and it seems intelligent and well thought out to me. My only concern is that it seems unlikely to be implemented any time soon. I propose to leverage a couple of the advantages that the UK system still has in order to try to expedite the construction of civil nuclear power here, if possible, and in the admittedly unlikely event that my suggestions land with decision makers.
Many thanks for the replies. I would be honoured if you have time to read a draft of my paper in due course - but I understand if you are too busy.
UCert (rebranded as the Nuclear Reorganization Act for those who are into bureaucratic organization charts) is not impossible. It starts with replacing LNT with SNT. In the US this can be done by an EO. Don't know what the UK rules are but I very much doubt LNT was mandated by Parliament. So I suspect this can be done administratively, That's where your focus should be. With LNT gone the whole system will start to unravel.
I think the evidence you put forward as to the inefficacy of LNT is compelling - but then I am not a scientist. The Taskforce, to their credit, highlight criticism of the view that LNT is accurate, but clearly feel obliged to endorse it anyway given current scientific consensus and the views of international bodies. I think it is very interesting that they mention that LNT has been criticised and that they also point out that the ONR have "Basic Safety Objectives" which are even more stringent than international guidelines. These are subtle hints, in my view, that they would welcome pushback. If you want a flavour of the scale of the challenge, check out this ONR press release: https://www.onr.org.uk/news/all-news/2025/07/our-assessment-was-not-bananas
The taskforce invites submissions pending their final report this Autumn (deadline: 8 September) and the guideline for such submissions is normally 3,000 words maximum. Email: nuclearregulatorytaskforce@energysecurity.gov.uk. A short explanation of your core arguments on LNT, with examples, could maybe have a real impact? I don't know, but it cannot do any harm, if you have time.
Many thanks for agreeing to take a look at the paper - what is the best way to send it to you in due course (I am still in the process of writing it).
I don't understand how ALARA can require these expensive changes. How can anyone think that saving 0.1 mSv per year is worth even 2 cents. It all comes down to the definition of "reasonable". Do we really need to abolish the agency, or can a simple change of leadership restore common sense?
Yes and no. Yes we must get rid of "independent" agencies to have balanced regulation. No, changing leadership will not make a material difference.
It's the system, not the people. The NRC, ONR, ASNR, etc purpose in life is minimizing radiation exposure. No other industry including some massive polluters requires an "independent" agency to regulate it. But radiation is a special horror, so we set up a special regulator whose job is minimizing radiation exposure, regardless of what the preamble to the enabling legislation says. The regulators are just responding to the incentives that we have given them.
When ThorCon tried to make the "let's balance nuclear risks against coal harm" argument to the Indonesian nuclear regulatory agency, one of the regulators had the courage and honesty to get up and say:
I don't care what the problems with coal are. I'm a nuclear regulator. My job is to make nuclear as safe as possible.
He's right. If ThorCon produces thousands of TWh's of dispatchable, cheap, clean, CO2-free electricity for Indonesia, he will get zero credit. But if ThorCon has a problem, his head will be on the chopping block.
There are leaders who are willing to take that risk. There are good people within the agency who will help. The rest can drag their feet, or even "leak" to the anti's that the agency is eliminating "bedrock safety regulations" to use Amory Lovins' words. They can't make that argument, if you are not eliminating anything, just changing the definition of "reasonable" to something that everyone will understand.
Sigh. If I had a dime for everytime I've heard this "it's not the system, it's the bad people" argument I'd be able to buy Musk. Would you call the Indonesian regualtor I cited a "bad person"? People responds to the incentives they are given. A good system recognizes that and takes advantage of it. I will have a piece on this "better pigs" argument soon.
I wouldn't call the old commissioners bad people, just misinformed or perhaps driven by their politics, or perhaps the bad incentives you cite. They can be replaced with people who are well informed and willing to risk the wrath of the anti-nuke industry. I wouldn't defend the system either. We are stuck with it, and calls to abolish it play right into the hands of guys like Lovins, who will claim we are eliminating bedrock safety regulations.
A bit of progress. The independent regulator is no longer a good system. It's now a bad system but "we are stuck with it". So we need to put self-sacrificing, public spirited saints in charge and they will make the other 3000 misdirected, wrongly incentivized bureaucrats risk their family's livliehood for the greater good. Youre offering us the counsel of despair.
We are not stuck with it. Pls read the Nuclear Reorganization Act in
I doubt that Chris Wright is "self-sacrificing", and I don't think it is necessary to get 3000 bureaucrats to go along with the new plan. I'm thinking maybe 10 members on a team, one team for each type of new reactor. The other 2900 can sit on the bench, if they wish, or perhaps run tests directed by the team members.
Replace LNT with SNT if you can get Trump's attention, but if not, the team should just insist that any allegation that a design is unsafe be supported by "risk informed" data and analysis.
I like your proposal for a Nuclear Reorganization Act, and all your other articles on an ideal way to regulate nuclear power. These ideals should inspire the people who will get things done within the agency. I just don't think these changes will happen soon enough in our political climate.
It's even worse than that. 0.1 mSv is 100 microsieverts. The UK ONR appeared to be concerned over 0.1 microsievert. Which is 0.0001 mSv. Which is 0.0000001 Sv.
Typical UK natural background radiation levels are in the range of 1000 to 4000 microsieverts. So the natural variation in natural background radiation is tens of thousands of times the level at which the UK ONR is making a fuss.
The dose models used to calculate the doses are certainly not accurate to 0.1 microsievert. The UK ONR is regulating the noise.
Yes, but if you believe in LNT, then one zoomie can kill you. How many zoomies in a microSievert? Quick Calculation: 1 eV is 10^-19 Joules. If the zoomies are 1 MeV average, then 1 microGray = 10^7 zoomies for every kg of you body. Now THAT is scary. :>)
If you believe in LNT and this Helen Caldicot line of reasoning, then 1 molecule of sugar can kill you too. The average adult consumes some 10e25 to 10e26 molecules of sugar per year, depending on your sweet tooth. Sweeties, we're doomed!
The UK regulator uses a variant of ALARA called ALARP. Mostly similar except there are 2 levels: a low level below which no action need be taken, and a high level above which no license can be given. In between the applicant would have to provide justification that the dose is acceptabable i.e. no further design or operating actions are reasonable. This is mostly grand standing, with the applicant providing some lame argument or engineering analysis to show it is all fine. After reviewing the dose limits it seems to me that the big problem is not so much the unlikely accidents, but the interpolation down to the normal operating level. Basically you can draw a more or less log line down with probability. You end up with a weird and unscientific standard that says it is ok for an unlikely accident to expose someone with 100000 microsievert, while not allowing 10 microsieverts a year of normal emissions. All because of probabilities that are made up and arbitrary on the unlikely end of things, and over conservative dose estimates on the normal emissions side of things. Partly this is because engineers are conservative bunch so they like to err on the cautious side. And partly because of LNT assumption. These amplify each other and you end up with death by a thousand cuts.
The main dose from turbine building effluent is noble gases. A regulator who demands a particulate filter to trap inert gases. So they don't understand the difference between a solid and a gas. Doesn’t instill much confidence in the UK regulator. And yet they allow dense packing and refueling with the containment open to air and both the active core and the hot buffer pool of freshly discharged densely packed spent fuel. Basically all large release risk is from the small fraction of time the containment is open during refueling, so it’s an unbalanced design. Plus station blackout is still a pretty ugly accident even with filtered venting. Regulator needs to get their priorities straight. I dislike a few aspects of the ABWR design but not having a solid filter for trace gas emissions in the non nuclear plant isn't one of them.
The real troublesome part is that the regulator gave the ok for them to proceed with a few ifs and buts that could clearly be met yet the applicant disbanded on vague "financial" reasons. This I suspect is where the crux is. If the investors and owners of plant are not convinced that going ahead with a licensed/licensable design then there is much more going on than a standard or a filter. The EPR builds show what is really the problem. The system ends up being so expensive that to construct it with the investors behind it requires endless financial guarantees, high interest rates, and massive CAPEX sandbagging to keep everyone and the supply chain happy to even get the go ahead. I wonder how many more Hinkley Points there must be for real change to happen. The Brits seem to have clink in their pockets yet so they seem unwilling to learn the real lessons.
As I understand it (may be wrong, reading between the lines a little here), the "financial reason" was the UK gov was willing to do a Contract for Differences (Cfd) where the gov guarantees the investors a price, usually well above current market. But this still leaves the investor with the risk that his cost ends up above the strike price.
GEH had learned how crazy UK regulation could be, and was uncomfortable with any politically acceptable strike price, so it demanded a regulated monopoly situation which transfers ALL the regulatory risk to the ratepayer. In fact, in such a situation the more regulatory costs the better for the investor. (One reason why there was so little push back from the US utilities during the disastrous 1968 to 1980 period when real nuke costs rose by a factor of five or better. All the US utilities were regulated monopolies at the time.) The UK gov at the time was not willing to go the RAB route, so the Japanese went home.
It appears EDF managed to negotiate RAB for the Sizewell C project, a clone of the Hinkley Point project. Yet the Japanese and Koreans don't appear to be warming up for new UK projects, it would seem.
The Japanese and Korean vendors' managements are still dominated by engineers. Engineers don't like doing stupid things, even if they can pass the cost on to the ratepayers. EDF is happy to work in this environment as is Westinghouse. Those will be the European choices. God help them.
I used to think the USA and the NRC were the most hopeless. I now think the US could lead the West out of this mess, but it has to come whileTrump is still in control. Problem is Trump has shown no real interest in nuclear. Somebody has to convince him that cheap nuclear power is essential for the "new Golden Age" and he can make that happen starting with the SNT for LNT EO.
That's definitely true. It is perhaps the saddest part of all - anyone, even Trump, can afford to ignore or spit on nuclear. And the US NRC can ignore Trump's letter. We got plenty of energy. And carbon emissions are nice for talks on conferences and birthday parties but don't amount to much else just yet. Plus most of the West has more money than sense. If necessity is the mother of invention, then we don't need to invent anything just yet. As Gordon Ramsey would say, "we ain't cooked yet". Hard to make real progress in this arena.
There is a huge risk of backfire in "repairing" NRC by EO. I have asked AI how many "infrastructure weeks" were announced from 2016-2020 and it couldn't give a straight number("many"), nothing serious in the form of infrastructure arrived anyways.
Now- EO with best ideas is announced tomorrow- SNT for LNT(hurray, finally some needed reason in that area!). How long it would take to see any benefits- any cheap plants building from the ground up? Fast enough before any changeover of power(unavoidable sooner or later)? Would vendors with licenses even appreciate such changes- higher CAPEX means higher profits (as long as one builds anything off course). Holders of licensed designs may scream against possibility of allowing cheap competitions, protecting valuable now "certificates"(isn't that best asset of say Nuscale?).
Change is needed, but to survive it must be demanded by the public, somehow bipartisan enough, not attached to controversial politician or single party(neither did much of value to cheap NP in last years, both could). If public will read it as "they want to poison us all for profit"(most because of being long conditioned to 2lies, some would use it to campaign) then backfire is all but certain.
You make a good point but offer only the counsel of despair.
Pls read the MNCA book.
After the EO, the second step is to redo all the EPA and NRC reg limits calculations repalcing LNT with SNT. This too can be done by quick administrative action while Trump is still around. All the limits would change by orders of magnitude. Yes the antis will scream and the administration will have to defend its actions. But this will lead to a much needed debate and for the first time the Two Lies will not be being parroted by the DOE, EPA and NRC as well as the antis.
It all comes down to do you trust the public common sense? I'm willing to take my chances. The antis are noisy but they are not representative. Latest polls are over 60% more nuclear. And we get 50 swings. It only takes a state or two to get the ball rolling, Are you listening Texas?
The alternative is do nothing and BAU. That's certain to get us nowhere.
There was a debate on twitter about whether the banana dose mitigation was specifically accurate. As I recall it wasn't exactly, but when I looked through the design assessment to find it the general critique was strongly supported, so I'll see if I can find it again.
The document I'm looking at is the "UK ABWR Generic Design Assessment" subtitled "Demonstration of BAT", which I think means best available technology. The section 5.2 about reducing environmental release including the filters starts at page 73 and starts out strong. Dedicated tritium and Carbon-14 off-gas mitigation were deemed "grossly disproportionate" to the public benefit and were not required. I'm not sure a more rhetorically broad criteria could be conceived of - saying something was not done only because it would be "spectacularly disproportionate" would raise eyebrows.
Compressed gas storage to hold discharges while radionuclides decay is interestingly mentioned as previously used by BWRs but no longer considered for new ones due to performance for cost. It says the ABWR would use activated charcoal beds. There is some discussion about the evolution of off-gas treatment where the compressed storage release reduction of radioactive noble gases is shown to be two orders of magnitude more than the charcoal beds if I read the chart right. The charcoal beds are said to reduce the release to 1/30,000 of the rate without treatment. This implies that the rate with compressed storage was 1/300. Effectively reducing this to zero with the beds means the radioactive release benefit of the charcoal beds is 1/300 - a minute fraction of a percent of biologically inactive noble gasses.
Other reasons are given for preferring the charcoal beds but this astonishingly weak justification in terms of radioactivity merits no comment in the report.
Section 5.2.3.2 "Evidence: HVAC Discharges" states: "Dose rates from HVAC discharges when HEPA filters are not installed are approximately 2.2ųSv/y" and that with them they would be 2.1ųSv/y.
"Thus it can be seen that the use of HEPA filters provides a very small benefit in terms of dose reduction during normal operation, however HEPA filters will be installed in the UK ABWR..."
Seeing this confirmation, I think I remember that the argument made was that it would help during a meltdown. No concrete justification for that appears in this section of the report, not even as a general reference.
Yeah. The banana dose is kind a silly. I have never tried to check it. More importatnly, its not dose; it's dose rate profile. Bad Jack. For the record, the biological half-life of K-40 in a human is about 30 days. The banana dose rate profile is 3% of the banana dose daily for 30 days.
BAT, Best Available Technology is showing signs of becoming the new ALARA. That's how EPA got rid of once thru cooling. I'll have a piece on this soon.
I was quite pleased to find out the banana dose was LNT correct, and that such a difference fit so well to the given job of the regulator that they felt no need to identify a broader justification than it was reasonably possible
I'll definitely use the banana dose line, keeping in the holster the fact that even that is a huge overstatement of how much practical dose rate it involves
Just a clarification, as this relates to BWR technology, there are three systems of relevance here: the offgas system, standby gas treatment system, and the turbine hall HVAC system.
Offgas system takes non condensables from the main condenser via steam jet ejectors. Since ABWR is a direct cycle there is some activity from leaked fuel pins in this offgas stream. Very little cesium, strontium and such as these stay in the water in the RPV, and the fraction that escapes is almost all retained by the steam separators and - dryers in the RPV. The activity is mainly inert gases - Kr, Xe, and some N16 from coolant activation. In the olden days, fuel failures were quite common so charcoal delay beds were installed in the offgas system to hold up any activity allowing it to decay. Nowadays due to improved process controls fuel failures are rare so the charcoal bed could be considered redundant. The offgas system also recombines hydrogen and oxygen, these are generated from radiolysis.
The standby gas treatment system is employed in case fuel damage has occurred, it can filter out fission products with filters, charcoal beds and the like, as well as recombining hydrogen and oxygen. Generally it can take suction from containment or from the refueling bay area above it (where fission products tend to leak to especially with containment open).
The turbine building HVAC is there to keep turbine hall air to reasonable temperatures, CO2 levels, and humidity. There is some activated nitrogen (N16) from coolant activation but that is contained and quickly decays without releases. The main issue being radiation shine which the ABWR fixes by having a concrete wall and access labyrinth around the turbine casings. After shutdown personell can enter this concrete wall enclosure.
Very little activity leaks into the turbine hall in normal operation. The turbine is sealed and any leakage is generally controlled via special vent paths or gland steam injection etc. So there is no point in having special filters here. Installing HEPA filters is pointless as almost all the activity is from inert gas. These are not stopped by a HEPA filter.
One thing LNT'ers feel comfortable with is using collective dose. The ICRP warns people against using it for risk assessment since it leads to alarmist predictions. However, the ICRP is OK to use it for risk comparison.
In the ABWR example, the dose to the most highly exposed person (usually someone who stands by the fence all year round, and in the location with highest dose) is 0.1 micro-Gy. The collective dose to people living in their homes is probably less than 1 micro-Gy from the HEPA. Compare this to the collective dose LNT considers to cause one fatal cancer, which is 20-50 Gy.
We need the government to change the mandate for the regulator. In the long term we would prefer a new dose-response model (SNT). In the short term the regulator must be forced to compare the societal cost of any requirement to the expected results of a release given its collective dose, probability, and the LNT estimate of harm (20-50 Gy required for a single fatal cancer).
I spent 4.5 hours on the NRC town-hall mtg on LNT just a few weeks ago. I can see It will be a steep climb to replace LNT with a better dose response model. Right now politicians hear that all of their constituents may get a lethal dose from something like a LLW site. That is horrible. Tell them that the most conservative model possible predicts 0.0001 of a single fatal cancer. That is obviously zero.
An EO replacing LNT with SNT would be much simpler and far more effective. Seeking out situations in which LNT predicts near zero incidence gets us no where whenwe have a release.
an LNTer should feel comfortable using collective dose. It is completely consistent with LNT and the ethical judgement that alllives are equal. What is totally inconsistent and nonsensical is ICRP's and others claim that LNT is a useful model but you cant use it when you need a model. It's a completely indefensible position.
By the way, an SNTer can use collective harm as a metric. All it takes is SNT plus the ethical judgement that all Lost Live Years are equal.
Outstanding as always, thank you. One minor quibble though. The Regulated Asset Base model was introduced by the previous Conservative government a couple of years after Wylfa fell through.
You are of course right that the new Labour government has adopted it without demur and it is the funding model for the second major reactor being built at Sizewell C.
The first one, Hinkley Point C, was funded by EDF and the government. Both are eyewateringly expensive at c. $14m per MW, despite Kepco apparently being around $2.5m per MW.
Rolls Royce are the preferred bidder for a fleet of SMRs, coming in at a chunky 470MW, but they're FOAK so it will be at least 10 years to first completion.
The existing fleet (c. 5GW)are due to be closed by 2030, with numerous extensions already, with one exception at Sizewell B. But the graph is pointing down not up at the moment, despite all the recent hype.
When you get in the weeds of everything the ONR has done, it all sounds superficially reasonable. It's only when your step back and look at the whole package that you can see how ridiculous the multiple layers are.
Anyway, more power to your elbow and thanks for all your insights.
Robert,
Thanks correction. RAB in its various forms has been around for a long time. If you think about it, the regualted monopoly is the only system that can work in a situation where the regulator's rules are unclear or the regualtor can change the rules as he sees fit. ALARA or ALARP is both. No rational investor can allow himself to be exposed to that risk. RAB allows him to trasnfer that risk to the ratepayer, who has no say in the matter.
The problem is now the limit on how far up the regulator can push nuke cost is the cost of alternative sources of electricity. If no other dispatchable source of electricity is legal as is becoming the case in the UK and elsewhere, that limit is very high indeed, impoverishingly high, civilization destroyingly high.
EDF made public claims that they had been required to make 7,000 changes from their pre-existing design which had been approved in France. They also argued that they had been required to use 25% extra concrete and 35% extra steel. In addition to the ventilation changes you mention, there was also a requirement to install a backup hardwired control system because the main system duplicated the software and therefore there could be a double failure. They also raised the sea wall and insisted on a second floor for the diesel engine buildings.
The ONR have pushed back quite firmly on some of these claims and argued that the overwhelming majority of these changes were approved during the GDA process and that most of the extra concrete etc was site specific due to local geography. They also argued that the sea wall changes and diesel engine elevations were post-Fukushima (the GDA was awarded in 2012) and that the hardwiring was internationally agreed as best practice and doubled down on the "hacking" risk. They basically denied imposing any real changes post GDA ('have remained largely unchanged since 2012'), instead blaming the Environment Agency and the Health & Safety Executive.
Less acceptable was the ONR response to the radiation filter which was, frankly, concerning: https://www.onr.org.uk/news/all-news/2025/07/our-assessment-was-not-bananas. The flippant nature of the press release suggests that the absolute catastrophe of GEH withdrawal from Wylfa has not been internalised by UK regulators. I accept that the primary reason was financing, but that does not absolve the ONR from some blame for the heavy handed regulatory environment which must have formed part of GEH decision making matrix.
What should really be happening is that the UK government should be in urgent negotiations with GEH to come back to Wylfa, pointing out that the very RAB model they were requesting is now in place. Of course, GEH may turn round and say that this kind of chopping and changing, plus some evidence of regulatory bone headedness creates a continuing unstable political and regulatory environment, and it would be hard to argue with that. Nevertheless, eating humble pie is something politicians and regulators need to do sometimes.
You are absolutely right that the deadline set by the Climate Change Act 2008 (as amended by Regulation in 2019) of Net Zero by 2050 creates a further looming problem for dispatchable power, particularly as the UK approaches grid level limits on non-alternative current sources of power in wind and solar etc. There is zero sense of urgency, despite government projections of peak demand increasing from 48GW to 72GW by 2030 and to ~100GW by 2035. Domestic non-intermittent supply is <40GW at the moment so, well, you can do the maths (no, not math, maths!).
ALARP is a pernicious ball and chain and has to be reformed. LNT is a equally ridiculous. But the good news is that there are straws in the wind that people in senior positions may be starting to question some of the most insane shibboleths. A new Nuclear Regulatory Taskforce recently issued an interim report. They even mention LNT - while they pay official lip service to its validity, if you read between the lines you can see they are at least aware of the pushback against LNT dogma. They are inviting evidence submissions by 8 September. I would encourage readers of this blog to submit evidence to that committee. I plan to do so.
I understand your cynicism about RAB, but when an industry is forced to invent a term for global project cost excluding financing costs, it is just not possible to continue with upfront models that are hobbled by the time value of money. The UK has a centralised regulatory model, so with a due nod to your concerns, RAB seems to be the least worst option to reduce the harm caused by the magic of compounding.
There are other potential options for improving the situation but that must await a paper I am currently drafting that I hope will be ready reasonably soon.
It's not cynicism. It's an indisputable fact. No sane investor is going to expose his capital to this nonsense, The solution is not to make the poor ratepayer bear all the costs of misdirected regulation. The solution is to change teh regualtory system to one that balances nuclear benefits and risks in a manner that is consistent with overall societal welfare. UCert is such a system. Pls read
https://www.amazon.com/dp/B0F32KLXRJ
Apologies, and withdrawn - even though I think deep cynicism would be an entirely appropriate reaction to the regulation of civil nuclear power in general and the RAB model in particular (whether in the UK or USA). I fully accept that your position is one of stating plain facts based on ample evidence, rather than cynicism.
Thanks for the link. I have read huge chunks of your blog, as well as the book you linked to, which happens to be sitting on my desk. I pretty much agree with your core theses, especially on LNT and ALARA, but also the insane DEGB and sequelae, your comparison with private shipyards and in general the over-regulation of the civil nuclear sector. I don't think the world is remotely ready to confront the second Big Lie though. That doesn't mean you are wrong of course, or that I disagree with you, but we are where we are.
I remain particularly struck by your account of the creative tension between regulators and constructors in the early years, and how nuclear competed successfully with coal when the latter was as cheap as it has ever been. I was also unaware until I read your work that the global limit for radiation exposure was 2mSv per day, and that this level is roughly 10% of the observable harm levels in the literature. I am also impressed by the 2km zone you suggest, for the reasons you give.
I have read your UCert suggestion and it seems intelligent and well thought out to me. My only concern is that it seems unlikely to be implemented any time soon. I propose to leverage a couple of the advantages that the UK system still has in order to try to expedite the construction of civil nuclear power here, if possible, and in the admittedly unlikely event that my suggestions land with decision makers.
Many thanks for the replies. I would be honoured if you have time to read a draft of my paper in due course - but I understand if you are too busy.
Robert,
UCert (rebranded as the Nuclear Reorganization Act for those who are into bureaucratic organization charts) is not impossible. It starts with replacing LNT with SNT. In the US this can be done by an EO. Don't know what the UK rules are but I very much doubt LNT was mandated by Parliament. So I suspect this can be done administratively, That's where your focus should be. With LNT gone the whole system will start to unravel.
Be happy to take a look at yr paper.
The newly formed Nuclear Regulatory Taskforce mention LNT in their interim report (https://www.gov.uk/government/publications/nuclear-regulatory-taskforce/nuclear-regulatory-taskforce-interim-report) and point the reader to IRR17 (https://www.legislation.gov.uk/uksi/2017/1075/contents) which is predicated on LNT. Such regulations are not full statutes passed by Parliament, and can be changed by administrative action usually, as you suggest. What is needed is the political will. I fear that in reality any change be may be slower and more incremental than that.
I think the evidence you put forward as to the inefficacy of LNT is compelling - but then I am not a scientist. The Taskforce, to their credit, highlight criticism of the view that LNT is accurate, but clearly feel obliged to endorse it anyway given current scientific consensus and the views of international bodies. I think it is very interesting that they mention that LNT has been criticised and that they also point out that the ONR have "Basic Safety Objectives" which are even more stringent than international guidelines. These are subtle hints, in my view, that they would welcome pushback. If you want a flavour of the scale of the challenge, check out this ONR press release: https://www.onr.org.uk/news/all-news/2025/07/our-assessment-was-not-bananas
The taskforce invites submissions pending their final report this Autumn (deadline: 8 September) and the guideline for such submissions is normally 3,000 words maximum. Email: nuclearregulatorytaskforce@energysecurity.gov.uk. A short explanation of your core arguments on LNT, with examples, could maybe have a real impact? I don't know, but it cannot do any harm, if you have time.
Many thanks for agreeing to take a look at the paper - what is the best way to send it to you in due course (I am still in the process of writing it).
I don't understand how ALARA can require these expensive changes. How can anyone think that saving 0.1 mSv per year is worth even 2 cents. It all comes down to the definition of "reasonable". Do we really need to abolish the agency, or can a simple change of leadership restore common sense?
Yes and no. Yes we must get rid of "independent" agencies to have balanced regulation. No, changing leadership will not make a material difference.
It's the system, not the people. The NRC, ONR, ASNR, etc purpose in life is minimizing radiation exposure. No other industry including some massive polluters requires an "independent" agency to regulate it. But radiation is a special horror, so we set up a special regulator whose job is minimizing radiation exposure, regardless of what the preamble to the enabling legislation says. The regulators are just responding to the incentives that we have given them.
When ThorCon tried to make the "let's balance nuclear risks against coal harm" argument to the Indonesian nuclear regulatory agency, one of the regulators had the courage and honesty to get up and say:
I don't care what the problems with coal are. I'm a nuclear regulator. My job is to make nuclear as safe as possible.
He's right. If ThorCon produces thousands of TWh's of dispatchable, cheap, clean, CO2-free electricity for Indonesia, he will get zero credit. But if ThorCon has a problem, his head will be on the chopping block.
There are leaders who are willing to take that risk. There are good people within the agency who will help. The rest can drag their feet, or even "leak" to the anti's that the agency is eliminating "bedrock safety regulations" to use Amory Lovins' words. They can't make that argument, if you are not eliminating anything, just changing the definition of "reasonable" to something that everyone will understand.
Sigh. If I had a dime for everytime I've heard this "it's not the system, it's the bad people" argument I'd be able to buy Musk. Would you call the Indonesian regualtor I cited a "bad person"? People responds to the incentives they are given. A good system recognizes that and takes advantage of it. I will have a piece on this "better pigs" argument soon.
I wouldn't call the old commissioners bad people, just misinformed or perhaps driven by their politics, or perhaps the bad incentives you cite. They can be replaced with people who are well informed and willing to risk the wrath of the anti-nuke industry. I wouldn't defend the system either. We are stuck with it, and calls to abolish it play right into the hands of guys like Lovins, who will claim we are eliminating bedrock safety regulations.
A bit of progress. The independent regulator is no longer a good system. It's now a bad system but "we are stuck with it". So we need to put self-sacrificing, public spirited saints in charge and they will make the other 3000 misdirected, wrongly incentivized bureaucrats risk their family's livliehood for the greater good. Youre offering us the counsel of despair.
We are not stuck with it. Pls read the Nuclear Reorganization Act in
https://www.amazon.com/dp/B0F32KLXRJ
The first step is an Executive Order replacing LNT with SNT. Trump could do that tomorrow. Let Lovins scream all he wants to.
I doubt that Chris Wright is "self-sacrificing", and I don't think it is necessary to get 3000 bureaucrats to go along with the new plan. I'm thinking maybe 10 members on a team, one team for each type of new reactor. The other 2900 can sit on the bench, if they wish, or perhaps run tests directed by the team members.
Replace LNT with SNT if you can get Trump's attention, but if not, the team should just insist that any allegation that a design is unsafe be supported by "risk informed" data and analysis.
I like your proposal for a Nuclear Reorganization Act, and all your other articles on an ideal way to regulate nuclear power. These ideals should inspire the people who will get things done within the agency. I just don't think these changes will happen soon enough in our political climate.
It's even worse than that. 0.1 mSv is 100 microsieverts. The UK ONR appeared to be concerned over 0.1 microsievert. Which is 0.0001 mSv. Which is 0.0000001 Sv.
Typical UK natural background radiation levels are in the range of 1000 to 4000 microsieverts. So the natural variation in natural background radiation is tens of thousands of times the level at which the UK ONR is making a fuss.
The dose models used to calculate the doses are certainly not accurate to 0.1 microsievert. The UK ONR is regulating the noise.
Yes, but if you believe in LNT, then one zoomie can kill you. How many zoomies in a microSievert? Quick Calculation: 1 eV is 10^-19 Joules. If the zoomies are 1 MeV average, then 1 microGray = 10^7 zoomies for every kg of you body. Now THAT is scary. :>)
If you believe in LNT and this Helen Caldicot line of reasoning, then 1 molecule of sugar can kill you too. The average adult consumes some 10e25 to 10e26 molecules of sugar per year, depending on your sweet tooth. Sweeties, we're doomed!
The UK regulator uses a variant of ALARA called ALARP. Mostly similar except there are 2 levels: a low level below which no action need be taken, and a high level above which no license can be given. In between the applicant would have to provide justification that the dose is acceptabable i.e. no further design or operating actions are reasonable. This is mostly grand standing, with the applicant providing some lame argument or engineering analysis to show it is all fine. After reviewing the dose limits it seems to me that the big problem is not so much the unlikely accidents, but the interpolation down to the normal operating level. Basically you can draw a more or less log line down with probability. You end up with a weird and unscientific standard that says it is ok for an unlikely accident to expose someone with 100000 microsievert, while not allowing 10 microsieverts a year of normal emissions. All because of probabilities that are made up and arbitrary on the unlikely end of things, and over conservative dose estimates on the normal emissions side of things. Partly this is because engineers are conservative bunch so they like to err on the cautious side. And partly because of LNT assumption. These amplify each other and you end up with death by a thousand cuts.
The main dose from turbine building effluent is noble gases. A regulator who demands a particulate filter to trap inert gases. So they don't understand the difference between a solid and a gas. Doesn’t instill much confidence in the UK regulator. And yet they allow dense packing and refueling with the containment open to air and both the active core and the hot buffer pool of freshly discharged densely packed spent fuel. Basically all large release risk is from the small fraction of time the containment is open during refueling, so it’s an unbalanced design. Plus station blackout is still a pretty ugly accident even with filtered venting. Regulator needs to get their priorities straight. I dislike a few aspects of the ABWR design but not having a solid filter for trace gas emissions in the non nuclear plant isn't one of them.
The real troublesome part is that the regulator gave the ok for them to proceed with a few ifs and buts that could clearly be met yet the applicant disbanded on vague "financial" reasons. This I suspect is where the crux is. If the investors and owners of plant are not convinced that going ahead with a licensed/licensable design then there is much more going on than a standard or a filter. The EPR builds show what is really the problem. The system ends up being so expensive that to construct it with the investors behind it requires endless financial guarantees, high interest rates, and massive CAPEX sandbagging to keep everyone and the supply chain happy to even get the go ahead. I wonder how many more Hinkley Points there must be for real change to happen. The Brits seem to have clink in their pockets yet so they seem unwilling to learn the real lessons.
AC,
As I understand it (may be wrong, reading between the lines a little here), the "financial reason" was the UK gov was willing to do a Contract for Differences (Cfd) where the gov guarantees the investors a price, usually well above current market. But this still leaves the investor with the risk that his cost ends up above the strike price.
GEH had learned how crazy UK regulation could be, and was uncomfortable with any politically acceptable strike price, so it demanded a regulated monopoly situation which transfers ALL the regulatory risk to the ratepayer. In fact, in such a situation the more regulatory costs the better for the investor. (One reason why there was so little push back from the US utilities during the disastrous 1968 to 1980 period when real nuke costs rose by a factor of five or better. All the US utilities were regulated monopolies at the time.) The UK gov at the time was not willing to go the RAB route, so the Japanese went home.
It appears EDF managed to negotiate RAB for the Sizewell C project, a clone of the Hinkley Point project. Yet the Japanese and Koreans don't appear to be warming up for new UK projects, it would seem.
The Japanese and Korean vendors' managements are still dominated by engineers. Engineers don't like doing stupid things, even if they can pass the cost on to the ratepayers. EDF is happy to work in this environment as is Westinghouse. Those will be the European choices. God help them.
I used to think the USA and the NRC were the most hopeless. I now think the US could lead the West out of this mess, but it has to come whileTrump is still in control. Problem is Trump has shown no real interest in nuclear. Somebody has to convince him that cheap nuclear power is essential for the "new Golden Age" and he can make that happen starting with the SNT for LNT EO.
That's definitely true. It is perhaps the saddest part of all - anyone, even Trump, can afford to ignore or spit on nuclear. And the US NRC can ignore Trump's letter. We got plenty of energy. And carbon emissions are nice for talks on conferences and birthday parties but don't amount to much else just yet. Plus most of the West has more money than sense. If necessity is the mother of invention, then we don't need to invent anything just yet. As Gordon Ramsey would say, "we ain't cooked yet". Hard to make real progress in this arena.
There is a huge risk of backfire in "repairing" NRC by EO. I have asked AI how many "infrastructure weeks" were announced from 2016-2020 and it couldn't give a straight number("many"), nothing serious in the form of infrastructure arrived anyways.
Now- EO with best ideas is announced tomorrow- SNT for LNT(hurray, finally some needed reason in that area!). How long it would take to see any benefits- any cheap plants building from the ground up? Fast enough before any changeover of power(unavoidable sooner or later)? Would vendors with licenses even appreciate such changes- higher CAPEX means higher profits (as long as one builds anything off course). Holders of licensed designs may scream against possibility of allowing cheap competitions, protecting valuable now "certificates"(isn't that best asset of say Nuscale?).
Change is needed, but to survive it must be demanded by the public, somehow bipartisan enough, not attached to controversial politician or single party(neither did much of value to cheap NP in last years, both could). If public will read it as "they want to poison us all for profit"(most because of being long conditioned to 2lies, some would use it to campaign) then backfire is all but certain.
msxc
You make a good point but offer only the counsel of despair.
Pls read the MNCA book.
After the EO, the second step is to redo all the EPA and NRC reg limits calculations repalcing LNT with SNT. This too can be done by quick administrative action while Trump is still around. All the limits would change by orders of magnitude. Yes the antis will scream and the administration will have to defend its actions. But this will lead to a much needed debate and for the first time the Two Lies will not be being parroted by the DOE, EPA and NRC as well as the antis.
It all comes down to do you trust the public common sense? I'm willing to take my chances. The antis are noisy but they are not representative. Latest polls are over 60% more nuclear. And we get 50 swings. It only takes a state or two to get the ball rolling, Are you listening Texas?
The alternative is do nothing and BAU. That's certain to get us nowhere.
There was a debate on twitter about whether the banana dose mitigation was specifically accurate. As I recall it wasn't exactly, but when I looked through the design assessment to find it the general critique was strongly supported, so I'll see if I can find it again.
The document I'm looking at is the "UK ABWR Generic Design Assessment" subtitled "Demonstration of BAT", which I think means best available technology. The section 5.2 about reducing environmental release including the filters starts at page 73 and starts out strong. Dedicated tritium and Carbon-14 off-gas mitigation were deemed "grossly disproportionate" to the public benefit and were not required. I'm not sure a more rhetorically broad criteria could be conceived of - saying something was not done only because it would be "spectacularly disproportionate" would raise eyebrows.
Compressed gas storage to hold discharges while radionuclides decay is interestingly mentioned as previously used by BWRs but no longer considered for new ones due to performance for cost. It says the ABWR would use activated charcoal beds. There is some discussion about the evolution of off-gas treatment where the compressed storage release reduction of radioactive noble gases is shown to be two orders of magnitude more than the charcoal beds if I read the chart right. The charcoal beds are said to reduce the release to 1/30,000 of the rate without treatment. This implies that the rate with compressed storage was 1/300. Effectively reducing this to zero with the beds means the radioactive release benefit of the charcoal beds is 1/300 - a minute fraction of a percent of biologically inactive noble gasses.
Other reasons are given for preferring the charcoal beds but this astonishingly weak justification in terms of radioactivity merits no comment in the report.
Section 5.2.3.2 "Evidence: HVAC Discharges" states: "Dose rates from HVAC discharges when HEPA filters are not installed are approximately 2.2ųSv/y" and that with them they would be 2.1ųSv/y.
"Thus it can be seen that the use of HEPA filters provides a very small benefit in terms of dose reduction during normal operation, however HEPA filters will be installed in the UK ABWR..."
Seeing this confirmation, I think I remember that the argument made was that it would help during a meltdown. No concrete justification for that appears in this section of the report, not even as a general reference.
Yeah. The banana dose is kind a silly. I have never tried to check it. More importatnly, its not dose; it's dose rate profile. Bad Jack. For the record, the biological half-life of K-40 in a human is about 30 days. The banana dose rate profile is 3% of the banana dose daily for 30 days.
BAT, Best Available Technology is showing signs of becoming the new ALARA. That's how EPA got rid of once thru cooling. I'll have a piece on this soon.
I was quite pleased to find out the banana dose was LNT correct, and that such a difference fit so well to the given job of the regulator that they felt no need to identify a broader justification than it was reasonably possible
I'll definitely use the banana dose line, keeping in the holster the fact that even that is a huge overstatement of how much practical dose rate it involves
Just a clarification, as this relates to BWR technology, there are three systems of relevance here: the offgas system, standby gas treatment system, and the turbine hall HVAC system.
Offgas system takes non condensables from the main condenser via steam jet ejectors. Since ABWR is a direct cycle there is some activity from leaked fuel pins in this offgas stream. Very little cesium, strontium and such as these stay in the water in the RPV, and the fraction that escapes is almost all retained by the steam separators and - dryers in the RPV. The activity is mainly inert gases - Kr, Xe, and some N16 from coolant activation. In the olden days, fuel failures were quite common so charcoal delay beds were installed in the offgas system to hold up any activity allowing it to decay. Nowadays due to improved process controls fuel failures are rare so the charcoal bed could be considered redundant. The offgas system also recombines hydrogen and oxygen, these are generated from radiolysis.
The standby gas treatment system is employed in case fuel damage has occurred, it can filter out fission products with filters, charcoal beds and the like, as well as recombining hydrogen and oxygen. Generally it can take suction from containment or from the refueling bay area above it (where fission products tend to leak to especially with containment open).
The turbine building HVAC is there to keep turbine hall air to reasonable temperatures, CO2 levels, and humidity. There is some activated nitrogen (N16) from coolant activation but that is contained and quickly decays without releases. The main issue being radiation shine which the ABWR fixes by having a concrete wall and access labyrinth around the turbine casings. After shutdown personell can enter this concrete wall enclosure.
Very little activity leaks into the turbine hall in normal operation. The turbine is sealed and any leakage is generally controlled via special vent paths or gland steam injection etc. So there is no point in having special filters here. Installing HEPA filters is pointless as almost all the activity is from inert gas. These are not stopped by a HEPA filter.
One thing LNT'ers feel comfortable with is using collective dose. The ICRP warns people against using it for risk assessment since it leads to alarmist predictions. However, the ICRP is OK to use it for risk comparison.
In the ABWR example, the dose to the most highly exposed person (usually someone who stands by the fence all year round, and in the location with highest dose) is 0.1 micro-Gy. The collective dose to people living in their homes is probably less than 1 micro-Gy from the HEPA. Compare this to the collective dose LNT considers to cause one fatal cancer, which is 20-50 Gy.
We need the government to change the mandate for the regulator. In the long term we would prefer a new dose-response model (SNT). In the short term the regulator must be forced to compare the societal cost of any requirement to the expected results of a release given its collective dose, probability, and the LNT estimate of harm (20-50 Gy required for a single fatal cancer).
I spent 4.5 hours on the NRC town-hall mtg on LNT just a few weeks ago. I can see It will be a steep climb to replace LNT with a better dose response model. Right now politicians hear that all of their constituents may get a lethal dose from something like a LLW site. That is horrible. Tell them that the most conservative model possible predicts 0.0001 of a single fatal cancer. That is obviously zero.
Ken,
An EO replacing LNT with SNT would be much simpler and far more effective. Seeking out situations in which LNT predicts near zero incidence gets us no where whenwe have a release.
an LNTer should feel comfortable using collective dose. It is completely consistent with LNT and the ethical judgement that alllives are equal. What is totally inconsistent and nonsensical is ICRP's and others claim that LNT is a useful model but you cant use it when you need a model. It's a completely indefensible position.
By the way, an SNTer can use collective harm as a metric. All it takes is SNT plus the ethical judgement that all Lost Live Years are equal.