Hacker News Reader

If the argument is, in part, about footprint then enough solar and battery capacity to output 300MW around the clock with the same uptime/reliability as a nuclear plant is surely going to cover a lot more ground.

I only bring that up because footprint was a point further up the thread.

There's still a fair amount of site work that has to happen here in the small modular concept, but I think when parts of it happen continuously in a factory you're largely immune from a lot of the jobsite nonsense that happened from contractors milking the job for every dollar they could get, plus economies of scale that you do get in a factory setting (citation: industrial revolution).

Nuclear reactor uptime is gonna be 90% at most, which is rather easy (battery capacity for several days) to beat even with just solar + batteries (in equatorial and most mid-latitude regions, at least).

And this BWRX design is based on ABWR which turned out to be only 70% in practice.

You don't get much useful notice for unplanned nuclear outages either, whereas both wind and solar output can be forecast with reasonable accuracy 24h in advance.

Five billion is just not all that much money. One guy spent eight times that to change the moderation policies of one social media website. Microsoft spent 14 times that to buy a video game company that they've largely mismanaged. Those two purchases alone could've paid for 20 of these reactors, maybe more if economies of scale kick in. Don't even ask how much Facebook is flushing down the toilet on VR or how much we're spending every single day in Iran. We could choose to spend our money on things like clean energy, but we choose other things instead.

It is for 300MW.

They're gonna throw in the factory that makes it for free tho.

No factory for civil construction projects. Nor I suspect for 300 MW (1000 MW-thermal) reactors. That puppy is getting constructed on site.

One could bring in modules and link them together. But then they tried that on the AP1000 to famously disastrous effect, at least at first.

Trying to minmax for the most cost effective renewable energy is fretting over spending nickles and dimes, while billionaires and megacorps are setting hundred dollar bills on fire. It's just not a productive place to be focusing your energy. We can afford both types of renewable energy, easily, we're just choosing to let others waste that money on garbage instead.

It's not nickles and dimes. It's 5x more expensive than a conservative estimate on 700mw of wind or solar

> It's not nickles and dimes

Yes it is. So far, we've flushed enough money down the Iran war toilet to pay for ten of these, and there's no end in sight.

I call this the "bigger rat" argument.

The rat, having been caught by the rat catcher, complains there are other, bigger rats. But he's still a rat.

The uneconomical project, having been called out, complains there are other, even more expensive activities. But that doesn't change that the project is too expensive.

Nah. Groups of people spending hours and hours debating whether it is smarter to buy the $1.95 or the $1.99 can of beans at the store is being silly. Just pick one. Or buy both! It's only two bucks and we spent $2,000 on flights to Europe last week.

What an absurd analogy. Of course it's worth spending time when deciding if billions of dollars are worth spending on something. It's worth entire individual lifetimes of effort if one can save that much money.

I identified ways to save 30 times that much money in this thread. It is more productive for you to focus your efforts there.

So, it's not going to cost $5B? Better inform those building it.

I just think it's silly to focus your efforts complaining about $5B being arguably somewhat inefficiently spent, versus hundreds of billions of dollars going straight into the garbage. There's far better things to focus on if you're concerned about money going to waste.

I mean it is, there is a project in Georgia for a new 1.4GW natural gas plant that has a $3.3 billion budget

But, the $5 billion here isn't purely an investment in 300 MW of capacity, it's an investment in starting up the nuclear reactor factory that can start churning these things out at $1-2 billion or whatever their goal may be

Also, have to take into consideration the overall operating costs. Nuclear costs significantly more up front but over time costs much less to operate (and is much more predictable) because you don't have to buy and burn natural gas forever

all cost modeling for nuclear takes this into account into the $/mwh of power. Fuel is cheap, but you have to pay back the enormous construction capex over time through power sales. Plus, staffing, security, insurance and maintaince are massive ongoing costs. Nukes remain the most expensive way to generate power.

The only cheap nuclear watts come from facilities built a long time ago in a completely different cost environment, and had construction and insurance subsidized by the state.

Natural gas plants all have security, insurance, maintenance, and staffing. Of course there are different levels of those but those costs are not unique to nuclear plants

Large natural gas plants have about 0.05 employees per MW. Nuclear plants have 0.5 to 1.0 employees per MW.

imagine how that number can be decreased as the plants are smaller, standardized, and more numerous. you can likely get a lot more efficiency out of the workforce when an engineer or operator can work across a dozen of these in a 200 mile radius instead of just one in the whole country. plus you need fewer training staff etc

Making plants smaller increases the number of employees per MW. Standardized might reduce the cost of construction, but how would it reduce the number of employees needed?

>but how would it reduce the number of employees needed?

A standardized reactor design that doesn't need a bespoke training control room at every single site means fewer training staff. It means the same maintenance workers and compliance workers and everyone else can be utilized more because they can cover more facilities instead of just the single site they are certified on and work at

or, you put a bunch of these small plants in one physical location and you get even more efficiency out of the workforce - only one facilties management crew, only one security team. And you'd even save money on just having one grid connection. Oh wait...

Those projected future savings are to be greeted with considerable skepticism. They are not locked in by contracts. They are the kinds of projections we've seen all too often fall apart in nuclear.

The underlying mechanic here is that cost projections are being used to sell a technology. As such, there is very strong incentive to underestimate the costs. This applies to FOAK plants and to projected experience rates.

The one thing that low carbon grids around the world share is high portion of nuclear (or hydro, which you can't build unless it naturally exists). Therefore, the only thing a SMR cost should be compared to is other nuclear reactors. Natural Gas may be cheaper (startup cost), may be more expensive (ongoing fuel), or will definitely be more expensive (carbon in the atmosphere doesn't go away), but regardless there is no reason to compare nuclear to natgas.

> The one thing that low carbon grids around the world share is high portion of nuclear (or hydro, which you can't build unless it naturally exists).

This is an historical artifact. When that generating capacity was built, nuclear was the alternative to fossil fuels. It isn't today; renewables are now cheaper and faster to install.

Your argument can be seen as a way to ignore the cost decline of renewables without at first glance doing so.

(As you say, there are hydro-dominated grids with no nuclear, for example Costa Rica, which gets 98% of its electrical energy from renewables.)

I was responding to the thread, which is comparing nuclear to natural gas.

Comparing to renewables is a different story - it can be done but it's much more complex (not dispatchable, seasonally variable) but certainly renewables makes sense in some or many places, (and in all places for a portion of the supply). I'm all in favour as long as it is actually done with the intent of getting to zero CO2, and doesn't just stall when it gets to the more difficult part of the transition. So many places get to 50% annual generation as renewable and throw up their hands about the remainder and leave it on coal/oil/natural gas

> I was responding to the thread, which is comparing nuclear to natural gas.

It's comparing nuclear to alternatives, one of which is natural gas. Others are renewables. One cannot make or justify a decision on nuclear without considering all the alternatives.

Weather dependent renewables (wind, solar) are not an alternative to nuclear, even when they are perceived in this way by the public.

From point of view of an electrical grid operator only wind/solar+fossil are alternative to nuclear, if you want to maintain stable electric grid without blackout or rolling blackouts.

https://en.wikipedia.org/wiki/Rolling_blackout

Sure. But the comment you replied to was that we spend more on things that remove Watts from the system.

> to change the moderation policies of one social media website

That guy wanted a particular US president elected; in this regard, the investment has likely already paid off.

With a nuclear power plant, the stakes are way lower, and the payoff is much, much more distant.

It is when you can get a 210MW S9G, along with an entire state-of-the-art fast attack submarine for 2.8 billion

I thought this was a sarcastic reply, so I checked the facts myself. I was stunned. You are spot on.

For those unaware, the submarine is Virginia-class[1] and is powered by a 210MW nuclear reactor called the "S9G"[2]. And yes, as of 2019 prices, each boat cost 2.8B USD. There are 28 of these subs active in the US Navy. This isn't some new, experimental reactor design. The US is pumping these out of their shipyards. Sheesh. 5B+ USD for a measly 300MW reactor... (without the nuclear sub!) looks way too expensive. What am I missing?

[1] https://en.wikipedia.org/wiki/Virginia-class_submarine

[2] https://en.wikipedia.org/wiki/S9G_reactor

You're missing the fact that a submarine tends to be submerged in the ocean, and that it is fueled with highly-enriched uranium.

Naval reactors and civil reactors look roughly the same at first glance, but as the folks at Three Mile Island found out the hard way, there are some rather crucial differences.

Weirdly, the "submerged in the ocean" part could be good for SMR to deflect the effects of a tsunami. This part is the most important: "it is fueled with highly-enriched uranium". Really, it makes me think, if the US military allowed it, what is wrong with a bunch of SMRs built in a low population density area using highly-enriched uranium? (For other readers, normally commerical nuclear reacors use low enrichment uranium: 3% to 5% uranium-235.)

That is 210 MW thermal. Which would produce about 70 MW electricity.

$2.8B for 70 MW is insanely expensive.

This one may be $5 billion, but the next one will probably be (made up number) $3 billion, and the next one $1 billion

Part of the point of these projects in particular is to get the machine spinning. Once it's running you start getting some of the economies of scale

This assumes learning rates never seen by nuclear energy. Within generations we've seen small learnings, and between generations the nuclear energy has been all negative learning by doing.

But that is literally why they want to get to assembly line levels of throughput. Dozens of identical reactors instead of dozens of bespoke reactors that can't use learnings from the last one

Also not said is the fact that going into more standardized designs you lower the operational cost because operations can be standardized. Today every single facility has different training from the next. Even at the same facility, if it's got new + old reactors, operators can't move between them without training on both

This is the biggest anticipated benefit of SMR. The US has 30+ licensed reactor designs. France has 3. Korea has 3. The economics of French and Korean reactors, built repeatably, are drastically improved over US reactors. If the NRC does its job and actually says "we know more about nuclear power than the local state energy commission and operator" then we the US can achieve those levels of repeatability and cost. We want safe reactors, we dont want infinitely customized reactors that are tailored to every state and operator's preference for how they want to polish fittings and lay out the pipes etc etc.

How many hundreds of billions or trillions should we waste to get ”learnings” when renewables and storage is already the cheapest energy source in human history?

This line of thinking seems to start with that we must have new nuclear, for some reason, and then try to rationalize it.

Hundreds of billions have already been spend on learning renewables in Germany.

https://www.researchgate.net/publication/321765366_How_expen...

In China you get the economy of scale for building nuclear power plants precisely because they are standardized. Thus, pivoting to SMR is unnecessary, since they produce significantly less energy.

    > In China you get the economy of scale for building nuclear power plants precisely because they are standardized.

Is this really true? Or how can we know it is true? For example: Why is the reason not explained by incredibly cheap construction labor costs compared to other rich countries?

Related: UAE had zero nuclear power plants before the Barakah nuclear power plant[1] opened starting in 2020. They built 4x Korean APR-1400s for only 32B USD. That seems incredibly cheap. How did they do it? I assume (near) slave labour prices for construction workers, like most other stuff built in that country.

[1] https://en.wikipedia.org/wiki/Barakah_nuclear_power_plant

How you raise finance is incredibly important for nuclear costs.

Higher lending rates can easily double the final cost and they multiply with any time delays.

That particular build had government finance from both the host and the building government.

Brandolini, is that you calling? There is so much bullshit here that I don't know where to start. The waters are murky...

    > How you raise finance is incredibly important for nuclear costs.

You really think the UAE, that has a sovereign wealth fund with assets of more than 2.5T (<-- trillion!) USD needs to borrow money to build nuclear reactors? Plus, Korea buys heaps of oil and gas from UAE. There are so many ways to "settle this bill".

> You really think the UAE, that has a sovereign wealth fund with assets of more than 2.5T (<-- trillion!) USD needs to borrow money to build nuclear reactors?

No, I'm answering your question as to why their build costs are surprisingly low. They didn't finance at the rate that other builds did and finance related costs can be 2/3rds of the total cost of a nuclear plant.

You cant make every part in the factory anyways. You need lots of on-site civil engineering, which accounts for a large chunk of the total cost.

And many parts of large (+1GWe) reactors have also been manufactured at off-site factories and then shipped on-site by barges in the past.

One issue here is that the structures containing the "nuclear island" are just as expensive as that island. Containment buildings are civil construction and are not cheap.

This suggests one should move to reactor concepts that don't need such large structures. The containment building size is dictated by the need to contain a certain volume of pressurized steam in an accident (and the requirement to contain the pressure of that steam dictates the mass of the building's structure). This is perhaps the strongest motivation for reactors cooled with molten salt.

Alternately, allow steam to escape in an accident, after filtering. Most of the radioactivity could be captured. But this violates current rules that require no release of radioactivity for 24 hours in an accident.

Agreed. The less nuclear-grade concrete you need the better.

The expensive part is not assembly, it is validation and documentation of that design and the lack of ability to spread those costs over multiple units. Site built units are fine so long as the design is sufficiently decoupled from site conditions that it can be exactly reproduced.