Powerwall Battery Degradation: Winter#3

Friends, as you are no doubt all aware: everything is getting worse. But the other day I noticed that one thing in particular was not getting worse quite as quickly as I had expected: the nominal capacity of my Tesla Powerwall 2 battery.

The Tesla Powerwall 2: Capacity

The Tesla Powerwall 2 is a home battery with a nominal capacity of 13.5 kWh when new. However the capacity of all rechargeable batteries declines over time, and it can be quite difficult to assess the actual capacity of the battery in use. I wrote about this at the end of the winter of 2022/23 where I estimated that since the previous winter the battery capacity had fallen by about 3.5%. It’s only the start of the winter of 2023/24 but so far it looks like capacity has fallen by less than a further 1%.

Click on image for a larger version. Measurements over the last three winters showing the amount of electricity discharged from the battery as it goes from 100% full to empty within a single day. See the text for more details.

What do we mean by Battery Capacity?

Assessing the capacity of a battery while it is in use in a home is tricky. And one reason for that is that it is hard to define even what one means by ‘capacity’. Really? Allow me to explain.

The nominal capacity of the Tesla Powerwall 2 is 13.5 kWh, but the battery can only be charged from AC power with an efficiency of about 95%. And it can only be discharged with an efficiency of about 95%.

• So to ‘fill’ the battery requires 14.2 kWh of AC electricity, 95% of which will be stored in the 13.5 kWh of battery cells, with the additional energy ending up as heat.
• Similarly, when the battery is discharged at 95% efficiency, only 12.8 kWh of useful AC electricity will be produced.

In practice, I can only assess the capacity of the battery in winter on days when the battery is re-charged to 100% capacity overnight and then we run from the battery until it is drained again. In this way I can obtain a figure for the total energy discharged from the battery.

Click on image for a larger version. Four screenshots from the Tesla ‘App’ showing the complete discharge of the battery during the day. The lower section of each screen shows the battery state of charge. The upper section of each screen shows charging from the grid, discharging to meet the household load, and a few brief episodes of solar recharging.

One small complication arises from small amounts of solar generation during these winter days. If solar generation exceeds the household demand then the battery will start to re-charge (at 95% efficiency) and this extra charge will then discharge at 95% efficiency.  I only consider days in which this solar re-charging is small, and simply subtract it from the nominal capacity, but the simplest measurements to interpret are on those on dull days when there is no solar charging.

Click on image for a larger version. Measurements over the last three winters showing the amount of electricity discharged from the battery as it goes from 100% full to empty within a single day. See the text for more details. Each blue dot represents a single full-to-empty measurement. The large black circles show yearly averages and trends are shows as dotted lines.

The graph above (the same as the one at the head of the article) shows all the results since 2021. Each blue dot represents a day of full discharge. And each blue dot with a pink outer circle is day of full discharge in which there was no solar re-charging.

• Based on the specification we might hope for a discharge of around 95% of the 13.5 kWh nominal capacity i.e. 12.8 kWh.
• During the winter of 2021/22, the average daily discharge was 13.1 kWh – rather better than we might have hoped for.
• During the winter of 2022/23, the average daily discharge was 12.7 kWh – a decline of 3.4% in one year.

This article is simply to mention the good news that:

• So far, during the winter of 2023/24, the average daily discharge has been  12.6 ± 0.2 kWh – a decline of less than 1% since last year. Which is pleasing.

What should I expect?

If we extrapolate two trend-lines, one based on the decline in Years 1 & 2, and the other based on the decline in Years 2 & 3 (so far), then one trend line indicates a 20% decline in capacity over 6 years, and the other suggests a 20% loss in capacity after 20 years. My guess is that the answer will be somewhere in between the two.

Click on image for a larger version. Measurements over the last three winters showing the amount of electricity discharged from the battery as it goes from 100% full to empty within a single day. See the text for more details. Each blue dot represents a single full-to-empty measurement. The large black circles show yearly averages and trends are shows as dotted lines.

When I bought the battery I guessed that the battery degradation might be similar to that seen in early Tesla cars (Model S and X). This data (now 5 years old) is plotted versus kilometres travelled below.

Click on image for a larger version. 2018 data from Electrek showing battery capacity (%) for Tesla Model S and X cars versus kilometres travel. The trend indicates about 10% loss of capacity after 250,000 kilometres. Both graphs show the same data but the right-hand side ‘zooms in’ on the data.

The data shows two interesting things.

• Firstly  it shows a relatively rapid decline in retained capacity in the first 20,000 km of life – perhaps the first year of typical use, followed by a lower rate of decline out to 250,000 km.
• Secondly, there is a lot of variability in the data. This data is from 2018 and so would feature battery packs installed in perhaps 2012. Some batteries don’t seem to perform well at all. The cells in my Powerwall are still of this design type (so-called 2170 cells), but I suspect manufacturing quality has improved substantially since 2012.

However the battery packs (i.e. collections of battery cells) in a car battery and a domestic battery are subject to quite different duty cycles. Car batteries are only rarely filled to 100% or drained to 0 % and avoiding these extremes inhibits many of the physical processes which degrade the battery. In contrast, domestic batteries are frequently filled to 100% and emptied to 0%: this probably happens about 100 times each winter.

So we might think that a domestic battery pack will have a much tougher time than a car battery pack. However, the temperature at which charging and discharging take place is also important, and the Powerwall includes a heating and cooling system and with the battery pack in a semi-sheltered location in the UK, I would guess the cells experience less extreme temperatures during charging and discharging than an EV battery pack.

Summary

So to summarise,  the battery degradation observed so far this winter is less than I expected – and that is a good thing. I’ll be sure to write an update at the end of the winter.