Hacker News Reader

Comment by retrac | Hacker News Reader

Heat storage has an aspect that was counterintuitive to me, but follows from basic geometry. It benefits greatly from large scale, since the ratio of the volume to surface area [1] decreases the larger you make a container. Accordingly, if a heat tank is large enough, the surface area becomes negligible relative to its volume, and it, in effect, becomes well-insulated by its own mass. For really big tanks, like might be used for an entire town to heat itself over a winter, practical self-discharge rates can be just a few % per month, which is better than most rechargeable battery technologies.

[1] https://commons.wikimedia.org/wiki/File:Comparison_of_surfac...

View on HN

> better than most rechargeable battery technologies

you can't compare heat storage to electricity, because you can't directly use heat for anything else other than for heating, where as electricity can be used to perform motion.

If you used heat storage as a battery, there's an additional loss when converting to electricity.

However, if the heat is cheap/free during summer, storing it for winter is a no brainer.

The sand is gravity-fed into a heat exchanger which transfers the heat to a fluid that drives a combined cycle turbine, in this case. I'm curious what the conversion loss is here.

To add to that, what is the energy expenditure of building the battery compared to sand containment plus heat exchanger and turbine - i.e. mining, refining, transport, manufacture, delivery?

Gas turbine efficiency targets for combined cycle are targeting 65%, so loss is at least 35%. To me I the loss sucks, but if it can be cheap to build and maintain it matters less. If solar is the energy source and that gets driven cheap enough then that loss could be acceptable.

No way you'll get 65%. This sand won't be stored at the same temperature as a natgas generator's combustion chamber. Efficiency depends on delta T.

I'd guess 20%

They're estimating over 50%, apparently. The sand is stored at over 1000 degrees. Not combustion temperatures, true, but also nothing to sniff at.

Nope.

The report says that the sand will be heated up to 1200 Celsius degrees.

This is much higher than the maximum temperature for steam turbines and equal to the temperature of the gas in the best gas turbines.

Therefore they will use a combined cycle, first a gas turbine will use hot air passed through the sand and the exhaust from the gas turbine will produce steam for a chain of steam turbines with decreasing working temperatures.

There should have been no problem in reaching a 65% efficiency for the conversion from heat to electricity, except that between hot sand and a gas burner there is the same difference as between an electric capacitor and a battery, while heat is extracted from the sand, it cools down.

Presumably, when the sand becomes too cold, the gas turbine is bypassed and the hot air just produces steam. When it becomes even colder, I suppose that the first steam turbine is also bypassed and only the low-temperature steam turbines are used.

This will lower the average efficiency, probably to around 50%. If the residual heat (after the steam turbines) had been used for heating or for cooling (i.e. heat-powered air conditioning), the efficiency could have been higher, e.g. over 80% in the beginning, while the sand is still very hot.

I did not read the report, but one way you can improve the round trip efficiency is to not let the sand cool below, let's say 1000 °C. You pay a one-time penalty of heating the sand to this temperature and then you operate the storage only between 1000 and 1200 °C. This reduces the capacity of the "sand battery", roughly by a factor of 6. But sand is (reasonably) cheap.

I have seen another proposal where instead of sand the material would be graphite, which can be heated to more than 2000 °C. Graphite also has a huge heat conductivity. But it's more expensive.

Sand has the added benefit over graphite of being nonflammable.

Sand will also benefit from greater insulation due to lower packing efficiency and therefore more air.

IIUC there is a physical limit from thermodynamics in converting back, based on the temperature. So when efficiencies are quoted, it could be either the proportion of input energy retrieved, or the proportion of the theoretical max efficiency achieved. I'm guessing you're quoting the former?

Doesn’t even need to be seasonal.

Plenty of times+places where you only need heating at night and still have some net electrical draw (because the sun isn’t shining at night).

Also, cold fronts move in with a lot of wind, but then it can stay cold a few days with calm winds until a warm-front moves in.

Nit: Rechargeable batteries have a loss when converting the chemical energy stored inside to electricity. There’s no free lunch.

Yes, but for Lithium and sodium batteries you are looking at 90+ efficiency.

Is a battery electricity or is it a chemical reaction?

you kinda can with sterling generators I think

> you can't directly use heat for anything else other than for heating, where as electricity can be used to perform motion.

Maybe today. But NASA has some interesting metal tires [0][1] which might change your mind for the future.

[0]: NASA info https://technology.nasa.gov/patent/LEW-TOPS-99

[1]: Neat youtube vid: https://www.youtube.com/watch?v=vSNtifE0Z2Q

> you can't compare heat storage to electricity, because you can't directly use heat for anything else other than for heating

You can use heat to generate steam that can spin turbines to perform work. But, I am not sure if it’s practical to generate steam from a sand battery, my background is in electrical construction. I’m guessing the sand battery isn’t nearly as hot as a natural gas steam boiler’s combustion chamber.

https://www.johnsoncontrols.com/en_sg/hvac-equipment/chiller...

Is absorption refrigeration bad enough to be permanently non-viable here? Or is it "just" a matter of scaling a niche technology?

Define "bad enough" in more specific terms.

I believe it largely depends on the application. If you have a lot of waste heat, it's potentially a way to get "free" refrigeration. (e.g., a paper plant that uses a lot of steam can use absorption chillers to make use of waste heat.) If fuel is much cheaper than electricity, it can be economically viable. Peak shaving can save lots of money. etc. But it's probably not competitive purely in terms of energy efficiency or GHG emissions.

If I had exact figures, I wouldn't need to ask for anyone's intuition. Thankfully, ChatGPT was willing to work with me, and I will summarize the results here.

Typical refrigeration COP (Coefficients of Performance):

Absorption refrigerator: 0.6-1.2

Compressor refrigerator: 1.5-4.0

Estimated TES economic advantage: 1.1-2.5x

Conclusion: yes, absorption refrigeration is probably inefficient enough to make it a long shot in this application. The only way I can see it becoming viable is if extremely hot TES can completely change the efficiency game, and then only just.

The temperature in TFA are outside the range of most commercial absorption chillers, so this is more about making cheap electricity. I would imagine absorption would be more applicable if the same tech was used to generate lower-quality heat that's not suitable for a combined-cycle generator.

> outside the range of most commercial absorption chillers

"Off the shelf" isn't a constraint here, sort of the opposite: I'm trying to imagine the space of things that are physically possible but not yet commercially mature. Theoretically, higher temperatures mean more possible efficiency. That's the one ray of hope in otherwise dismal efficiency figures that are currently only viable, as you point out, if the heat is ~free.

>Theoretically, higher temperatures mean more possible efficiency.

I think you might be conflating a few things here. For a cycle to produce electricity, that's correct. But the mechanism of absorption chillers is fundamentally different. The chemistry of the materials and their phase change temperatures are definitely a constraining factor. Could you, in theory, develop some other absorbent/refrigerant that works at those higher temperatures? I suppose, but I would suspect there are much easier ways to get efficiency gains.

because you can't directly use heat for anything else other than for heating

There's Stirling Engines. If the solar collection is pure thermal, and if that collection and the storage can be made dirt cheap, then the 37% or so efficiency of conversion to electricity stops being a problem. But what are currently problems with Stirling Engines -- Hardly any of the industrial optimization has been applied to them Re: Wright's Law. So they are quite costly! Heat pipe solar thermal could be made dirt cheap through economies of scale, and it works very well, even in climates like England's.

I could envision house construction changing to include sub-basements which are just polystyrene insulated boxes filled with sand. By over-provisioning storage by 4X, houses in cold climates could have huge electrical power stores, especially in summer. (Especially if the house uses heat exchangers which can draw directly from the thermal store.)

But that doesn't really matter before all heating demand is served from storage. Sure, heat will never be the be all end all of energy storage, but there's a lot of demand in places that have winter. On top of this, when the conversion to heat is done by heat pumps, you not only get the benefit of a warmer baseline during the conversion, you also get some free cooling while charging.

Heat can be used directly not only for heating, but also for cooling.

There are air conditioning systems that are powered by heat, not by electricity.

There are places where the power plants use the residual heat from the generation of electricity not only for heating during the winter, but also for cooling during the summer, by producing chilled water.

> There are air conditioning systems that are powered by heat, not by electricity.

How does that work?

https://en.wikipedia.org/wiki/Absorption_refrigerator

Either absorption refrigeration or a steam turbine that compresses the working fluid.

Half of all primary energy usage is for heating purposes. Heat is the best format to displace, as well, since a significant chunk of all that primary energy use is direct fossil fuel combustion right now.

Especially since you can use heat to drive a heat pump, which ends up actually achieving above unity efficiency effectively since the heat input is used to move additional heat from the environment. Even if it's inefficient, if the heat would otherwise have been wasted, it's a net benefit.

The Pernell with heat compared to electricity is portability. It is easy and efficient to transmit electricity but heat is not

> The Pernell with heat

Did you mean to type "problem"?

yes, but somehow it got "corrected" to Pernell. I'm not sure who Pernell is or why they would be relevant to the discussion. :-)

Thermovoltiacs, converting heat dirctly to electricity. IIRC, At least one of the "box of rocks" thermal battery startups (interviewed on the Volts podcast) intends to have both a two way heat exchanger and electricity play.

An extrapolation directly to heat transfer is the Biot number

https://en.m.wikipedia.org/wiki/Biot_number#:~:text=The%20Bi....

Gravity seems much more natural to me than heat; eg pumping water up a dam and releasing it again when you need the energy.

It strikes me that heat has the problem that it always loses energy in its “stable state” because the surrounding environment absorbs the heat, and gravity doesn’t have this problem.

With water you lose some due to evaporation.

That was my first thought too. But lakes lose ~20% / year to evaporation, and with the use of shade balls that is cut by ~90%, so we are at 2% / year - which is about the same as very efficient daily loss from heat storage.

If the heat is being turned back to electricity, a heat sink is needed and if this is done by evaporation the water loss will greatly exceed that of natural evaporation from a PHES facility.

It's not the evaporation per se that matters in the pumped hydro, it is the evaporation loss of water you already invested in pumping, so it is just an efficiency loss. The vapor loss for the heat sink is indirect, you usually just calculate the turbine efficiency. I don't know if the (generally much smaller) pumping requirements for turbines is already included in their efficiency calculations, but it would need to be of course.

Either way, it just goes to show further that pumped hydro would be more efficient, when and where it is feasible.

With a mineshaft gravity storage you could actually try to spin up a convective loop powered by geothermal that dries the sand over time, allowing you to lift dry sand when charging and abseil heavier wet sand when discharging.

The issue of finding a location that has dramatic elevation change, a basin capable of storing vast amounts of water, and a suitable source/sink for pumping make pumped hydro difficult to deploy. Then there are additional logistics challenges such as environmental damage and proximity to human settlements for maintenance and engineering teams (prior challenges mean you don't really get to select the locations).

Heat sinks meanwhile can be built wherever you have a big rock by drilling some holes. Additionally, gravity definitely does lose stored energy in its "stable state", through evaporation and water entering the water table. Losses depend on geology and local climate, but it's not negligible.

Not to say that pumped hydro is a bad technology, it's just got it's own challenges and uses. It's most applicable in the form of electrical grid storage. But specifically on the scale of keeping towns and cities warm, heatsinks outperform almost across the board.

And, if you inundate an area of wilderness to create reservoir you have to also count the lost carbon capture of the growing plants, and the significant methane emissions of the now dead and rotting plant matter under the water. In dry deserts this may be negligible, but the mountains where people most want to install hydro projects are generally very forested.

The "vast basin" part of that is an exaggeration.

Look at https://www.whitepinepumpedstorage.com/ and the sizes of the upper and lower reservoirs there. This is to be a 8 GWh, 1 GW facility.

"vast" is a pretty non specific word so I can maybe see why you would think that? I feel like it's pretty obvious from context I'm not talking about an inland sea here. The quoted 5000 "acre-feet" in that project is a considerable amount of water for a man-made structure!

Either way, the details page[1] supports all of my above points. It even comments on the page how rare it is to find a suitable site like the one they've chosen.

[1]https://www.whitepinepumpedstorage.com/project-details

> This is to be a 8 GWh, 1 GW facility

That is big for time-shifting daily energy usage, but way too small for seasonal energy storage

Right. PHES is not a good fit for seasonal storage.

Pumped hydro is great, but the potential energy from gravity is actually really low. Stacking blocks will probably never work, and pumped hydro only works due to scale and existing geography. It’s also not modular and can’t be co-located with generation or loads unless the geography works out.

Terrament is working on a modular gravity storage solution that uses deep mine shafts to gain 20x more height than stacking blocks above ground. So you don’t need water or mountains. And since gravity storage uses ballast that is really just dumb weight, it could even be economical to make that ballast a secondary storage like thermal storage.

Pumped hydro is great, but only in certain areas where you already have the elevation gain available.

Compressed air storage is another one that's pretty good, but it's only particularly good if you can store it in underground caverns or unused mines, so it's also geography dependent.

For longer term storage, producing hydrogen can be a good one.

And batteries are actually fast becoming competitive with some of these options from the other end, generally better for shorter term storage but getting better at longer term. Even shorter term, flywheels can be a good option.

There's room for several different types of grid-scale energy storage, based on how efficient they are at different energy storage periods and numbers of charge-discharge cycles, and also in some cases on local conditions, like the availability of terrain and water sources for pumped hydro.

There's a good paper here which shows what the most efficient energy storage systems are for various combinations of length of storage (hours per discharge) against number of discharges per year, and for each one shows the current cost, and a predicted cost based on trends in technological advancements: https://www.sciencedirect.com/science/article/pii/S254243511...

There's a lot of the chart that is dominated by pumped hydro currently, but plenty of other storage technologies that are more cost effective on different timescales and numbers of discharges. But it looks like it's predicted for prices of battery storage and hydrogen storage to fall relative to the others, causing a different predicted landscape in 20 years.

And some of these, like pumped hydro, are dependent on geographic features, or access to certain resources, so even when one dominates overall, there can be others that dominate in particular geographical regions.

When you can store electricity more directly instead of the (inefficient) conversion to and from heat it is just crazy to not do so. With electricity, thanks to heat pumps you can "move" heat instead of producing it, which makes heating a lot less expensive since the "efficiency" can often reach 300-400%. For that reason you want to keep using electricity for as long as possible in most cases. Of course, if you have a huge source of cheap heat the equation might change. In such cases you might still want to use some of the heat to get higher temperature heat. (https://en.wikipedia.org/wiki/Absorption_heat_pump)

And no, it will not help too much if you use a heat pump before storing the energy in this facility. You still have huge conversion losses, the volume for low-temperature heat storage will be huge. If you want high temperature storage your heat pumps will need more stages and run less efficient.

> the ratio of the volume to surface area decreases the larger you make a container

Did you mean to write the reverse? i.e the ratio of the surface area to volume decreases the larger you make a container.

Yes.

I also think this is a very interesting approach to take. This might be a pipe dream but it would be really cool if we could just turn a big chunk of desert into a battery by plumbing some heat exchangers through it.

Or maybe just putting a gigantic Fresnel lens in the desert and pointing it at the ground. I almost wish someone would try this just to see what would happen.

Reply on news.ycombinator.com