Comment by ben_w | Hacker News Reader

Comment by ben_w | Hacker News Reader

Parent Full thread

In 2021, we added this many new gigawatts: 183 PV[0], 100 wind[1], 7.88 nuclear (but 8.6 nuclear was also shut down)[2]

28 times that, ignoring the shutdowns (I don't actually know how long wind lasts, IIRC nuclear is "about 50-70 years" and PV is "about 30 years"), gives a total of: 5124 GW PV, 2800 GW wind, 218 GW nuclear.

But that's a lower bound for PV, as that's growing with what looks like an approximately exponential long-term trend (since 1992) of 36.5% per year.

While such a growth rate is physically possible to maintain for 28 years even without going into space, given that if we did make so much we wouldn't merely give everyone an American lifestyle but also be able to get the entire human population continuously flying[3], I'm assuming it will turn into a sigmoid before then.

[0] https://en.wikipedia.org/wiki/Growth_of_photovoltaics

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

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

[3] https://www.wolframalpha.com/input?i=%28105*1.365%5E%2828%2B...

The thing is, 183GW from PV is like 20-30GW of nuclear, because nuclear's capacity factor is near 100% and PV's is often near 10%.

Storage seems hard and expensive. We're going to need a whole lot of approaches:

- Wind taking a bigger share of renewables

- Whatever storage we can reasonably build, including some exotic stuff like power-to-gas-to-power.

- Nuclear filling in a little bit

- PV comically overbuilt and routinely throwing away a lot of power midday on most days.

- Still peaking off natural gas, a little bit.

Renewables' capacity factor is much greater than 10%, in practice. If you need falsehoods to make your case, you have no case.

> > The thing is, 183GW from PV is like 20-30GW of nuclear, because nuclear's capacity factor is near 100% and PV's is often near 10%.

> Renewables' capacity factor is much greater than 10%, in practice. If you need falsehoods to make your case, you have no case.

Residential/commercial PV -- e.g. rooftop solar-- very often is 10% or less.

If you want to only look at utility-scale fixed tilt, instead, recent deployments average near (not quite) 20%.

https://emp.lbl.gov/pv-capacity-factors

Nuclear's capacity factor is 93%. In which case, 183GW from utility-scale, fixed-tilt PV is like 183/.93 * .20 = 39GW. Which, admittedly, is outside of my range of 20-30, but also relies on generous assumptions (utility scale projects only-- which are only ~25% of US solar wattage).

If you consider only tracking, you can get all the way to 49GW, which still doesn't eliminate the point.

Perhaps the bigger point is looking at the eastern seaboard. Shipping power 2000 miles is not super practical for various reasons, and the capacity factors for a whole region of the US are pretty poor.

How much solar capacity is actually residential? I'd say quite small since solar is hard to deploy residential + lack of home value improvement.

Nuclear should be part of the equation (it has it's own issues with wastewater) but if adding nuclear means not investing equally or more in solar/wind renewables, that's completely misguided.

> How much solar capacity is actually residential?

I'm not sure how much is residential, but about 25% is utility-scale projects. The rest is stuff on commercial buildings and homes-- which have an average capacity factor around 12%.

> Nuclear should be part of the equation (it has it's own issues with wastewater) but if adding nuclear means not investing equally or more in solar/wind renewables, that's completely misguided.

I totally agree. I think our priorities should be, in order:

- Wind

- Pumped storage

- Improved utility interties

- Demand-side management

- PV

- Nuclear

- Battery storage

- Electrolyzers, etc

But even the lowest priority item should be aggressively pursued.

Utility scale solar is greater than small scale (< 1 MW) solar in the US, and this is disparity is projected to increase through 2050. This is good, because utility scale solar is cheaper per unit output than small scale solar.

https://www.eia.gov/todayinenergy/detail.php?id=50357

> Utility scale solar is greater than small scale (< 1 MW) solar in the US

Your graph shows in total generated energy per year it's roughly even--- not in nameplate power. Because the capacity factor is worse, you need a lot more nameplate power at non-utility installations to get the same energy out.

> This is good, because utility scale solar is cheaper per unit output than small scale solar.

Yup. Small scale solar really needs to stop, other than for installations that have special requirements (e.g. need for power backup / diversity).

You would forbid me to buy panels?

Since we know your use of panels isn't going to be efficient compared to utility scale--- I sure wouldn't subsidize you, and would subsidize the utility scale stuff instead.

Better that production go to installations that have good capacity factors than to your roof.

It's also time to reconsider how good an idea net metering at retail electrical prices is. Grandfather existing customers, but ... sure looks like we need a grid, so maybe only buy power from new end customers at wholesale rates (this further incents storage!)

Shipping power 2000 miles is absolutely super-practical.

China has a project to send it 8000 miles from solar farms they are building in Chile, for use during nighttime in China.

> Shipping power 2000 miles is absolutely super-practical.

DC interconnect makes it more practical, and it's better than storage, but losses and costs accumulate.

> China has a project to send it 8000 miles from solar farms they are building in Chile

Chile has proposed a project to do just that. It is in exceptionally early conceptual planning stages-- but look at you acting like it's a fait accompli!

Chile does not drive massive energy projects. China is in the driver seat.

I doubt it will end up built, but that will not be because it is impractical. It will be because local generation and storage will turn out to be cheaper and less vulnerable to geopolitical upset.

Last time I looked at this, antipodal HVDC was several (2? 3? Can't remember) orders of magnitude cheaper than batteries at the scale of continental electrical requirements; the limit is needing so much metal you become a dominant player in the world metal mining industry just for this project.

Batteries are the most expensive storage. A negligible fraction of utility storage will be batteries unless some new battery chemistry is very, very cheap.

Almost all current utility storage is gravity. That probably won't change.

Hmm. Last time I suggested that[0], I was quickly shot down. I still don't really know why, but as I'm no geologist or civil engineer, I'm willing to believe either a yes or a no at this point.

I definitely don't see any other gravity batteries being useful on this scale.

[0] specifically I suggested that there was a lot of room for a lot of additional pumped hydro — I don't mean my more recent claim, which is merely that existing pumped hydro is cool

There is, in fact, scads of room for add'l pumped hydro. What there isn't room for is more native hydro generation, which needs watershed and a dammed-up river. But pumped hydro only needs an elevated depression. That has actually been a dike around the top of a hill. Natural topography is cheaper, e.g. a box canyon with a dam at the mouth. An alpine pond with no outlet is cheapest of all.

People like to insist pumped hydro is badly limited by geography, but hills are very common, and need not be nearby: transmission lines work very well. Transmission losses matter little when top-line generation has zero opex and minimal capex. You just build a little more of it.

> Chile does not drive massive energy projects. China is in the driver seat.

Chile proposes it, hoping that Asia invests a very large amount of money into Chile to make it happen. Thus far, this latter bit has not happened.

https://www.gob.cl/en/news/president-pinera-presents-proyect...

Any evidence it has progressed past the Chile-showing-other-countries-Slideware-phase?

You tend to uncritically read the most positive, tentative stuff about PV/storage and accept it as absolute fact.

Nukes' utility factor will fall rapidly as renewables and storage undercut them. Each decrease in utility factor multiplies their per-kWh cost, making them increasingly uncompetitive, and that at a steadily increasing rate.

Nukes will shortly be mothballed, to perhaps be fired up once in a while in response to spiking demand.

You continually do this: reply to the same comments multiple times from different angles hours apart; make unsupported assertions that storage will fix everything, etc.

It doesn't lead to productive conversation.

Even at 10% capacity factor, and with a sudden cessation of the long term exponential trend in PV, and assuming the 2021 numbers for new nuclear power while ignoring that more was shut down, that still leaves more than twice as much new PV as new nuclear.

Sure, but I think we already know that we build a lot more PV than nuclear. The question is, should we build a bit more nuclear? Even a small change would be equivalent to a lot of PV production, and provide power at times when PV can't.

Well on that point I agree completely, although mainly because I think it's a good idea to have a diversity of supply and not because it's the most cost effective.

From what I've seen, the cheapest solution is a few square meters' cross section of HVDC encircling the planet, but that would almost certainly take a big increase in global metal mining even on these time scales and even when ignoring the political issues.

For the USA, utility scale PV runs at a capacity factor of over 25% in the sunbelt and mid teens elsewhere. I suppose that may qualify as "often near 10%" but I think you are overstating it a bit, as really household PV is the only one that really is around 10%.

See https://news.ycombinator.com/item?id=31755603

Note non-utility scale solar is 75% of the installed watts, too-- so you have a huge share of it pulling towards that 10-12% number.

PV throwing off excess during peak is fine - if we have storage. Storage is has only recently become feasible, it would likely be utility-grade storage, Powerwall/home storage or Vehicle2Grid integration by the 2050 timeframe.

Storage cost being excessive means it's early adoption timeframe. As the need for it becomes clear options will open up.

> PV throwing off excess during peak is fine

PV throwing off excess during peak is fine even if we don't have storage, because it means PV is adequate for a bigger fraction of the morning and evening (and in turn, the needs for storage become a little less because of the over-provisioning of production).

Storage is cheap and getting cheaper even faster than solar.

Batteries will only ever be a tiny fraction of storage.

Reply on news.ycombinator.com