Comment by Turing_Machine | Hacker News Reader

Comment by Turing_Machine | Hacker News Reader

Parent Full thread

>now that solar energy is far cheaper than coal ever was

I'm sorry, this is simply not true. Usually such claims are the result of ignoring two basic facts:

First, solar is only online 50% of the time, at most (in most regions, considerably less than 50%). That means that you need at a minimum twice the nameplate capacity for a solar plant than for a coal one.

Second, being offline half the time means that you also need sufficient (very expensive!) storage capacity to cover the half the time that the solar plant is not working at all.

In other words, to replace a 2GW conventional plant you're going to need at least 4GW worth of solar cells, plus 24 megawatt-hours of storage.

If you have a source for reliable figures that take these factors into account, and still show solar being "far cheaper", please provide it.

Edit: oh, and no woo-woo battery or solar cell technology that's not currently in mass production. Your statement was that it's "far cheaper" right now, not using future fantasy batteries or solar cells that may (or may not) be on the market in the future.

You nailed it. What do you do when you have cloud cover for two weeks in a row? It’s not unheard of when some areas have 200+ days of at least some cloud coverage.

Nuclear is essential to avoid using coal and not having regional rolling blackouts due to weather. Solar alone is not realistic anytime soon in all regions of the US.

Nuclear is a horrible complement to cheap intermittent renewables. Running it as a peaker multiples your costs and running it as base loads means selling power for negative prices when the sun is shining.

Solar/wind plus batteries for short term storage and pumped hydro for long term storage is the cheapest way to get zero carbon energy. Pumped hydro is more expensive than fossil peakera so build out of that hasn't happened yet.

How do you scale up to producing far more energy in that approach? What about the efficiency to produce large amounts of energy on large spacecraft and other planets and in the ocean?

are you asking how you scale up to producing far more energy with solar panels? world marketed energy consumption is about 18 terawatts, total terrestrial insolation is about 128000 terawatts, and current mainstream panels are about 23% efficient, so if you put solar panels on 50% of the earth's surface, you get 15000 terawatts, which is almost 1000 times more than the humans are using now. on the bottom of the ocean you probably need a different approach, maybe nuclear, or egs geothermal, or maybe running a cable up to the surface, or periodically receiving shipments of thermite in a submarine. some other planets will have no trouble with solar panels; others will need nuclear reactors

Getting only 1,000 times more than humans are using now but requiring 50% of the earth's surface seems like an awful deal. Not only do you need much more of the earth's surface, taking away from trees and habitats and other uses, but you need to significantly increase mining activities to produce the panels and their associated infrastructure. Whereas to get 15,000 additional terawawtts from more nuclear reactors, you could do that with 1,200 - 1,900 additional nuclear reactors occupying just the size of Rhode Island.

mostly you'd be floating them on the oceans (ideally the currently-nutrient-depleted parts of the oceans that aren't teeming with algae), but yeah, that's roughly the limit for solar, and as you're approaching that limit, you need to be thinking about space-based solar power, nuclear power, geothermal power, etc. maybe when you're at 64× of current energy consumption, say

since 02000, total solar installed capacity has gone from a gigawatt to 1.6 terawatts (see https://www.ise.fraunhofer.de/content/dam/ise/de/documents/p... and https://en.wikipedia.org/wiki/Swanson%27s_law#/media/File:19...), which is roughly 10 doublings, one doubling every 2.3 years. but that's peak capacity, and 1.6 terawatts peak is only 220 gigawatts of actual production at a presumed capacity factor of 21% (see https://en.wikipedia.org/wiki/Solar_power_by_country). that's 6 doublings away from world marketed energy consumption; adding the other 6 doublings to get to 64× gives you 12 doublings, and thus about 28 years before this starts to be a concern

probably we should think of this as a lower bound, though; adoption is likely to slow down as solar moves into application areas that are not already electrified or indeed yet done at all by humans, and the last 24 years have been, historically speaking, unusually peaceful

But we wouldn't need 50%, we'd need .05%

That depends on how much more efficient solar panels get and what future energy requirements are, but in any case it'll be worse to scale energy creation with solar than with nuclear.

We'll never substantially increase our energy demand if we don't reduce our energy cost. And since nuclear is horrendously expensive, it won't scale.

i would instead say that as long as nuclear energy is horrendously expensive, it won't scale. but nuclear energy is not inherently horrendously expensive; it's just that human technology is very primitive still

do you have a source on the pumped hydro costs? i thought it was cheaper, just impossible most places

Costs are high because gas peakers exist and pumped storage generally doesn't. For rarely used long term storage, capital costs dominate, and already built sites don't incur additional capital costs.

All you need is a hill and some water for pumped storage. Those sites are very plentiful.

https://scholar.google.co.uk/citations?view_op=view_citation...

oh, yeah, i was thinking of places where you already have a reservoir suitable for hydroelectric production, not places where you have to build a dam

> You nailed it. What do you do when you have cloud cover for two weeks in a row?

Never mind the volcanic winter that seems to happen every few thousand years.

no place has as high a capacity factor for solar pv as 50%; the world average is 14% (i.e., 14 watts average output per 100 watts nameplate output) and the country with the highest is egypt at 35%

(this doesn't mean that you get zero power for 65% or 86% of the day. it means that you get lower than max power at all times except noon. to a great extent you can compensate by that by just installing more panels, but at night you need a better strategy)

but the capacity factor is irrelevant to the fact that solar energy on the power grid generally sells for about half the price that coal-generated electricity sells for, on the same grid, when storage capacity is insufficient. in fact, prices used to go negative at night (to avoid shutting down slow-ramping baseload plants) and now they go negative in the day

https://pv-magazine-usa.com/2020/05/28/record-low-solar-ppas... is an article from four years ago giving some specific prices: a solar ppa had just been signed for 15 dollars per megawatt hour, while the cost of production with coal at the san juan generating station was 44.90 dollars per megawatt hour, even though it was built right on top of a coal mine to save on shipping costs. that's why the san juan generating station has been decommissioned. if you look, you'll find stories like this all over the place, and solar panels now cost half of what they did when that story was written

now, it's true that a ppa that includes battery storage will be more expensive than the 1.5¢ per kilowatt-hour ppa in that article. (https://emp.lbl.gov/pv-ppa-prices has a queryable database of all the ppas signed in the usa, although the usa is less reliable as an indicator of true costs because of how its prices are artificially inflated by protectionist tariffs.) how much more expensive depends on how much utility-scale storage is needed; you suggest 12 hours, but a much more typical number in practice is 3–4 hours, partly because there are still coal plants and partly because electrical demand drops a lot at night

also, you are incorrectly assuming that 'conventional' plants have a capacity factor of 100%, when a more typical capacity factor for a coal plant is 60%

so let's consider a kind of worst case: replacing your suggested 2 gigawatts (nameplate) of coal plants in the usa, where construction costs are ridiculously inflated. before we swung the wrecking ball, those coal plants were generating 1.2 gigawatts (real) of power (10.5 billion kilowatt hours per year), so we need 1.2 gigawatts (real) of solar panels. in the usa the average capacity factor is 21% (the article i linked above is from an area with more sun than average) so that's 5.7 gigawatts peak. typical costs for utility-scale fixed-tilt solar plants in q1 02024 were 98¢ per peak watt https://www.seia.org/research-resources/solar-market-insight... including costs like permitting, design and engineering, etc. so that's 5.6 billion dollars. utility bond yields in the usa are currently at 4.35%, and often these things are amortized over 25 years. i think the amortization calculation is that you have to pay 372 million dollars a year, which works out to 3.5¢ per kilowatt hour or 35 dollars per megawatt hour. definitely too cheap for coal to compete with, but still a lot more expensive than the price that ppa came in at

so, suppose we need 4 hours of storage for our 1.2 gigawatts. that's 4800 megawatt hours. six months ago lithium-ion batteries have fallen precipitously to 139 dollars per kilowatt hour https://about.bnef.com/blog/lithium-ion-battery-pack-prices-... so we need to spend another 670 million dollars on the batteries, which adds about 12% to the cost of the project. except that in real life you need more than just a pile of batteries, you need to pour concrete and run wires and connect inverters and so on. and the batteries won't last 25 years, maybe 8, so you have to amortize this capex over a much shorter period. but it should be clear that this is not a crushing cost that dwarfs the cost of the solar farm

(theoretically lead-acid might be cheaper by a factor of 1.5 or 2 or so, but lithium-ion's advantages seem to have driven it out of the utility-scale market)

aha, here we go. https://atb.nrel.gov/electricity/2023/utility-scale_battery_... says that a 4-hour 60-megawatt lithium-ion battery system costs 446 dollars per kilowatt hour and has 240 megawatt hours of storage. so our required 4800 megawatt hours cost 2 billion dollars. that's about three times the cost estimate above for just the batteries, but that's an estimate from before the batteries dropped in cost by half, so 1.3 billion dollars is a better estimate. this plus the 5.6 billion dollars for the solar plant gives us a total up-front cost of 6.9 billion dollars

since it's getting late, i'll just link you to this four-year-old profile of a ppa for 20 dollars a megawatt hour for generation plus 20 dollars a megawatt hour for storage https://www.energy-storage.news/battery-storage-at-us20-mwh-... and this lawrence radiation lab brief https://eta-publications.lbl.gov/sites/default/files/utility...

the main reason for the difference seems to be how sunny the location is; the technical brief explains:

> Aided by the ITC, most recent PPAs in our sample are priced around $20/MWh (on a levelized basis, expressed in real 2021 dollars, and including bundled energy, capacity, and RECs) for plants located in the West, and $30-$40/MWh for plants elsewhere in the continental United States.

the itc is a subsidy, so the real cost is a bit higher (due to the tariffs)

> when a more typical capacity factor for a coal plant is 60%

This is, of course, patent nonsense. "Not producing maximum power because it isn't needed at the moment" is an entirely different thing from "not producing maximum power because you can't".

A coal or nuclear plant that's producing less than its maximum output because the power isn't needed at the moment can be ramped up if needed (not super quickly, hence the need for peaker plants, but it can be done).

A solar plant that's producing less than its maximum output because it's night, or because it's cloudy, cannot.

> https://pv-magazine-usa.com/2020/05/28/record-low-solar-ppas... is an article from four years ago giving some specific prices: a solar ppa had just been signed for 15 dollars per megawatt hour, while the cost of production with coal at the san juan generating station

1) No storage costs are mentioned. 2) Most places are not New Mexico. 3) I'd bet money that there's some heavy government subsidization involved here. Ah, yes: "EPE will also receive the associated renewable energy credits (“RECs”) bundled with the purchased energy."

> you suggest 12 hours, but a much more typical number in practice is 3–4 hours, partly because there are still coal plants and partly because electrical demand drops a lot at night

Wait: you're claiming that the nighttime zero production from solar plants doesn't matter because there are still coal plants?

I'm sorry, that is a benefit of the coal plants, not the solar plants.

What happens in your scenario when there aren't any more coal or nuclear plants? Again, that might work in New Mexico, but in most regions people would prefer not to freeze in the dark.

for someone who started out demanding that people 'please provide' a 'source for reliable figures', your comment is astoundingly devoid of any sources or indeed factual information. you are very much not living up to the standards of debate i was hoping for; please try to do better

> "Not producing maximum power because it isn't needed at the moment" is an entirely different thing from "not producing maximum power because you can't".

your replacement solar plant also doesn't have to produce the un-needed power, so you have to take the capacity factor into account when you're calculating the replacement size

> No storage costs are mentioned [in the article about the palo verde trading hub prices]

yes, that's why other parts of my comment explore storage costs in great detail

> Most places are not New Mexico

most places that people live are sunnier than new mexico or close enough to someplace that is

> there's some heavy government subsidization involved here

yes, you may note that the comment that you're replying to discusses those subsidies and actually names another one and computes what the unsubsidized cost would be for your suggested 2/4/1.2 gigawatts. as i mentioned, there's also some heavy government taxation involved here; solar panels in the usa cost twice what they cost in the rest of the world. also, as i mentioned, the price of solar panels has dropped by half since the article was written

> you're claiming ... because there are still coal plants

technically what i claimed was that the typical number in practice was 3–4 hours not 'because there are still coal plants' but 'partly because there are still coal plants and partly because electrical demand drops a lot at night'. probably i should also mention wind power, nuclear plants, hydroelectric power, gas peakers, and distributed storage such as car batteries, all of which currently play a role in compensating for the intermittency of solar

the current average is about 3 hours; that will presumably increase as solar becomes a larger fraction of the grid. to estimate how much it increases, we need to start by understanding the current situation, but we also need to predict how extensive demand response will be. you posit that the storage requirement will increase to 12 hours, but that seems unlikely to me. if it did increase to 12 hours, that would make the battery system cost as much as the solar farm. unless batteries somehow got cheaper

what's at issue is not 'freezing in the dark'. people generally do want it to be dark most of the night, because they're asleep, so not being in the dark only requires storing about 40 watt-hours per person, which is a single usb power bank. and you can reliably avoid freezing with a so-called 'sand battery'; 12 hours of 6000 watt heating can be provided by a tonne of sand heated up to 250° when the sun is shining. that's about fifty bucks of sand and a few meters of nichrome wire. if you live in a normal sized house with maybe some insulation you need a lot less than that

so relax, you're not going to freeze in the dark

a mean man scared you with a scary story, but it's not real

rather, what's at issue is industrial process plants like blast furnaces and haber–bosch nitrogen fixation, which are traditionally pretty intolerant of being shut down; it takes a long time to bring them back to steady state after a perturbation. in some cases there are alternative batch processes that compete with continuous-flow processes, which have conventionally been economically uncompetitive, though there are exceptions such as electric arc furnaces. in other cases, it may be possible to design continuous-flow process plants in such a way that they can ramp up and down efficiently, so that they can take advantage of the unprecedented abundance of free energy during the daytime

Reply on news.ycombinator.com