Solar has an average capacity factor of roughly 20%, meaning over the course of a year 1W generates an average of 0.2W.
You are using a metric of TWh/year, so multiply 1 TW * .2 * 24 hour/day * 365 day/year and that 1TW of solar generates 1700TWh/year.
Given an average lifespan of 25 years, a 1TW/year productive capacity supports 44,000TWh/year production.
Given that solar will be so cheap, I expect us to be throwing away 20%-50% of solar electricity (curtailment, in current jargon). For example, production may be sized to produce enough at the seasonal minimum (winter in Northern countries), and summer time might have ridiculous abundance.
> but is it power in = water out, plus a bunch of facility maintenance and operations?
Basically yes. I'm mostly familiar with Reverse Osmosis, where electricity to drive the pumps is about 40% of the TCO. There's another big chunk for consumables - chemicals, failed membranes, etc. Then the rest is capital costs to build the thing.
It's not "free power == free water", but the power is a significant factor.
> Desalination will be the new pumped hydro.
Eh, no. Desal is a great place to dump surplus energy, but you can't get the energy back out. It pairs well with nuclear power, so you can use extra nighttime capacity. It's less useful with solar, unless you've severely overbuilt panels and can't find anything else to do with them on sunny days. It's a better tradeoff to buy fewer panels and use that money to pay for storage (pumped hydro, batteries, whatever).
It's power in, water + brine out. You have to do something with the brine- dribble it into the ocean over a large area, pump it into old salt mines or below freshwater aquifers.
It's too concentrated to just put back into the ocean in one spot without killing everything around it, and you don't want it contaminating what little fresh water you have under ground. Not an insurmountable problem, but it does add another cost in addition to maintenance and power
The dilution can be done in holding tanks. Fill a tank a tenth of the way with the hyper-saline brine (or whatever the right ratio is), flood the rest of it with sea water, stir it around for a while then dump it into the ocean. If the holding tanks are built in the tidal zone and are sized appropriate relative to the brine output of the desalination plant, then you wouldn't even need to do much pumping. The salinity of the tank could be tested before dumping it, an extra layer of safety in the system.
There is public opposition to building desalination plants in part because a lot of people don't like that solution. For starters, it just sounds bad; it sounds like a company brushing aside pollution concerns by saying they'll spread it out where you can't see it. That characterization isn't fair; but fair or not, poor public perception still stalls land development.
More significantly, it requires more infrastructure under the water where it will be difficult to inspect and repair. Furthermore there is a financial incentive for plants to neglect these pipes. Suppose there is a break in the pipe that discharges concentrated brine straight into the ocean. Slowing or ceasing production until the pipe could be repaired would be very expensive to the plant operator. On the other hand if they ignore the problem, they wouldn't have to slow production and it would cost them nothing (unless they got caught.)
Yeah we need good things to do with cheap power that don't always have to run and aren't too capital intensive (Making them need to use more expensive power). I'm not sure what is best between Desalination, Aluminum smelting, Crypto mining, or what other good alternatives are.
Desalination is higher priority overall then crypto mining. There are several factors which make something a good candidate to use excess solar power. Maybe it makes sense for desal to always be running? Crypto can more quickly be spun up or down to respond to the energy market. Crypto mining can be done anywhere with cheap power but desalination needs to be near salt water and where water is needed. We will probably need a mix of several industries to effectively use the cheap power.
Indeed, we're seeing this happening. Bitcoin mining will pay $$$ for your waste power, and an increase in waste power is the problem with wind and solar.
We need a shitload of green generation infrastructure, and Bitcoin is going to pay for it. Sometimes the world is weird, sorry.
this is why I think hydrogen will be a huge part of the system.
even if the round-trip efficiency is 20% and all you do is mix it into the gas turbines, eventually the numbers are going to work out to "why would you not, why would you leave the money on the table"
even if the tanks leak like sieves a 10 - 100Hr buffer will go a huge way toward smoothing out winter lulls.
Liquid hydrocarbons can be stored safely and at virtually no cost for years in all sorts of climates in cheap, lightweight containers. Transport is likewise cheap, simple, and safe. Dealing with leaks is relatively safe too.
We have a hundred years of experience dealing with them with all the deep knowledge and huge mass of deeply embedded systems that that implies.
And they are nearly as energy dense as liquid hydrogen.* There doesn't seem to be much "there" there with hydrogen, except generation efficiency if it is generated at the point and time of use.
Reliability, safety, operational life, risk management, availability of expertise, and cost will probably trump technical efficiency, just as when buying servers. I expect to see synthetic hydrocarbons being made with green hydrogen (and green carbon) before I see the hydrogen itself being transported around in any significant way or stored for any significant time.
* Fun semi-related fact: there's more hydrogen in a liter of gasoline than in a liter of liquid hydrogen.
I agree about not being transported or stored "significantly" but thats the beauty of it. The very places that would consume it (gas gen stations) would also be the ones who could produce it when they're not running because solar is overproducing. produced and stored on site, never transported, pre-exising grid connectivity. and again i'm only talking about the 10 - 100 hour market. the <10hr market is already solved by li-ion, and the >100hr market is an unsolved problem, we're just gonna have to burn gas for 4 - 6 weeks in the winter.
We store and transport millions of tons of LNG -- liquified natural gas, dirty methane -- in the same way as liquid hydrogen would be handled. LH2 is colder, but not so you would notice.
LNG goes point-to-point from liquefaction plant to regasification plant. There's no reticulation to city neighbourhoods.
Gaseous hydrogen is much more dangerous than natural gas (flame speed, range of air fraction for combustion, ability to escape seals, etc., etc.) and you need three times the volume to deliver the same energy.[1] I doubt there will be much reticulated use.
I'm hoping we can convince gas companies to convert into district heating companies. It's still the heating business, and uses pipes and liquids. They could make heat pumps far more efficient by using ground sources of heating and cooling, and maybe even shipping process heat from one building to another.
It would be a it of a shift, but it uses the same core competencies and would allow them to survive.
However I don't think there's a single gas company with management with the vision, innovation, and smarts to make the switch. It would take a new generation of management, that's smarter and more forward looking.
Maybe a shareholder revolt at one could cause a change in management that would pioneer this change... how much does a gas company cost?
Most ex-Soviet/Warsaw pact countries tried to move from district to direct gas, or small CCGT plants (as in one for about 10000 people) exactly for this reason: maintenance is hell and losses are enormous. And that's in high-density urban. In low-density suburbs it won't ever fly.
Leaked hydrogen, all told, traps 200x as much heat as CO2. We really should not leak any more than we must.
By itself, it is just 6x. But it has lots of other interactions, such as increasing lifetime of methane.
But round-trip efficiency of hydrogen will be rising fast. Electrolysis by itself, which used to be 60%, will go well over 90%, maybe over 95%, in short order.
Hydrogen storage is such a pain that I don't think we will be keeping hydrogen around except for chemical feedstocks.
We will need to convert hydrogen into something else to make it easier to handle. Perhaps ammonia, which is useful both for fertilizer and perhaps as a liquid fuel.
There will be a hell of a lot of synthetic ammonia. The shipbuilders and ship fuel suppliers are already tooling up for ammonia-powered shipping. It will take a long time to get electrically synthesized ammonia production up. GW-scale plants under construction now won't start delivering until 2026.
There will be a lot of hydrogen banked in underground caverns and tapped-out fracking wells. But industrial users who have been buying it, or LNG, will take to electrolysing locally instead. Airports will electrolyse and liquify locally, international hubs first.
You can't compare TW and TWh in that way (one is power, the other energy). With solar the way that TW translates to TWh is a little complicated, but it sure isn't 1:1 (probably closer to 1:2000).
Just curious, did you get 2000 as in eight hours a day, five days a week? It just reminds me of the rough estimate of the number of hours worked in a year which I find funny to think about the sun going to work in an office for five days a week, eight hours a day.
I think more accurately the sun puts in incredibly long hours, but isn't working at full productivity for a lot of it (maybe if it cut back to a normal 40 hour work week it would get more done).
A single TW of solar would produce a TW for every hour that the sun is shining, so it could be 3,000 TWh for every TW installed. Obviously the exact value depends on the insolation.
You're confusing TW with TWh. If running at full capacity that would be an additional 8760 TWh generated every year. Of course it won't be 100% utilised but it takes more of a chunk out of it then you suggest.
https://www.statista.com/statistics/280704/world-power-consu...
EDIT: Misread power as energy.