Comment by nayuki
3 days ago
> The tried-and-true grid-scale storage option—pumped hydro [--> https://spectrum.ieee.org/a-big-hydro-project-in-big-sky-cou... ], in which water is pumped between reservoirs at different elevations—lasts for decades and can store thousands of megawatts for days.
> Media reports show renderings of domes but give widely varying storage capacities [--> https://www.bloominglobal.com/media/detail/worlds-largest-co... ]—including 100 MW and 1,000 MW.
It looks like the article text is using the wrong unit for energy capacity in these contexts. I think it should be megawatt-hours, not megawatts. If this is true, this is a big yikes for something coming out of the Institute of Electrical and Electronics Engineers.
> big yikes for something coming out of the Institute of Electrical and Electronics Engineers.
Besides the unit flub, there's an unpleasant smell of sales flyer to the whole piece. Hard data spread all over, but couldn't find efficiency figures. Casual smears such as "even the best new grid-scale storage systems on the market—mainly lithium-ion batteries—provide only about 4 to 8 hours of storage" (huh, what, why?). I could also have used an explanation of why CO2, instead of nitrogen.
> provide only about 4 to 8 hours of storage" (huh, what, why?)
Because the most efficient way to make money with a lithium ion battery (or rather the marginal opportunity after the higher return ones like putting it in a car are taken) is to charge it in the few hours of when electricity is cheapest and discharge it when it is most expensive, every single day, and those windows generally aren't more than 8 hours long...
Once the early opportunities are taken lower value ones will be where you store more energy and charge and discharge at a lower margin or less frequently will be, but we aren't there yet.
Advertising that your new technology doesn't do this is taking a drawback (it requires a huge amount of scale in one place to be cost competitive) and pretending it's an advantage. The actual advantage, if there is one, is just that at sufficient scale it's cheaper (a claim I'm not willing to argue either way).
It ought to be cheaper at scale. Batteries' cost scales linearly with storage capacity. Cost for a plant like this scales linearly with the storage rate - the compressor and turbine are the expensive part, while the pressure vessels and gas bags are relatively cheap.
The bigger you build it, the less it costs per MWh of storage.
6 replies →
i think it had something to do with CO2 can be made into supercritical state relatively easily, not for nitrogen or other common gases.
This pretty much
You can liquefy CO2 at a higher temperature than N2
You can do it easily with something like propane, or other larger molecules. But CO2 is non-flammable, largely non-toxic, and easily available.
I'm sat here thinking: why not compressed or liquefied air?
The basic issue is that they need a phase change at a reasonable temperature. Liquifying air requires much lower temperatures than CO2.
> only about 4 to 8 hours of storage" (huh, what, why?)
Or it's just so obvious - to them! that it doesn't need to be mentioned, which then doesn't make it an ad.
Lithium ion battery systems are expensive as shit, and not that big for how much they cost.
Because CO2 is a magic word. It can open free money doors. Or at least it used to.
Power plants are often described in terms of (max) power output, i.e., contribution to the grid. So, I can see how it might confuse a writer to then also talk about storage inadvertently.
But also, the second paragraph already describes the 100 MWh vs MW nuance.
It is not a nuance in an article that focuses on storage from the supposed premier professional association. As an engineer I would expect typical energy content (median/average) of the top 10 hydro pump projects and also some discussion about the availability of suitable sites. I think one should strive for at least high school level physics. There is no need to push out texts that can be easily surpassed by any current llm.
If 1 watt is 1 joule per second then, honestly, what are we doing with watt-hours?
Why can’t battery capacity be described in joules? And then charge and discharge being a function of voltage and current, could be represented in joules per unit time. Instead its watt-hours for capacity, watts for flow rate.
Watt-hours… that’s joules / seconds * hours? This is cursed.
I believe it's just a matter of intuitively useful units. There's simply too many seconds in a day for people to have an immediate grasp on the quantity. If you're using a space heater or thinking about how much power your fridge uses kilowatt hours is an easy unit to intuit. If you know you have a battery backup with 5 kilowatt hours of capacity and your fridge averages 500 watts then you've got 10 hours. If you convert it all to watt seconds the mental math is harder. And realistically in day to day life most of what we're measuring for sake of our power bill, etc. is stuff that's operating on a timetable of hours or days.
There are two types of jobs, the ones which require you to know that a day is about 8.5x10^5 seconds, and those which don't.
3 replies →
True. Otherwise we would be using square meters for measuring gas mileage instead of miles-per-gallon (or litres-per-km) [1].
[1] https://what-if.xkcd.com/11/
11 replies →
Plenty of people use Joules or rather kilojoules or megajoules or even gigajoules for various purposes.
Watt hours is saying, how long will my personal battery pack last me that powers my 60 W laptop? Which is also fine in that context.
1 Wh = 3600 Ws = 3600 J
It is not more cursed than km/h (1 m/s = 3600 m/h = 3.6 km/h)
Both those units are more convenient than their SI equivalent and their "cursedness" come from the hour/minute/second time division.
If we had decimal time, as it was initially proposed with the metric system, we wouldn't have this problem, but we weren't ready to let go of hours/minute/second.
Yeah. I get this is all kind of silly. I think what trips me up is that a watt doesn’t represent a timeless amount of something the way a meter does. A watt involves a unit amount of time.
Imagine if the distance between you and I was 438 kiloflerp-hours. And to get to you in one hour I have to drive at a speed of 438 kiloflerps. It works, it kinda makes sense. It just feels inconsistent with all the other units I work with.
3 replies →
It's easier to figure out for people that measure power in watts and time in hours ... 1 kW for 1 hour is 1 kWh.
That camel's nose was already in the tent with the mAh thing in phone/etc batteries, now with electric vehicles we're firmly in kWh land.
Not to mention that's what the power utilities used all along ...
A watt of power multiplied by a second of time has an agreed upon name called joule, but a watt second is also a perfectly valid SI name.
A watt is a joule of energy divided by a second of time, this is a rate, joule per second is also a valid name similar to nautical mile per hour and knot being the same unit.
Multiplication vs division, quantity vs rate, see the relationship? Units may have different names but are equivalent, both the proper name and compound name are acceptable.
A watt hour is 3600 joules, it’s more convenient to use and matches more closely with how electrical energy is typically consumed. Kilowatt hour is again more directly relatable than 3.6 megajoules.
Newton meter and Coulomb volt are other names for the joule. In pure base units it is a kilogram-meter squared per second squared.
So when I torque all 20 of my car's lug bolts to 120 n-M, I've exerted 2/3 of a W-h? So if it takes me 4 minutes, I'm averaging 10 watts? That's neat. I wonder what the peak wattage (right as the torque wrench clicks) would be; it must depend on angular velocity.
4 replies →
Of course it can be. Nobody does it in practice because it's inconvenient.
Watts = volts * amps and the people working with batteries are already thinking in terms of voltage and amperage. It'd be painful to introduce a totally new unit and remember 1 watt for an hour is 3.6kj instead of... 1 watt-hour.
Don’t stay there: EVs are even reporting consumption in terms of kWh/100km or kWh/100miles instead of just average kW.
What people care about when talking about EVs and consumption is generally how much distance they can cover. If you take away the distance factor and just report power, it becomes meaningless/almost useless.
4 replies →
Heh. To borrow an idea from xkcd (measuring gas consumption as area): The kWh measures energy, right? And energy is force times distance. So energy divided by distance is force! Let’s all start measuring EV consumption in newtons, folks. It even makes intuitive sense: It correlates well with how hard you need to push the car to get it going at the usual travel speed. But it sucks if you need to figure out how far you can travel on a given charge.
Yep, it's stupid from a units consistency pov. A bit like using calories instead of joules.
But on the other hand we also use hours for measuring time instead of kiloseconds...
Yeah, if only we would define seconds to be 13.4% shorter than that are, we could have 100ks days. Also, ksecs would be a really convenient unit for planning one's day: a ~15 minute resolution is just right for just about any type of appointment.
Oh, and 1Ms weeks, consisting of 7 working days and 3 off days sound nice too.
One can dream! :-)
1 reply →
California found out pumped hydro isn't so "tried and true" when it was shut down during a drought due to lack of water behind the dam.
Not every area is as messed up as the Colorado river watershed...
All users (states) were given an allotment which, when it was set, was more than what would ever be the yearly supply.
From the outset it was essentially a free for all. Everyone was happy, they kinda got what they asked. It's just that they were all living in a paper reality
I should have explained in my original comment why I think those sentences are wrong. I'll do so now.
> pumped hydro [...] can store thousands of megawatts for days.
You can't "store" a megawatt – because you can only store energy, not power.
But another interpretation is, if you actually store thousands of megawatts (i.e. gigawatts) for days, then at the very least, 1 GW × 1 day = 24 GW⋅h. If we take "a few" to mean 3, then 3 GW × 3 day = 216 GW⋅h. I'm not sure there exists a large enough pumped hydro plant in the world that stores 216 GW⋅h of energy. So I think the article meant to say, "store a few gigawatt-hours to be released over a period of a few days".
> Media reports show renderings of domes but give widely varying storage capacities—including 100 MW and 1,000 MW.
Once again, you can't store megawatts of power, full stop. You can store megawatt-hours of energy. The linked article at Bloomberg said that a project in China is building 600 MW of wind power, 400 MW of solar power, and 1 GW⋅h of energy storage – which is the correct unit.
I'm old enough to remember when IEEE Spectrum was a respected technical publication.