Hacker News new | past | comments | ask | show | jobs | submit login
Megapack: Utility-Scale Energy Storage (tesla.com)
601 points by jjallen 24 days ago | hide | past | web | favorite | 262 comments



The energy regulatory model in Australia is complex, and broken. And rather than simplifying it, it is going to be made even more complex, and probably more broken, to resolve the problems.

Tesla is not the problem, the 30 minute bidding model, and the state/federal divide on energy policy, and the monopoly capital rent-seeking behaviour of encumbent privatized generators using coal and gas, and guaranteed yields on capital investment, and weak regulatory oversight, and a lack of national coordination, maybe they are the problem.

But Tesla exposes the problem: the Neoen battery stack has been fantastic, but needs to be kept in perspective: To use this for more than FCAS (frequency stability services) nationwide, means re-building the national transmission grid to favour battery models. Parliamentarians got up on their hind hooves and brayed about how it couldn't keep lightbulbs on for more than 20 minutes and was a waste of money: They didn't for one minute admit it wasn't designed as a longterm power store, but as a frequency-control store, stability, and for peak-price bidding to cap off supply-price disfunctions, its working fine: Something like 60% of its capacity is for private bid to earn money, 40% contracted to the state for their goals. It is a real game changer but needs to be understood, its not like a Pumped Hydro unit, or a synchronous condenser (although it is more like that, than PHES) It is its own beast. Its response speed is so fast, the regulatory behaviour models have to adapt to cope with it.

A Five minute bid pricing model is coming in under 2 years I think. I hope the majors don't work out a way to "game" it because my hope is, other mega-scale battery and PH systems will be onstream then, and be able to survive in it, and drive some more coal and gas out of the system.

We have a fictional belief in "base load" power here rather than responsive power. We also lack useful tools like demand management which can be used to bid alongside supply for pricing to avoid having to turn on nasty dirty generators, when the option to time-shift load exists (supermarkets and places with huge thermal mass can time-shift their AC and cooling or heating budget, and so avoid load)

Meantime, people are arguing for a de-centralized model using household batteries where resilience is kept in the customer distribution net, as well as at the transmission net level. There are competing pressures around "how shall we fix it" which are not simple.

If you want to read a good overview of this, look for stuff written by John Quiggin, an economist in Queensland who has written on the stuff-ups in privatized energy supply in Australia and the market failure.


I suspect every model is currently broken, and they are all distorted by: regulatory capture.

My guess is that without storage systems, wind and solar and distributed power will always be "regulated away" by wealthy incumbents.

But with storage systems, new forms of power will be valid participants and maybe there is a chance of an efficient energy market.


"Regulatory capture" hits that sweet spot of being easy to grasp, somewhat smart, and sufficiently cynical.

But while it is definitely a valid concern, and probably applicable to energy markets, it is unlikely to be the single reason for any and every problem with energy markets and production.

Using it as a highbrow euphemism for "corruption" probably causes more harm than it is useful: it gives license to ignore actual technical problems and the painstaking search for solutions, opting instead for an all-encompassing, nihilistic, and cynical narrative where nothing really can be done, except trying to tear down ever more existing institutions.

I also don't quite understand how storage systems are supposed to bypass problems arising from "regulatory capture". If incumbents get to write the rules in their favour, they could just as easily keep newcomers using storage out of the market as those without.


I really appreciate your comment but I think another element we need to consider is what type of leadership is going to force change.


Whenever I see people throw around “regulatory capture” when talking about immensely complex and finely tuned systems for delivering critical services, I think of newbie programmers who approach a big legacy codebase with the idea “oh this is all crap, we should just rewrite it.”

The electric grid is the product of a century of technical and regulatory co-design. If you were starting from scratch with an eye to accommodating renewables, you wouldn’t design either the grid or the regulations the way they are. But you’re not. You’re slowly evolving the system from point A to point B. You’re dealing with expensive physical infrastructure built on 30 year planning horizons, in many cases with express guarantees about revenues (because at the time, you needed that capacity and sought to induce someone to build it).


When I think of distortions to the energy market, I think of screwed up regulation. And it is usually not to the benefit of the customers.

A friend in texas pays about .06/kwh for residential power. In California, PG&E gets power at 0.03/0.04 per kwh, and resells it to residential customers for .22/.28/.49 per kwh. (you pay more when you use more, unlike any other commodity)

I hope these markets benefit from robust competition, and maybe projects like this help. Allow alternative energy, route around distribution problems and shake things up a bit.

(yeah, "throw out the old code, rewrite it")


I ask you - Why does the entire grid need to be rebuilt for batteries? Australia needs to split generation and transmission/distribution apart - the utilities shouldn't be able to make money of their own generating assets similar to a majority of US jurisdictions.

I'll counter that I believe the energy systems are moving in the right direction and will likely get to a better spot over the next 5-20 years as the infrastructure upgrades are necessitated and alternatives products that are more cost-effective options become more viable / pressure builds on policy makers.

And yes, energy markets are incredibly complex - especially in highly dynamic areas with fluctuating demands and supply possibilities. I believe the CA-ISO was integrating electricity for 10,000-35,000 MW every 3 seconds about 7 years ago -- probably way faster at this point. It's amazing and humbling what our grids operators are able to do.


I deleted a rant. I think you make excellent points. Its an amazing system, we should laud the engineers who keep it running. I am also hopeful for the future but wary of the resistance to structural change from the incumbency. Thats why I ranted: I'm so frustrated by a cowed regulator.


What do you mean by integrating every 3 seconds? As an ISO in the US, CAISO is pretty standard in that they have a day-ahead hourly market and a 5-minute real-time market. Of course they have very large ramps due to solar (duck-curve).


For those who aren't aware, the US power grid has been using 5-minute bidding for many years.


For what it is worth, Australia has 5 minute generator dispatch, but 30 minute settlement. The market operator is presently implementing a project to bring settlement to 5 minute as well.


15 minutes for some (most?) of Europe


I can also recommend this analysis by Jess Hill in mid-2014:

https://www.themonthly.com.au/issue/2014/july/1404136800/jes...

Quoting:

"Since 2009, the electricity networks that own and manage our “poles and wires” have quietly spent $45 billion on the most expensive project this country has ever seen. Allowed to run virtually unchecked, they’ve spent vast sums on infrastructure we don’t need, and have charged it all to us, with an additional fee attached. The spending was approved by a federal regulator, and yet the federal government didn’t even note it until it was well underway."

As you intimate, the whole AU vs state vs private interests approach to power supply leads to a breathtakingly poor result for citizens / consumers.


Do you have ongoing guaranteed yeld for coal plants? That's insane. Is it some old contract that is about to expire?

Anyway, I wonder how the grid will cope after the lion-share of the generation is done by synchronized inverters, instead of high-inertia rotational generators. I am concerned we move too fast and make our grids instable worldwide, with generators amplifying deviations instead of absorbing them.


Do you have ongoing guaranteed yeld for coal plants?

No, that is referring to capital investment in the distribution assets, not generation.


5 Minute Settlement is due to start on July 1, 2021.

https://www.aemo.com.au/Electricity/National-Electricity-Mar...


And a wholesale demand management market looks like it'll be coming around the same time.


Whats wrong with a bidding model for the bulk electricity market?


This is a bidding market, it's just the granularity of the demand blocks in that market that's changing. It already uses 5 minute dispatch, but the settlement is on 30 minutes.


>We have a fictional belief in "base load" power here rather than responsive power.

What are the time-frames to qualify as 'responsive power'? I can imagine a battery acting like a capacitor and smoothing out supply on a second by second basis, but big pumped hydro like snowy 2.0 can (will) provide 'base load' within 90seconds of being turned on.


Excellent comment.


I’d like to give it the HN equivalent of Reddit Gold... maybe HN could implement something along those lines but instead of keeping the money and removing ads, the money could get donated to something like the EFF?


My favorite thing about HN is that the votes don't matter that much and comments stand on their own. I would not look forward to that addition, unless only the receiver was aware of it.


A cryptocurrency address in poster's profile would do the trick


You can click "favorite" on the story itself and save that to a separate space. Can't favorite individual comments though (yet).

Edit: Yes you can. Click your profile and click favorite comments at the bottom.

  To add a comment here, click on its timestamp to go to its page, then click 'favorite' at the top.


I think the words "Excellent comment" alone show the parent comment as truly an excellent one. I agree, upvotes for both!


Base load is just the minimum power demand. What's fictional about that?


Minimum in what situation? At what price level?


> Minimum in what situation?

Low demand starting at night into early morning. e.g. https://energymag.net/daily-energy-demand-curve/


This is excellent, thanks.


> means re-building the national transmission grid to favour battery models.

You literally can't store enough energy in batteries to power the network. Yesterdays demand for SA was [0] 218,247,760 kWh. At even the most optimistic battery price no one has ever managed to achieve as scale of $100 USD per kWh betteries that's $21,824,776,000 (21 Billion Dollars) for one state, in Australia.

You need to double or triple that to deal with summer heat waves, having enough storage to deal with intermittent renewables etc. You'd be looking at $200 billion for SA alone, with 1.7mil people, to move to battery power, without any of the infrastructure change.

Oh and lithium ion batteries lose half their charge every 10 years or less. So you will need to invest $100 billion every decade. That's $128,381 per person in SA to move to battery power once, and then $6,419.05 per year to replace the lost capacity.

This is a pipe dream wilder than cold fusion.

[0] https://www.aemo.com.au/


> Oh and lithium ion batteries lose half their charge every 10 years or less.

Most Teslas sold today will still be driving with their original batteries and well over 60-70% of their capacity well beyond ten years. Tesla warranty actually covers them up to 8 years up to 70%. Meaning they are fully confident they can promise that and keep their money in the bank.

> even the most optimistic battery price no one has ever managed to achieve as scale of $100 USD per kWh

Cost keeps dropping: https://about.bnef.com/blog/behind-scenes-take-lithium-ion-b.... Ten years is a lot of time and Tesla just announced the desire/goal to increase production capacity to about 1 Twh/year (i.e. about a 35x change relative to production today). Whatever the cost is going to be in ten years, it's going to be nowhere near what it is today. Well below 100$/kwh in ten years certainly.

> Yesterdays demand for SA was [0] 218,247,760 kWh.

You seem to assume a complete lack of wind and solar on a day (unlikely) is going to require 100% battery reserve. That scenario is very unlikely. It might happen locally a couple of times per year but you could simply import power for regions elsewhere in Australia via a cable. Or you could install some over capacity solar to ensure that even at reduced efficiency you still get some power and use the surplus that you end up having for other purposes (clean water, producing hydrogen, etc.).

In any case, nobody is talking about rolling out that much battery. So we're talking way less battery that will cost less than you assume and lasts much longer than you assume and that will allow for the retirement of some very expensive aging coal/gas plants (which you should factor into your numbers).


And Tesla batteries are living a harsher life than grid storage cells. Thermal management, charge levels, and charging speeds are much more difficult to balance in a car than in a grid battery.


There is wind at night. It's extremely unlikely that all renewables produce no power at all for a whole day. Batteries are perfectly fine to smooth over a 50% dip in power generation over a few hours. For longer periods (i.e. winter) you want power-to-gas and gas plants.

Also, 21 billion is actually not that much. Coal power costs about $2k per kW (or more[1]), so having enough coal plants to satisfy that demand would cost about 18 billion.

[1] https://www.synapse-energy.com/sites/default/files/SynapsePa...


Your total energy demand figure for SA is incorrect.

For 29/7/2019 (midnight to midnight), total energy demand was 34,700 MWh.

(Go to https://www.aemo.com.au/Electricity/National-Electricity-Mar... and download the SA data for Current Month. Divide the TOTALDEMAND in each half hour settlement interval by 2 to get MWh, then add them up for 48 intervals for the day in question.)


> You literally can't store enough energy in batteries to power the network

You need that only when renewable production goes to zero, that happens pretty much never. Your PV production goes to 0 at night but demand is lower at this time.

So yes, it is completely feasible to do battery load averaging if you do the math correctly


One can't run an industrial society on an agrarian power source. If we want to go on a holiday for winter, like the good old days, brilliant. But be honest. And be honest what will happen when China realizes that you can't power the grid and all those military bases only work at 50% efficiency in winter.


China has the biggest hydroelectric dam in the world, I think they will be just fine (though I'm not sure it does pumped hydro - they probably don't bother).


I think the point was that China would take advantage of the energy weakness of other countries during the winter somehow.


Ironically that is /really/ obsolete thinking. Pax atomica says hi for one.


Nitroglycerin will make war too horrible to fight.

Wonder where I've heard that before.


I sincerely hope that Tesla can increase cell production to keep up with demand for all their products.

When you look at product lines like this it makes you realise the scale and breadth of Elon Musk's influence and even if only half of his endeavours prove to be as successful as they seem like they will be he will be written about extensively in history books.


It's an interesting study in economics (from my layman's perspective), because on the one hand they're multiplying production of batteries several times (and their competitors are also trying to keep up), while at the same time massively increasing the demand for them with both a system like this, Powerwall and the Model 3 for consumers (they plan to produce hundreds of thousands a year). It's mad to think about how they scaled that up in... pretty much the past five years (Gigafactory 1 started to produce powerwalls / packs in 2016 and battery cells themselves in 2017).


They've acquired a company called Maxwell Technologies which has some new dry electrode battery technology that should allow them to expand cell production with dramatically less Capex per unit of capacity


I hadn't heard about that acquisition. I hope it they don't stop Maxwell's offerings in supercapacitors.


They haven’t commented publicly but I’m guessing they’ll play a key role in regenerative braking for semi and pickup and also help higher powered super charging


On the last Tesla earnings call, they talked about an investor day early next year [1] where they'll talk in detail about their plans for terawatt-hour per year level production, directly as a result of their Maxwell acquisition.

[1] https://electrek.co/2019/07/25/tesla-plans-massive-productio...


Regarding cell production, at the Q2 earnings meeting Tesla talked about increasing it to 1-2 terawatt hours per year, about a hundred times the current rate. No, I'm not kidding.

https://cleantechnica.com/2019/07/24/liveblog-tesla-q2-2019-...


They also used to talk a big game about how much power generation capacity they were going to install at SolarCity. Now that business unit is a shell of its former self.


Musk said (about an hour before you posted your comment) that SolarCity is hoping to manufacture 1000 solar roofs per week by the end of this year. [1]

Most of Musk's company goals are aspirational, and the means to the end often change. For example, Falcon 9 first stage was supposed to be reusable by landing it in the ocean with a parachute, but they found it didn't work. Instead, they lengthened the first stage booster dramatically so it would have the propellant margin to land propulsively.

[1] https://twitter.com/elonmusk/status/1156005185656782848?s=20


"Musk said" ....

Dan Telvock, a local reporter in Buffalo who covers the solar gigafactory which taxpayers bought for Tesla, is already calling out Musk on his "aspirational" BS. https://mobile.twitter.com/DanTelvock/status/115621452719479...


Uh huh. A former employee claims that the goal as of February was 200/wk by end of year, but a tweet by the CEO in July states publicly the goal is 1000/wk by end of year. I have no reason to trust a former employee over the CEO.

Regardless of which one is correct, the fact remains that they are scaling up a new technology. Both 200 and 1000 are ambitious goals, and both contradict your implication that the solar department is "a shell of its former self". It has dramatically shifted directions (from market standard solar panels on roof, to integrated solar roof tiles).


It's just one among many tweets from Telvock questioning the veracity of Musk's tweet. A rampup of the kind Musk is promising doesn't happen in a vacuum -- there should be new hiring happening at the Buffalo Gigafactory already (they laid off 50 employees in January, has that headcount even been restored?), raw materials procurement should be underway, and there should be a healthy roster of orders for the $50,000 roofs. Where are they?

Of course normal companies, with a level of corporate governance ranking somewhere above "drunken frat house," publish these sorts of production forecasts in official channels after careful vetting and consideration. It is neither admirable nor acceptable for a public company CEO like Musk to simply pop off on Twitter with "ambitious goals" on which investors might rely for trading decisions if they have no basis in fact. Which is why questions are already being asked about whether Musk's tweet violated the terms of his amended SEC settlement agreement, which requires legal pre-approval for any public statements he makes about matters relevant to Tesla, including production forecasts. https://www.bloomberg.com/news/articles/2019-07-30/musk-twee...


AndrewBissell, I have a question for you.

First let me say Musk has a habit of doing big, new things that the experts and the fossil-fuels-forever gang said were absolutely impossible.

That being the case, what are your views on global climate change? Do you think it is a huge danger and it is urgent we get off fossil fuels as soon as possible? Or do you think it is the greatest hoax in the history of the human race? Or what?


The question about views on climate change is a common red herring tossed into these discussions, but let me answer it this way: If Tesla turns out to be the second and much bigger coming of Solyndra, as I think it will, it cause a huge wave of public skepticism toward investment in clean energy technology. As far as personal premium/luxury BEVs themselves are concerned, they're a piss-poor, horrendously expensive way to reduce carbon emissions, mostly sold to people of means who want to believe we can consume our way out of the carbon crisis with flashy greenwashing gestures like buying a new car every couple of years and slapping an HOV lane sticker on it. There's a reason Musk reacts with such vitriol every time someone suggests that mass transit infrastructure (or any one of a number of other public initiatives) would be a better use of the money that has gone to EV subsidies.

With the essential help of enormous government subsidies, Musk succeeded in creating a cult brand around EVs and little else. It's an achievement of celebrity and branding, not engineering or manufacturing. Which is not to take away from it necessarily, he has proved the EV concept, but he has yet to show that EVs can be financially sustainable over the long run without constant public money injections.

Meanwhile, his application of Silicon Valley management techniques to industrial manufacturing is a total disaster of scrapped alien dreadnought plans and cars being taped and zip-tied together in a tent. The machine-learning-on-wheels he calls "Autopilot" is dangerous and undercooked, and could be easily rolled out by any of Tesla's competitors if not for the fact that they have at least some scruples about beta testing their software on public roads. His impact on home solar generation has probably been net negative as Solar City was a debacle which had to be bailed out by Tesla shareholders and has been in a tailspin ever since, despite massive subsidies forked over to build the Buffalo Gigafactory by the state of New York.

If the whole point is to build a green energy future, why has Musk used Tesla as a piggy-bank to prop up his celebrity billionaire lifestyle by borrowing enormous sums against the value of his Tesla stock? It has created perverse incentives harmful to the long run health of Tesla itself. A restructuring to shed debts and shore up its balance sheet would be the best possible move they could make if the goal is to remain a viable entity replacing ICE vehicles with EVs into the indefinite future. But Musk will never do this voluntarily, because his goal is to keep the stock price propped up.


The reason I brought up the matter of global climate change is that once it was discovered it was a danger, it was also clear that the human race needed to get off ICE's and on to EV's, and do so as quickly as possible.

The problem was the EV's at the time were not remotely up to the job. Where Musk comes in is that, thanks to his brilliant efforts, we are getting there many years earlier than we would have otherwise.

The people who nonetheless are screamingly angry at Musk and Tesla seem to fall mostly into three classes. One is global climate change deniers who think we should stay on fossil fuels forever. The second is shorters who don't care if the world goes to hell, and only want to make money. The third is haters, by which I mean people who are not happy unless they have someone to spend their life hating.

My impression from your comment is that you are in the third class.


I'm tired of a sociopathic CEO who treats his employees and customers like shit, puts the driving public at risk with untested software, flogs unsustainable greenwashing scams, and gets away with blatantly defrauding the investing & taxpaying public time after time (fake solar roof tiles to push through the SCTY buyout, "$420 funding secured," 1 million self-driving robotaxis by 2020, etc) because prosecuting him would cause the price of a particular stock ticker to go down.


I want to live in your world, where benevolent megacorps don’t cheat on emissions tests, or self-validate relaxed stability planes that require additional features to avoid crashing, or turn a blind eye to overseas suicide factories.

And where the only bad person is Elon Musk.


Ah yes, the old "if you don't call out every single corporate bad actor then you also cannot call out Tesla" argument. Maybe your comment would make more sense if there were HN threads full of commenters repeating talking points from Boeing or VW execs.


No, I am not engaging in whataboutism, it is just that alleging that Elon Musk is the single worst person in the world is ludicrous. Please reread this sentence of yours:

>could be easily rolled out by any of Tesla's competitors if not for the fact that they have at least some scruples about beta testing their software on public roads.


At least they realize that they have to plan big if they want to decrease CO2 emission of the world significantly.

Their plan can fail, but I don't see any better plan.


I hope that other companies will start producing similar solutions instead. We can't rely on just one company to provide essential infrastructure worldwide. Those cells will need replacing in a few years and if there's still monopoly we may have series issues. It makes for a great history book material, but I like my utilities predictable and boring.


Here's one: https://spectrum.ieee.org/energywise/energy/the-smarter-grid...

It's based on a novel battery. One of the interesting claims about XNRGI's battery is that it can be manufactured at scale for a much lower capital cost because you can use existing semiconductor wafer manufacturing infrastructure:

https://xnrgi.com/technology/

So let's see if Cross Border Power's grid battery works. It'll be good if it does.


Someone on HN had a wonderful writeup previously about ALL the companies Musk is building are directly applicable to colonizing mars...

He is piloting and testing them all as colonization building blocks here first.

However - I wish that he would start by building a moon base first...


Waitbutwhy has some interesting posts about what they see as the role of each company. The Neural link one summarises them at one point (though is an enormous article, and a great one).

https://waitbutwhy.com/2017/04/neuralink.html

As an aside, they have some really wonderful long form content. I love the 1 to Graham's number posts (a two parter) https://waitbutwhy.com/2014/11/from-1-to-1000000.html


That's good news for earth. You often need a closed loop product lifecycle in space, since you can't throw away and source new parts and materials without great cost. Any spinoffs should help us close open loop systems down here on earth.


I believe the standard working hypothesis is that Elon Musk is a lonely alien trying to get home, and started a car company because it seemed nicely inconspicuous.


s/started/joined/


Okay, but no quibble re. the lonely alien part?


I'm agnostic on that :)


As much as I love what Tesla is doing there, I'd like to point out that they are not the only or even the biggest player ine the field:

https://en.wikipedia.org/wiki/Battery_storage_power_station#...

There are other tech than lithium ion (namely, Sodium-sulphur) that seem more appropriate for applications where weight is not an issue.


They're not claiming to be the only or the biggest player in the field. They're claiming to be cheaper and faster to build than the incumbents.

... to achieve significant cost and time savings compared to other battery systems and traditional fossil fuel power plants.


The downside on those ones seems to be that they need to be heated (and kept at a heat) of 300-350 degrees and they are quite corrosive if there is an emergency. Not saying li-ion batteries are safe, but they and how Tesla has modularized them give me a lot more confidence to leave in an unattended location. Of course, it's all situational so there's a place and purpose for these things as well.


If weight/volume is not an issue, there are lots and lots of technologies that will work well for energy storage. Compressed air, flywheels, heated water, lead-acid.


Are those (e.g. flywheels, compressed air) production-ready or just start-up ideas? Also flywheels have a pretty destructive worst-case failure mode...

> heated water

How does that work (specifically for storing electricity)?


Re energy storage generally, heating is a top energy user in many parts of the world and I think relevant to the conversation. In places with CHP plants and district heating systems, the thermal battery concept is tried and tested. Since the system is already based on circulating hot water around, putting some storage tanks/caverns in the system is straightforward.


> flywheels have a pretty destructive worst-case failure mode...

Do they? If their casing is solid enough, the worst that they can do is damage themselves internally and heat up a bit. There is no buildup of pressure, leakage of chemicals or risk of a chemical fire.


Many data centers use flywheels


Not if the flywheel is in space!


In utility scale the only usable thing now is hydro pump station like https://en.wikipedia.org/wiki/Kruonis_Pumped_Storage_Plant. Requires luck with geography: natural elevation and water not too far away.


Indeed. For this kind of discussion, it is good to have this table in head:

https://en.wikipedia.org/wiki/Energy_density#Table_of_energy...


There was once a guy called Johan Coetzer who stumbled upon more funding than he could spend and ended up with a project called Zebra. [1]

Their biggest success, I am told, was however single use batteries and the (unverified) rumour that I was told is that they built a matchbox sized singe use battery that in theory could power a car for 400 kms.

[1] https://www.batteriesinternational.com/2016/09/22/battery-pi...


>the (unverified) rumour that I was told is that they built a matchbox sized singe use battery that in theory could power a car for 400 kms.

Unless it was run off of antimatter this is just flat out made up. Even if you start with the assumption that that matchbox sized battery was 100% solid aluminum that was oxidized to aluminum oxide to release energy and 100% of that energy was usable power that still only gives you 2.2MJ of energy. To put this into perspective a 15 gallon tank of gas contains ~1940MJ worth of energy and while an internal combustion engine is relatively inefficient, it's not like it's going to use almost a thousand times as much energy to go the same distance.


Why aluminum?



Does anyone else have a turnkey system like this? From here it looks like they are the only player in the field, at least for a little while. It looks like a product that can be directly sold to federal, state and municipal organizations around the world. Perhaps even to for-profit investment funds. Or high energy consumption industries like aluminum switching to solar or wind production. A product on the market that you can buy, rather than a technology you need to build something with.


> they are not the only or even the biggest player ine the field

Well, from your link, they indeed _are_ the biggest player by power ?


More power only means they run out of energy quicker :)

Energy is the more informative metric here. You can easily construct a low-energy storage solution that generates huge amounts of power for a very short time.

The power column is only a result of the number of households and outage duration that the plant was designed for.


More power also means higher bids and more frequency control which is very useful too for 1st layer ancillary services.

Energy is good, but power is also very important. You can stabilize the frequency (with active power) and the voltage (with reactive power). Both are traded and needed.


One thing puzzles me. LI batteries are appropriate for cars because of their high energy density. But for grid applications, density is not much of an issue. They can have large, heavy batteries. So why not use a cheaper/joule battery technology for grid applications?


That might happen long term but short term the lack of production capacity is the problem. And right now a lot of companies are investing in vastly expanding production capacity for lithium ion batteries. We've simply never needed batteries at this scale before EVs started happening.

Tesla seems early in the grid storage market. Probably some companies are trying to come up with cheaper ways to do that. So, long term, I think cost will go down by quite a bit. Ten years from now, there will probably be a few companies cranking out cheap batteries for grid and domestic use that are optimized for $/kwh instead of kg/kwh.


> We've simply never needed batteries at this scale before EVs started happening.

Is that really true?

There a a LOT of lead acid batteries out there, considering there's been at least one sold alongside nearly every internal combustion engine larger than a handheld tool.


I can think of a few reasons (as a layperson). They're heavily investing in li-ion battery production, really optimizing for that one in particular.

But I think they can also produce the battery packs in huge amounts and sell them in multiple products - Powerwall, Powerpack and now this one.

And I suspect these ones are more easily manageable.


Because both demand and supply aren't there yet. Grid storage isn't needed yet and when it is lithium ion is cheaper because its cross subsidized.

This leaves the grid battery manufacturers with no market so they don't expand production.


This talk by Dr Sadoway is pretty amazing regarding his team's design for utility-scale liquid metal batteries: https://www.youtube.com/watch?v=NiRrvxjrJ1U It's an hour long, but it's well worth the time investment.

This is not merely a theoretical technology. A company has been created around this technology: http://www.ambri.com

(I am not connected to any of this, but I am quite excited by it.)


In addition to what others have commented on: You can think of it as a way to boot-strap the market. That's easier with existing established technology. Once they're delivering massive quantities of grid storage batteries, it might be relatively easy to justify doing a big investment in a different chemistry and hopefully you can just swap it in for the old one (with some changes to charging electronics probably)


No one is building those batteries at scale yet. Battery chemistry -> efficient production line -> market is a VERY long game (decades).


Power curves are one reason lead isn't more used (they spike a fair bit). The other one to look at is how the tech itself scales, almost all batteries all catch fire easily due to the chemical processes. That becomes a bigger problem the more you pack them in and there has been more research in getting Lithium cells to not catch fire than lead acid lately.


Apart from economies of scale on lithium batteries - Perhaps the extra weight makes construction more complicated even though the heavier batteries are cheaper. Like, unless they're being built on site, you've still got to get them there, etc.


Economy of scale which should help bring costs down. It's easier to take them to Mars.


This is actually huge. Tesla is again thinking and executing way ahead of anyone else. I don't attribute it all to Elon Musk. But there is some life force in Tesla that's pushing it along even when naysayers, sometimes criminally, try to destroy Tesla.


There are other players in the modular utility scale energy storage space. GE and Fluence are two examples.

https://www.ge.com/renewableenergy/hybrid/battery-energy-sto...

https://fluenceenergy.com/


Yes, but Tesla is actually doing equally good or better than the incumbents. And this is on top of a car company. It is hard to overstate what Tesla has achieved. There is a reason why Tesla has crazy fans.


What do you think is a mistake Tesla is currently making?

I may agree with you in part, but overall your tone is just so overwhelmingly positive it makes me curious where you have dissonance?


Not sure if my comment seemed otherwise, but I am overwhelmingly pro Tesla.


"naysayers, sometimes criminally, try to destroy Tesla"

Who & what did you have in mind, exactly? The only example which comes to mind is the skabooshka lawsuit, which Tesla had to drop when their megacorp legal resources finally ran up against an opponent with some funding, thanks to crowdsourced donations.

Given that essentially every major executive (other than Elon Musk) and huge chunks of the Tesla board have all left the company and furiously dumped their stock, I'd say there's plenty of naysaying coming from within as well. Not exactly the signs of a vital corporate "life force" at work.


What Tesla did -- use the contract from Australia as a MVP for their generic utility scale product offering. Took their lessons learned from the Australian offering and streamlined it so they can make an "easy" product for deployment reducing custom costs. A good marketing piece for any utility considering installing a product of similar specs. Also distracts from some of the other narratives taking place.


> A good marketing piece for any utility

This seems like a win-win-win to me. Tesla gets paid, the utility gets to show off how cool they are (and how they are investing in going green) without breaking the bank, and the rest of us (hopefully) experience less carbon emissions.

Ideally, that utility runs the battery for a while and it proves to be useful/profitable enough that they decided to change some of their future plans to include more renewable/battery tech.


I don’t always agree with the way he’s conducted himself and I don’t understand why people overshare on Twitter when the upside is minuscule and downside is huge. But Elon does seem to be pushing make the world a better place. I’ve a feeling he will be viewed more favorably by future generations. I work in energy and I am super excited to see alternatives to fossil fuels come online.


The impressive thing is that this is competitive even before we have a carbon market in place which would make all gas peaker plants too expensive if they had to pay a market rate for dumping/storing their carbon.


I wonder why Tesla is selling these systems instead of deploying them directly and profiting off the arbitrage? If Autobidder was as good as they imply, then realistically Tesla should be in position to capitalize off global energy markets on their own, earning far more than just the sale of the system.


To do that would require them to _become_ a power company in each jurisdiction that they wanted to do this in. Those markets are generally highly regulated - the amount of distraction & overhead this would create would be staggering, _immense_.

I think Tesla already has enough on it's plate.


This is my answer. Selling into unrelated markets creates a need for duplication. At lower ranks that's 'fine' but becomes problematic at the level of senior management. Companies often sell off profitable divisions because management either can't or doesn't want to give them enough attention.


Actually they will do this by selling to consumers directly as well with the same technology. They are getting ready to sell solar + battery directly to home owners at a huge scale (bottle-necked on some production capacity issues currently). Once they get there, they will have a virtual power plant that they can operate and control. So, Tesla is going there but just not in the way people expect. In the same way vehicle to the grid technology is a substantial battery resource that people overlook. Not a stretch of the imagination that a battery + solar solution for your home would play really nice with the Tesla parked in front of it. Vehicle to the grid technology already exists. There's just no way that they are not planning for that.


That seems fairly obvious to me - SolarCity/Tesla have always talked about building a "virtual" utility via residential solar+battery, with the cars working as an (occasionally) mobile battery component.

But the OP's question was why don't Tesla install MegaPaks that they own themselves, into diverse energy markets around the world - and become a traditional utility.


Because there isn't much money to be made. Frequency control used to be provided inherently in every generator. With renewables that isn't the case and no consumer wants to see a new line item on their bill for all the extra cost of accommodating energy markets and renewables so it is all hidden in the network connection fee which is controlled by regulators. So the return on providing grid services that aren't MWH (energy) is Essentially fixed to what regulators will allow, which isn't that attractive, so better for tesla to make the lump sum profit up front on the equipment sale and let the utility run it for 5 years until the batteries are worn out and the utility have collected their ROI.

In terms of actually bidding MWH in to the market when supply is tight so the price is high, it will end up lowering the price by increasing supply, and the intervals of high prices are few. By virtue of existing the system will reduce frequency of high price intervals since the market participants will know that capacity will come on at some price so if they try to bid really high their is will not be accepted. So batteries at this scale is nice for the system operator and politicians to put a lid on high price intervals but won't actually sell enough MWH back to the grid to amount to much.


You are correct.

Due to certain tax credits, utilities can break even and even make a little money off of these devices, but it is certainly no slam dunk.

One of the most recent FERC orders deals with directing all the ISO/RTOs to add energy storage participation to markets as long as the battery is 0.1 MW or greater if I recall correctly.

A lot of folks are interested, but there isn't much money in ancillary services or energy for the reasons you described above.

The large-scale storage is pretty sweet though.


Why does Boeing sell planes instead of operating them directly? Why does John Deere sell tractors instead of operating farms? Companies focus on what they’re good at.


Interestingly (and admittedly unrelated to your point), Boeing used to both build and operate planes, as well as build components for planes. The US government passed a law to break them up in 1934 - the plane manufacturing became Boeing, the air operations became United Airlines, and the eastern components manufacturing became United Aircraft Corporation, later United Technologies.


That is interesting! I guess that’s not such a great example.


Arbitrage isn't as profitable as the other revenue streams in energy storage and, quite frankly, isn't really profitable or attractive. Arbitrage doesn't add any true value to the world - it just lines your own pocket. Not to mention that developing software to monetize the different revenue stacks is demanding and fickle. So many different integrations and different market requirements and dimensions. There's a good reason they aren't doing it - too much work with minimal upside and that whole Tesla/Elon has way too much on their plate thing.


Arbitrage does provide value but indirectly and abstractly usually by boosting availability and allowing for some efficiency boosts. The power role is one of the times when it had the most direct utility.


Because they need cash, probably


You can always find debt for profitability in the energy markets, and this seems pretty low risk as energy investments go.


But you can only take on so much debt as a company, regardless of market, and telsa is kinda already stretching that limit (or at least they are according to HN comments)


They are. Tesla is debt leveraged to hell which is why they're saying they're going to raise another $2B in cash from equity and bond sales to pay some of it off.

Tesla needs cash, and lots of it.


Debt hurts your fundamentals and hurts your stock price. Cash helps your stock price.

If you're concerned about stock price, sell for cash.


It depend on how you use the debt.

If are able to make slightly less ROI (due to cost of capital, cost of servicing debt) at larger scale, it's worth having debt and will have positive effect on stock price.

The relationship is not as simple as you claim.


True, but we're not talking about an anonymous company in a vaccuum here, but Tesla, which already has plenty of debt.


No. They are free cash flow positive, and by $600+ million per quarter already.



"on a three-acre footprint"

Is there a reason these installations are always shown laid out very flat? Is there a disadvantage to stacking them up or at least putting them on different floors of an appropriately-reinforced building?


Probably just the simple economics of cost of land vs cost of construction of a new building + ongoing maintenance.

I'm sure if they ever saw need to place this in a city they would construct a building. But for now, it's probably cheaper to just build out where land is cheap (and possibly closer to generation source) and then transport the energy to where it's being used.


How safe would it be to put large quantities of a potentially unstable alkali metal in close proximity to lots of humans?


I know your comment is meant as a potshot, but it actually points out an interesting side-point: safer than putting in a natural gas plant.

In California, the Puente gas plant made headlines last year because they found it was cheaper to create battery storage than an equivalently-sized gas peaker plant. A large part was permitting and land rights. People don't want an emissions-spewing gas plant in their backyard, but an array of batteries in what looks like every other industrial warehouse is much more palatable. (Time-to-go-live was another factor: battery arrays are modular and can be distributed across multiple sites to further minimize risk, a gas plant is a massive multi-year centralized organizational challenge with all the associated costs and all-or-nothing risks)

https://www.greentechmedia.com/articles/read/sce-picks-major...


Another side of the organizational improvements that batteries have: redeployable.

Underprovision? Pour some more concrete, truck a couple more batteries in.

Overprovision? Load em up again and resell them to someone who wants them. Can't quite do that with a gas plant.

Sure there will be some overhead lost, but it just make soooo much sense to go with battiers from a risk perspective.


Definitely, I saw that these units are sized to standard container dimensions as well so they could even load them onto a truck for temporary load balancing or supplying power.

They also look (but I'm a layman) easily maintainable, more like server racks than high tension electricity stations. I wonder if the batteries would be hot swappable.


I mean, we all sleep with phones under our pillow, so


I'll have you know that my phone is at least eight inches away from my head at night.


Haha noooooo way. Even if the EM doesn’t cause any health issues, battery combustion due to overheating right under my head might.


Quite a bit of alkali metal ions exist in your body, your kitchen, and your food.


Probably pretty safe with an appropriate fire suppression system in the building.


I dunno, Teslas going up in flames sometimes cause pretty dangerous fires - iirc in one instance they put a car wreck in a container full of water so the lithium could burn out overnight.

But, if something starts to go wrong it should have monitoring systems and an engineer can swap out the faulty module, and if it does go wrong I hope they have good containment.

But, not much you can do against a proper lithium battery / electrical fire I suspect.


Replace all air in the facility with Halon. That'll put out most fires.


If it's stacked up in a building maybe they can use the energy to lift-off and self destruct elsewhere ;)


Besides land being cheap outside of cities, it might also be a safety thing. Like, you might not want batteries close enough that if one catches fire it could catch the rest on fire too. You might also want easy access for fire crews.


Yes, it's more expensive. They're not normally situated where land is expensive.

You can ask the same question about car parks. Same answer. Building up or down is expensive.


Space isn’t an issue in the places you need to locate these things, with a few notable exceptions (eg Japan).

Building up is a lot more expensive, and there is no shortage of available land on Earth at the moment.


Well I would assume they weigh a ton and moving them and/or replacing them will be less troublesome if you can just bring a crane or something similar


Amusingly, heavy things carried up and down by cranes are a promising way to store electricity.


I can't really agree with 'promising'. A tower that can lift 100 tons 50 meters is not exactly cheap, and only stores 13.6 kilowatt hours.


Here’s a company pursuing that very idea.

https://qz.com/1355672/stacking-concrete-blocks-is-a-surpris...


Ah, if you actually build the structure out of the items you're lifting, then that certainly changes the math. Compared to just lifting with a crane you get a thousand times the capacity at the same infrastructure price.


Exactly. It’s really the same principle as pumping water uphill except using concrete blocks and a crane. Potential energy is potential energy.

I’d just love to see it in action.


consider off shore gravity batteries that can drop mass down 2 to 4km. Major fixed cost is Hvdc out 100miles, but it’s extremely scalable after that.


Lead-acid batteries are pretty damned heavy and would make a great kinetic battery as well as a chemical one.


I wonder how much more energy this setup would store if they added cranes. Also it could be possible to make a circular vacuum maglev train out of these and add energy to it by accelerating.


There was an episode of Nova about energy storage that included a company that made kinetic energy storage devices. IIRC they had 5000 pound flywheels that spin in a vacuum and can be installed in your home. Expensive to purchase and install but essentially unlimited lifespan and totally renewable.


I've heard about either a working system, or a calculation of something like that but with regular trains. Long story short, it can't actually store that much energy.

More efficient (and less mechanically complicated) would be pumping water up a hill into a lake.


This is fantastic, incremental progress like this adds up quickly.


I'd imagine that this minimum size of 3MWh is ideal for fast charging locations, combined with solar panels on the roof. A by-product of SCv3 development? If it costs around $200/KWh or less, it's very attractive for a variety of commercial applications with PV capability and non-uniform electricity needs.


I would not be surprised if Megapacks are what will power the Megachargers for the forthcoming Tesla Semi. A simple-to-deploy solution for carriers that doesn’t require significant site prep. Inverter power output of 1.5MW would align very well with the stated charge rate and buffering 3MW of energy is probably sufficient for 2 full charges without requiring a massive and rapid draw on the grid.


Fast charge is DC, they wouldn't use an inverter. Theoretically if the semi can handle it and the voltage were the same they could connect the semi pack directly to the megapack battery to charge. It would only get you to 80% charge but it would be simple to implement.


The inverter could be used to recharge the Megapack from power mains. Though if it is connected to a solar array the inverter might be skipped. I'm sure they can exclude that component for such installations.


For sure; I foresee that Tesla will be able to build off-the-grid charging stations with one of these and a roof of solar panels. I mean assuming cars with a 100 kWh battery, you could do up to 30 charge cycles (not counting losses of course).


I'd be interested in an examination of how the purchasers and vendors of these battery storage installations manage the balance of capacity and life expectancy vs wasted cost. If you build an installation that exactly matches your needs, you're putting a full charge cycle on the batteries on a regular basis, which reduces their ultimate lifespan and limits your ability to cope with unusual power demands.

On the other hand, if you overbuild for your needs, you increase the total lifespan of the installation and your emergency capacity but at the cost of greater initial investment as well as the inescapable fact that even the unused portion of the capacity has a finite lifespan which is always (very slowly) expiring regardless of whether it's being used, so your per-dollar lifespan could be lower with a larger capacity.

I'm also interested in how well real-world installations are matching the expected figures when it comes to amortized cost and life expectancy. Does the operating budget for something like the Hornsdale Power Reserve include the cost of replacing the entire thing when its cells are worn out in 10-15 years?


It likely wouldn't be a big bang replace job, just a few cells / year. That won't do much for the total cost of ownership, but would avoid bursts. Also, I don't know what that would cost, nor how much the alternative solutions currently in place cost. I'm sure Tesla will spin it so that their TCO looks a lot more attractive than their competitors.

Also, who knows where battery technology might be in ten years; given that these are modular batteries, they could replace them over time with higher capacity ones.


This is good news! Are there competitors in this space or is Tesla going to absorb marketshare for the next few years? It seems like it could be a very big market.

I'd be happy if we end up seeing more inexpensive battery production.


I assume they are just using a bunch of 18650 cells?


probably 2170, but yeah, small cylindrical batteries


Tesla has been quietly installing 'powerpacks' in New Zealands power grid for the past two years.


For comparison, 3 MWh allows a MacBook Pro to run for about 15 years continuously.


Anyone have any estimate on the cost of these Megapacks?


The best number I've heard for the Australian battery is $500/kWh after full installation. Individual battery cells are less, but inverters, packaging, and heat control add up.

I'd guess that for a big install they could probably get closer to $350/kWh these days, but that's a WAG.


So one of these units would cost 1-1.5 million (if my math is correct) - that doesn't actually sound too expensive for utilities like this, especially given how operating costs would be lower than the gas peaker wotsits that were mentioned earlier.


Awesome. I hope they get a ton of orders. These will probably use the dry battery electrode tech they got from Maxwell?


They say it scales up to 3MWh but whats the smallest purchasable unit?


The Powerwall.


When they have large purchasing power they have price leverage on cells


When they come to Europe (CH at least), Tesla's gonna eat up the lunch (in the ancillary services field at least) of everyone. Even more so with the new renewable capacity added every day.


I dunno, it seems like the function of these Megapacks is to replace natural gas driven peaker plants? Europe doesn't have nearly as many of those as the US does.

AFAIK Switzerland in particular has exactly zero, thanks to all that lovely hydropower that can be turned on and off to match solar/wind fluctuations.

In fact, Switzerland alone has hydropower equivalent to 10 000 Megapacks. This basically means you can install all the renewables you want.


And with hydroelectricity often comes pumped-storage opportunities.


The coal and nuclear phase outs across Europe will lead to a need for new flexible capacity. That should at least partly be covered by battery peakers.


Nuclear isn't flexible capacity; plants are always run at max output. If it is to be phased out the replacement needs to have more long-term output than a battery, unless the renewables become incredibly overprovisioned.


> Nuclear isn't flexible capacity; plants are always run at max output.

In France and Germany, nuclear reactors are configured and permitted to run in load following mode. It's harder on the valves and similar equipment, but it can and is done. [1]

> unless the renewables become incredibly overprovisioned.

Which is the likely outcome, as solar and wind are still declining in cost and in some places the cost is as low as 2 cents/kwh. The generation will be cheap (what's the old saying? "too cheap to meter"?), and it'll be the storage that'll cost a bit more.

[1] https://www.oecd-nea.org/nea-news/2011/29-2/nea-news-29-2-lo...


There's an ambition to integrate the huge European market with HVDC lines, which should allow excess capacity to be shared more efficiently.


Always run at max output? This doesn't seem right to me, they have moderators for a reason don't they?


US companies are not permitted to eat the lunch of European state-owned companies in Europe. If it's not illegal now, it will be before it's profitable.


Why would this be illegal?


I believe that implied protectionism perhaps through vague regulations as the threatened would come up with some pretense for it.


The particular laws used vary from business to business.


Lithium batteries seem like a weird choice for large-scale power storage that just sits in one place. Lead is a lot cheaper, and its heaviness doesn’t matter in that application.


Take a look at ggm's comment -- the machine wasn't designed for power storage, it was designed for short bursts of frequency regulation.

When designing for grid-scale capacity, there are companies looking into flow batteries and chemistries that use cheaper (non-lithium) materials.

Tesla however built out tons of lithium manufacturing capacity for their cars, and given the design constraints of this project, their existing products were a good enough fit for the job at hand.


Indeed. And it's worth remembering that one of those design constraints for the South Australia project was Musk's "Done in 100 days or it's free" offer. While surely done for publicity, it was in fact included in the contractual agreement with the state government.


Lead batteries have pretty horrible characteristics compared to Lithium, especially around depth of discharge. They're good to start a car (lots of amps, fast), but not very good to run a computer (low amps, for a long time). Discharge it below 50%, and you'll significantly shorten their lifespan. Below 25%, better try that only once or twice. Below 10%, might as throw them out.


Lead batteries have a shorter lifespan and require more maintenance. The additional weight and volume also increases shipping costs.

Oddly enough, you can get a lithium ion car battery jump starter for less than the cost of a car battery.


Gotta look at the total cost of ownership. The lead-acid battery is fairly recyclable.



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

The amount of effort spent reducing Lithium Ion costs has resulted in dramatic price reductions. Technologies that looked extremely promising just 10 years ago are having a hard time keeping up.


Correct for "new" batteries; however, sometimes "used" Li-Ion batteries are "recycled" into static storage, and they can be cheaper than lead batteries.


Not sure if any insurer will cover used lithium batteries of unknown provenance in a multi-ton cube.


They are almost never from "unknown" provenance. e.g. used Tesla car batteries, you know everything about them already (tech specs, years in operation, and number of charges).


ex-automotive-service might be the worst case provenances.

Would make sense if Tesla kept a life-time score of subjected G-forces and vibration for the pack so they could bin salvaged cells.


>All Megapacks connect to Powerhub, an advanced monitoring and control platform for large-scale utility projects and microgrids, and can also integrate with Autobidder, Tesla’s machine-learning platform for automated energy trading.

Oh. My. God.

pow(What_could_go_wrong, 2);


How do you think the grid currently works? This is nothing new, really.


It's just trading. Maybe the ML makes some mistakes and someone loses some money while someone else gets some free or negatively priced electricity.


Hmm, valid point. I hadn't thought of the risk factors on the downside.

That just leaves the common attack-vector concern for any grid adding this into their ecosystem.


It's just trading and can flash-crash the grid.


I'm glad to hear that Elon Musk solved options trading, the complexities were really bugging me.


4 meter radius bore, go 2km down, then bore a 10 km tunnel. fill it with water to generate power, pump water out to store power losing 20% roundtrip.

(pi * ((4 meters)^2) * (10 km)) * (1 (kg / liter)) * (9.81 (m / (s^2))) * (2 km) = 2.73946879 gigawatt hours


Another thing to consider with the 'dig a hole' based pumped storage approach is that you can only 'suck' water about 15ft vertically (well technically you can't really suck water any appreciable distance, but the atmosphere can push it up about 15ft) so it means that your pumps need to be at the bottom of the hole.

You might be able to get away with some sort of remote actuated pump head like that used in an oil well (energy for pumping provided by a reciprocating metal rod) but the efficiency of such schemes is never going to get anywhere near the 80% you were targeting. Better hope that the pumps don't fail while the hole is full!


Yea, I’ve been thinking about the requirement that the pump is in the bottom is probably the Achilles heel. I’m more keen on dropping heavy blocks of metal now :)


Do you know how hard/expensive it would be to even dig out a hole that wide and deep?


nope, but I know that someone at Tesla knows (who also works in the boring field). Probably costs less than one rated for humans, where they quoted $10M per mile. So 10km ($62M) we're at $22 per kwh. Losing 20% of the energy every time will add up to about $40K per cycle so after like 6000 cycles it will be roughly the same price as $100/kwh lithium.

I'm not sure about the technical issues regarding narrowness and depth, though a quick search indicates the deepest hole was a Russian made a 9 inch diameter hole 40,230 feet deep and had to stop because it was too hot. Anyway, I bet that offshore gravity batteries can also scale up very cheaply once you get over the HVDC cable fixed costs (this almost certainly means building 10GWh level capacity to justify a $100M cable to a place where ocean depths are 2-4km)


Costs to dig go way up 2km underground. This might be less so with your example as you don’t need access tunnels, but it’s still significant.


I guess I'm just trying to introduce the idea that cheaper boring will reduce capital costs for pumped hydrostorage (see https://en.wikipedia.org/wiki/Pumped-storage_hydroelectricit... )

I don't think lithium ion batteries, which are optimized for their energy density (mass per joule), belong in grid storage situations where we should care mostly about lowest cost total capacity (regardless of round-trip losses) and lifetime.


Due to economies of scale resulting in continuous price declines they are the cheapest long term solution.

They also last much longer, can regularly use a wider range of their capacity, and take less maintenance than lead acid while being easier to transport and install. Part of this is better power management, but it’s also because they are useful even after significant reductions in capacity.


No, pumped hydrostorage is enormously cheaper, if you already have the dam. I also think some of your assertions about lead–acid batteries are wrong, but they're plausible.


Unfortunately, only finite amounts of existing dams with huge changes in elevation exist.

Beyond that “The round-trip energy efficiency of PSH varies between 70%–80%.” That’s a major ongoing cost, at say 6c / kWh * .25 * 365 * 15 years = 82$/kWh which is a large fraction of the Lithium Ion batteries costs. Though this increases if it’s used more than once a day, and decreases with unused reserve capacity.


Indeed, I think the finite amounts of existing dams are the reason people are looking to batteries.

Your point about the efficiency is interesting, although I didn’t understand it at first. I think you’re saying that the capital cost of the lithium-ion battery storage is partly defrayed by the higher efficiency of the storage system? Like, for each kilowatt-hour of Li-ion storage (with, let's say, a round-trip efficiency of 95%, although I think that's too high), you “get back”, say, 0.25 kWh every time you use it, that you would have lost if you'd stored that energy in pumped storage instead? So over, say, 15 years, you “get back” US$82 or so, at US$5.48 per year?

https://electrek.co/2018/11/20/tesla-gigafactory-battery-cel... claims that the battery cell cost is US$111 per kWh at the moment (though other manufacturers are still stuck around US$140), so that would work out to about a 5% annual IRR if the cells were the only cost; I think that in fact they are on the order of half the cost (though Tesla's blog post here doesn't actually list prices!) and so that would be a 2.5% or so IRR. Not enough to justify the battery investment on its own, but it would definitely be a significant boost to the project's ROI.

I have a couple of objections to that line of reasoning, one trivial and one serious.

The trivial objection is that the wholesale cost of electrical power, although it varies a lot, averages about half of the 6¢/kWh you're imputing. https://www.zmescience.com/ecology/climate/cheapest-solar-po... talks about the just-signed Atacama project at 2.9¢/kWh, which I think includes the cost of some storage. So the numbers are more like 1.25% IRR rather than the 2.5% I suggested above or the 5% you suggest.

The more serious objection is that, when you're filling up your utility-scale storage during hours of excess power production, you're not paying 6¢/kWh or 2.9¢/kWh. In fact, due to non-dispatchable “baseload” plants like coal and nuclear, it's common right now for the power plant to pay you to take the power, with the price typically around -4¢/kWh, which is the cost of burning it up in giant resistors. When instantly-dispatchable solar plants come to dominate power production, we can expect to see a price floor of 0¢/kWh. Maybe if a storage-plant operator is paying a solar-plant operator to leave their PV plants running, they'll have to pay 0.01¢/kWh or 0.1¢/kWh. But they won't be paying anywhere close to the average price of electrical energy. They'll be paying the marginal price of generating electrical energy when it is cheapest.

So that means that the amount of money you make from a utility-scale energy storage plant isn't going to be determined by how much energy you need to charge it up. Your round-trip energy efficiency could be 10% or 5% and you still wouldn't pay a significant percentage of your revenues to obtain that energy. What determines your revenues is how much energy you can release once you are selling energy rather than buying it. (And the quality of your trading strategy, of course; if you decide to wait to sell your energy until your LMPs go above US$45/MWh, and they sit at US$42/MWh all night long, you don't make any money.)

Round-trip energy efficiency only matters at all in the sense that it diminishes your effective storage capacity — if theoretically you have “1 MWh” stored, but when you turn it on, only 0.9 MWh flows to the grid, you only get paid for that 0.9 MWh, and that's what you need to pay your capex and opex with. But it only matters very marginally whether you had to buy 1.1 MWh or 2 MWh or 5 MWh or 10 MWh to charge up your storage facility.


Wholesale prices are location specific. The northeast has quite a bit of pumped storage, but it’s a poor location for solar power, and not that good for wind either. That wholesale market looks very different from say California or Texas. Further, you need infrastructure to transport power to and from the pumped storage and generation. Several ways or model this, but increased costs are the simplest.

Price inversions are also generally rare and location specific, pumped storage wants to operate every day and can’t wait for unusual events thus increasing their average prices. Further as you increase storage you change the local market reducing price swings. More to the point a company that’s building both generation and storage for say an island needs to build extra production to be guaranteed to fill up that storage. It’s only useful with surplus generation and guaranteed surplus is not free.

As to solar production prices that’s completely independent of wholesale prices. These deals involve low prices specifically because of the intermittent nature of solar power and it’s close correlation with other solar production. The fear for these operators is the wholesale market changing over the next 20 years with ever more and ever cheaper solar not how the current market operates.


It seems strange that you would reply to a comment where I'm talking about strategies for trading LMPs and how storage and highly dispatchable resources will change the energy markets and reduce price swings by informing me that “wholesale prices are location specific” and “you need infrastructure to [transmit] power to and from” and “as you increase storage you change the local market reducing price swings”. It suggests to me that you didn't understand what I was saying. Maybe you don't know what LMPs are?

Nothing I wrote was specific to a single kind of storage resource.

I haven't been involved with any of these solar IPPs or seen the terms of the PPAs, but while I think your characterization of the PPAs is mostly right, I don't think it's accurate to say that they're “completely independent of wholesale prices”. You've addressed the reasons for the IPP to seek a PPA, but those same reasons are disincentives for the counterparty. The buyer is specifically betting on the PPA price being lower than the LMPs over the life of the PPA, or at least not too much higher. That's why some of these PPAs include storage—it reduces the buyer's risk.

Also, BTW, negative LMPs are a different phenomenon from price inversions. Not sure if you're confused about the phenomenon or just the terminology. Negative LMPs are not rare; they happen nearly every night in some markets.

Thank you for exploring these questions with me!


By location specific I mean the geography and generation types heavily impact the market.

Areas with significant elevation drops conducive for pumped storage also tend to have a lot of hydro which is generally very dispatchable. The Great Plains have a lot of wind and are in significant need for energy storage, but they don’t see a lot of elevation drops.

Thus negative LMPs being rare for pumped storage.


Deep drilling is actually a thing for deep geothermal. Still early days but can produce a lot of energy on its own, no solar required. Expensive to go that deep though

http://www.thinkgeoenergy.com/drilling-finlands-deepest-well...


Hang on a sec, you're assuming that the water is smoothly accelerating at 9.81ms^2 for the entire 2km way down.

You forgot to subtract energy losses due to drag which cancel out your exponential advantage.


> Hang on a sec, you're assuming that the water is smoothly accelerating at 9.81ms^2 for the entire 2km way down.

No, that's the raw potential energy calculation.

For power, you use the basic relationship that P=FV (force times velocity). The system would operate in a steady state for any given rate of power input/output, but that potential energy calculation governs how long it would take to charge/discharge.

The acceleration due to gravity only matters if you're trying to discharge all of the energy of a parcel of water on impact at the bottom (via smashing it into something). A potential energy storage system would instead want to charge and discharge close to steady state, with very little acceleration throughout.


This has been tried. So far unsuccessfully, for practical challenges see http://www.ipsnews.net/2013/08/geothermal-energy-stuck-in-a-...


Pretty sure you can't make a bore that narrow and that deep without it collapsing pretty soon.


They make the claim the giant lithium battery grid is “cleaner” than the “dirty” natural gas “peak power” plants.

I’m curious, what’s the environmental damage of this overall? These battery packs likely degrade, lithium has to be mined, etc. Natural gas also has a cost to mine, to burn (though if I recall, the off gases are co2 and water vapor, lower levels of NO2, etc)

Looking from the outside it looks like both options honestly may pollute or have the same environmental impact long-term. One is just a cost we don’t see immediately (or is don’t me in another country)


> Looking from the outside it looks like both options honestly may pollute or have the same environmental impact long-term.

Not sure how you come to that conclusion. Lithium is not a fuel that is consumed in the process. It is a recyclable component. And there is no dangerous imbalance in the global lithium cycle. Which can't be said about the carbon cycle.

Just because every technology entails some environmental impact does not not mean all impacts are the same.


>It is a recyclable component.

In theory, batteries are not currently recycled in any scale equivalent to these proposals. Plastic is also recyclable but most of it ends up in landfills due to it being uneconomical.


We also tend to put plastic into contact with food and small packaging, create many different types and form factors, and ship it consumers. I imagine the recycling problem for a known environment (e.g. a power facility's used batteries) vs an unknown environment (e.g. consumer houses) is significantly different.

I do wonder (and a quick google search didn't turn much up) how the rates of industrial plastic recycling compare with consumer rates.


Time will tell. Right now no industrial-scale deployments have reached EoL, so it's not surprising that there's little recycling. Plus those things also contain cobalt, which is significantly more expensive, plastic doesn't have that incentive.


No cobalt in the new dry battery tech they got from the Maxwell acquisition.

I hope Tesla has a plan for recycling the batteries they manufacture


Tesla is already recycling batteries at Gigafactory 1.

https://www.greencarreports.com/news/1122631_tesla-launches-...


Lead acid batteries, a mature battery technology, have a 99% recycling rate in the USA [1] (and similar rates in other developed economies).

[1] https://www.recyclingtoday.com/article/battery-council-inter...


>batteries are not currently recycled in any scale equivalent to these proposals

They're already recycled at >50% https://www.pv-magazine.com/2019/07/12/lithium-ion-recycling... although there's still no need to recycle them in volume (since we are only starting to deploy Li-Ion batteries for EV and stationary storage, which last for more than a decade)


> [Lithium] is a recyclable component.

But lithium is only 5%* of the value of a battery, so it's the other 95% of the impact of battery production that matters.

* the market for mined lithium raw material may be worth $20 billion, compared with $43 billion for refined products and $424 billion for battery cells - https://www.bloomberg.com/news/articles/2019-07-28/the-lithi...


> And there is no dangerous imbalance in the global lithium cycle

Interesting, I had never thought of a "global lithium cycle" before – can you elaborate?


That was facetious to contrast a rather important thing to something of little significance. Some processes must obviously enrich lithium over geological timescales, otherwise we couldn't mine it, but to my knowledge it's not important to the biosphere as a whole.


Pretty sure they’re not assuming they are the same, but are curious to know how similar they actually are.

In other words, does Tesla saying they’re the better option than gas make it so?


Once you have mined the material to use in the battery it can (and will) be recycled into making new batteries (or other uses). It is essentially mine once.

Natural gas is extracted and then burnt. Because it is consumed it requires continual extraction, i.e. endless mining.

Furthermore, batteries are of course energy storage, while natural gas is generation. If you couple batteries with renewable energy generation (solar, wind, etc) you no longer require the use of an extractive fuel, nor do you incur combustion pollution costs. This has to be where we end up because it is the only sustainable way for us to operate. Better we get there sooner rather than later.


There are tiny amounts of lithium in the batteries and we have tons of lithium on the planet. The real environmental / human rights issue is around cobalt and they're allegedly close to completely removing cobalt from their battery chemistry.


1kWh of li-ion batteries has a carbon footprint of around 110kg CO2[0].

Tesla batteries put in cars are estimated to last 800-1000 cycles to 80% capacity[1].

Given these figures batteries add around 138g/kWh of CO2 on top of whatever source was used to charge them(say wind at 14g/kWh).

Meanwhile natural gas plants emit around 550g/kWh. The battery would have to last less than 250 cycles to match that.

[0] https://www.sciencedirect.com/science/article/pii/S136403211...

[1] https://electrek.co/2018/04/14/tesla-battery-degradation-dat...


Why are you assuming that 80% batteries are discarded?


You are forgetting the methane emissions from leakage: https://theconversation.com/the-us-natural-gas-industry-is-l...


It also depends on how much you value CO2 emission reductions to the non-global-warming environmental damage (e.g. toxic chemical leakage) caused by battery production and decay.


Natural gas is methane. The pumping and transport of methane involves leakage. Methane is 30x worse than CO2 in terms of heat trapping.


...Autobidder, Tesla’s machine-learning platform for automated energy trading. Tesla customers have already used Autobidder to dispatch more than 100 GWh of energy in global electricity markets."

Tesla is in Enron's business?


I used to work for a company in California that does this. I'm not sure how commonplace this is, worldwide, but there are definitely companies trading energy.

https://www.ferc.gov/market-oversight/mkt-electric/californi...

The California Independent System Operator (CAISO) operates a competitive wholesale electricity market and manages the reliability of its transmission grid. CAISO provides open access to the transmission and performs long-term planning. In managing the grid, CAISO centrally dispatches generation and coordinates the movement of wholesale electricity in California and a portion of Nevada. CAISOs markets include energy (day-ahead and real-time), ancillary services, and congestion revenue rights. CAISO also operates an Energy Imbalance Market (EIM), which currently includes CAISO and other balancing authority areas in the western United States.

CAISO was founded in 1998 and became a fully functioning ISO in 2008. The Energy Imbalance Market launched in 2014 with PacifiCorp as the first member or EIM Entity. The EIM serves parts of Arizona, Oregon, Nevada, Washington, California, Utah, Wyoming and Idaho.


Wasnt the whole business model of Enron was to trade energy, and price fix to their advantage?

I recall in january or feb of 2001 - right when Enron was at their peak douchebag-ness... I lived in San Jose ca - and my energy bill jumped to $900!


gah I watched The Smartest Guys In the Room the other day and that was only one of the things they tried to trade.

Interestingly, their broadband trading thing doesn't look as insane a decade or so later. They tried to do a deal with Blockbuster for on-demand video streaming but couldn't get it to work.

A decade or so later and we've got Netflix, which can make the technology work. Now we just need to wait one more decade and someone will figure out the "make money" bit.

As far as losing cash hand over fist goes, the only thing that's really changed is you boast about it instead of hiding it with an elaborate system of frauds these days.


> A decade or so later and we've got Netflix, which can make the technology work. Now we just need to wait one more decade and someone will figure out the "make money" bit.

I hear people say Netflix isn't profitable all the time, and I believed them until one day I decided to look it up. They are very successful judging by their net income.

https://www.statista.com/statistics/273884/netflixs-quarterl...

I hear people say they're losing money so often though that I wonder if I'm misunderstanding something.


Netflix's switch from licensing to producing original content has meant higher capital expenditure. If you were to account for this in a single year as an operational expense rather than deprecating it over a number of years it might look like Netflix was losing money.


I think that argument is advanced by a fair few companies to justify losing money indefinitely.

A tweet (which I now can't find), once said that "the ability to sustain losses indefinitely is the 'moat' of this economic cycle"


Sounds like the Hollywood accounting methods where Star Wars is yet to turn a profit.


I still don't understand. The last time they lost money was over 6 years ago...if they make money every quarter how can you say they're losing money?

Can you give me some hypotheticals with numbers to help me understand?


This is because Netflix account for the spending over several years even though on a cash flow basis spending exceeded income some of those years.


So you're saying they might lose money some quarters...but then wouldn't they make that much more other quarters?

Ultimately their bank account is growing, and it's not like their war chest is so small they can't ride out some oscillations.


No I’m saying that they are spending more per year than they are making but the content they create with that spending has value beyond the year it was made so though they are cash flow negative they are still profitable.


So is the future value of that content reflected in the net income I linked? That explains my misunderstanding. I'd ask how they calculate that number but I guess I should just dig into the earnings reports myself.


Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: