[cupofpython]

emphasis on massive scale.

Moving 500,000 kg (over 1 million pounds) 7.5 meters (~25 feet aka the height of a house) will give you about 10 kWh of energy. This is equivalent to running a 425W device all day, like a small air conditioner. The relationship is linear. Double the weight or the distance to double the energy. All of the metal at a scrap yard I know of amounts to less than half that weight, for reference.

I'm also a fan because pumped storage is a really interesting storage method, but it is beyond niche. It is very tough to move that kind of weight around efficiently for what you get back. Pumping water to great heights is not easy either. (see also: moving rail-carts up a mountain)

[SECProto]

> All of the metal at a scrap yard I know of amounts to less than half that weight, for reference.

That's not a great reference point when you're trying to visualize to pumped storage, as water is 1t/m3 while steel is up around 7 or 8. Also, 500t of steel at a scrapyard seems very small - 70m3?

A better reference might be a back yard pool, which might be in the 30-40t range - so like lifting 15 back yard pools the height of your house to power a tiny AC.

[cupofpython]

good point, and yes its a small scrap yard. I was trying to emphasis that it's possible for a full-time commercial operation moving heavy metal to involve less weight than what was referenced. The backyard pool paints a better picture though

[xyzzyz]

> Moving 500,000 kg (over 1 million pounds) 7.5 meters (~25 feet aka the height of a house) will give you about 10 kWh of energy.

In dollar terms, 10kWh is worth around $1. 1 million pounds is the weight of 2-5 residential homes, depending on size. Think about it: the cost to lift a couple of entire houses three stories up into the air is literally just one dollar. That’s why gravity energy storage only makes sense at a massive scale.

[cupofpython]

it's also why "storage" is a very loose term for gravity based energy storage. at a massive scale it is still only best at storing/discharging the difference between demand and supply - while still trying to keep actual energy production as close to demand as possible at all times. It really should never be used to power a city the way we would use a battery to power our phone. As in, spend significantly less time charging it than discharging it

[PaulHoule]

They are building a lot of them in China

https://en.wikipedia.org/wiki/List_of_pumped-storage_hydroel...

[cupofpython]

Makes sense, it is definitely a useful tool. I just think it is insufficient to act as storage. It can be good at producing variable amounts of Watts on demand but not so good at storing enough Watt-hours to keep things running for very long. I can see a great appeal for it to help with load-balancing for a significant amount of choppiness between supply and demand on the hour timescale.

For something like solar, where we will want to store over half our daily energy production at peak storage (ideally 2-3 days worth I think) - I don't think it holds up. Additionally, it doesnt seem like a good bet as a primary mechanism for either storage or on-demand generation if energy consumption continues to increase due to the rather large coefficients involved for scaling it up.

"The United States generated 4,116 terawatt hours of electricity in 2021"[1]

4,116 TWh/year = 11.2 TWh/day

The storage capacities for the largest items listed on the wiki is on the magnitude of GWh. The scale goes kilo-, Mega-, Giga-, then Terra. So we are talking about a need on the order of a thousand pumped storage facilities per country. The US would need over 50 of them per state (on average) in order to keep everything running without production for 24 hours. Doesnt matter how many solar panels we have, if we get 1 dark day then we would run out of power. If we tried to rely on solar entirely, we'd also still need very roughly half that amount of storage just to get through the night.

lithium batteries are obviously much better suited for overnight storage, but I have no idea what the numbers are on how much lithium is physically available to use as such storage.

If we want to get on the order of monthly to yearly storage to allow, for example, solar panels in alaska to provide enough energy for a resident to get through months of darkness - I have no idea what the leading storage options are, probably lithium still

[1]https://www.statista.com/statistics/188521/total-us-electric...

[Schroedingersat]

Sodium ion is expected to sharply take over cost limited applications some time in the next couple of years. There are pilot mass production programs designed to avoid scarce materials that drop into existing processes. Natron have products on the market (at presumably high cost) targetting datacenters for high safety applications.

For longer scale storage it's a tossup between opportunistic pumped hydro, CAES where geology makes it easy, hydrogen in similar areaswith caverns, ammonia, synthetic hydrocarbons, sodium ion, and one of the emerging molten salt or redox flow battery technogies. Lithium isn't really in the running due to resource limits.

Wires also have a lot of value for decreasing the need for storage. Joining wind and solar 1000s of km apart can greatly reduce downtime. Replacing as much coal and oil with those, and maintaining the OCGT and CCGT fleet is the fastest and most economic way to target x grams of CO2e per kWh where x is some number much smaller than the 400 of pure fossil fuels but bigger than around 50. Surplus renewable power (as adding 3 net watts of solar is presently cheaper than the week of storage to get an isolated area through that one week where capacity is 1/3rd the average) will subsidize initial investments into better storage and electrolysis with no further interventions needed.

[cupofpython]

Awesome response. I've come across the molten salt option but havent researched in depth. I saw it referenced as something a lot of scientists are hyping up, but I am not sure what kind of engineering challenges exist for implementation and maintenance.

Second paragraph is a bit too information dense, I had trouble following some of it. Renewable energy deficiencies will be localized, so i understand how wires help here. A larger connected area produces more stability, makes sense. Agreed with the carbon reduction priority to tackle coal and oil first. Surplus renewable power acting as a subsidy checks out, but that is skirting around the energy storage problem imo. Sounds like you are saying "instead of storing renewable energy, get more than you need and sell it back to the grid and then use those funds to buy the energy back later". This would certainly work for local consumers, but doesnt do too much to help the power grid itself manage what to do with the surplus energy. Sell it to neighboring power grids? Ties in to the first point about connecting a larger area - but what are the limits here? Can we physically connect the sunny side of earth to the dark side? (ignoring that it seems logistically/legally prohibitive)

the question really comes down to what should we be spending money on to get "better storage"? What are the best solutions for long-term local storage?

[Schroedingersat]

> the question really comes down to what should we be spending money on to get "better storage"? What are the best solutions for long-term local storage?

The solution I'm proposing is basically 'the best place to spend your money on storage is to not spend it on storage yet'

If the goal is to reduce emissions asap, then focusing on the strategy that removes x% of 100% of the emissions rather than 100% of y% of the emissions makes sense unless there are enough resources/money that y% is more than x%. And storage is currently expensive enough that you need many times as much money for this to be true to 99.9% confidence.

Getting a wind + solar system that has at least y watts at least eg. 90% of the time is remarkably affordable already and still going down.

In excellent climates new solar costs less per MWh than fuel for a gas turbine (and is not far off fuel for a nuclear reactor). Wind is not much more. Distribution, dealing with less than ideal sites and oversupply increase the cost, but an ideal mix has very little storage (4-12 hours) which can be delivered by lithium batteries.

By relying on the existing fossil fuel/hydro/nuclear/whatever to pick up the last 10% for now, you can replace more coal/oil more quickly than other strategies. During this build all storage technologies where they make the most sense so that when that last 10% is needed, prices will have dropped. I'm fairly sure some mix of green hydrogen and green ammonia burning in those same turbines will be one of the winners (ammonia in particular has negligible marginal cost of capacity allowing for a strategic reserve, and will be needed to replace fossil fuel derived fertilizer anyway).

In the unlikely case that there's an overnight $2 trillion investment in new wind/solar/powerlines and production capacity to match in the US then choosing a dispatchable power source from some or all of: expensive green hydrogen, expensive abundant existing batteries, expensive pumped hydro, and expensive nuclear or immediately going all in on commercialising every vaguely promising electrolyser tech becomes the priority.

[cupofpython]

Completely agree with the hybrid approach wrt reducing emissions. I am talking more towards work that would be done concurrently with that.

> During this build all storage technologies where they make the most sense so that when that last 10% is needed, prices will have dropped

this is kind of the point of what I'm getting at. Without any investment, none of the storage technologies are going to make much progress. If not financial investment, then at least a time investment from research/science teams. then again, maybe opportunism/free market will take care of this and we can assume any progress that can be made will be made by people trying to make a name for themselves or be first to market. I'm still curious to size up what that progress might look like for discussion/entertainment purposes in any case

Good storage solutions would immediately pay dividends through arbitrage, which would keep electric prices stable, and then anywhere renewable energy generation is more than demand and storage is sufficient, that stable price point could come down below the cost of using coal/oil as well as any other continuous production method. We would be able to consolidate power generation over time, not just space, and realize gains from that. As in, use massive bursts of energy production to top off storage and use them to exactly meet demand. Maybe this opens the door for more alternative energy production methods as well (that are better suited for burst than steady)

[kenhwang]

It's not "beyond niche", it accounts for 95%+ of worldwide stored energy and is the de-facto energy storage mechanism that all new battery storage technologies are compared against. It also has round trip efficiency comparable to the li-ion batteries (80-90%), which is incredibly hard to beat.

[cupofpython]

95% stored energy by what measurement? See my other comment. It is not accessible to everyone, nor can it be made accessible to everyone, and the current storage capacity is a marginal fraction of what we actually use. It's a short term load balancing tool that operates within a small energy window.

[kenhwang]

95%+ by total energy stored/provided.

There's little to no water use in the storage or discharge of pumped hydro, water goes from one reservoir into another. The limiting factor is how much water can be pumped/discharged, not how much water is available in storage (which tends to be significantly more than the amount pumped around). So there's little reason why they wouldn't currently be fully utilized.

It's true that it requires specific geography (water and a place to put water), but it turns out population centers tend to be developed near water sources, already store water for the sake of storing water, and water can feasibly be stored in large quantities underground as well. Which means there's practically many viable large capacity sites near the places that use electricity.

[cupofpython]

the 95% is misleading. it is barely storing or providing energy but it is a passthrough akin to plugging your phone into the charger 24/7 and saying your phone battery is providing 95% of the energy just because the wall outlet charges the battery first then the battery powers the phone (not an exact metaphor). If you unplug your phone and the phone dies 10 minutes later, you wouldnt say your phone has a good energy storage solution.

Pumped storage is great at what it does, no denying that. And what it does is allow energy production to remain near average while demand varies, and consequently allows energy production levels to be adjusted a bit slower. You aren't addressing the raw numbers though. It serves best as a compliment to a continuous energy production system. As an actual battery/storage solution, it is weak. So it will not be the solution used to store a massive amount of energy generated over a short period of time in order to be used over a longer period of time.

I agree they should be fully utilized, but I am trying to explain that if you fully utilize pumped storage you are still going to have an incomplete energy storage problem. Of course the water levels dont get near max or min capacity - it is designed to take out exactly what you put in as soon as possible or else there is too much risk. The raw storage capacity is small to medium sized - about 10 hours at max discharge (and max discharge might not be enough to keep up with demand entirely on its own).

Basically, the more energy you need to draw the faster you need to drain it and the more energy you want to store, the more massive your reservoir needs to be.

These things cannot be made 100 to 1000 times bigger, nor is there capacity to make 100 to 1000 times more of them. We are better off having them vs not having them but it isnt enough, and if we find a better solution it may become obsolete

[kenhwang]

The reservoirs we already have are rated in the thousands of MW. There is no storage problem, we already have large enough reservoirs where we can practically store years of energy indefinitely; especially now in drought conditions where reservoirs are regularly well below historical levels. So if there's ever a need to store solar energy from summer for use in winter, pumped hydro is the only energy storage solution that works.

They also don't have issue with storing energy quickly, they can all store energy at a significantly faster rate than they can discharge. We can run pumps as quickly as possible and install as many as you'd like, but the discharge has to be controlled (thus limited) because releasing massive amounts of water at once. So their main use case today is storing massive amounts of energy generated in a short amount of time and releasing slowly across a long period of time.

What the grid actually needs is faster discharge than charging, because that more accurately matches summer energy use patterns. This is what chemical batteries excel at which pumped hydro cannot easily do.

So it's unlikely we'll be able to make them 100-1000x bigger, but they're already 100-1000x bigger than other battery solutions. We should be able to make 100x more of them because the reservoirs already exist and very few of them currently are used as both power sources and energy storage, we simply need to add pumping capability to them in most cases.

[cupofpython]

Agreed on charge vs discharge comments.

We seem to disagree on the storage numbers. Genuinely curious if my math is wrong on this. I did research a bit more about recent advancements in pumped storage since my first comment and found that my original numbers were almost an order of magnitude smaller than what would likely be built today since I had referenced older tech. So admittedly, pumped storage is much more feasible than my original attitude suggested - which is great because id love for it to be all we need. However, I'm still not sold on it's ability to act as sufficient storage, and I do not see in any way how it could possibly keep things running for multiple days, let alone years of energy as you suggest.

There is a reason we only talk about pumped storage in terms of its discharge rate rather than its storage. We dont really use it for storage. We use it to store the difference between peak and average energy demand, not the total actual demand. You keep the generators running near average all the time, fill the reservoir during the demand valleys and drain the reservoir during demand peaks. Discharge effects ability to actually reach the peak demand, while storage effects how long you can sustain the demand. My point is even if we could discharge as fast as we need to, the reservoirs would empty in less than a day if we needed to rely upon them while energy production was down.

There is a new project (snowy 2.0) in Australia that will have a notable storage capacity of 350,000 MWh .

Current energy usage in the US is over 10 TWh per day. 350,000 MWh = 350 GWh = .35 TWh. So we would need 28 of this brand new top-end pumped hydro stations to hold 1 days worth of US energy demand in reserve. It's ballpark feasible, but lets keep in mind that this plant is costing Australia ~$5-10 billion and is working with two dams that already exist. Very much still in short-term load balancing territory.

This would also lock up 500,000 liters of water per 10kWh. 1 days worth of storage for US: 10TWh / 10kWh = 1 x 10^9; then x 500,000 liters = 5 x 10^14 liters of water = 100 cubic kilometers* (26 trillion gallons). Storing 1 years worth of energy would be 100 km^3 * 365 = 36,500 km^3; which is 3 times the size of Lake Superior (12,000 km^3). I still dont see this as an energy storage solution. MAYBE if use seawater and find a cost-effective way to build facilities into the coastline?

*(1 x 10^12 liter = 1 km^3)

Also to keep in mind that all of this is assuming CURRENT demand, which excludes the incoming energy demand increase for electric vehicle adoption. that's about 2-4 kWh per gallon of gasoline. US uses about 369 million gallons of gasoline on vehicles per day. We can add almost another 1 TWh for that, and then still whatever is necessary for increased usage in general.

[jcrben]

It also doesn't consistently depreciate the way that batteries do.

[fy20]

That's around the weight of a fully loaded A380.

[im3w1l]

What if you go down instead of up? Drill like a kilometer or two and then build a huge cavern at that depth.

[cupofpython]

that is something people are doing. Also when you go down into rock, you are able to leverage pressure as energy storage as well - which is similar to what this article is about.

There was 1 design I saw where they have a large cylinder cut out of the ground but left in place (so it is loose). Pump water underneath it to raise the cylinder up, then flip the valve and the cylinder squeezes the water back out for power through gravity. I am not sure how the sealing works on that, probably similar to hydraulics

[merely-unlikely]

Old mine shafts have been used.