Hacker News Reader

Comment by photochemsyn | Hacker News Reader

Costs for nuclear don't really benefit that much from economies of scale, each installation has to be heavily reinforced and the required equipment is very expensive (you really don't want primary coolent pumps to ever fail, so they're over-engineered to essentially eliminate accidental failure - although exceptional situations e.g. Fukushima are still always a possibility). Also, the global uranium fuel pipeline is limited (US imports most of its uranium fuel doesn't it?).

Building 500 1 gigawatt solar/storage plants would be much cheaper - by at least a factor of ten - and could be done much faster.

View on HN

1 gigawatt solar is not the same as 1 gigawatt nuclear though. Also, building solar in Alaska or in Northern Canada would be a waste of money.

But yes, a solar power plant is cheap and can be implemented almost instantly, it is a huge advantage.

The right comparison is roughly 4GW of solar plus 6-12GWh of storage for 1GW of nuclear, but even then nuclear is about 2x as expensive.

>4GW of solar plus 6-12GWh

It will use a lot of land and I'm not sure if anyone ever did 6GWh storage, not even mentioning 12 GWh. The Li-ion battery technology that is currently being used for storage is very expensive and will last maybe 15 years at best before becoming chemical waste, while the nuke is normally planned to last 60+ years.

And that is just to match 1GW of nuke. Now, if you want 100GW, you multiply that by 100. Now solar isn't looking all that advantageous or even technologically feasible (with storage factored in) to me.

I don't think you have a good idea of scale here 4GW of solar and 12 GWh of storage are a tiny fraction of what we produce every year, barely worth mention.

Whereas even a single GW of new nuclear is newsworthy because it was such a monumental effort, over budget and behind schedule by factors of 2-4x

Wind, solar, and storage get delivered on time, under budget, and at a scale that nuclear can not even dream of achieving for decades. Claiming that 100GW/year of nuclear is even in the realm of possibility is beyond delusional, whereas 400 GW/year of solar is close to happening, and we will deploying TW/year in the 2030s, before a second site of nuclear could even come online in the US.

Land for solar is similarly a red herring. Solar "land use" is easily dual purpose, if need be, but it need not be. Using even a fraction of land that we currently use to grow corn for ethanol would completely power all of our transportation.

According to a recent report, there was only 13.4 GWh of grid storage in the US 2022.

Even with projected growth, each addition 12 GWh is not trivial.

https://www.smart-energy.com/regional-news/north-america/us-...

12 GWh is a fraction of a single factory's output from a single year. We are producing >50GWh total for storage per year, and the only reason more of it isn't going to the grid is that grid storage is a much less lucrative use of our storage capacity compare to cars, which have higher margins.

The entire industry is growing at 10x capacity every five years!

Don't forget that this single 1GW of nuclear took more than a decade.

We simply can not scale nuclear to meet our needs, it is not a technology that scales well at all.

For example, we have roughly 100GW of reactors in the US rapidly reaching their end of life. Even if the industry put all its effort in rebuilding fast enough to replace existing reactors, it would require a scale up that we can't possibly imagine right now. Meanwhile, we are already scaling storage, by building factory after factory.

> The entire industry is growing at 10x capacity every five years!

Grid storage is projected only to grow 10-17 GWh per year according to the aforementioned article.

Is that anywhere enough, if the aim is to decarbonise?

And there's zero new expected nuclear after vogtle 4...

If we add 10-17GWh in 2023, it would be foolish to think we will add the same amount in 2024. The amount added per year will grow too.

We are at the beginning of the logistic curve, where growth is exponential. A few years ago, there wasn't even a single GWh on the grid.

For more than a century we have balanced the grid by keeping energy stored in fuel; as we switch to cheap renewables, that's the only time that grid storage becomes necessary, and you have to get to pretty high levels of renewables before it makes economic sense to add storage. So the market is small now, and will grow. Until then, those lithium ion batteries are being shoved into cars, because car makers can not acquire them fast enough to meet demand.

There are several other very promising and quickly advancing battery technologies as well, based on iron or sodium instead of lithium. We will see which get cheap, but until there's high amounts of renewables.

> If we add 10-17GWh in 2023, it would be foolish to think we will add the same amount in 2024. The amount added per year will grow too.

According to the Woodmac report I referenced, the amount of grid storage added yearly is going down.

12 GWh in 2023 10 GWh in 2024

Based on what do you think the report is wrong and where is the exponential growth in grid storage going to come from?

Now be honest and tell us what Woodmac is projecting for storage by 2030. I'll wait.

Why are you asking me? I am not Woodmac. Furthermore, their projections only go to 2027.

I’m sure they will commission a custom report for you if you compensate them enough.

Either way, 2030 grid storage projections are highly speculative. Projections for the next five years are based on announced and planned projects. Nobody is announcing or even planning 2030 deployments yet.

Besides that, what’s your point? Are you trying to say that the previously touted exponential growth is only going to start after 2027? In 2030 perhaps?

As an aside I find it amusing that you or somebody else downvoted me for quoting a report and asking where the exponential growth that contradicts the report findings are going to come from.

Kinda; cars need to decarbonise as well as the grid, so the batteries are still really important even if none end up on the grid for whatever reason.

Casually suggesting we add the equivalent of our nation's supply of grid battery storage to get the equivalent of a single nuclear plants worth of energy is a new one.

In an industry that is an exponential growth phase, and basically a side product where the majority of its product goes into cars, yes, this is indeed a very casual suggestion.

Far more causal than the suggestion of building a single new GW-scale reactor, when they have just proven to have a 50% success rate on build, and come in several times over budget and behind schedule.

Until the nuclear industry shoulders the construction risk themselves, there's little chance of another utility making the decision to build a new reactor.

Is it?

Given how fast reactors are getting built in the US, I'd say the batteries look like they're already doing much better than the reactors.

Of course, both can change as the number are, while interesting, small.

Even if we really could ramp up those factories and the corresponding mines for all the battery materials, those batteries will last maybe 10 ish years. Then they need to be recycled and we have basically no capacity for that today. A nuclear plant can easily go for 60 to 80 years with good maintenance.

> Then they need to be recycled and we have basically no capacity for that today

Do you have any reason to believe making a batteries starting with batteries will be harder than making batteries starting with rocks?

What do you mean "even if"? Can you name a single battery factory that has seen delays of cost overruns like nuclear regularly has? Can you point to a single investor that's willing to back new nuclear in the US like they are backing new battery factories all over?

Your usage of numbers is off and extremely biased. Warranties for grid storage batteries are already longer than 10 years, and they don't suddenly stop working, they just experience finished capacity over time. Whereas 80 years for nuclear has no basis in reality, and most are getting shut down at forty due to excessive costs to continue operation.

This sort of rose-colored view of nuclear, cherry picking the very best case from one's imagination, and then taking the very worst possible estimate of the comparator, is the type of self-deception that results in nuclear losing again and again. Without a clear view of why these projects fail, and what they actually can deliver, the entire industry has lost any credibility.

> Can you name a single battery factory that has seen delays of cost overruns like nuclear regularly has? Can you point to a single investor that's willing to back new nuclear in the US like they are backing new battery factories all over?

I mean, those aren't the important parts in my view but yes to both. Berlin gigafactory had significant delays and the proposed expansion is on indefinite hold. Bill Gates is a large backer in Terrapower and they have a pilot plant planned in Wyoming. But I'm sure you will simply move the goal posts.

> Warranties for grid storage batteries are already longer than 10 years, and they don't suddenly stop working, they just experience finished capacity over time

I did say ten-ish, and good luck collecting on those warranties since the battery manufacturer corps are pretty ephemeral. Also, I kind of doubt operators will want to keep old batteries in their facilities since you're at increased risk for fires and those tend to be pretty dramatic in BES stations. Especially since in theory newer batteries should be better and cheaper right?

As for the 60-80 year lifetime, the Gen II plants that are worth keeping around are just now getting close to their 60 year birthdays. Most of the shut down plants in the US are gen I with small capacities, usually less than 200 MW. And the ones that aren't often fell to local and private interests that wanted them gone for other reasons. The problem with nuclear in the US is not a technical one but a socio-political one. I have no idea if the turning tide of sentiment will be able to fix these issues but I find more people are coming around fairly recently. Thank god for France, South Korea, and China.

We're building batteries and panels at scale now? When did that start?

The US will add more nameplate power capacity of batteries to the grid this year than it will new gas generators:

https://arstechnica.com/science/2023/02/us-will-see-more-new...

(Noting of course that power characteristics of entirely different tools can be an apples an oranges comparison... but look at solar versus fossil fuel generators in that.)

Depends on the definition of scale, but the 2010s was where it was in GW added every year:

https://www.pv-magazine.com/2023/04/17/the-world-installed-2...

And are "we" building that capacity? Who is "we" for you?

There are two definitions of "we" in the studies I linked, one for the US and one for the world.

What definition of "we" did we have in mind, and if you have a different one why didn't you state what you mean instead of asking questions that have clear answers?

Define "at scale".

If you take into account load factor, you will need closer to 2,000 such solar plants. And you need far more land too. Also, you’ll end up being most of them in deserts, far away from any human habitations. So you will then build a vast network of power distribution.

Now, to be fair, we could do that. It will just need trillions of dollars, and will have its own set of huge complications. It’s mostly just a fantasy that this will be all that much more cheaper.

I'm not quite sure what your 2000 could be referring to.

If it's land use, that's off by orders of magnitude.

Nuclear at this site: 3100 acres [1], 1/3 of which is a nature preserve, so 2000 acres. It will be 4.5GW at 90% capacity factor, so roughly 500 acres/GW.

NREL puts the least land efficient solar PV at 6.1 acres/MW [2], which with a 0.2 capacity factor corresponds to 30500 acres/GW.

This is a factor of 61 for land use, which is nowhere near the 2000, but I'm not sure if that's what you were talking about.

But land use is not a concern for powering the US by solar. It doesn't take that much land, and even the land it does use is easy to use simultaneously for pretty much anything else.

You talk about "trillions of dollars" as if we should be scared of big numbers, but for some reason forget that it would be far more trillions of dollars just to continue with our current system! That's a very biased way to look at things. And if we were to use nuclear instead of fossil fuels like we do now, it would be far more trillions.

Seriously, this is just one GW at a cost of $15B. Average US load is ~500GW, but will be much higher in the future after the energy transition. A nuclear grid is easily a $15T prospect! And that's without the absolutely massive amount of storage that nuclear needs in order to deal with load following.

[1] https://www.bechtel.com/projects/plant-vogtle-units-3-4/

[2] https://www.nrel.gov/analysis/tech-size.html

It comes from 20% capacity factor versus 90% capacity factor. Replacing 500 GW of nuclear will vaguely require 2,000 GW of solar.

The problem is you want to build it in the southwest of the US, and has a long ways to go to reach the rest of the US. That will be a challenge.

If you think nuclear has a load-following problem, you have no idea how big the problem for solar will be!

And cost will go down, not up when we build more. The cost is because we build them so slow and only one at a time. If we built 500 at once, cost per GW will drop dramatic. Anti-nuclear people are just being dishonest about economics. Ultimately, even if we had to spend a few trillion dollars on this, it will still save money in the long run.

> Anti-nuclear people are just being dishonest about economics. Ultimately, even if we had to spend a few trillion dollars on this, it will still save money in the long run.

On the contrary, the only people who are being dishonest about nuclear are those advocating for a large nuclear build!

Use your own numbers here, actually run them, and you will see that nuclear is more expensive than solar, and indeed even more expensive than solar plus storage at today's prices for storage.

Your assertion that building hundreds of nuclear reactors at once will decrease prices drastically has been proven decisively wrong, again and again

But in contrast, building lots of solar and storage has proven to decrease costs of these technologies over decades.

Literally ever testable assertion in your comment is false, based on thoughts from behind rose-colored glasses from more than a decade ago. It's time to revisit the data and the numbers and see where the world's technology is on costs.

"A few trillion dollars" is enough for a global HVDC grid sufficient for global energy demand (why yes, I did do the maths, it's in an old comment of mine somewhere on here), and you previously gave me a link saying a hydrogen pipe was x10 cheaper, and electrolysis doesn't care if its nuclear or PV, and PV is much cheaper than nuclear per MWh when you don't have to add on the storage which you obviously don't in this scenario because that would be double-counting.

And the US already has north-south lines (with a Texas-shaped asterisk), so even without a global grid the "winter?" question is: what's the capacity of those lines and substations, and how much loss are people willing to face when the price at the plant is only 1-2¢/kWh.

First of all, that is unlikely to ever happen. It is likely geopolitically impossible. But also it is very fragile as you still do not have any real backup power generation. Ultimately, it will make more sense to build system of hydrogen distribution and storage. As you already known, this is 10x cheaper than your HVDC idea.

> But also it is very fragile as you still do not have any real backup power generation

Don't need one. Always light somewhere.

Have repeatedly agreed about the political difficulties here since before you created your account.

Just need the ability to have one part of the world be able to power the rest of it at any given time...

I've had similar conversations before this website. I know what I'm talking about here. What you're suggesting is the same story of massive overprovisioning. Just build making times more renewable energy than we need, and a ridiculously large supergrid to use it all.

The problem ultimately is that it is incredibly expensive, and worse you will end up wasting the vast majority of your energy production. The alternative solution, of using hydrogen energy storage, eliminates this problem, and makes any hypothetical solution more efficient.

And in the end, this is also a zero emissions solution. And works entirely with renewable energy. Like seriously, why do you oppose a solution that flat out works? Are you really that prideful that you cannot let go of your ideas?

> Just need the ability to have one part of the world be able to power the rest of it at any given time...

which we do

> I've had similar conversations before this website.

The meme with Principal Skiner asking "am I out of touch? no it's everyone else who is wrong" comes to mind

> I know what I'm talking about here.

no you don't

> What you're suggesting is the same story of massive overprovisioning.

no i'm not

> Just build making times more renewable energy than we need, and a ridiculously large supergrid to use it all.

not ridiculous; probably politically unviable, but I've done the maths and it's fine from a simple engineering POV

0.5-2.2¢/kWh even if such a grid needs to be given a trillion dollar overhaul every decade, not that you have shown any capacity to integrate new information in this topic…

> The problem ultimately is that it is incredibly expensive, and worse you will end up wasting the vast majority of your energy production.

even a genuinely global (40,000 km) multi-TW grid is cheaper than nuclear, and even bog-standard current HVDC (where you'd be making it of thousands of different segments with combined cross section of 1-2 m^2 instead of a more efficient configuration) isn't a majority of the power being lost, it's at most half.

Also, looking up the best-case efficiency of hydrogen fuel cells[1] and electrolysis[0], you get the same losses as for the dumbest possible HVDC solution, even if h2 transport is lossless (which it isn't).

> The alternative solution, of using hydrogen energy storage, eliminates this problem, and makes any hypothetical solution more efficient.

An alternative solution. One of many.

> Like seriously, why do you oppose a solution that flat out works? Are you really that prideful that you cannot let go of your ideas?

I'm not, you are.

"Look in a mirror".

[0] 84% last year, and higher numbers are misrepresentation https://www.sunfire.de/en/news/detail/worlds-largest-high-te...

[1] 85% when waste heat is captured for use in co-generation, but only 60% directly, which matters a lot for transport

Like I said elsewhere, what you’re proposing is geopolitically impossible. And the type of overprovisioning need to even come close to solving means you are wasting the vast majority of your power. Ironically, this gives you basically free energy for hydrogen production, rendering all of your counter arguments against it invalid. It’s just a doubling down of the same delusion idea you’ve been peddling for multiple posts now. It’s time to get real.

Not to mention no one has ever built a ocean spanning HVDC line, nor have you seriously calculated how much resources it will take. The alternative is already proven to work and has no such problematic weaknesses. Your opposition to it is just stubbornness.

Finally, as fuel cells are electrochemical systems, they will continue to reach higher efficiencies, eventually approaching 100%. Even now, SOFCs can exceed 70% and even higher if waste heat is used to power a gas turbine. By the time you get around to building the HVDC system of your dreams, the efficiency of the alternative will render it obsolete.

It is cheaper, and more and more solar, wind, batteries, and hvdc lines are built everyday.

Reply on news.ycombinator.com