Hacker News Reader

Comment by new_realist | Hacker News Reader

For long term energy storage batteries are not ideal. Their costs increase linearly with time of storage and amount of storage. Peak shaving is an OK use for them, though.

View on HN

What percentage of our global storage needs do you believe are "long term" in this dichotomy?

Is this at the "electric cars can't do my daily commute" or "electric cars can't tow my boat 600 miles without stopping, once a year" stage?

Seems like most of the world just needs solar to last until the next day with a decent chunk of that power needed immediately after the sun sets and any wind power available could get smoothed out and/or stored for the morning by the same batteries.

Short term = diurnal

Long term = weeks to seasonal

Batteries, especially that sort of cheaper battery, would be great for diurnal load leveling.

Long term storage is probably more hydrogen and mass thermal storage. The tradeoff would be capital cost per unit stored energy vs. efficiency.

Storage competes with and complements two other approaches: overinstallation of generation capacity w. curtailment, and dispatchable demand.

Curtailment involves installing more wind and PV than you need, and curtailing its output when supply > demand. A study in Minnesota found that this was cheaper than adding storage up to about 70% of the electrical power supply. Of course this creates dirt-cheap power during the curtailed periods that people can find use for (industrial thermal storage, production of hydrogen in low capital cost membraneless electrolyzers).

Dispatchable demand would involve redesign of industrial processes to operate in fits and starts. For example, today's aluminum cells do not respond well to inconstant power sources (their either freeze up or overheat and damage the lining). This was fine when the cheapest power sources were baseload sources. But perhaps an aluminum technology could be developed that would be more forgiving of variation and outages, to exploit the new situation where intermittent power is becoming the cheapest.

Of course, another approach is to move those industrial processes down close to the equator, where there is little seasonal variation in insolation. High latitude countries may become industrially disadvantaged in a renewable powered world.

> Long term storage is probably more hydrogen and mass thermal storage.

I was thinking about how multiple big heating packs[1] would stack up. Seems the Wikipedia reference[2] has the topic well covered. Tricky but possible. And apparently (duh) not a novel idea by any means[3][4].

Though the references I could find were mostly for home-size installations or for actual heating packs. Here in Norway we're blessed with lots of mountains with lakes, so we can do pumped hydro for latent storage. But for flatter countries (say our neighbors Denmark), is grid-size PCM storage viable?

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

[2]: https://books.google.com/books?id=EsfcWE5lX40C&q=latent+heat...

[3]: https://www.sciencedirect.com/science/article/abs/pii/S13594...

[4]: https://ideas.repec.org/a/eee/appene/v221y2018icp522-534.htm...

I was thinking more in terms of artificial geothermal. The time constant for a sphere of bedrock to cool off goes as the square of the radius, and can easily be made to be many years. These would be very large installations, of course.

Another option for seasonal variation is to overbuild. Build enough solar for winter and just discard or figure out something to do with the excess power supply in summer.

(I'm not sure that would be cheaper, it's just another possibility.)

That's curtailment. A twist on that is to install some of the solar at an angle appropriate for winter instead of summer.

Multi day storage is always an issue. X kWh stored over Y hours takes X kWh of storage but X/Y kw of generation capacity. Start talking 50 hours and the cost difference becomes huge.

PS: Though batteries inherently have some emergency storage capacity as daily deep discharge is inefficient.

Methanol is an alternative to hydrogen. With current technologies it is less efficient to produce it and then generate electricity in fuel cells, but storing it is trivial.

All countries already have infrastructure for strategic gas reserves that are sufficiently sized to cover demand for long periods. So storing renewable energy via power-to-gas processes is probably the way to go.

If so,then it stands to reason that they would be unlikely to move to renewable power.

  What percentage of our 
  global storage needs do
  you believe are "long
  term" in this dichotomy?

Solar output in the UK varies substantially between summer and winter [1] - as does demand for heating.

As a Brit with solar panels, it would be super neat if I could generate power during the summer, store it for 6 months, and use it for heating during winter.

That would reduce my direct carbon emissions relative to the status quo, which is burning natural gas for heating during winter.

However, replacing my (fairly modest AFAIK) £222-per-year gas bill would need ~150 Tesla Power Wall packs, costing several times the value of the home being heated.

[1] https://pvoutput.org/aggregate.jsp?p=0&id=4768&sid=3764&t=m&...

https://en.wikipedia.org/wiki/Power-to-gas is probably feasible for consumer energy use (at least for people with room for solar panels). We have the facilities to store and distribute the gas already, and to the extent that additional storage would be necessary, it is fairly economical.

Replacing industrial and shipping energy with renewables is a bigger problem.

> store it for 6 months, and use it for heating during winter.

I wonder when we will get high voltage transmission systems between the northern and southern hemisphere? Actually solar farms in Southern Spain, Morocco or Arabic countries would make more sense.

Also seasonal thermal energy storage does that. Avoids solar electricity altogether, although has other limitations.

https://en.m.wikipedia.org/wiki/Seasonal_thermal_energy_stor...

Another "paradigm" is that full and empty batteries are shipped between regions by boat or truck as needed instead of just waiting 6 months for the planet to move through the cosmos.

Whew, that doesn't sound feasible, but I'd love for someone to do the math for me. It seems much more realistic to build HVDC lines from equatorial regions to the temperate zones, but even that is not very popular anymore as far as I know due to cost and NIMBYs.

I haven't done the math either, but I've seen serious discussions of it.

It requires zero infrastructure investments to load a hundred Tesla batteries on a truck and drive it to place with a different climate situation.

Compare that to long distance HVDC lines, which for the US are desperately needed, have been discussed for decades, and show no signs of ever getting built.

A HVDC transmission line might be a more practical way to do this.

For this, wouldn't it be better to pump water onto the top of a bunch of mountains in Wales?

https://en.wikipedia.org/wiki/Dinorwig_Power_Station, but more.

Pumped water does not scale well, there are only so many mountain valleys to block. And I'd rather keep a few in a somewhat natural state (a magic "no more mountain valleys in exchange for instantly solved climate change" would be attractive nonetheless, even though I write this from a very nice, unblocked mountain valley).

But when your concern is just seasonal storage for heating, you don't need pumped storage. Just great water in a big, isolated tank, it's not even new technology. The surface per volume shrinks the bigger the system, so very large tanks can have quite nice loss numbers.

Actually, there is enormous potential capacity for pumped storage around the world.

Pumped hydro is generally used for diurnal storage. The infrastructure of the dams and turbines is darned expensive; it makes practical sense if you're filling and draining the basin (or half of it) every day, but paying pretty much the same cost to get just a single lake volume per year ... not so much.

That's the point, isn't it? Batteries are not a panacea for energy storage needs, even with a doubling of their capacity.

There is no requirement for a panacea or super neat developments that allow individual households to store seasons worth of privately generated electricity.

For seasonal and sporadic production shortfalls in solar and wind generation, biofuels and electrically synthesized fuels can fill in. On the occasions where carbon-neutral fuel resources are insufficient - gas generation can remain as fallback (increasingly rarely) until sufficient carbon neutral fuel production is developed.

And with atmospheric CO2 capture, CO2 released by intermittent fallback fossil combustion can be scrubbed at a rate well below the peak release rate at the turbines.

The entire UK energy use is about 2% of the world total. So I'm guessing that the storage needs of your specific (and oddly non-grid connected) home is only a small fraction of that and doesn't affect the larger, global, back of the envelope figure I'm looking for.

Using numbers to quantify the situation is part of why I'm asking, asking (and answering) the right questions so we don't get silly answers is the other important part though.

Having short days in the winter, long days in the summer isnt a UK specific thing. As you get closer to the equator it may become less of a problem as summer air con load overtakes winter heating as the dominant load, but we are talking about large swathes of the planet that will have higher heating demand that can't reasonably be covered by solar power alone.

If you could put some kind of number on how much of a not-UK thing it was on a global scale, that would answer my original question.

Like, apoarently air con is currently 10% of all electricity and set to grow. 10 AC units will be sold every second for the next 30 years.

It seems solar plus batteries has that covered, but I'd be interested to know the ratios.

According to this [1] it looks like about a third of energy used in commercial and residential buildings is for space heating. That only accounts for some 32 Pwh of 157Pwh [2] of energy use but is greater than the 21-23Pwh of total electricity generation.

So judging by your figure of 10% of electricity usage that's 2Pwh (10% of 21-23Pwh) for air con v 10Pwh (third of 32Pwh) for space heating.

But that energy demand is mostly based on combusting fossil fuels directly, decarbonisation would require a move to electric heating which would probably involve heatpumps, and quite a lot of insulation, so I would hope that that figure would drop quite alot in the future.

[1] https://www.sciencedirect.com/science/article/pii/S136403211... fig 3.

[2] https://en.m.wikipedia.org/wiki/World_energy_consumption

The sun does not shine every day. And not equally throughout all seasons. As the percentage of renewables increases so will the gaps on bad days/weeks.

The differences can be dramatic and must be compensated some way. https://www.energy-charts.de/power.htm?source=solar-wind&yea...

I'm familiar with the concept of seasons, just as I'm familiar with the concept of yearly boat trips.

But lots of people live near the equator, just like lots of people don't really ever go for long drives so rather than focus on one particularly difficult case and then giving up I was looking for some kind of rough quantification of the problem.

In analogy I'm just wondering whether the equivalent of the "daily commute" for most of the planet can be easily covered with current tech at a cost less than curent alternatives and we only need to worry about a comparitively small fraction that remains, which we could apply those savings to dealing with.

Equatorial weather patterns are not all that relevant because most of the energy consumption is at higher latitudes which experience significant seasonality.

And your daily commute/boat towing "analogy" is not really helpful because those problems are not comparable. Boat towing is an optional luxury and can be worked around at a slight inconvenience by visiting charging stations. Lack of long-term storage on the other hand can't be worked around, it requires double-spending to keep dispatchable fossil fuel plants in standby.

And I don't think there is a dichotomy either. Variability exists on a continuous time-scale from seconds to seasons, so storage solutions can expand along that continuum.

95% of the world population lives below 50 degrees N/S.

And I suspect if seasonality really does become a problem that's costly to solve, heavy industry will move down toward the equator. Ship the finished products rather than try to make them where it doesn't make sense to.

I intentionally wrote energy consumption instead of population because those distributions differ.

Short term solution: natural gas peaker plants could be repurposed to provide power during lulls in production (better than most fossil fuels, but very expensive and still produce greenhouse gases, so can't be the solution for long)

Medium term solution: long-distance transmission, overprovision, and even more grid-scale storage (whole continents don't get cloudy & calm at once for long periods of time)

Potential long term solution: space-based solar (power is 24/7, just need lots of landing locations + distribution, because microwaves are attenuated by atmospheric moisture)

Edit: tl;dr is that for the issues, there are many potential solutions

Wont introducing net energy to the earth add to the warming?

Yes, it would (though an interesting comparison would be ground-based renewables vs. carbon capturing land use + space based solar).

Another possible long-term, space-based alternative to reduce global warming would be space-based manufacturing. It's equally as plausible - if the economics work for space-based solar, it means either space-based-manufacturing is a thing, or launch costs have dramatically lowered

Barely. The temperature of the earth is a function of the balance between energy inputs and energy outputs. The solar input is much much more power than total human utilization, so adding "human" amounts of power won't shift the balance very much.

We're talking “long-term” here, as in when the humans are using over a thousand times as much energy as they are now, so that it's cheaper to launch solar panels into fricking space than to upgrade the ocean-carpeting solar panel rafts blanketing much of the oceans to be 90% rather than 75% efficient. So yes, even before that point, solar energy harvesting will begin to have noticeable climatic effects.

1000x would make human use ~20% of the earth system, but earth side solar wouldn't get counted twice, so it would be less. I agree that would have an impact, but I doubt the other poster was imagining the 1000x scenario.

The numbers are something like 20 terawatts for current human utilization and 100 petawatts for solar power.

Yeah, that's about what I calculated in https://news.ycombinator.com/item?id=20426368 — 130 petawatts of terrestrial insolation and 18 terawatts of world marketed energy consumption.

I think solar panels will tend to have lower albedo than the surfaces they shade (sunlight they reflect is not available for energy production, and even the 16%-efficient ones that are common today are pretty black), especially before the humans start floating them on oceans. Essentially all the energy they successfully harvest will later be converted to heat on Earth. So they will represent a positive thermal forcing in the climate system. I don't know what you mean about getting counted twice?

I don't know why anyone would propose orbiting solar panels when it's still easy to put solar panels on deserts. So I assumed that was what they were imagining, even though that's not a long-term scenario; it could happen before 2089, which is very short-term.

Solar energy gathered and utilized inside the earth energy system doesn't add energy to the system.

So if we were using 5 Petawatts of ground based solar power, we wouldn't be adding much to the overall system. Unless we stored a lot of it and shifted the heat to a different time of year.

Oh, I agree that you don't add energy to Earth just by harvesting energy that would have been absorbed and converted to heat. So a solar panel on the ocean doesn't warm up the Earth any, because the ocean is just as black as the panel. But you do add energy to the Earth by harvesting, or merely absorbing, energy that would have been reflected into space. So the same solar panel on the Salar de Uyuni will warm up the Earth. That's what my comments above about albedo are about; I'm not double-counting in the sense you thought. Does that clarify?

None of this is significant until solar power generates substantially more power than current world marketed energy consumption, which is to say, for at least a decade or three.

Thank you for helping me to explore these issues!

I may be incorrect, but space based solar would catch energy that would reach the earth anyway, basically shading us. So it would probably be more or less adding the same energy, but sending it to the ground rather than dispersing some in the stratosphere.

That's incorrect, sorry.

ok, but can you explain why, please?

The satellites would be in wide orbits (typically imagined to be geostationary) that would almost never be on the line between the Sun and the Earth. And if they were between the Sun and Earth, they couldn't beam power to the Earth's night side.

Not substantially. The issue is that energy that would normally escape does not.

If an electric car can do 200km on a charge, that's 99.9% of my annual driving needs met.

>Seems like most of the world just needs solar to last until the next day

as long as it's not cloudy for a week

Did the person sizing the solar panels in this hypothetical forget that it was sometimes cloudy? Seems a rather silly mistake.

Wouldn't they have access to data predicting to within fairly narrow margins how often this would occur statistically?

I mean if cloudy weather reduces the output by a factor of 4 then you don't really need to bring long term storage into the matter to have trouble if you're going size everything on the basis that will only provide the required energy on the very sunniest days.

You are going to build panels that will provide enough energy for one day on the day it is the least sunny at the site? What do you do with the excess on sunny days?

You can turn it off. Solar panels can be turned off in nanoseconds with nothing more than a transistor. Or you can invest the zero-cost energy in something potentially useful that doesn't need a constant power supply: electrowinning metals, removing corrosion, heating the hot tub back up, making ice to run your fridge or air conditioner when there's less sun, running low-cost machine tools (the kind where the depreciation is insignificant compared to the cost of power and consumables), generating hydrogen by electrolysis to convert into methane or methanol or ethanol for fuel, baking some bread, washing some laundry, doing a short pottery firing in your kiln (<8 hours), recharging your electric car, in short any of a wide variety of processes.

or you could have some batteries, save the energy for a cloudy day, and have a smaller installation of panels.

Right, the problem is that a week’s worth of batteries is an immense cost that is reducing only slowly, while waiting a few days to wash your laundry is close to free.

Let's say you use 1 kW on average in your house. You'll need about 4 kWp of panels to supply that with a typical capacity factor of 25%—a bit more panels than that if you're in Scotland, a bit less in Perú, but about that. At current prices that's US$800 of low-cost panels.

Now consider a couple of alternatives. You can overprovision the panels by a factor of 5 so that even on cloudy days that block 80% of the light, you can continue rocking your one-kilowatt lifestyle. This costs an extra US$3200 of panels. Or you can buy, say, a week of lithium-ion batteries; say you're paying US$150/kWh for the batteries, which is in the ballpark but not exact. You need 7*24 = 168 hours of batteries, which at one kilowatt is 168 kWh. This costs US$25000. In both cases you need some power electronics (charge controllers, inverters, etc.) but this is not a significant cost difference between the two scenarios.

So 5x overprovisioning of solar panels costs you about an eighth of a week’s worth of battery storage. You can overprovision by a factor of 50 for less than the cost of that week-long battery backup. There's a trade-off between the two—more battery storage means you can get by with less panels, and vice versa—but this strongly suggests that in most cases the optimal trade-off is not going to involve multi-day battery banks.

How about a few years in the future? This year, panel manufacturers have apparently sewn up the market in a cartel, preventing prices from falling over the last few months. But sooner or later, the cartel will break down, and panel prices will return to their long-term learning curve, say 20% per year until the raw material costs dominate. In 5 years, panel costs would then be 40% of what they are today: 8¢ per peak watt instead of 19¢. No such dramatic cost reductions are on the horizon for batteries.

But in a lot of cases, shifting demand to times with cheap energy is still going to be cheaper.

Often with residential solar you run out of space. At that point you have to start adding batteries (or thermal or something) for long term storage, or stay connected to the grid.

True! But batteries are very much the most expensive of those options.

Whatever seems useful. Or do the inverse of the power satellite idea: send the excess energy to space as microwaves.

I'm not sure long term storage will ever make sense since the capital costs are amortized across far fewer cycles. Once a year rather than once a day means 365x the capital cost. Instead over-provisioning of green generation and increasing grid capacity (to avoid risk of local week long lulls in generation) seems a much better bet.

So what is needed for that is a solution that trades lower efficiency for lower capital cost. Hydrogen burned in gas turbines may be that solution, if the cost of mediocre efficiency electrolyzers can be made low enough.

Luckily, there's plenty of carbon to offset (peakers) by deploying batteries before we have to worry about long term (seasonal) storage.

However, I wouldn't be surprised if the electrification of transportation ends up making batteries be the default "long term" storage option by using EVs as flexible grid storage. There's going to be 100x more batteries rolling around than what you'd need to stabilize the grid (e.g. peakers), so I wouldn't be surprised if the grid asks more from EVs than just stability so it doesn't have to spend the additional capital for seasonal storage.

Somehow, I think people waking up to find out that their EV doesn't have the charge they expected it to have will put that idea in its place.

The idea that someone with an EV will allow the power company to randomly deplete its battery is a fantasy.

Edit: We're much more likely to see home standby generators put to that use, or home batteries (connected to solar) put to that use.

Keep in mind that we're talking about single digit percentages of a battery's capacity being used. Also, its likely that many of the EV-to-grid arrangements will be for fleet vehicles (company/city cars, plus self-driving fleets when they happen), and they will be technically sophisticated enough to negotiate that capacity sharing with the grid operators.

So you won't ever actually have a scenario you describe where a private EV will have it's battery depleted.

"likely that many of the EV-to-grid arrangements will be for fleet vehicles"

Wont these be out and about during the day when the sun is shining and need to be charged during the night? Seems like these are the opposite of storage, they would exacerbate the problem? Creating a lot of demand(lots of buses in lots of cities) charging at night, and driving during the day.

"will be technically sophisticated"

This is an assumption, though it does seem technically possible.

"scenario you describe where a private EV will have it's battery depleted"

Im an environmentalist myself, but if I wake up on the morning of my vacation, and my car is at < 100%, I'm not going to be happy. You might say you could provide a way to signal that your car needs to be at 100%, but I see a non trivial amount of people enabling that all the time.

I guess if we require every parking lot with > x spaces to have plugs for all vehicles to charge/store during the day. Still requires people to actually bother plugging them in. Of course, they would also have to plug them in during the night, and unless they are getting paid to, I dont see people going through the minor hassle if it is no benefit to them. I'm just not really optimistic about the vehicles as storage idea, sadly.

Sorry for the pessimism, but I feel like there is too much naive hand waving among environmentalists. I think we need a more realistic, critical discussion, focused on likely, probable solutions. Theres too much: "Everybody can just go out of their way too do x." Theres still people who think putting an aluminum can in a different bin is too much work. We have to work with these unfortunate realities, not just dismiss them.

> if I wake up on the morning of my vacation, and my car is at < 100%, I'm not going to be happy.

That will never happen, unless you intentionally jump through several hoops to shoot yourself in the foot.

For example, there's a company called OhmConnect that does aggregated demand response. You can sign up for it, optionally connect your smart thermostat/plugs, and bank account. When there's going to be a high load period, OhmConnect notifies you asking if you want to participate in an upcoming "Ohm Hour". If you do, they will pay you $5-10 for your trouble. So, if you had a vacation the next day, of course you wouldn't opt-in to the demand response event.

Disclosure: I am on a nonprofit board with OhmConnect.

Is there any storage mechanism whose costs don't increase in the same way?

Anything that involves volumes and abundant resources.

E.g. storing hydrogen or compressed air in a salt dome decouples the power rating of your storage (MW) from the storage capacity (MWh). Similarly hydroelectric dams contain water by volume but you only need to dam a cross-section of the valley.

> Similarly hydroelectric dams contain water by volume but you only need to dam a cross-section of the valley.

I would caution that the cost of a dam does not increase linearly, either (and that most resources that are profitably dammable at utility scale have already been dammed)

Wouldn't the above description also fit flow batteries?

It's a little less clear-cut since you still need to pay a linear cost for the electrolyte, but it should be much lower than the same amount of Li-ion storage, at least in raw material costs.

Also batteries lose their stored energy over time.

Reply on news.ycombinator.com