The South Australian grid was particularly degenerate, making this more profitable there than elsewhere.

It does make at least some sense in most places, though.

I remember there being a day earlier this year where the prices for wind-generated electricity in west Texas flipped. If you have to pay people to take your electricity (instead of spinning a generator down), why not store it for a later demand period?

The amount of energy that is stored in batteries like this isn't that big in grid-scale terms. For example, on the night that Texas wind generation flipped electricity prices into the negative back in September 2015, their wind turbines alone were pumping out almost 11,500 MW of power. This has a capacity of 129 MWh. It's useful for grid stabilization because it can start up almost instantly and fill in the gap until another generation source which can run for more than the 90 minutes the battery lasts can start up, but less practical for longer-term storage.

(The 90 minutes capacity at full output is pretty much an inherent property of lithium battery tech, as I understand it. There's a slight capacity vs power tradeoff which can be tweaked, but only within a relatively narrow band, and the maximum charging rate is if anything even less flexible than the discharge.)

Also, it looks like there's already a small-scale battery bank providing frequency stabilization using Samsung lithium ion batteries in West Texas: https://www.energy-storage.news/news/minimal-downtime-younic...

> The 90 minutes capacity at full output is pretty much an inherent property of lithium battery tech, as I understand it.

Even if that is true, that only means you need to expand your battery capacity and split into multiple banks.

It doesn't make sense to store something for later unless the storage costs are lower than the cost of reproducing what you need. A few periods per year of negative prices won't be enough to offset the investment in batteries.

I was under the impression that large scale energy storage is one of the big problems with renewable energy that has yet to be truly solved.

Don’t forget the stochastic control problem that needs to be solved to stop the whole grid blowing up!

Quick! Mine bitcoin for a day!

Superwog ep.5

This is immediately viable anywhere you're operating an inefficient single cycle peaker plant. The longer your battery can discharge for (4 hour, 6 hour, 8 hour, 10 hour), the more likely it's able to replace a peaker duty cycle when called upon.

Note that the amounts are in AUD dollaridoos, it's $48m USD for saving $29m USD.

The savings are because you can use the energy generated by green sources to charge the battery if you can't use it immediately. That's free energy you'd otherwise throw away.

It's not in AUD, it's in USD. From the linked article for cost -

"The French energy company revealed that the entire system cost 56 million euros (~$66 million USD)"

> The savings are because you can use the energy generated by green sources to charge the battery if you can't use it immediately

Not true. The system is used for grid stabilisation rather than large scale power storage. It doesn't have the capacity to power South Australia for more than a few minutes at most.

> Why is this not more wide spread? These figures suggest this is a no brainer. Am I missing something?

Grid infrastructure isn't exactly agile. It's politics.

At least in Switzerland (Swissgrid loco), this is not true at all (for AS and frequency control)

The complete process is documented end to end, free to see, and the enroll process in documented and "cheap" (~150kCHF, 100kCHF being for the warranty). You can start playing in the ancillary services field with 5MW, which is not that much.

PS: If a company in CH needs help in that field (process approval or swissgrid-compatible IT systems), let me know !

The particular grid economics of Southern Australia are not reproduced everywhere else. It is all about the dynamics and regulations of the spot market, existing short term generating capacity, and the capabilities of the other market participants.

The batteries didn't "save" $40 million of electricity, they enabled one entity (Southern Australia's power operator) to trade more efficiently against the other participants, who were the losers in this battery deployment in the form of reduced profits.

It is beginning to happen in the US. Battery adoption on the grid has been slow because the utility companies like solving problems by building new generation plants (natural gas plants mostly) or by building new transmission lines. They have historically always done this and they make a guaranteed rate of return on this type of infrastructure. This has led to lots of un-needed generation capacity and expensive transmission projects passed on to rate payers.

Regulators in many states are now finally forcing utilities to examine the use of alternative solutions to fix grid issues. They are finding that battery installations can do a better job than new generation or transmission in many cases and the batteries can be cheaper over the long run. It's just a matter of getting the utilities to try something new and get out of their typical playbook.

So it paid for itself in 1.65 years, which comes out to an ROI of ~52%.

Why is this not more widespread? There must be some barriers.

There are a limited number of grid topologies where a battery can add this kind of value. It's really unusual to be this extreme.

Correct, at the current price.

I saw elsewhere that battery price is going down 15% per year. So, over time, more grid topologies will be able to leverage this technology.

All grid topologies. Grids will be rendered obsolete absent forced support.

I think part of the savings don't go directly to the battery owner, but instead prevent the electricity suppliers from price gouging the customer.

Happens a lot with renewables. You personally don't capture all the value you provide. This is part of the reason that rooftop solar should get paid more than the going rate but even 1:1 rates get attacked as some kind of scam, even as they reduce peak load which may be literally 1000x more expensive than average.

Correct. A proper grid system would pay a premium for the size of the solar grid that fulfills solar-induced demands.

After that, solar should get paid grid rate.

I have no idea if anyone has bothered to calculate what the solar-induced demand is.

Unfortunately, they don't really reduce peak load anymore. With the solar-adjusted demand curve looking duck-shaped, the peak load is now in the late-afternoon/early evening

I'm in Australia and we recently committed to putting solar panels on the roof of our small office building. Pay-off projection in the estimate/quote was all of 2 years!

Nice! I had my solar panel installed on the roof of my house in 2015, and they have a pay-off projection of just over seven years. At the beginning of the year, my neighbor had solar panels installed, and we calculated his pay-off projection of just over four years (both of us took advantage of available subsidies).

And battery storage is only getting cheap, I believe at a rate of roughly 15-20% per year.

Yet we also hear that battery production is a bottleneck and the reason why automakers other than Tesla are so limited in their EV offerings.

Other automakers are limited in their offerings because they didn't commit with any enthusiasm until a few years ago, and the development time for a new car is at least that long. There are a whole bunch of new EV options coming to market in the next year or so.

It's new and the procurement cycles for these things are long. Also, Australia is coal country.

Judging from the article I think it's a proof of concept or similar stage.

But its not like battery technology is advancing so quickly, and prices dropping so fast, that this is only just becoming profitable. Based on these figures this would have been profitable years ago.

The only plausible reason I can come up with is that the 'savings' are very hard to attribute correctly, but then I still cant explain why no one before today noticed the potential savings and took a chance.

Anyone with experience in the industry have any thoughts?

(a) It takes several years to plan a project and then build it. So, this decision was made years ago, not "today".

(b) Battery prices (and pricing for the balance of the system) have dropped pretty rapidly, I'm not sure why you're claiming otherwise. There have been recent years where these systems have dropped in price by ~25%[1]. That's not insignificant.

[1] https://www.utilitydive.com/news/not-so-fast-battery-prices-...

By "that quickly" I was thinking orders of magnitude per year. This would have been competitive at 4 times the price. Depending on the details it might have been competitive at 10 times the (battery) price.

Was the decision taken years ago? This is the 100 days or free battery. But yes I accept the power companies may have been slow. But still it isn't just Australia that was slow, its the US, Germany, Denmark, the world.

>This would have been competitive at 4 times the price.

Not exactly.

Australia's high prices were not due to a technology failure, but a regulatory one. The power generators were manipulating the market causing extremely high peak prices, and essentially being allowed to get away with it. Most other countries don't allow the market to wildly fluctuate so much, so the payback is far longer. Think decades and not years.

4 * $66m = $264m / $40m = 6.5 ish years payback.

Not many investments offer that rate of return.

You may be right to suggest it may not be competitive in other locations though. I don't know.

He's right. Unless this was a media stunt for Elon's benefit (and it really doesn't look that way), the decision was made in a few days and the turnaround was 100 days: https://twitter.com/elonmusk/status/840032197637685249?s=20

And (B) is an issue. It's essentially deflationary economics. Why spend money now when it will be cheaper/more profitable next year? Once pricing on batteries stabilizes, you'll probably see more people willing to commit to big battery projects.

First off you don't know its going to fall next year. 2nd someone may sign a contract first.

The old sunk cost fallacy strikes again. You invest today because by not investing you forego the savings in the mean time. That something would be cheaper next year is not a factor, only the current expected ROI.

The sunk cost fallacy is most often stated as, it isn't worth getting rid of something because you've already spent the money on it. Its unclear to me if this is an example of that. I agree with your general thrust though.

Yeah, this is more of a reverse sunk-cost. If there is a term for that I don't know it.

(a) didn't take that long

https://www.greentechmedia.com/articles/read/tesla-fulfills-...

I worked on software for trading energy/transfer/balancing capacity in EU market a few years ago.

EU is pretty good in that regard, it demonopolized the energy market, so if you build this battery yourself on your land you can just entered the market and start selling balancing offers. You will easily outcompete traditional producers and consumers on that market if the article is right.

The widespread rule of thumb was - batteries on grid level don't work. Seems this isn't true anymore.

>But its not like battery technology is advancing so quickly, and prices dropping so fast, that this is only just becoming profitable

I wonder if there was even the capability to manufacture such a large order at the given price point before Tesla built their gigafactory.

guess: Its a politics driven industry and no-one wanted to take the chance to try something that hadn't been tried anywhere else in the world at this scale before.

Battery manufacturing is actually advancing fairly quickly. The costs has been dropping at a pretty steady, steep rate for last 15-20 years.

> I assume most of the saving is from grid stabilisation, rather than supplying/storing energy but still.

Yes, the entire reported savings was by avoiding building a fast-response generator, which would cost tens of millions to maintain each year.

From the article:

> Has contributed to the removal of the requirement for a 35 MW local Frequency Control Ancillary Service (FCAS), saving nearly $40 million per year in typical annual costs

The article several includes other significant benefits as well that aren't even included in the $40 million.

How long do these batteries last?

At least a decade. This is why most grid services/frequency response revenue is going to be cannibalized from legacy thermal by utility scale battery storage.

Energy arbitrage is still a thing, but a smaller percentage of revenue compared to grid services (at least in Australia).

Disclaimer: TSLA investor

Long enough for this to be a no brainer.

Also the biggest issue with consumer batteries is typically heat management. Industrial batteries with proper thermal control systems will exhibit much better lifespans than a typical iPhone. Thus you should adjust your expectations for these batteries compared to what you’ve seen in consumer electronics.

That's true and the reason that electric cars will work much better than most expected, with no need for a battery change during the life of the car. Here though part of the reason for the huge savings in one year is the market was in such dire need they've worked the battery hard. The return on investment is still stellar I'm sure, beating even the best initial estimates. But the battery will most likely not last as long as in the original estimates either.

if it lasts even 2 years it was already a good investment

...if disposal is free

Actually, lithium is quite valuable and can be recycled into new batteries. I don't have the numbers, but AFIAK, disposal is cashflow positive.

Edit: see hwillis's comment down below, he/she is obviously more knowledgeable about the topic than I am. I was wrong.

I read on Wikipedia quite the opposite: we don't recycle Lithium it because it costs 5 times more than freshly extracted lithium. https://en.wikipedia.org/wiki/Battery_recycling#Lithium_ion_...

If my math is correct, the battery needs to last at least 10 years to break even if you factor in the cost of recycling the lithium.

Recycled lithium might cost 5x more. Not that the disposal cost of a battery is 5x the purchase price.

If the battery pays for its purchase price in 2 years and the disposal cost is 1/2 the price of a new battery (doubt it's anywhere near that), then the it would still pay for the total cost of ownership in 3 years.

From what I understand, used-up lithium batteries are fairly non-toxic and can be safely dumped in a landfill. Still not zero-cost, but probably about as low as you can get. And they can probably recover some of that cost by recycling some of the other metals from them.

I’ve heard quite the opposite: recycling lithium is difficult.

Maybe the other metals in the batteries make it worth recycling, but not for the lithium recovery (with today’s tech).

EV batteries can be repurposed for small-scale grid storage applications, and therefore can have a second life for many years before they are finally broken down to raw ingredients for recycling:

https://www.bloomberg.com/news/features/2018-06-27/where-3-m...

Do aged lithium batteries lose charge efficiency, discharge efficiency or capacity?

Each loss would encourage different kinds of secondary uses.

Yes. Eventually, Li-Ion batteries will "puff up" due to chemical degradation and may even spontaneously catch fire.

https://www.reddit.com/r/askscience/comments/55q7py/why_do_l...

This always makes me nervous. If I have some old laptops lying around, how worried should I be about them exploding? Does it only apply if they're in use?

I believe they experience all of those. But yes, there are many secondary uses, especially in situations where high energy density isn't required.

After the last economically viable use for electricity storage, however, there has to be an economic incentive to actually recycle the raw materials. Presumably this is when value of the raw materials exceeds the cost of their reclamation.

>> Do aged lithium batteries lose charge efficiency, discharge efficiency or capacity?

> I believe they experience all of those.

Hmm...would that explain the CPU performance loss of an aging cell phone battery? Lower discharge efficiency = less power available for the CPU = CPU throttles itself to lower power usage?

That is exactly what happened, yes. Without the throttling, the battery can't deliver peak power, and the device crashes.

We are missing mostly production capacity. Also - what is the longevity of the battety? It may break even in two years time but if cells need to be recycled at the fourth year it may not be as good as it looks like.

My car is almost four years old and its battery has lost about 1% capacity. These things don’t wear out that quickly.

> My car is almost four years old and its battery has lost about 1% capacity. These things don’t wear out that quickly.

Probably because Tesla undersells their batteries. During hurricanes or other disasters Tesla routinely unlocks the software limit, allowing customers more range (https://www.teslarati.com/tesla-unlocks-full-battery-capacit...).

It's the same with HDDs and SSDs, which always carry a decent safety margin to account for media wear.

Only certain models are software limited, and the 85 isn’t one of those. They actually substantially overstated the capacity: the 85 means 85kWh, but the actual usable capacity is only about 77.5kWh. (https://electrek.co/2016/12/14/tesla-battery-capacity/)

The SoC window is not 0 to 100%, it's might be eg. 15 - 75 % or something. That's also why you don't notice battery degradation, since the window might be moving (keeping constant energy limits for 0 and 100% charged).

For kWh capacity it's just a matter of counting the cells. Actual kWh will vary abit by measuring technique but I assume there is a way to drain the nominal Wh:s from them.

That is good to hear as I typically only consider used cars and other than range, concerns about how the battery capacity holds up long term have kept me from really considering EVs. What do you drive? Tesla? Leaf? Something else?

It’s a Model S 85 with not quite 60,000 miles. From what I hear from other owners, this is slightly better than average but not much. I haven’t done anything special to treat it well, either. I charge it to 90% nightly (the default), charge to 100% before long trips, and have used Superchargers for maybe 20% of those miles.

Not sure what the parent drives but there are a number of variables to consider, for example with Tesla supercharing often will degrade the battery faster as will charing it in trip mode for the same number of miles driven, but there is a good amount of data out there for this car.

https://www.greencarreports.com/news/1110149_tesla-model-s-b...

"But, overall, the data offer some basis for confidence that a Tesla Model S will lose—on average—less than 15 percent of its battery capacity over the average 150,000-mile (250,000-km) life of a vehicle."

I have a 2011 volt and have had negligible degradation. Thats an N of 1 but from GMs data, degradation hasn't been an issue in volts.

Tesla gives their powerpacks a 10 year full warranty but projects a 15 year lifespan.

Even if it is scrapped after four years, if it keeps generating this kind of savings for that period, it’s still a 600% return on capital. Turning $40MM today to $240MM 4 years from now is a deal any business would take in a heartbeat.

I think your maths is off.

$40m * 4 = $160m

$240m / $40m = 6 years

Edit: Mixed up cost and income?