Click on image for a larger version. A younger slimmer me back in March 2021 with my newly-installed Tesla Powerwall 2.
Friends, just over 5 years ago in March 2021 my Tesla Powerwall 2 was installed. Within a day or two the battery charged itself from my 4 kW-peak solar PV and my house went off grid: it felt magical and I wrote about it a lot back then:
Click for larger view. Re-published from my “Week One” article in 2021. After the battery was installed, we quickly went “off-grid” for the summer.
Since then I have written about the slow degradation of the battery which is disappointing to experience, but which is just part of how batteries work. However, day-to-day, the battery remains a thing of wonder – enabling us to go off-grid in summer and operate the heat pump at very low cost in winter.
Click on image for a larger version. Cumulative consumption of electricity from the grid through each year from January 2022 to April 2026. Between early May and late September we consume very little grid electricity. In 2024 I experimented with a different way of operating the battery – but it wasn’t a good idea!
In this article I just thought I would summarise how the Powerwall is performing half-way through its guarantee period.
Round Trip Efficiency
Round Trip Efficiency is an important parameter for a battery – but one that few people take time to measure. It answers the question:
If I charge the battery using 1 kWh of electricity, and then discharge the battery, how much of that energy is returned as discharged electrical energy?
The answer varies depending on the rates of charging and discharging, the temperature of the battery, the existing state-of-charge of the battery and how long the energy is stored for – amongst other things. At the moment the best one can hope for is around 90% i.e. roughly 5% loss on charging and 5% loss on discharging. The losses end up as heat.
The graph below shows the cumulative charging and discharging of the battery since March 2021. Fitting trend lines to the data we see that on average we put 3.34 MWh of electricity into the battery each year – but only extract 88% of that i.e. 2.93 MWh/year.
Click on image for a larger version. The graph shows cumulative charging and discharging of the battery month-by-month since installation.
Plotting the charging and discharging data month-by-month shows up seasonal variations with the summer being a bit worse than the winter: I’m not sure why.
Click on image for a larger version. Graph showing the the monthly averages of the amount of electricity charged per day into the Powerwall 2 and discharged per day from the Powerwall 2 since installation. The battery is used more in the winter than the summer but the decline in winter peak capacity is clear.
Click on image for a larger version. Graph showing the monthly round trip efficiency of the Powerwall 2 since installation. The red curve shows the average of the cumulative round trip efficiency.
I have not seen any other evaluations of Round Trip Efficiency but I am reasonably satisfied with the cumulative figure of 88%.
Capacity
The capacity of battery is an important parameter for a battery – but one that few people take time to measure. It answers the question:
If I charge my battery to 100% and discharge it to 0%, how much electricity (in kWh) is discharged?
I have charted the slow decline in the capacity of the Powerwall in a series of articles. The last article I wrote on this subject has links to the previous articles, and noted that Tesla had recently changed the software that controls the battery in way which was likely to extend the life of the battery (good) by reducing my ability to charge it fully! (not good).
Click on image for a larger version. Data gathered from winter days in which the Powerwall discharged from practically full to empty. The blue dots show all the data, and the pink circles around the dots show days on which there was no solar re-charge. The large black circles show the seasonal average
The Powerwall guarantee states that it will retain 80% of its capacity (i.e. 13.5 kWh × 80% = 10.8 kWh) after 10 years. Based on the data above I think it will be close to that that in 2031.
However, even a 10 kWh battery is actually a useful size so it is not obvious that 10 years represents the useful lifetime of the system.
Is Electricity from the Powerwall Free?
No.
In the summer the proximate cost of electricity is zero – electricity is generated by the solar PV during the day and used to charge the battery – and the battery discharges overnight. But it’s not really free.
We should really take account of the cost of the capital tied up in the slowly-degrading battery. If we take the cost of the battery (~£10,000) and divide it by the number of kWh it will supply over its lifetime (say 30 MWh or 30,000 kWh) one comes up with a cost of about £0.33 per kWh.
So, far from being free, every kWh of electricity delivered by the battery is actually quite expensive! One might consider the battery cost as the pre-purchase of a large amount of electricity.
If the battery lasts longer and delivers more than 30 MWh then the allocated cost per kWh also comes down, but for my particular battery system the figure will be in that ballpark. For battery systems purchased now the cost is definitely lower and for some systems it might be approaching £0.10 per kWh.
I am comfortable with these relatively poor financial returns. When I installed the battery it was not clear to me that it would work at all! But in terms of reducing carbon dioxide emissions, the battery is big help.
- We use no grid electricity in the summer – and hence there are no proximate emissions. Of course we still benefit from the mere presence of the grid – enabling our society to function and acting as a back-up. This is one reason why I am happy to pay standing charges – I am still benefiting from the grid even when I am not personally using electricity from it.
- In the winter the battery allows us to operate the heat pump at low marginal cost – and the heat pump is the lowest carbon way to heat the house.
Financialistas would argue that instead of buying the battery I could have invested the money and used the interest to pay the electricity bills. But Carbonistas would point out that that would have done nothing to reduce carbon dioxide emissions.
When the battery eventually reaches its end of life, I am modestly confident that the materials will be able to be recovered, and that new battery systems will be better in almost every way. And cheaper.
Protons for Breakfast
Happy birthday Michael’s battery !
> Plotting the charging and discharging data month-by-month shows up seasonal variations with the summer being a bit worse than the winter: I’m not sure why.
One possible cause for the reduced efficiency in summer is the rate at which you are discharging your battery.
In winter you will be discharging your battery much more heavily to run your home, heat pump, lights, etc so the ‘overhead’ energy required to run the battery gets spread out over a higher load, versus in summer when the battery is much more lightly used and so these same losses are more pronounced.
In essence the battery is less efficient when operating under lighter load.
> Is Electricity from the Powerwall Free?
> In the summer the proximate cost of electricity is zero – electricity is generated by the solar PV during the day …
> We should really take account of the cost of the capital tied up in the slowly-degrading battery
I take a different approach.
Agree that the electricity you self consume from solar generation is not free, I consider that it costs you 12p or whatever your export rate is, because that is what you could have earned if didn’t use it yourself.
I then compare the cost of electricity I pay with having the solar/battery system to what I would have paid were I on the standard variable tariff, and use this saving difference to work out the ‘payback’ on the money I spent on the system.
This is giving me about a 7 year payback. The solar panels should continue to operate without issue for 20 years or so. The inverter/battery, who knows how long they will work as my manufacturer has gone into administration and the warranty is now worthless, but will cross that bridge if I have to.
Yes could include cost of compound interest on the capital spent were it to remain in the bank, but I don’t.
There are of course multiple ways to consider the economics as well as the carbon dioxide emissions.
[heart] Ciro Alberto Sánchez Morales reacted to your message:
Well done Michael! I have updated https://en.wikipedia.org/wiki/Tesla_Powerwall#Technology to cite this blogpost rather than a 2023 one (in which you had estimated RTE at 87%).
Clark: Thank you! you are too kind! Being a wikipedia reference is the culmination of my career!
Best wishes
Michael
The Marstek Saturn B2500-D 2,24 kWh is available for roughly 300 Euros. With 6000 cycles this works out as 2 pence per kWh. The battery only works with DC electricity. The panels are connected to the battery and a small inverter then is used to convert to AC. This means that the battery cannot be used in winter to store off-peak electricity.
Balcony plug-in solar is a nice system for people in flats. Four 500 Watt panels are also about 300 Euros and can be mounted on a balcony. The 800 Watt inverter is available for about 60 Euros. With a bit of mounting material we are talking about 750 Euros for 2 kWp and 2.24 kWh of battery. The 800 Watt inverter can be plugged into a standard socket, so there’s no installation cost. Assuming 1500 kWh of yield per year and 20 years, that is 30000 kWh for 750 Euros of investment, or 2.5 Cent per kWh including storage of part of the electricity for the night. 6000 cycles times 2.24 kWh works out as 13440 kWh, which should just about be what needs to be stored over a 20 year period.
Heiko, Good Morning.
Thank you for running those numbers – I had not realised that things had got quite so cheap so quickly. The numbers warm my heart!
Best wishes
Michael
Hi Michael,
Interesting point you make about the financial cost of battery stored energy – I thought I was a financialista but having previously failed to take that into account I cannot be.
But there is also a carbon cost of the battery manufacture – undeclared by the manufacturer- perhaps true carbonistas should estimate that before deciding whether to make the battery purchase.
The saving grace is sought by overnight rechargers, who are time-shifting night generation into the day and might think they are earning notional carbon credits from the reduction in emissions that that entails. A true carbonista knows that carbon credits are an idea from the Devil himself; should the overnight rechargers (I confess I am one) be cast into purgatory (EPC G rated)?
Best wishes
Nick
Nick, Good Morning,
Regarding embodied emissions in the battery, I have (of course) estimated this. In this 2022 article I showed my calculation which is summed up in the graphic linked below:
Of the top of my head, I estimated that all the embodied CO2 in my retrofit amounted to about 11 tonnes of CO2: I think the battery was about 1.5 tonnes if I recall correctly. But the emissions reductions per year from using the heat pump are very significant so that in 2026 I think we are well into the “net reduction” phase.
On the graph I show th effect of my CLIMEWORKS investment – which didn’t work out. In fact it has probably led to a net emissions (as all spending does) from executives flying the world to pre-sell their broken technology.
I have learned that nothing is easy or simple: a good lesson!
Best wishes
Michael
If you’re storing energy in a battery to later use for heating for less than 24 hours it’s worth giving some thought to the cost of just storing the heat.
Obviously not something you’d actually do but for comparison I once calculated that simply heating lead-acid batteries to 80°C (like a night storage heater) would store as much energy as charging them fully.
Equally obviously, you need a lot more thermal mass to store equivalent amounts of energy at the sort of output temperatures you’d get from a heat pump with multiplication by the COP taken into account. Phase-change materials (as used in the Sunamp, amongst other things) spring to mind in this context.
Never the less, I think it’s worth considering heating strategies based around storing as little energy as possible in batteries.
Ed, Good Morning.
Yes, that’s a very good point. When I am discussing individual projects this does come up, particularly for people who already have electrical storage heaters. For many people these work very well, and last for a very long time. And compared to gas the emissions benefit is less dramatic.
The amount of useful heat stored depends on the material and the storage and discharge temperatures. I wrote about this a long time ago in my article about Sand Batteries.
A single wall-mounted unit (volume 123 litres and costing about £1,000) can store up to 23 kWh (!!) of heat by heating special bricks to (I think) roughly 700 °C. That’s a serious amount of storage! And being stored at such a high temperature it is all useful heat. In contrast storage of heat in water is limited to much lower temperatures and for a storage temperature of 70 °C and a discharge temperature of 40 °C the stored thermal energy is only 3.5 kWh per 100 litres.
My Powerwall (volume of about 85 litres) stores (say) 10 kWh of electrical energy which runs a heat pump with a seasonal average COP of 3.8 so stores about 38 kWh of heating “equivalent”. ANd the battery has additional utility outside of the heating season.
So while there are cases where the battery/heat pump combination may not make sense, for most people who can, the battery/heat pump combination makes a lot of sense.
Best wishes
Michael
Thanks Michael
I agree with Geoffrey – I’ve seen much improved charged/discharge efficiencies over my first winter with a heat pump. The BMS overhead is shared over a much larger kWh throughput.
My battery is 9.6kWh (4x Uhome LFP 2.4 kWh).
Pre heat pump summer figures showed charge/discharge efficiency in the 75-80% range whereas my Dec, Jan, Feb figures were 87-89%. Adding the additional inversion loss of say 2.5% reduces the efficiency slightly to 85-87%.
Over the winter though I also need to add the 5% rectification loss, as I’m paying to charge the batteries, so further reduction to 81-83%.
This doesn’t matter in the summer as power is largely free.
Fiancialistas may well say invest the money, but they will be tempted to take annual long-haul holidays!
Cheers, Graham
Graham, Thank you for those numbers – which are very interesting But also indicate how complictaed the whole things is !
This is a sad truth that I am personally reluctant to acknowledge. I suspect that a large fraction of the financial savings that people make are spent on foreign holidays, overall making the climate situation worse!
Nothing is simple.
Best wishes
Michael
Hello Michael,
Thanks for another really interesting blog and set of data! We have a Solar edge battery nominally 9.7KWhr battery which in late spring and summer (usually) fills by the evening and lasts through until morning. In the winter we fill it from the grid overnight. It sits on the DC side of the system. You lose the back-up ability when the grid goes down that comes with the Powerwall – but reduce losses by only doing one inversion on leaving the battery. In our case it has a secondary advantage that we have about 8kW PV – an accident of panel availability at the time – but only a 6kW inverter. (Our DNO wouldn’t allow more unless we installed 3 phase and that was prohibitively expensive.) However, this means that in the middle of a sunny day even at this time of the year we can get >7kW going into the battery but of course only for a short period. It has become a game on such days to try to have the battery low by noon to maximise our generation. This has recently become much easier as I have acquired an EV. It has a 77KWhr battery which is massive by comparison with the SolarEdge and although the car doesn’t do that many miles a year, it’s usually possible to put in 5-15% and still keep the car battery below the reccommended 80%. So almost as good as a having a second dedicated PV battery. This has led me to think that the future is not so much in bigger home batteries, which I was previously coveting, but vehicle-to-home technology. I am hoping the electricity companies and grid operators get their act together on this soon.
Storage of energy on a national scale also seems to me to be an increasingly urgent problem. On days when its windy and sunny we already have a negative spot price for electricity but this rises dramatically after dark. A 0.5 GW solar farm – 500 acres – is planned very near where I live and there are two others of similar size within an hour’s drive which are more advanced. Although I’m a huge fan of PV I am not convinced this is the way to implement it – especially if you don’t have good storage.
Lastly, curious to know if you are really ‘off-grid’ in summer? We generate much more than we use, but the system is constantly using the grid to balance power flows so taking a few W to the grid and then returning some as the panels change generation level.
Interested in your thoughts.
Best Wishes,
Steph
Steph, Good Afternoon.
How interesting! Three things.
Thing 1: EVs
Yes! It’s sort of shocking just how big (and relatively cheap) EV batteries compared with domestic storage. Even ny 10 year old LEAF has 20 kWh for less than the cost of the Tesla Powerwall. As you say, when the various companies involved can get their head around how to deal with this it could contribute very significantly.
Thing 2: Grid-Scale Storage
Yes! But I am sanguine over the current state of affairs. To provide fully renewable grid power with high reliability the peak renewable output needs to be many times (probably 3 or 4 for the UK) peak grid need. I think building out generation first makes sense because storage needs over-supply! Over-building generation provides opportunities for storage to be built at the optimal places – and batteries are only getting better and cheaper.
Storage using technologies beyond batteries is seriously lagging. We really need lots of storage that can work for longer periods at low cost. The problem is that batteries are eating everybody’s lunch – they are already cheaper than many interesting alternatives – using high pressure gas or similar.
The current plan – which is to use gas as a back up for an every reducing fraction of generation seems to make sense to me for the next decade or so.
Thing 3: Off-Grid
I wrote about this in this article back in 2022
That article includes a link to this video which explains the effect.
Basically, the battery monitors the power requirements in the home and adjust it’s output to match demand. But it takes a few milliseconds to respond. (One cycle of 50 Hz AC is 20 ms). So whenever things are switched on or off there can be transient balancing currents drawn from the grid. The Powerwall reacts pretty quickly and this is a small effect. In my statistics I count a day “off-grid” as one in which the amount of energy drawn from the grid is 0.1 kWh or less. Some battery systems respond more slowly and can draw up to 1 kWh on an “off-grid” day.
Of course this graph is really the only thing that matters
We have a switch that would allow us to go fully off-grid with full electrical isolation but I have never done that! If we did then the kettle or equivalent would switch on a few milliseconds slower.
Best wishes
Michael
Thing 1 and Thing 3:
Thanks Michael, I hadn’t seen your 2022 blog and the video is indeed very clear. After a bit more research I think that your system has a further advantage of allowing you to use power from solar through your main inverter and from the battery’s inverter simultaneously whereas we are limited to 6kW from the only inverter. Is that correct? It hasn’t been a problem to date as in the summer the heat pump is off and usage never exceeded 6kW, whereas in winter we are using mainly grid energy anyway. However, with the EV I can see times where it might be more of an issue – or require some planning. We don’t have the option of going completely off-grid either. My limited understanding is that the powerwall can act as a voltage source and be grid-forming under these conditions. Not essential most of the time but a quite nice-to-have.
Thing 2:
Yes, I take your point about needing over supply to make storage viable but this is in part why I think it’s better, as a general rule, to put PV on buildings than in fields as it automatically generates the pressure on the owner for storage to smooth power availability. Hence our discussion about domestic batteries/EVs. There is not the same pressure on the large PV generators – and it’s very wasteful of farm land. We definitely need to build significantly more renewable generation of all types but both grid infrastructure and storage to deal with this seem to lack financial drivers. I hope the high pressure gas technology takes off as it’s both less materials intensive and seems it could eventually replace gas turbines as rapid response generation.
In the meantime I enjoy your graphs and finding new ways to look at our data very much. Thanks!
Best Wishes,
Steph
Stephanie, Good Morning.
I am not sure quite why the system we have works as well as it does. I had always ascribed it to the Tesla Powerwall being “top notch” – and responding faster to transients than less capable systems. But it may be also because of the way solar feeds into the system.
ANd yes, the Tesla does have “grid forming” ability which would allow us to off grid if we chose. But since I’m paying 50p per day for access to the grid, that seems a bit pointless.
Regarding the split between domestically-installed battery installations and large scale battery installations, I feel sort of neutral. In a brutal analysis, things happen whenever it makes financial sense for them to happen. This is independent of any “need” or societal requirement. Over time I think all useful niches will be filled up as the price of batteries falls and their performance increases.
Personally I think there is a role for mid-scale batteries and solar PV on a neighbourhood or village scale. But that may be just be a communitarian fantasy of mine. But one interesting application I have seen lately is their incorporation into induction cookers. A 2 kWh battery will allow a lot of cooking while not requiring any major re-wiring.
Anyway – I will keep working on the graphs!
Best wishes
Michael
Michael,
Looks like your battery is inside the house? If so, perhaps you might for fun calculate a modified cop for your heating system, taking into account that the rount trip losses are lost as heat to the property.
Matt
Matt, Good Afternoon.
The Powerwall is in the porch which is unheated. Why did I put it there? There was no clear guidance about the optimal place to put it.
It could have gone outside but the pictures seemed to show it typically placed in a garage (which I don’t have). It was not recommended for inside the house because it is not completely silent. The unit contains a small heating and cooling system and I reckoned that if it used that too much it would form an overhead that would be difficult to measure but which might impact its performance. So I chose the porch as a compromise – not getting as cold as outside in winter, but probably being a bit warmer than the outside in sunny weather.
I have thought it would be great if manufacturers could produce a combined heat pump/battery unit – it makes sense to me to have a single unit that would look after itself – but I am not sure anyone will buy that until batteries are a bit cheaper and the whole tariff issue becomes more stable.
Anyway: top marks for integrated thinking.
Michael