Hacker News Reader

> Nuclear and various fossil fuels create heat that boils water to generate steam that turns a turbine to generate power. This adds cost and complexity that you can never get away from.

That part is almost negligible. A General Electric LM6000 turbine costs about $20 million and generates about 50 MW of electricity. That translates into $400 MM per GW.

Solar comes to about the same price, but it has a capacity factor of only 30%, vs 98% for the GE LM6000 turbine.

98% not counting periodic overhaul. Then, 80-90%.

But you have to add the cost of the heat source. When that is a nuke, cost balloons out of sight. Geothermal is better, but drilling is still expensive; that might come down if microwave drills work out.

But the cost of periodic steam turbine overhauls will be hard to bring down to match dusting off PV panels. Pressurized steam is just very corrosive stuff.

That is why the mirror-array solar projects all failed. Steering thousand of mirrors, keeping them clean, and then driving a steam turbine just cannot match a static array of PV panels. The technology might have a place for supplying process heat, but it will be hard to match PV-powered fuel synthesis for cost and convenience.

Here’s a study by the DoE that shows that the capital and operating cost of a steam turbine that generates 15MW is about $666 per kW.

That would scale to $0.7 BN for a 1 GW plant, which is negligible for a nuclear power plant.

The steam turbine is definitely not what keeps the nuclear power plants from being economical.

[1] https://www.energy.gov/sites/prod/files/2016/09/f33/CHP-Stea...

nuclear doesn't use cheap small turbines like that. the output is far too large and they're nowhere near efficient enough.

a 1gw plant is currently about $7.5 billion, most of which isn't hardware but cost of capital, so the price you're claiming is 10% of total, which is hardly "negligable." Typical US buildout is 71% cost of capital, 15% equipment, 12% construction, 2% permitting, so, the number you're claiming would represent half of the entire plant equipment cost, which isn't correct

nuclear plants typically use either two or three very different kinds of turbines, which are an order of magnitude different in price - msbs, lsbs, and maybe elsbs if the plant is big enough

EDF's entire nuclear turbine business - they're one of the largest on earth - is only valued at 1.1 billion, which at your price is ... less than two turbines

https://www.nasdaq.com/articles/edf-seeks-to-lower-price-of-...

I can't follow your argument. You seem to imply at the same time that my numbers are too small and too high. Which one is it?

In any case, here's a separate reference to corroborate my numbers. It's a study by EIA ( US Energy Information Administration) published in 2020 [1] where they compare the capital costs of different power plants.

One such power plant they consider is a nuclear plant with 2 AP1000 reactors with a total capacity of 2.2 GW. The conventional part of the power plant is listed at $1.4 BN (page 107), which comes at about $640 MM/GW, in line with the estimate from my previous comment. Their overall estimate for such a power plant is about $14 BN, which means the conventional part of the plant is about 10%.

Of course, in the case of Vogtle 3/4 (which uses 2 AP1000 reactors), the overall cost is estimated at about $25 BN, due to various cost overruns.

[1] https://www.eia.gov/analysis/studies/powerplants/capitalcost...

> I can't follow your argument. You seem to imply at the same time that my numbers are too small and too high. Which one is it?

The devices are too small and the numbers are too high. (These things can be looked up.)

You know, kind of like how if you build a $300k machine out of A100s, and then try to match it with 3090s, you end up spending $800k.

.

> The conventional part of the power plant is listed at $1.4 BN (page 107), which comes at about $640 MM/GW, in line with the estimate from my previous comment.

Sorry, no, your comment was about the turbines, and the document you're linking to is about *the entire plant*.

Please stop trying to google-fight your way through this. You can't learn with your finger in the air.

The issue with turbines is not their capital cost (which is on par with the still falling all-included cost of PV) but rather operating cost. Steam turbines need frequent, expensive overhauls because superheated steam under high pressure is very hard on everything inside.

Renewables suffer no opex of such scale. They age out and need to be replaced, on a time scale of decades, but will be much cheaper to replace, then, than they cost up front. Probably perovskites will be used to replace silicon.

Both their initial and their replacement cost is wildly higher than nuclear.

Comparisons without a null hypothesis are defunct by nature.

> 98% not counting periodic overhaul.

the reason for the null hypothesis is that if you had run one, you'd realize that solar has a 20 year periodic replacement, as compared to nuclear's 80 year periodic maintenance, so this argument (like all arguments) is against solar

.

> When that is a nuke, cost balloons out of sight.

Not when compared to literally any other power source

But if you hold it up in isolation, yeah, sure, it's expensive. Once every 50 years because of bad laws, you have to go in there and pay about a quarter what the plant was worth.

And it'd still be cheaper if you had to do it every ten years, plus more reliable, and it wouldn't waste all the copper and rare earths, and it wouldn't leave you with China as the new Middle East.

.

> But the cost of periodic steam turbine overhauls will be hard to bring down to match dusting off PV panels.

You don't dust PV panels off. That's daily cleaning.

Maintenance for PV panels is replacing them every 12-23 years depending on what brand you bought.

An honest discussion would really help.

.

> That is why the mirror-array solar projects all failed.

No it's not. Also, the mirror-array solar projects do better than the PV ones, if you take the time to look it up.

The reason the mirror-array solar projects failed is, like the PV projects, they bought into the myth of storage.

.

> but it will be hard to match PV-powered fuel synthesis for cost and convenience.

This is the only good reason to use PV, and it's because the end product that's desired is a chemical battery

Batteries are the only thing PV can fill, and gasoline is the only kind of battery anyone actually wants to buy

In the meantime, the idea of using PV for this is absolutely absurd. It's 1/5 the price to do it on nuclear, and you know, it turns out that people are sensitive to the price of gasoline.

If you do an actual comparison with actual numbers where all the solutions are checked for every claim, instead of just the one you want to make look a certain way, suddenly you realize why nobody's building this despite that all the parts are cheap and unregulated, and despite that the market demand is vast and well established.

If this was true, instead of posting about it, you should be building a proof of concept, then applying for a bank loan, so that you can get rich saving the planet.

Cheap off the shelf parts don't do world-saving things, in general, or else someone would have already done it. If you believe that you can do something the world's scientists and engineers are desperately struggling to do using stuff you can buy at WalMart, there's a very solid probability that you're missing some key pieces in the evaluation, and should look up one of the publicly available studies.

> Nuclear ... create heat that boils water to generate steam that turns a turbine to generate power.

One of my greatest disappointments as a kid was learning this. I'd thought nuclear power somehow got the power of the atom directly to a wire/grid.

Maybe RTGs [1] that are used on some spacecraft will reappoint you.

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

Love the comment and the use of the word “reappoint”. Never thought of the root of the word “disappoint” prior

+1

Aneutronic nuclear fusion with direct energy conversion is the dream

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

the kid in you wants to hear "it's basically a steamboat"

Now I hope this is the dystopia with nuclear paddle-wheel steamboats.

i mean. that's kind of what us nuclear submarines are, just, all jules verne-y

and without the action of a casino

Solar power is intermittent. This make it far more expensive for practical, real world applications in large scale. Nuclear is the only thing that can realistically substitute fossil fuels. Solar is at best a niche due to the storage needs.

Re-check your sources. A lot of pro-nuclear material out there is pure advertising from nuclear lobbies. Nuclear is a hugely expensive and high-maintenance way of generating energy. There's risks that don't show up in neat, averaged numbers, think Russia shelling a nuclear plant in 2022, that was on none of the yay-nuclear bingo cards.

GP poster is correct.

Instead of talking about propaganda, you should go check the sources you're telling other people to check.

South Korean energy does cost circles around subsidized wind and solar, and doesn't need ridiculous battery infrastructure.

Gen-3 PWR nuclear is, in reality, by far the cheapest form of power on Earth, and you'll never find any hard evidence from respectable sources saying otherwise. (Probably that's why you're poisoning the water with claims of lobbies and bingo cards.)

.

> There's risks that don't show up in neat, averaged numbers

No, there aren't. Nuclear has killed fewer people all of in history than any other form of power generation in its worst month. Almost no industry in human history, power generating or otherwise, has a human death safety record anywhere near as good as nuclear power's. (I can't think of a single specific counter-example. Can you?)

The worst nuclear disaster in history, Kyshtym, which most people have never even heard of, didn't kill as many people as a bad bus accident.

Please bring hard evidence with your next set of claims. No, I don't mean estimates from the 80s by non-doctors that never panned out.

.

> think Russia shelling a nuclear plant in 2022

Zero deaths. You wouldn't expect that from other forms of power plant being shelled.

Your examples work against you.

There's a reason why when we talk about deaths from other forms of power we talk about actual deaths, but when we talk about deaths from nuclear power, we talk about risk that hasn't occurred and what might maybe happen someday

If you look at the numbers, and say "yes, but in my imagination," then you're doing the wrong thing

You say "check your sources" but you haven't actually done that yourself, yet

You talk about "advertising from nuclear lobbies" but nuclear fans' sources tend to be the government

That is of course wholly false. It is why solar farms are going up all over the world, counted in thousands, while nukes are not.

Storage is cheap. We have myriad alternatives, according to local requirements. Efficiency used to be a worry, but solar has got so cheap, you just add panels to make up the difference.

> Storage is cheap.

And yet it's solar being built all over the world, not storage.

Storage is also being built.

But storage will lag in volume until intermittent sources like solar produces enough in any given region that there are significant periods of excess that aren't better to deal with by bringing other plants offline.

Right, it would be stupid to spend big on storage that cannot be charged up from renewables that exist. First you build enough renewables, to displace fossil fuel (and opex), then build storage when you cannot displace much fossil fuel, any more.

All this is easy for an operator to figure out, daily: add solar, fuel cost falls in exact proportion. When you have to turn off enough banks of panels enough of the time, you start building stuff to absorb that energy for later. But not until building more panels cuts your fuel cost less than building storage cuts your night-time fuel cost. (Wind input makes this calculation less deterministic.)

Exactly. There is some benefit to building some storage before then to reduce the need for peaker plants, and that is largely what happens today when storage gets built. E.g. Tesla's battery project in Australia a few years back for example.

Plus, if this catalysis system works, then excess electricity can power the process to convert methane into methanol as a chemical battery store.

That, combined with the ability to extract CO2 from the atmosphere and convert that into Methane would result in carbon neutral (or even negative, if any of the converted CO2 carbon breaks away from the O2 in the burning process as soot) any time power delivery.

Yes, you would be burning the methanol in an engine to turn a turbine but if the carbon for that process came from the atmosphere to begin with then who would be upset by that?

You could be using the excess power to actually sequester the carbon you extracted, and help fend off looming climate catastrophe.

But selling it let you pocket a few bucks, so the choice was obvious.

They should tax half the carbon you extract as if it were mined, unless you sequester it instead.

Burning hydrocarbons extracted from the atmosphere is carbon neutral assuming the electricity required for the conversion is solar/wind supplied.

These would also count as "direct" power generation, I think (how should it be defined?). None of them work at scale, yet.

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

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

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

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

> Nuclear and various fossil fuels create heat that boils water to generate steam that turns a turbine to generate power. This adds cost and complexity that you can never get away from.

Back here in reality, nuclear is much cheaper than solar, once you ignore the subsidies and the US-specific legal overhead

More importantly, it's base load, meaning you don't need the batteries that are riotously more expensive than water boilers

.

> I actually think solar is and will be the most important method of power generation in the coming centuries

Wind and solar are older than the grid, and have never been able to contribute

Energiewende failed

Humanity has no realistic non-nuclear future, and people should be required to look at the economic numbers before discussing this

.

> that will culminate in space-based solar power collectors.

Found the person who grew up on SimCity

"Boiling water adds too much complexity so let's build an infrastructure in orbit"

.

> Creating methanol is essentially a way of storing excess energy

One of many. There are much better ones. We generally don't use methanol as a chemical battery, even though it can be, because we have such radically better options.

Methanol is almost solely consumed as a fuel or as an intermediate product from fuels

Besides, the great thing about base load is that when you have it, you don't need to store excess energy. You just create the amount you need

You're solving problems that don't need to exist, while calling the normal way "too complicated"

> reliable

This is not usually a word that one uses to describe something that works when it wants to, as opposed to things that work when you want them to

And even more so if the plants which produce solar products use solar power to offset. I wonder if they do that or not. I'm guessing it's not quiet that simple and distributors of the individual components vary in their methods.

Well, hydro or wind power don't have their own heat engine (the planet of course does stuff with sunlight that ultimately moves the turbine blades, yes)

Dyson sphere style