I think its important to note, for those who complain Tesla can't compete against pumped storage or other utility scale storage methods, that Tesla is able to have this deployed in 3 months (this is partly because Tesla can just drop racks of batteries on site and be up and running, and partly because of the regulatory environment after the Aliso Canyon natural gas storage complex leak fiasco).
Edit: This will replace the need for peaker plants first (generators of last resort, very expensive, only run a handful of hours a year), and as the cost drops, will slowly push out base load coal and natural gas (by increasing the capacity factor of solar and wind). "Batteries are the new peaker plants", as it were. [1] [2] [3]
This is what it looks like when batteries are used to offset fossil fuel generators (instead of curtailing excess wind and solar, it'll be soaked up by utility batteries such as these). Frequently, depending on renewables output, the spot of price of power can go negative. This means someone gets paid to use that power. This is where utility scale battery storage shines, as its happy to gobble up that power, being paid to do so, and can later be paid to release that power when demand is high.
Edit 2: If Tesla can book this revenue in Q3, combined with their vehicle sales push, they're going to be GAAP profitable for the quarter, which will allow them to close the Solar City acquisition. Well played.
So this is only financially viable because of Californian risk aversion after last years gas leak. If you have a reliable natural gas distribution network, like the rest of the world, batteries don't make any financial sense versus gas power plants.
I mean, this entire plant can supply 20 MW and will cost somewhere around $20 million. For that price you can buy two General Electric LM2500 gas turbine generators that can supply 47 MW and will fit in the same space as three 40-ft shipping containers. (These are basically just aircraft engines modified to produce electricity.)
So while this particular sale is of course good for Tesla, I don't see how one can justify this as Tesla "transitioning to a clean energy company". They saw a profitable, but likely one-time, opportunity localised to CA and they took it.
Why are people comparing and debating "batteries vs combustion turbines"? Batteries only store energy and you still need to generate it somewhere. Probably from a CCGT plant.
And it's not risk aversion. Its buried subtly in the article but part of SoCal's gas infrastructure is shut down because of the leak. They literally can't provide enough gas to generate enough electricity for peak demand. So they will generate over capacity at night, store in the batteries, and discharge during the day.
Also what seems to be lost here is that Tesla created it's utility battery products as a renewables play, but are just taking advantage of extra-ordinary circumstances in this case.
People are comparing them because they serve the same purpose here.
Base and peak loads on the grid can be wildly different. Without storage, if you want to avoid brownouts or blackouts, your generating capacity needs to match the peak load, even though you might only hit peak load a few times per year. The traditional way to handle this is to have power plants that can be spun up rapidly but sit idle 99% of the time. Because they're idle most of the time, the electricity they produce is extremely expensive.
Storage (including batteries, but also many other technologies) can substitute for these plants. You fill the storage when demand is low, then drain it when demand is high. Yes, the energy still needs to be generated somewhere, but you can generate it using existing plants during periods of low demand. If this is cheaper than maintaining peaking plants that mostly sit idle, it's a win.
Yes I know how all that works becuase I worked in the utility industry for over a decade.
My point was
(1) when comparing peakers vs storage you need to consider the cost of:
Off-peak generation + storage cost vs peaker cost.
(2) This wasn't a cost driven thing. This is a demand driven thing.
Off-peak generation is really cheap, though. It doesn't affect the equation that much. In some places (I'm sure SoCal isn't one of them) it's cheaper than free sometimes.
And of course it was a cost driven thing. This is the cheapest way to meet the demand. If there were a cheaper way, they would have taken that.
Energy density of gas (at conventional pressure) is 100x lower than that of a charged battery. Storing gas instead of electricity might not be such a good option.
Otoh, a gas tank requires only O(E^(2/3)) material, while a battery requires O(E) material, where E = stored energy.
I'd encourage you to spend 15 minutes researching what the costs are of running a peaker plant. There's a reason they're the most expensive generation method per MWh (peakers run for under 100 hours per year).
Revenue is revenue. If Tesla is able to cannibalize the peakers to contribute towards its enormous capital expenditures, and continue to reduce costs with the Gigafactory, it only makes utility battery storage more affordable.
There is no scenario where the cost curve per kWh of battery storage goes up.
Batteries are now cheaper than peakers everywhere, not just CA. That's what I'm saying.
Tesla's PowerPack installs (battery storage receives federal incentives, nat gas does not) will replace peakers, leaving them as stranded assets (the cannibalization I refer to).
Edit: unepipe: Tesla packs perform the exact same role as a peaker plant: energy on demand.
They are distributed, dispatchable loads and generators, available to draw or release energy from/to the electrical grid as needed, when commanded to do so by a grid operator.
I assume I used "cannibalization" improperly in this context. Mea culpa.
holy cow that is an amazing pdf file! Or maybe I just especially like it bc it's exactly what I've been looking for in the last few weeks.
I'm really very interested in the economy and technology associated with deploying clean energy. I'm an idealist and want to see green power, but only in a way that makes sense. And I'd like to profit it from tracking the process.
Have you got any other resources or directions to go to find more information like this? I've found this: https://sam.nrel.gov/download
Since they're not technically a "plant" - just storage - is there any advantage to be had to build in many smaller (say per-neighborhood) battery plants? Seems like they could act not only as peak providers, but also battery backup in the case of storms and whatnot. Plus it reduces a single-point of failure.
As you can see in the graphs I linked in upstream comment, in the UK alone, CCGT power plants have a capacity of 20 GW. That's 1000x this battery plant; meaning a $20 billion capital investment. And I don't see any way that Tesla (or anyone else) has the manufacturing capability to deliver that. Remember this is just for the UK, so for the US and Europe in total you're talking at least 10 000 plants like this.
Not all gas generators are peaker plants. Gas is displacing coal for what is often termed "baseload" (though I don't like the framing implied by the term).
The batteries will nibble away at the most expensive peaks first, the plants that run for only very short periods per year.
Meanwhile solar and wind will roll out to displace fossil fuel usage as well. Ideally they'll meet in the middle at some point, but it appears financially viable for this process to start now, and await further drops in pricing to expand later.
I thought about this some more, and I realised, Tesla and other EV manufacturers better be praying that this is not a financially viable replacement for gas turbine peakers in general.
Otherwise there will be a big jump in demand for batteries, as power companies will be looking for hundreds of GWh of battery capacity. If they make a yearly profit for the power company that's significant enough to warrant the CAPEX of replacement, increasing the battery price by 10% only means one or two more years before break-even. So power companies could easily afford that if we assume the investment at current price levels is profitable for them.
That is going to significantly (negatively) affect the price developments of batteries and constrain supply even more than today, so EV manufacturers will be faced with either losing (more) money on each car or increasing prices and selling less cars.
Unless we assume Tesla is looking to pivot out of EVs in general, but I think that's a rather absurd assumption given their current actions.
Tesla seems to be making the opposite gamble - that the grid-attached storage market will give the large-battery market economies of scale and help the EV market (and the other way around). They're building a very-large-scale battery factory, and probably cannot utilize it fully just on building their own cars.
Even if massive demand for batteries driven by grid storage makes EVs uneconomical, Tesla is setting themselves up to build and sell the batteries, so they win either way. (Well, aside from their goal of having the entire planet transition to EVs. They win financially.)
Musk has stated publically that they're betting the company on the battery price coming down for the Model 3 to be profitable. That's why they're building the Gigafactory, to reach $100/kWh. If suddenly power companies start buying lots of batteries, and prices stabilise at or above today's levels, it would mean (by implication) that the Model 3 won't be profitable (at least initially).
Sure car batteries and power banks are different, but they're all Li-ion and I believe Tesla (unlike others) is packing 18650's in both.
If prices don't come down, especially due to higher demand, and the car isn't profitable then the battery factory will be instead. Sounds like they win either way.
If this really ends up being the case they can just slow the model 3 deliveries to a trickle-- All those pre-orders are really just an interest free loan with fungible time tables on actual delivery, and the "financing terms" only get better for Tesla the longer they kick the can down the road.
(What is the net present value of $1000 in 2020? 2025?)
Extra demand for that giga-battery factory Tesla is building? What a dreadful problem that is(?!)
Also, there are other vendors happy to sell more batteries, including alternative technologies which are unsuited to cars but very suited to static power battery stations. (eg, Flow Batteries)
That would only be a problem in the short term - if that. In the long term, if the demand is as high as you say (hundreds of GWh per year), then you'll see Tesla, Panasonic, Samsung, LG, Bosch, and whoever else is making batteries step all over themselves to build new mines or mining capacity and more battery pack factories like the Gigafactory.
This would be a great thing for EVs. At worst it will be a few years of slowed down growth for EVs just like we're seeing now with cheap oil. These are all temporary, and neither will "kill EVs". And after the dust has settled, you'll probably see $50-$70/KWh batteries in cars.
The cost of batteries is falling at a current rate of 16% per year and it's accelerating. As such, battery storage will become increasingly competitive with peaker plants.
"From around 1995 through 2010, energy storage costs had fallen by about 14% a year – a rate that means the price will halve in a little more than five years. In 2010 – that rate sped up to 16% a year, and while a 2% increase might seem small it actually means that instead of taking five+ years to halve – it would now take slightly less than four years, more than 20% faster. Tesla’s Gigafactory has now increased the rate of price decrease even faster than the 16%. It is estimated that the Gigafactory, by doing nothing more than moving all of the supply chain arms under a single roof, has decreased the price of battery packs by 30-50% – while doubling the global volume."
Bookmarked to come back to show my math in the morning.
The gist is they can absorb power at a cost that is pretty close to free (when there is excess power flowing through the grid) versus burning a fuel whose price is volatile, a steep declining cost curve for batteries, and aggressive federal government tax incentives.
That's entirely fair, but then we should also include the full costs of peak power, including all the externalities due to emissions and such that the utilities don't currently have to pay for.
Peakers aren't "profitable" per se, they're a result of the government granted monopoly given to power utilities. In california, at least, a brown or blackout caused by insufficient capacity can cost the utilities tens or hundreds of millions in fines so the peaker plants are a relatively cheap way to avoid having your margin wiped out in the summer. That said, I'm not sure 20MWh would really help replace any peakers although in the future they definitely will.
They're probably not profitable in the usual sense of the word. Instead they're just required by the regulators that enforce the utility monopoly and set the utility rates such that the power company doesn't go out of business.
Peakers are profitable precisely because the supply is close to non-existent at peak consumption. What batteries are doing is "moving" excess supply from one moment to deal with excess demand at other.
> Sure peakers are expensive, but they are profitable, otherwise no-one would build and run them.
i don't know anything about power, but this is faulty business logic. we do plenty of stuff that isn't profitable, but rather is value-add to either win or retain business.
Lithium is a common element. The difficulty is always that it is rather even distributed making it expensive to concentrate. It's the kind of thing that is solved simply by scaling up. It is in fact more abundant in the Earth's crust (17ppm) than lead (10ppm) and according to Wikipedia <https://en.wikipedia.org/wiki/Abundance_of_elements_in_Earth... 2012 production of lead was 5.2MT while lithium was 37kT.
Abundance in Earth's crust doesn't help at all during a supply crunch. Bringing a mine up takes at least 5 years. And your mine is going to be on an orebody, not just processing random rocks. Cobalt is another critical element for batteries I believe.
Lithium is "mined" by pumping underwater brine to a surface pool and letting it evaporate. The cost effectiveness of this apparently caused most Lithium mines to shut down inthe 90's because they couldn't compete[1]. Lithium is also present in seawater, but they apparently don't have a cost effective way to extract it from that yet. Apparently the evaporation process is 18-24 months per batch though, so there is some lead time, just not 5 years (good), and it doesn't necessarily lesson by all that much if you have the facility already set up (as in you can't necessarily throw more people/machines at it for a short term increase, which is bad).
A commercially viable way to extract it from seawater would likely solve any supply problems indefinitely.
This, of course, doesn't address the need for Cobalt, but I'm not sure how much is required for the process.
Just to inform the discussion wrt. timescales and magnitudes of power variability, here is a real-time plot of power generation in the UK showing daily/weekly/monthly variations in consumption and in production from nuclear, coal, gas turbines, wind, bio and (pumped) hydroelectric. This is with zero battery storage on the grid (I believe).
> "If you have a reliable natural gas distribution network, like the rest of the world, batteries don't make any financial sense versus gas power plants."
"The rest of the world" also has measurably more expensive gas. ($10/mmBTU for large scale users around here).
> If you have a reliable natural gas distribution network, like the rest of the world, batteries don't make any financial sense versus gas power plants.
In the US, gas companies seem to be having issues expanding natural gas distribution recently. Plenty of cancelled projects, lots of protests on construction sites, etc.
I wouldn't say it's one time. There are many islands (like Hawaii) that have to import diesel to run their generators, which is very expensive. Replacing diesel generation with renewables + storage will be very attractive for a long time.
Gas power plants and batteries have different timescales. Grids with a lot of renewable energy need a slice of capacity that can ramp up faster than a gas turbine.
The fact is, most parts of the country don't have good energy storage options. Plenty, like Florida for example, are almost exclusively coal + natural gas and some nuclear. The advantage of these battery facilities is that they enable transitions away from coal and natural gas.
One of the things I wish I could buy would be a Bloom Energy 2kW natural gas fuel cell and a 50kWh LiON battery pack with whole house inverter.
The nice thing about the Bloom fuel cells is very efficient conversion of natural gas to electricity, the weakness is that it's response time is long (an hour or more to change its output by 50%). The nice thing about LiOn battery packs like the ones in Tesla cars are that they respond instantly to various power demands, can deliver massive amounts of power in a short period of time, and recharge again and again.
This combination would let me supply all of my house power under all circumstances using nothing but natural gas. That would take my house completely off the grid infrastructure for PG&E (although I would still be a gas customer).
If you can get the Bloom fuel cell, you could already do this with several PowerWalls ganged together and the proper inverter. Why the fuel cell though instead of solar? Trying to sidestep the monthly electric utility charge? If so, I'd still go solar, and rely on a natural gas generator as your supply of last resort.
As you may have gathered by now from another section of this thread the fuel cell is meant to be supplementary to an existing solar energy system whose usability is variable.
Not only does Chuck want to own & charge an electric car, but he also wants to be able to take advantage of many days of sunny weather in a row. A smaller battery means a bigger gas bill.
It's pretty important to realize that this isn't the problem that needs solving.
Batteries aren't great for large amounts of power for long periods of time (yet). What they are good at is smoothing out the supply curve from potentially intermittent suppliers like solar and wind, and that is one of the things that causes price spikes.
Having real-time pricing of electricity coupled with users who can adapt their usage automatically to real-time prices will also considerably moderate power demand.
In the Netherlands there is a lot of reclaimed land that needs a lot of water pumped out regularly (though not 24/7, during normal weather).
I know there have been pilots with some of the larger pumping stations to incorporate the current real time price of electricity, as well as the forecast rain within the next few hours to decide when to turn on. Not sure if it's actually used in production now.
That sort of thing is a great fit for that problem.
This seems outrageous to many of us in the West, but here in Australia our power utility already cuts power to neighborhoods in half hour blocks during peak power consumption during summer (I believe they exclude hospitals and some other sites like that from this policy).
The infrastructure savings made by this are huge (think of 99% reliability vs 99.99% in software), and the few that really need it can invest in backup solutions for those times.
How huge are the savings? Do you have any numbers or sources?
Do those power utility savings translate into savings to the people who are cut off? I for one would not want to trade electricity price for reliability.
It's surely less expensive per capita to have reliable electricity for everyone than to have everyone who wants reliable electricity to have to install a generator.
99% uptime means 7 hours of downtime every month. And it's not random downtime, it's downtime under peak demand.
If the choice was free electricity at 99% uptime and you have to take care of the rest yourself, or $0.13 / kWh for 99.999% uptime, I would go for the 0.13 / kWh.
Of course in AUS they probably charge more than 0.13 / kWh and make you deal with the brownouts anyway. Actually some quick searches reveal AUS pays about the highest rates in the world.
Reliably not being able to handle peak demand isn't a cost saving measure, it's an excuse for a major failure of the infrastructure.
To be clear, in Aus we don't have 99% reliability. It's about 45 min every 2 years (whatever that works out to). That number was to explain the cost savings.
In much of the world those cost saving are directly applicable. Most of the world's population doesn't have 99% reliable electricity, and so they build reliability at the edges with everything from batteries in phone towers, to batteries in lighting.
It isn't at all clear to me if this is more or less expensive for new infrastructure. Given the increasing popularity of roof-top solar, the pricing model for 24/7 reliable wiring doesn't work out now in places where it is already built out. Network operators are trying various legislative measures to get subsidies for the networks, because no one wants to pay the rates they cost to maintain.
If I had free (or very cheap) electricity 99% of the time, spending $5K to get 99.99% reliability via a battery system is very tempting (and we are getting close to that point now). $1K for 99.9% - maybe.
One interesting thing is that the distribution of the downtime matters.
Rooftop solar is on average a $25,000 - $35,000 investment, and that's not even counting energy storage required to actually be off-grid. I don't think that only the rich should have reliable energy, and the poor should suffer brownouts to theoretically save some infrastructure costs.
For countries who don't yet have a reliable electric grid, I think investing in providing quality / reliable electricity to their population is about the best ROI (after providing reliable drinking water) investment they can make.
The economic cost of blackouts and brownouts are extremely high [1] -- for example, the rolling backouts in California back in 2000 - 2001 were estimated to cause GDP loss of 0.7 - 1.5%! (GDP was ~1.2T, so we're talking economic losses on the order of $10 billion).
The Tesla powerwall is currently $3000[1]. There are installation costs etc on top of that of course though.
Solar is separate - I'm talking about charging the battery from the grid.
Brownouts are expensive, but so is the cost of making sure they don't happen. Peaker plants cost more than $10m/year without even turning them on[2]. There are over 40 in California[3] currently available. Hopefully they don't all cost $10M each to keep available, but it doesn't take many years to surpass that $10B cost...
Rooftop solar is on average a $25,000 - $35,000 investment, and that's not even counting energy storage required to actually be off-grid. I don't think that only the rich should have reliable energy, and the poor should suffer brownouts to theoretically save some infrastructure costs.
I'm assuming you are talking USD, if so? Are you sure you have your numbers correct? That seems ridiculously expensive compared to say Australia. First link on DuckDuckGo: https://www.solarquotes.com.au/panels/cost/
If you need your solar to generate your AC in California or another southern state of the US, sure.
Here in the Netherlands, our complete solar system with installation cost us EUR 4,500, and we do run some AC in the hottest summer months. We are currently generating more than we use, with a household of 2 adults and 4 kids.
Commercial users do see this, with higher costs for energy in peak hours as well as "demand charges" based on their highest rate of consumption (typically measured in 15 minute increments). Charging more for those gives an economic incentive to trim down the energy consumption bursts that require peaker plants to run.
Buildings can design load leveling into their systems with things like thermal energy storage that shifts air conditioning consumption off-peak by freezing ice overnight and using it to cool the building during the day.
People have always tried to come up with a reason to connect the refrigerator (and other appliances) to the internet. Finally, we have a reason - to moderate the power consumption of the refrigerator based on the spot price.
Sensible price conscious refrigeration control would just be different temperature targets based on the price, not turning off the system. If l33t d00dz wanna make my milk one degree warmer, it's not a big deal.
Most peoples electricity usage is relatively inelastic and no one wants the cognitive overload of checking electricity prices before they put the TV on.
Oh, I disagree. A/C is very power hungry, and can be adjusted to cool the house lower than normal during cheap power times and let it go a bit higher during expensive power times. It can also "pre-cool" when power is cheaper.
The same goes for an electric water heater.
The charger for your Tesla can also adjust when it charges
based on power prices.
The dishwasher/washer/dryer can be set to come on at night when the power rates dip.
This has been the promise of "smart-grid" attached devices; the implementation hasn't caught up with the dream yet, but in theory, yeah, there are a lot of very elastic loads that can be used to smooth out the demand curve, or to match a jittery supply curve.
It ain't gonna work if you need people watching the electricity prices and deciding what to do. It need to be devices talking directly to the power grid. And that need standards for things like communication protocols. Does anyone here know how are those coming along?
> And that need standards for things like communication protocols.
All that is needed is a way to get the spot price of electricity from the internet, similar to how I can get the current temperature in Anchorage. Devices don't need to talk to the power grid.
If you're going to be billed by the minute, presumably you won't be taking meter readings manually, so your meter will be connected to some sort of communications network anyway.
And if you're /not/ getting billed by the minute, what's in it for the buyer to choose an AC that sometimes turns itself off?
The power company has been upgrading the meters for years to support variable pricing so they can bill by the minute. None of what I've proposed is in any way a difficult or expensive problem to solve.
Of course, the power company having billing by the minute will then create the incentive for such an A/C. That's the whole point!
Moving electricity is "hard" so effectively there are many marketplaces in which the spot price is different. A natural design that's been considered/proposed is to get that info directly from the power utility itself.
If I can get the temperature online for my neighborhood (I can), then how can it be hard to get the local spot electricity price online? I don't see anything unnatural or difficult about this. No infrastructure changes are needed.
What's the behavior of the system when the home cannot access the marketplace via their normal ISP? I think power utilities don't want to bind their QoS fate to a company outside of their control or specification.
Um, if it can't connect to the internet, it reverts to the default behavior it has now? Or it could revert to the pattern seen yesterday? Besides, an internet connection is critical infrastructure for about everything else. What's so special about this?
I'm with you all the way up to affecting appliances. The key is the power shift should be virtually invisible to consumers. You don't turn off A/C, you just bump up the set point by 1 degree for that 15 minutes. The individual impact is minuscule, the aggregate impact is leveling out the peak.
But you need everyone to be wired for this in order for a small change to add up to a big impact. Which isn't worth the cost unless it's mandated. Which is a political mess.
So we solve the problem another way; with batteries to meet the peak power demand at a reasonable cost, and not demand a massive IoT network, constantly monitoring usage and with override control authority.
You don't need everyone wired up, just a sizeable chunk. All that is really necessary is for the power companies to make spot pricing available via the internet. Appliance companies will find it worthwhile to add this capability because customers will want it because it saves money for customers.
The system will convert over as people normally replace their appliances, just like bluray has pushed out dvd players.
They actually already do it, we enrolled in Boston this summer in "RushHour". They pre-cooled our house 6 times over the summer before a expected peak demand and gave us a $40 or $50 credit. It was great, we were not home 3 of those times and the others we didn't notice at the time at all, so it was basically just free money.
That's a fine idea. I didn't think of that. But I'm sure there's a LOT of low hanging fruit like this that would greatly mitigate the need for the power companies to store power.
Most of benefits come if you have a system which can regulate slightly your use of air conditioning and heating (and a few other things like pool pumps).
In some cases you can save 50% of a daily powerbill by turning off the compressor in an air conditioner for 15 minutes.. but the right 15 minutes.
Few people would get up and turn off the air con for that, but a system that let you say how much you want to spend on it that you setup once makes a lot of sense.
It's a different beast. The price is high during those 15 minutes because of supply and demand. If you implement a smart grid that regulate load, you upset the supply and demand balance and the power might become so cheap that new self regulating systems are not worth the initial cost.
Once the smart-device interface is properly defined and the software is written, the hardware cost should not be a problem.
You will end up with some kind of "smart device" router that connects to your internet/home network to talk to the power company, and uses a commodity wifi interface to talk to your smart devices. The cost of an ESP8266 is already sub-$7 at hobbyist volumes, you can expect the additional hardware cost to easily be driven to trivial levels with widespread adoption.
The problem is purely in getting a standard system out there and designing devices that incorporate the new control model.
It is completely unnecessary for appliances to talk to the power company. They only need to talk to the internet. The power company just posts the spot price on a web site.
The appliances won't talk to the power company, they'll talk to the appliance hub. The appliance hub will be the one with the logic to handle temporarily raising the thermostat or whatever, and will handle the internet connection.
Also, using a website or other resource run by the power company is still talking to the power company. Not sure ho else you could have parsed that, but you clearly had some crossed wires there.
However, wouldn't it make sense to e.g. adjust heating/air conditioning depending on electricity spot price? That's then just a gadget that gets the information and guides the home system.
It'll probably be more meaningful in some industrial plants though, if there's any such energy-intensive industry.
I'd be pretty happy to pause an EV charge many times. Since Teslas are data-connected, they could possibly do this with no grid interface beyond the power connection.
This is one very nice aspect of a grid when there are a lot of electric cars which charge at night, and have big enough batteries that they don't need to fully charge: you can absorb a lot of variability of overnight wind energy.
And if electric cars also plug in at work... jackpot.
Even better than that, it could be set up so that your house drew power from the car battery when rates were high! You could set it to, for example, draw from the car battery until the charge drops to 80% or some settable limit.
Then all the rechargeable cars become part of the battery moderation of the whole grid!
The Tesla engineering head said giving power back isn't likely, as it would degrade the battery too much, but he stressed how much benefit to grid stability could be delivered just by altering when the batteries drew power to charge.
Sure, there are some research projects for that, but note the wildly different cost: charging a little different doesn't degrade the car's battery. All it takes is an Internet connection, a good prediction of how much charge the driver will need, and a micropayment of some kind.
That's a no-brainer compared to paying the owner for shortening the life of the battery.
Most load is predictable; only a little bit is spiky. You don't really need to power 2,500 homes for a full day; you need to power five or ten huge factories at temporary overcapacity for a few minutes at a time, as reserve plants spin up/down.
Consider EC2: you have reserved instances, and then you have autoscaling. If you know the total amount of work is going to predictably rise, you just buy more reserved instances (i.e. build more real power plants.) The non-reserved "elastic" VMs (like the batteries) are just to soak up your highly unpredictable excess load.
Or, think less of a power plant, and more of a capacitor: if a capacitor is way more than is strictly needed to soak up any possible load spike in a system, then we'd say that the system is overengineered, even if a later version of the same system might have enough load to blow that same capacitor. It's "massive" in terms of the predicted load spikes it's facing, not massive in terms of the long-term consumption trend of the system.
But put in another perspective, it's enough to power ~3,6000,000 homes for a minute,or 720,000 homes for 5 minutes. If that can be used to prevent a brownout or blackout, that's a net win in my book.
This of course assumes average use, which isn't really likely in the case you are protecting against the system being unable to cope, but at the same time it's supplemental to the current system, just as an extra few percent of capacity for an area, not meant to power homes directly or without other sources.
A more useful perspective would be how much of the gap between peak generating capacity and peak load it covers. It's not intended to run SoCal for a day. Unfortunately, the "2,500 homes for a full day" statistic is really misleading, since it seems to imply that's what it would actually do.
So California gets big rechargeable batteries, interesting.
Does it mean Solar City (pending acquisition by Tesla) will speed up building Gigafactory and that's how Tesla will deliver it?
And does it mean that the market was wrong with recent Solar City stock drop?
I'm assuming this is just a "pilot" and, if executed happily, can keep doubling capacity every x months, driving battery prices down, leveraging it as an further advantage over fossil alternatives.
I'm assuming I'm completely wrong because I don't see any spikes in Solar City stock prices?
ps. Bit off-topic but when opening this article I've decided to disable adblock for bloomberg, just because they made bucklescript :P
There are two different Gigafactories. Tesla is building the one in Nevada for batteries, and that's the one you hear about in the news the most often. SolarCity is building a different Gigafactory for solar panels in Buffalo, New York.
There was a really neat class I took in college on the economics of elctricity markets, and one of the things you discuss is base vs peak power generating plants (as discussed other places in this thread) https://en.wikipedia.org/wiki/Capacity_factor seems to be a nice introduction to the idea.
A few years ago, one of the more interesting responses I had to an announcement of a project in to modern asset-generic, topology-generic, risk model-generic, actor-perspective oriented transaction protocols[1] was a US academic engineer interested in applications to electrical grids and embedded power systems in resource-constrained environments. I would like to have continued further and still believe in the currency of such a system (especially given rapidly evolving, complex risk models in infrastructure systems of all types), unfortunately my employer had other priorities and I have not had reason to return to the project. The IFEX protocol notes in particular may be interesting reading for those in the area, however.
You could also frame this as his companies align well with government initiatives because their success is largely in the public interest (or at least the success of the markets they are in or trying to establish).
Is this really unusual? I've seen that said before, but I'd expect everyone to take maximum advantage of such things, aside from a very few extremely principled people.
SpaceX rockets don't have the most "advanced" engines, but they sure are cost-effective. Tesla's done some pretty good things with 18650 cells and cars, and now they're hitting some serious volume. I'm surprised that that cell works for utility-scale deployments, but...
I've seen a bunch of HN comments over time claiming that Tesla has no battery technology, because they don't build the basic cell. Turns out that battery packs are more than just the cells they're built out of.
At utility scale I expect it is a very simple design problem in fact. It's harder to design them for confined spaces or for crashworthiness but for a fixed installation with no geometric constraints it would be a piece of cake. Anyway Panasonic already has this expertise since they deal in utility projects elsewhere (and in America according to the other comment).
Because the world we live in has certain types of people -- politicians, lawyers, business executives -- that don't solve problems for the world as in, "what is the most efficient way to accomplish this;" rather, they solve problems as in, "what is the most efficient way I can enrich myself.
I guess it's maybe rational in a self-interested way--no longer do we live in a world of many competing tribes advancing ourselves for our own survival among competition for scarce resources. We live in an increasingly globalized world with a class structure and an overabundance of resources. Or, in other words, "the scarcity problem," as Keynes put it, has pretty much already been solved--at least in terms of survival we have enough resources to feed, cloth, provide shelter and basic healthcare to every person on the planet, but we don't. The answers to that are some of the same reasons.
In a general sense is it not better for all people to be in he business of creating value than to quarrel over transferring it from one to another? I believe that if you care about human welfare, then it is obvious that it is.
Because distribution of value is important, and people believe that distribution of value is important. After all, there are limits to how much I can feel your well-being, and vice versa.
I actually support wealth redistribution, and I even buy in to the concept of diminishing returns in utility to increases in wealth, like you said--though, economists do debate whether that's actually true.
I find it weird how technological process sometimes plays out. We have been complaining for decades about how green energy is unreliable and that we need innovative storage solutions. And now batteries somehow seem to be good enough and we are like: "Hey, why don't we put a bunch of batteries in a container?" Proving ones again that economy trumps technology.
The next really big thing after green energy will be recycling. Not that it matters to our generation.
I think this is very much a success of technology. I see it as battery tech having advanced enough (at a slow, but exponential rate) to bring us to a point where this is technologically possible AND economical
I guess what I am trying to say is that the future is rather lame and predictable. When imagining the future we are sometimes thinking there will be some giant jump forward, while in reality it is just decades and decades of minimizing microchips, increasing photovoltaic efficiency, improving batteries. This technology (except the microchip) is from the 19th (!) century. On top of that we are using some statistical methods and are trying to emulate neurons.
This has a positive and a negative side: Everything is already there, we just need to implement it, but this does not really encourage you to dream. There is no magic. For me that is part the reason why careers like research or medicine were not attractive to me: You can progress further or save more lifes by just using stuff that has been here for centuries. I believe everybody can make a difference, because we are wasting our potential by not implementing stuff.
I know that this view is oversimplified and there is indeed fundamental research (e.g. new materials), which leads to some of these gradual improvements.
The day a room temp superconductor is developed (and is scalable to mass production) you'll see that change in a crazy hurry. Assuming we're alive then, which is probably not a safe assumption. There are a few things like that which could change the game almost overnight, but a room temp superconductor would be one of, if not the biggest.
Room temp superconductors will probably also take a while to come down in price enough to make sense and will see some limited deployment for costly applications before that. It's unlikely that we'll go from "no superconductors except in MRTs" to "superconductors as an essential part of the power grid" in a time frame that is measured in years instead of decades.
If the future is predictable to you, use that ability to make trillions of dollars and then use the resulting power to impact the future. Whining about it on the internet just makes you seem dismissive.
I predict that solar will be used in the future. How do I make trillions with that wisdom? You can only make money with predictions you are right on but others are not.
> Whining about it on the internet just makes you seem dismissive.
That was not my intention at all. It was more the child in me that wanted to see flying cars and jetpacks instead of massive adoption of stuff, which has been here for 150 years. I still welcome that it is done.
"The deal fits into Tesla Chief Executive Officer Elon Musk's long-term vision of transforming Tesla from an an electric car company to a clean-energy company. That's the same motivation behind his pending deal to acquire SolarCity Corp., the rooftop solar company founded by his cousins, of which he is also chairman and the largest shareholder."
Yeah man, what a lucky coincidence your cousin can help you saving the planet!
Solar City was retconned into a vague document that takes a mile-high view of the problem. This was unambiguously poor corporate governance, unless you completely ignore the business side of...the business?
A positive development, sure. But 80MWH is a little less than the amount of energy a 1GW power plant produces every five minutes.
I want alternative energy to work, but it's sobering to see the scale of the problems involved. It doesn't seem to me that laptop batteries scale up so well. If I were forced to place a bet on the future of grid-scale storage, I'd look for something else.
I don't understand why there are so many power grid issues in the USA. It's not that hard to generate power and make it go from A to B, yet for some reason (political? commercial? geological?) there seem to be decades of general issues in generating enough to meet demand. Does someone know what the actual issue is?
Famously the US had an extraordinary power grid for many decades. Among the most reliable of any nation. Blackouts used to be very rare in the US, now they're still rare but not nearly so.
Since the 1980s there has been a significant under-investment in the grid. Since the 1970s there has been a significant under-investment in production. The outcome to that is what you see now.
Good article on the rising blackout phenomenon from 2010:
Musk was commenting on this point in his YC interview yesterday. There can be no assumption that technology will always get better. If people aren't focussed on improving it then the current level of the technology will fall.
It isn't mentioned in the article, but one of the advantages of something like battery storage is the ability to regulate consumption and production on a millisecond timescale to regulate grid frequency. Anyone know if that is a part of the project here?
This was brought up in a Fresh Air interview I listened to recently. Nowadays most of our devices use DC, and our grid is also on the verge of becoming much less centralized.
You could reap some easy efficiency gains by omitting the DC-AC-DC conversion of a solar panel powering a house that charges a laptop.
Heat in wire is ~ current squared times resistance. If you use low-voltage DC, and you're supplying any significant power, then the current is high and the losses to wire heating are high. Inefficient and dangerous (fires).
AC at high voltage can be low current, you can use much smaller wires for a given power.
DC doesn't have to be low voltage though. Before high power solid state electronics became reliable, it was difficult to do DC-DC voltage conversion without going through AC first, but that isn't true today. Plenty of power is delivered across different grids using HVDC.
Without reading this somewhere or having anything to back it up: This, combined with the acquisition of SolarCity, looks like a move to get Tesla into a different Asset Category to make it even cheaper to loan money. If they can be seen in the same risk/asset-category as a power utility, they can rely even more heavily on loans.
Does anyone know if that would be a viable strategy?
OP seems confused, claims these battery packs will 'replace fossil fuels'. Under some misconception that the batteries get charged for free or something.
Solar and Wind aren't fossil fuels. That said, we need a whole heck of a lot more of them to replace fossil fuels. But, as other commenters have stated, if you can charge these batteries during the day and use the power at night, then yes, you can replace the 24 hour nature of gas/coal.
They do displace fossil fuel generation to some degree, since utilities use natural gas turbines to respond to rapid load shifts. If all these batteries do is buffer short-term demand spikes, then base load can be supported by e.g. nuclear power, which is very slow to spin up or down.
The biggest need for storage is to offset the intermittency of solar and wind, and to help shift peak solar generation (typically 12-2pm) to match peak demand (typically 4-8pm). Reliable and affordable storage is a prerequisite to wide scale solar and wind generation.
Sure you want to smooth out intermittencies, but this plant is way too tiny to actually store energy when it's windy and releasing it when it's not. It's more likely to handle peaking on a minute-by-minute basis.
Considering that a single big wind turbine produces 8 MW, this 80 MWh storage facility can only store 5 hours of production from two wind turbines!
And one of those turbines only costs about half of the Tesla storage facility.
Chicken meet Egg, we have to do both things does it matter in what order it is done. Oil can be stored in a tank, renewables really are playing catch up in terms of storage.
Thermal power stations (oil, coal, gas, nuclear) cannot start up or shutdown quickly so they have to run enough to meet peak demand 24/7. Lots of energy is wasted in the small hours.
Sure - but I think alternative source folks are excited about this because at scale batteries allow solar and wind to be used to power larger swaths of the grid more dependably. That's a pretty exciting concept.
That's 99% true, but Southern California also has periods of time when there's too much _solar_ power and they need to take some solar systems off-line!
What a bullshit. I'd call this Russian way of doing business, where you benefit from unhealthy connections with state. 2500 houses? Go find out how many houses are there in California.
Edit: This will replace the need for peaker plants first (generators of last resort, very expensive, only run a handful of hours a year), and as the cost drops, will slowly push out base load coal and natural gas (by increasing the capacity factor of solar and wind). "Batteries are the new peaker plants", as it were. [1] [2] [3]
This is what it looks like when batteries are used to offset fossil fuel generators (instead of curtailing excess wind and solar, it'll be soaked up by utility batteries such as these). Frequently, depending on renewables output, the spot of price of power can go negative. This means someone gets paid to use that power. This is where utility scale battery storage shines, as its happy to gobble up that power, being paid to do so, and can later be paid to release that power when demand is high.
Edit 2: If Tesla can book this revenue in Q3, combined with their vehicle sales push, they're going to be GAAP profitable for the quarter, which will allow them to close the Solar City acquisition. Well played.
[1] https://en.wikipedia.org/wiki/Peaking_power_plant
[2] http://www.greentechmedia.com/articles/read/dueling-charts-o...
[3] http://www.bloomberg.com/news/articles/2015-12-22/batteries-...