23 Comments
User's avatar
David MacQuigg's avatar

Why are we still building PWRs? What is the "should cost" of an APR1400 (materials and labor)? Leaving out the turbines, generators, and all that is common to every power plant, what is the should cost of an MSR?

Expand full comment
Jack Devanney's avatar

David,

You can't leave out the turbine, etc. in your costing. A Light Water Reactor is limited to lousy steam conditions, so you end up with a thermal efficiency about 32%. Molten salt and otehr high temperature designs can produces supercritcial steam conditions, and you end you with

an efficiency around 45%. Not only do you have a much smaller turbine for a given electrical output but in order to achieve the 32% the light water turbine needs massive demisters and reheaters. The turbine hall, the condensers, and any cooling towers are quite different for a LWR and a high temperature reactor. And thats before we get to other power cycles such as supercritical CO2 etc.

The Czechs and others are building APR1400's because they are the best of what is actually available. It's a smart choice.

Expand full comment
David MacQuigg's avatar

Good point on cost of turbines, condensers, etc.

I will take a guess on should costs, scaling up from ThorCons $800 - $1000 per kW current cost estimate, adding 20% for a reasonable overhead and profit. Let me know if you have a better estimate.

Let's assume the nuclear part is 30% of the total plant cost. For a PWR, let's add 25% for the materials and labor on the non-nuclear parts.

On the nuclear side, we have to add a big containment dome, a massive pressure vessel, and a bunch of redundant safety systems. Let's call that 100%.

With these numbers I get for a PWR a should cost of $$1593 per kW.

Here is a link to my calculations, if anyone would like to edit or comment.

https://docs.google.com/spreadsheets/d/18p5IN51i6myPjO0bce4oFw8Bz6IJofSGsBF1pIBZH9U/edit?usp=sharing

Expand full comment
Mike Conley's avatar

I wonder what KHNP would bid to build an APR-1400 here on US soil.

Expand full comment
User's avatar
Comment removed
Jul 18
Comment removed
Expand full comment
Mike Conley's avatar

Actually, NRC policy reform would probably be the best and most direct way to clear the log jam.

Expand full comment
Jack Devanney's avatar

Brain,

Mike's right. The IRA may make the situation slightly worse by using taxpayer money to allow the NRC to push nuclear's cost even higher, but the core issue is an regualtor with no checks or balances.

Canada has much the same system. They have not built a nuke in over 40 years. I

Furthermore, this is not the place for partisan name calling. Reword your comment or it will be deleted.

Expand full comment
Jack Devanney's avatar

Mike,

We dont know what the terms of the Czech deal are

but I have learned that the UAE contract was firm, fixed price

and the Koreans not only met all the milestones,

but earned significant bonuses. There was a delay

in actual startup. It is rumored that this was due to

a struggle between two parts of the ruling family,

but that's just a rumor.

The Koreans are not stupid. They wont sign a fixed price deal

unless they are insulated from regulatory uncertainty and changes.

Such a deal would be impossible under NRC style regulation.

Expand full comment
Mike Conley's avatar

Thanks, Jack. Pretty much as I thought. But could you please expand on the last part, where you wrote: "[The Koreans] won't sign a fixed price deal unless they are insulated from regulatory uncertainty and changes. Such a deal would be impossible under NRC style regulation."

How would this play out, exactly, with an experienced firm that works from complete blueprints of a design approved for US build, if the firm could also ensure themselves that they would have a reliable supply chain?

Or should I ask: How in the world could / would the NRC (inevitably) screw this up? The Koreans obviously know what they're doing, and can deliver as promised. They're as close to a "should cost" for LWRs as it gets.

Expand full comment
Jack Devanney's avatar

The NRC system is unacceptable in two ways.

1) The NRC can simply change the rules in mid stream. This happened to the AP1000 in Georgia with a new aircraft impact rule. The NRC says they wont change the rules after you get your COL but the NRC can change that rule too. Section 187

of the AEA explicitly gives the NRC the power to change the rules whenever they want. A recent example was the rescinding of the Turkey Pt extensions they had already granted when the Commissioners under pressure from anti-nuke NGO's reinterpreted the word "initial" to mean "not just initial". There's no appeal from this arbitrary nonsense.

2)Even if you have a compete detailed design and well established rules there are always gray areas. When we build a ship, there were constant "discussions" between our inspectors and the yard guality control guys about whether this or that job was to spec. Was this weld ground smooth enough? Was this surface blasted to SA-3? It looks a little dark to our guy, but the yard thinks its good enough. Almost all these disputes are resolved on the spot. Occaisionally they are bucked up to owner and yard management where the issue is thrashed out. The reason why these issues almost always get worked out is the arbitration process. If they cant be resolved, we go to arbitration, owner gets 1 arbitrator, yards gets 1, and the 2 arbitrators pick a third, and the three make a decision by majority vote. If either side is being unreasonable they will lose the arbitration.

But in the NRC case, the NRC's inspectors opinion is ex cathedra. It cannot be challenged. And if you piss him off by pushing back, it will just get worse. The Koreans are way too smart to put themselves in that position.

Expand full comment
Jesse's avatar

Apperently APR1000's, or a reduced output version. So I think more like $4/W if the price holds...

Expand full comment
Jack Devanney's avatar

Jesse,

That part does not make sense to me. I don't think the Koreans have a 1000 MW design, and why would they derate the 1400? Will see, but maybe the reporting got garbled.

Expand full comment
Jack Devanney's avatar

I stand corrected. The Koreans do have a 1000 MW design and it has been certified by the EU. So it is looking like $4000/kW. Still unclear why they did not offer the 1400 when the Czechs have an appetite for 4GW's.

Expand full comment
Jack Devanney's avatar

Another mistake. EUR Certified, not EU. EUR is a collection of European utilities.

Expand full comment
Jack Devanney's avatar

Yeah, it gets worse and worse. So now we are at 8.6 billion for a single APR1000.

I should have stayed in bed.

Expand full comment
Niels Harksen's avatar

It is the APR1000 and 8.6 billion is per reactor. Unfortunately. For your numbers, we need to build multiple reactors per year.

Expand full comment
World of Abundance's avatar

$4000/kw x 8000 (8760 Hr/yr @90%) = $0.50/yr cap cost or $0.025 over 20yrs. Seems cheap for what you get.

Expand full comment
Jack Devanney's avatar

World,

We are up to $8600/kW after correcting for my multiple mistakes. Even if that number holds up, you have to discount the future earnings, so we are north of 5 cents/kWh for CAPEX alone.

Expand full comment
Jack Devanney's avatar

Dan,

The Westinghouse position is a total crock, but an interesting example of the US nuclear establishment (DOE, incumbent vendors, NRC) colluding to protect themselves. See

https://jackdevanney.substack.com/p/the-westinghouse-kepco-suit

Expand full comment
Jack Devanney's avatar

Better version at

https://gordianknotbook.com/download/he-westinghouse-kepco-suit

Here's the paragraph that was left out of the substack piece,

The APR1400 is based on the Combustion Engineering (CE) System 80+ design.

In 1997, Kepco licensed that design.The Koreans demanded and got a Total Technology Transfer Agreement.No strings attached.They could use the IP anyway they wanted.

CE's back was against the wall.They had become embroiled in an asbestos suit,

and were going bankrupt.The cash from the Koreans kept CE alive for another couple of years.In 2000, Westinghouse acquired CE and became the licensor.

Kepco has bought the unfettered right to use the System 80+ IP.

How can Westinghouse keep them from exercising that right?

Expand full comment
Daͷ Müιιεя's avatar

Insightful piece!

So there are two vectors for Westinghouse to pursue for blocking the deal:

1. Possible IP law

2. Part 810 requirements

Even if the Total Technology Transfer Agreement gets disputed, Point 1 should no longer be valid, since patents expire after 20 years from application.

And it looks like Westinghouse already had their IP case dismissed last year:

https://world-nuclear-news.org/Articles/US-court-dismisses-Westinghouse-case-against-Korea

Regarding Point 2, you write "No one can export any American nuclear technology without complying with the Atomic Energy Act, Part 810".

So the question is, does expired IP still count as "American nuclear technology"?

In the extreme, this could mean that even an old sodium-graphite reactor still would fall under Part 810?

Expand full comment
Jack Devanney's avatar

Dan,

I'm no lawyer and if I were the Koreans I'd simply ignore this nonsense. But if for whatever reason the Koreans decide to fight this in American courts, I'd point out that the DOE claim that only an American can sign an 810 letter, then the "export" in 810 must be referring to "export from the USA". The Koreans are exporting nothing from the USA.

Expand full comment