Your edit is spot-on. Utilities typically use natural gas "peaker" plants to meet mid-day spikes in demand, while nuclear plants are better suited to running at a steady state.
Seems like that would be charging and discharging 80% of the 10kWh battery's capacity every day. Would that affect lifespan? Also, it would seem to render the device less useful as a battery backup system during most of the day, since it would be mostly depleted.
Lithium batteries are fairly tolerant of deep discharge, but yes, I would expect that the pack would be worn out after max 7-10 years if used like this.
Yes, using the pack like this would make it much less useful for battery back-up, but where I live (SF Bay Area), I haven't had an outage that lasted longer than 2 hours since we moved into our house, and this would cover the basics (probably everything but A/C) for that time. You could always split the difference, for example using 4 kWh per day -- this would pay still give you 6kWh+ of backup power and let you get the battery at half off or better.
Economy 7 KWh cost : 8.4p (Best deal I saw was 9.4 for me, but this was a figure on MSE)
Standard tariff KWh cost : 14.5p (Best deal I saw was ~12 for me, but this was a figure on MSE)
That'd be a £306/year cost on economy 7 and £530 on a standard, taking some more extreme figures.
A $3500 (~£2300) battery would then have a payoff time of about 10 years.
Using the other figures (9.4 and 12) gives a £95/year saving or a ~25 year payoff.
That's also assuming the batter is perfect, no extra costs in installation and you use exactly 10KWh per day. Economy 7 tariffs are about 22p/unit during the day so any overrun would quickly eat into savings.
I do think this is all very interesting, but with current offers in the UK it doesn't look too beneficial. However, if someone were to offer cheaper rates on a more granular level (rather than always midnight-7am) then this could work out better. Perhaps an energy company owned battery would be a good idea, they offer a flat rate back to you knowing they can smooth things out.
The battery is going to wear out well before 10 years. If you oversize your battery so you are only pulling out half the charge in the battery[1], you can probably get 1500 cycles out of it[2]. That's still less than 5 years.
(This is a common pattern if you are trying to price-compare an electric car to a gas car. For the electric car, the electricity cost is a rounding error compared to the capital cost of the battery, so you can just about assume electricity is free, and worry instead about how long the battery will last.)
You will get a much bigger bang-for-the-buck by demand-shifting: if the biggest load is cooling your building in the heat of the day, run the A/C overnight to superchill a heat sink that can be accessed during the day. No battery installation needed.
[1] Toyota did this with the Prius. They purposefully did not use the full range of the battery because they wanted the battery to last 7+ years.
[2] http://batteryuniversity.com/learn/article/how_to_prolong_li...
3000 deep cycles is really pushing what the industry knows to be state of the art.
Maybe they are just taking the economic chance that most people won't be deeply cycling these batteries, and planning to do replacements for those who actually put it through its paces.
That's a good question, and it makes a big difference.
> That you can still get 20% capacity levels in year 10, or that it will still turn on? This might be considered "normal wear and tear."
Given that you can optionally warranty it for an additional 10 years, I doubt it's nearly that bad. If people are confused as to how they how they hope achieve 10 years reliability, and they are willing to warranty 20 years, they must have something up their sleeves.
> 3000 deep cycles is really pushing what the industry knows to be state of the art
I suspect that the powerwall's true capacity is higher than it's rating, and it uses that reserve so it's not doing deep cycles, similar to another commenter's assertion to how the Prius gets it's 7+ year rating,
> Maybe they are just taking the economic chance that most people won't be deeply cycling these batteries, and planning to do replacements for those who actually put it through its paces.
I think a combination of most users not fully cycling every day, extra reserve capacity to keep it from deep cycling, and some subset of people not using failing warranty conditions may all contribute.
I guess the energy companies have a financial incentive to run a more consistent supply if it could reduce the peak load.
Just reminded me of a brilliant BBC documentary from a while back - The Secret Life of the National Grid. I think you probably have to piece it together from clips on youtube these days.
The solar tie in is quite nice, but current feed-in tariffs are huge in the UK (30p per kWh perhaps), so you're getting more out of paying it back than you would for storing.
Government incentives obviously change the calculations, but there aren't any at the moment.
Consider the power output characteristics of some common energy sources:
Gas, Nuclear, Hydro, Wind, Solar
Our "sustainable" sources of power - hydro, wind, and solar - are also the ones with the most irregular and unpredictable power output. This creates a problem, because even if there was enough cumulative power output from sustainable sources today, it most likely wouldn't align with our usage schedule.
So there are two solutions: 1) find a way to map the sustainable power output to our usage 2) find a way to store energy
Today, our most readily available form of energy storage is gas. We just put it in a container and burn it when we need to. "Spin it up," if you will.
An alternative to gas is a crazy network of Powerwalls. The hope is that, eventually, instead of spinning up gas turbines during peak demand, we can just draw from our Powerwall.
There was another alternative on HN recently. Basically a super-deep hole in the ground with super heavy object falling into it, but suspended by a rope. You run a motor to lift the object during off-peak hours, then let it fall and spin a generator during peak hours. I can't find it right now but I remember thinking they had a great name.
---
That said, power output is only part of the problem. If you're looking to factor the potential disappearance of arbitrage opportunities into your buying decision, you also need to consider the time until regulatory changes allow for perfect pricing, and the time until these Powerwalls have precise enough information to decide exactly when to draw from the grid versus the battery.
tl;dr: "quickly" is probably optimistic. We're a long ways away from economies of scale.
Or a train going up and down a hill...
http://www.aresnorthamerica.com/
Both are basically the same concept as pumped hydro which has been in use for a long time... pump water up a hill when energy is cheap, let it fall down and turn a turbine when energy is expensive. And hey, if it rains, free energy!
The rail/hole in the ground approach have the advantage of taking up less land I suppose, but I suspect are less efficient.
Now, vanadium-redox flow batteries... That's always been my pet large-scale energy storage. That or hydro.
However, there is a bunch of companies that buy and sell electricity on the open market, so they have a direct financial interest to play with this arbitrage.
In the UK, another reader has calculated the best case scenario was profitability in about 10 years. (his numbers ring true to me, I calculated recently that there was no reason to take an economy 7 contract for my electricity usage using similar data) That's a lot of time even in the utility market. So if batteries not attached to solar panel are going to be profitable (in the UK), that would be a future generation.
Probably right around the same time they're no longer necessary, as the cost of distribute solar continues to plummet. Its actually truly amazing how quickly the price continues to drop for panels each year.
Considering the rate at which solar power is getting cheaper, I'd say a solar power based economy is within reach.
Chart here: Check out monthly/weekly energy production for 2014 https://www.energy-charts.de/energy.htm
http://www.greentechmedia.com/articles/read/Bring-on-the-ene...
I'm also wondering if the same software is going to be rolled out to Tesla's cars so that they act like Tesla Energy when plugged in overnight.
---
Edit for why: On-demand power is expensive. "Spinning up" additional power during the day is more expensive than having certain types of power generation running all the time. In theory, pumping power into a battery at night during non-peak hours will cost less than consuming directly off the grid during the day.
I'm forgetting all the details of energy economics, but IIRC there would also be far less emissions by having consistent load on the grid all day long. Something about the ones we "spin up" also being the worst for the environment.