[bhuber]

I'm certainly not an expert on nuclear vs solar energy, although I do have some background knowledge. As with many things, I make up for my lack of expertise by seeking the analyses and opinions of real experts. No offense, but I trust them over a rando on HN who built a solar array in his backyard :) . A simple google search of "nuclear vs solar cost" strongly suggests that you are very wrong - nuclear is far more expensive than solar, and the gap is growing as solar gets cheaper. Of the first four results[1][2][3][4], only one argued that nuclear was even remotely competitive[3], and it's 6 years out of date and the most convincing data it cites is from 2005.

I found [4] to be the most straightforward and pithy explanation. Basically, nuclear wins on capacity factor, but solar makes up for it by a) still being cheaper even when you have to build 4-6 times more capacity, and b) it takes 10 years to build a nuclear plant vs 1 for solar, so you get to start using your electricity and paying off capex (and reducing CO2 emissions) much sooner.

Now it's possible all these sources are so deeply flawed that they came to the completely wrong conclusion, but the onus is on you to present evidence of your assertion and provide a convincing argument. Saying (and I'm obviously paraphrasing here), "the calculations are simple and you all are idiot sheep" doesn't cut it.

1. https://www.literoflightusa.org/solar-vs-nuclear/

2. https://www.reuters.com/article/us-energy-nuclearpower/nucle...

3. https://www.greentechmedia.com/articles/read/the-problem-wit...

4. https://earth911.com/business-policy/solar-vs-nuclear-best-c...

[robomartin]

> A simple google search of "nuclear vs solar cost" strongly suggests that you are very wrong

Kindly show me where I was making a cost comparison between nuclear and solar in my original comment:

https://news.ycombinator.com/item?id=31430747

You see, what happened here is that @ncmncm masterfully diverted the conversation into cost and project failures. A typical political argument form when you can't discuss the actual subject.

You spent a lot of time researching and composing an argument against something I didn't even touch in my comment, at all.

I can definitely get into relative cost discussions. I am not in the habit of making comments unless I devote a serious amount of time to understanding what I am talking about. In this case my research into this was triggered by trying to understand the realities of converting our entire ground transportation fleet (US, 300 million vehicles) to electrics.

That led to creating a series of mid-sophistication models to try to arrive at parameters, from technical to financial. For example, my power requirement model, done about five years ago, predicted we would need between 900 GW and 1400 GW of new, additional power generation. I other words, we would have to double what we have now. That's what led me to try to understand how we could go about doing something like that. Solar isn't going to do it. It can be a part of it, but solar and wind are not what people seem to think these technologies are in real life.

So, my claim was simple: In order to build a solar system that delivers power equivalent to that of a nuclear power plant you need a system with at least 7 times the peak generation rating. This is a matter of physics and it requires understanding how real-world solar systems work, not imaginary pink unicorn systems.

My favorite saying, by Mark Twain:

"A man holding a cat by the tail learns something he can learn in no other way".

A corollary to this is to listen to someone who has before believing it's easy.

[ncmncm]

If you do not compare cost, you promote irrelevancies. If you do not take into account real-world circumstances, your conclusions are meaningless.

In this case, we need not rely on a single solar farm of a size to match your nuke plant, sited where the nuke plant would be. Instead, we have many solar farms scattered widely, thus not all affected by the same weather, connected by long-distance transmission lines and augmented by similarly widely distributed wind farms, and hydro power. In the near future, we will be able to import synthetic ammonia from tropical solar farms to fill in shortfalls. So, whether your straw-man installation would need 7x peak capacity is irrelevant; nobody deploys that way. You don't need all the sources to add up to 7x equivalent; the industry figure is closer to 2x, although there will be good economic and practical reasons not to stop building at that level.

[robomartin]

Nice attempt, but you provide no calculations and conveniently ignore the fact that my multiplier isn't based on some seat of the pants hand-wavy idea but rather the most fundamental physics and math related to making solar energy. At the start of that chain of calculations is the fact that a fixed solar array will, at best, only deliver 66% of the energy of a nuclear power plant during a 12 hour solar period. That alone requires one to overbuild the array by a factor of 1.5 in order to get the same energy output. And the analysis continues from there. You double yet again to account for night time generation. You add another 25% to account for energy conversion losses. And so on.

My estimate was "over 7x". The US Department of Energy's own estimate sets it at a minimum of 10x and up to 20x.

So, no, you are wrong. And, yes, costs will sky-rocket if you have to overbuild at these scales. Even worse if we need to to double our power generation capabilities --which is what we need in order to be able to transition to electric cars. That would require 1200 nuclear power plants in the 1 GW range. If this was done with solar (using DoE numbers, not mine) you would need a minimum of 12 Tera Watts. Not sure we have the land and resources to do that in, say, 25 to 30 years.

Again, you are wrong. Do the math.

[ncmncm]

Again, what matters is cost. Do you need to spend 7x as much on renewables as you would have on your (to date massively subsidized) nukes? No. How much does a GW of nuke really cost, all told, in the US? Current numbers look bad. Disaster insurance alone would price them out of the market. Decommissioning cost is never included in the ticket price. Nobody can quote a reliable price for a nuke in the US. Then, we have operating cost.

The more total power you need, the worse the nukes look. Renewable costs are still in free fall, so setting out, 10 years hence, to double total capacity costs much less than it cost to get to that point. Nukes you started on today, meanwhile, would be just beginning to come online, after ten years shelling out for mined carbon they have not displaced yet. Does it seem unfair to charge that to your nukes?

Calculations divorced from real-world conditions do not enlighten. We are nowhere near short of land to site panels on -- they coexist, synergistically, with crops and pasture, and industrial rooftops, parking lots, reservoirs and canals -- or of silicon to make them out of. We need not discuss the amount of concrete that would be needed to build out your nukes.

[bhuber]

Thanks for taking the time to reply.

> Kindly show me where I was making a cost comparison between nuclear and solar in my original comment

Ok. "There simply isn't a comparison between nuclear and solar. Nuclear is a far better solution on all fronts... The comparison is brutally tilted in favor of nuclear." Two of the most important factors in choosing a grid-scale energy solution are cost and time to deployment, so they're included in your statement. Perhaps you intended to convey a different assertion, but based on any reasonable interpretation of what you actually wrote, @ncmncm didn't divert the conversation, he focused it on a subset of your claim.

> The simplest calculations clearly show this...

> That led to creating a series of mid-sophistication models to try to arrive at parameters, from technical to financial.

Ah, so we've gone from the "simplest calculations" to "a series of mid-sophistication models" :)

> [I] predicted we would need between 900 GW and 1400 GW of new, additional power generation. I[n] other words, we would have to double what we have now.

I haven't done any research to verify this, but based on your reasonable assumption of transport fleet electrification, these numbers seem reasonable. So we agree we'll need more electricity generation in the future. I don't see how that's evidence that nuclear is a better source for it than solar.

> ...done about five years ago...

You may want to update your models, the cost of utility-scale solar has roughly halved in the last five years worth of data points: https://www.nrel.gov/news/program/2021/documenting-a-decade-...

> Solar isn't going to do it.

Why not? None of the evidence you've provided supports this assertion.

> [Solar] can be a part of it...

Your original assertion was that, "The comparison is brutally tilted in favor of nuclear." An obvious corollary is that we should invest all our resources into nuclear deployment instead of solar. By saying solar can be a part of it, you're implicitly changing your original position.

> So, my claim was simple: In order to build a solar system that delivers power equivalent to that of a nuclear power plant you need a system with at least 7 times the peak generation rating.

The sources I cited above say a factor of 4 - 6, but it obviously varies a lot depending on climate and latitude. So sure, let's say 7 conservatively. So what? Even taking that into account, solar is still a fraction the cost of nuclear. It has myriad other advantages such as being faster to deploy, safer, has unlimited fuel, doesn't have any significant waste products, has much more predictable costs, etc. So why should we bother investing in new nuclear deployments?

I realize you're saying you're not making an economic argument, but you haven't made any other kind of argument either. The biggest thing nuclear wins on is steady output - as everyone knows, we can't rely on solar generation 24/7. So perhaps that's what you're trying to get at? As sister threads have discussed, however, a) we have a lot of solar to install before we have to worry about excess peak capacity, and b) there's been great progress in utility-scale energy storage systems which mitigate this problem.

[robomartin]

Nice try. I'll give you that.

The mid-sophistication models, as I clearly stated, were intended to try and understand what it would take to convert our entire ground transportation fleet to electric. This required a medium level of sophistication as I had to code models to simulate average behaviors across six time zones, different driving habits, slow and rapid charging, business and personal use, etc. Tons of variables to play with. The model predicted the need for additional power generation in a range between 900 GW and 1400 GW, doubling what we have today. I did this about five years ago.

This has since been confirmed by other sources.

That led to trying to understand how we might be able to do it. Hence looking at the various technologies and focusing on solar --something I had devoted a considerable time and investment into just the year before-- and a comparison to nuclear. Solar lost.

> I don't see how that's evidence that nuclear is a better source for it than solar.

You have to do the math. If you are not willing to do that there's nothing I can say here that will convince you of it. In order to do the math you do have to have a good level of experience in construction. I have, so I understand how things are built. Most people don't. I understand this, of course. This makes it very difficult to have a conversation because people don't have a sense of proportion to what it might cost to, for example, prepare a ten square mile site for the construction of ground mount structures and the installation, operation and maintenance of a solar array. Simple example: If you just leave untreated dirt on the ground in some places you are going to lose half of your generation capacity inside of a week or a few weeks as winds cover your panels with dirt.

This isn't as simple as magical solar panels making magical energy. Not even close.

> solar is still a fraction the cost of nuclear.

Have you actually done the numbers?

I'll take the Department of Energy's baseline number that indicates you need 10 GW in solar panels in order to match a 1 GW nuclear power plant.

And, BTW, that also means you need at least the ability to store 12 GWh of energy in batteries in order for this to actually replace a nuclear facility. That scale is massive.

Let's just look at the panels. How many do you need for 10 GW?

Assuming 325 W panels, which is a reasonable assumption today:

10 GW = 31 million panels

Let's say you can buy the panels at $300 for easy math:

That's $9.3 billion just in the panels.

Now you have to add the land, preparing the land (bulldozing, grading, leveling, etc.), the concrete, mount structures, wiring, inverters, installation labor, vehicles, transportation costs (what does it cost to move 31 million panels, wires, steel, concrete, etc.). The list goes on and on.

Too much to throw into a text comment. This is spreadsheet territory. If you understand construction and do the math, the numbers quickly click up into the billions and the total cost of the installation is easily in the tens of billions of dollars.

And that does not include the batteries and related technology.

Even worse. You just installed 31 million solar panels that will suffer that will degrade at a rate of 0.5% per year. Your 10 GW facility becomes a 9 GW facility in twenty years.

Either you overbuild it to an 11.1 GW facility so you have 10 GW by year twenty (at the cost of another 3+ million panels and all else that goes with it) or you have to replace millions of panels, likely starting somewhere around year ten and on a constant basis for the next ten years. You'll probably have to replace the entire array somewhere between year 20 and 30.

Same for the batteries. What does it cost to replace 12 or 20 GWh of batteries? Where do they go after 20 or 30 years of service?

And we haven't even accounted for the oil, gasoline and diesel you'll need to burn to build and maintain this monster. How much fuel are you going to use to dispose of panels and batteries gone bad?

And here's the kicker: In order to be able to switch our ground transport fleet to 100% electric we would need 1200 of these facilities. My not-so-humble opinion is that this is both crazy and impossible.

Nuclear isn't without issues, of course. However, if you build a 1 GW reactor it will produce 1 GW 24/7 for at least the next 50 years. Last I checked that only requires somewhere around 1 square mile (vs. at least 75 square miles for the same output with solar, according to the US Department of Energy).

That is a no-brainer. We just need to get good at building them and build next generation clean and safe reactors. If it becomes a national mission to do this with no political bullshit in the middle, we can do it. Otherwise, forget about it.

This is what will happen. As electric cars start to become more common our grid will be taxed to the breaking point. At that point the cost of upgrading our infrastructure will be even worse than it is today and we might not be able to afford it (the US is already broke). This will mean that economies who made heavy investments in nuclear will have huge advantages while we keep talking about pink unicorns in the form of solar, wind and whatever else.

I obviously like solar, I invested a non-trivial amount on it (my entire project cost over $100K). However, I choose to be a realist about what this technology is and is not. I only learned these things after owning such a system for a few years and looking at it as an engineer devoid of any cult-like attachment to the technology. Math, physics, engineering. The numbers don't lie.

> The biggest thing nuclear wins on is steady output

It's a lot more than that. It's nearly 100% output capacity, 24/7 for at least 50 years with no serious degradation and not having to replace half the reactor every 15 to 20 years.

The weather is a huge factor. The nuclear reactor keeps going, rain, windy or calm. A solar array can get ripped to shreds by a strong wind event and cut down to 25% energy output by rain or clouds. It can be damaged to the tune of billions by hail. It can literally produce half the energy output for days or a whole month. Which requires heavy over-building of the storage portion in order to effectively survive a one week brown-out due to weather, etc.

Solar can be great at home to run your air conditioner and lower your bill. At a massive scale, to supply 1200 GW over and above what we have today. I am not sure. Right now, I don't think so.

> So perhaps that's what you're trying to get at?

Well, cities don't work with intermittent unreliable energy. So, what I am trying to get at is what you actually need for society, industry, life to function.

Either we are talking about things that have to be equivalent or we are changing the rules. A factory needs consistent and reliable power. So does a hospital, school, office and home. So, yeah, 24/7 performance isn't just important, it's a requirement.

> we have a lot of solar to install before we have to worry about excess peak capacity

No. If we are going to make the claim that solar is the path to our energy future, we have to answer the question. We need at least 1200 GW of additional capacity. What is the best way to achieve this? Solar or nuclear? My argument is that nuclear is likely the bulk of it and solar will play a lesser --yet important-- role.

It's a mathematical reality, not my opinion. If you have enough command of the basic science you can verify this yourself, this does not require a wall of links to studies, its super-simple math and physics. Most people can't do it or don't care to do it, because it is always easier to just believe what you are told.

[aiProgMach]

Your whole argument will fall down because of this weird mistake:

> Let's say you can buy the panels at $300 for easy math

You'er assuming ~1$/1W for solar not installed, you're off by almost 5 order. I bought 330w panels ~2 years ago for 165$ each, which is exactly 0.5$/W and currently I can find 0.38-0.45$ / W for more modern panels (> 400w, Mono Perc... where 0.38 or less for price per pallet not containers even) and this is for home usage without subsidies and in the middle of the price hikes we're facing. For utility scale you can expect it to be 0.18-0.3$ / Watt (0.18 is a real price offered from some Chinese companies for wholesale without shipping prices), so all in all 10GW of solar panels will cost $1.8-3 Billion give or take, so your $9 Billion figure could make ~ 30-40 GW or even more given the expected price reduction and/or efficiency increase of the solar panels.

[Gibbon1]

> No offense, but I trust them over a rando on HN who built a solar array in his backyard :) .

That same rando built one in his backyard isn't nothing burger.

[bhuber]

Agreed, it's actually fairly impressive. Beware the false equivalence fallacy however - there's little overlap in the expertise needed to design and install a home solar array, and that needed to analyze the economics of complex industrial technologies. Based on my links above, I think op is a far better electrician than economist.

Also, I doubt op ever built a nuclear reactor in his backyard :D

[robomartin]

> analyze the economics of complex industrial technologies.

Except I have been responsible for the design, construction, installation and operation of large complex industrial installations. That is what I did 40 years ago. Not solar, of course, but beyond a certain scale a project is a project, whether you are building a bridge, road or chemical processing installation. Of course, the private sector has different dynamics when compared to government projects.

Also, as I mentioned in my other reply to you, I never made financial claims about solar. This is a branch introduced by someone who could not argue against what I was saying and chose to divert the conversation. My claim was simple:

https://news.ycombinator.com/item?id=31430747

"if you want the equivalent of a 1 GW nuclear power plant, the minimum size of your solar array is at least 7 GW peak"

So, please, don't charge me with something I did not enter into the argument.