Powerwall % Degradation

[holeydonut]

Simply dividing yearly powerwall numbers discharged/charged from the tesla app:
2021 - 90%
2022 - 89.5%
2023 - 88.9%

RIP my Powerwalls…

(Yes I know this is too small of a kWh sample to mean anything)… just sad to see sub 80% heh

 

[offandonagain]

Simply dividing yearly powerwall numbers discharged/charged from the tesla app:
2021 - 90%
2022 - 89.5%
2023 - 88.9%

I'm pretty sure this is measuring Powerwall round-trip efficiency (which is expected to be ~90%), not degradation. I get similar values, and my batteries have not degraded based on nominal_full_pack_energy values.

 

[wwu123]

As Tesla's warranty states 'aggregate' (and I assume Tesla would interpret in which way works best for them), I read it as incoming + outgoing. Unless someone has received something specific from Tesla on the contrary, it's the way I would expect it to go. Once again, this would only affect those using PWs outside of the first category.

Yes, I meant 17mWh, thank you for catching that.

The word "aggregate" makes sense in your interpretation, it's the word "throughput" that doesn't make sense to me. But I was thinking of the battery cells, if I think of the inverter, then it's true that the energy must pass through the inverter twice, once on input, then again on output - so I think you're right, and it's in that context that Tesla uses "aggregate throughput"....

 

[tomuo]

I'm pretty sure this is measuring Powerwall round-trip efficiency (which is expected to be ~90%), not degradation. I get similar values, and my batteries have not degraded based on nominal_full_pack_energy values.

I'm implying that the small fall-off each year is indicating the degradation, since the round trip efficiency shows it took more kWh to reach 100% in the later years.
We don't know how the powerwall measures its own nominal full pack energy, but it must be based on the same base data, how much power did it take to charge to 100%.

 

[veganjoe]

Are any of you doing energy arbitrage, and can you provide some idea of what you think your ROI will be, along with your Peak and Off-Peak rates? I'm trying to decide whether I should get a system to try this out, but since my home isn't a candidate for solar, I'd have to be battery only and use arbitrage to try to make it pay for itself / pay for my other energy use... Thanks!

 

[buckets0fun]

Are any of you doing energy arbitrage, and can you provide some idea of what you think your ROI will be, along with your Peak and Off-Peak rates? I'm trying to decide whether I should get a system to try this out, but since my home isn't a candidate for solar, I'd have to be battery only and use arbitrage to try to make it pay for itself / pay for my other energy use... Thanks!

That’d be a hefty order.

You’d want to pull your specifics to start calculating for that one.

Your rates
Your usage during your peak/partial peaks
Cost of powerwall install -rebate
How many tens of thousands of kWh you’d have to use + push back to wipe the cost of the install out.

 

[eml2]

Here are the numbers of my two 6 years old PW2:

nominal_full_pack_energy: 13421,
energy_charged: 15183020,
energy_discharged: 13327580,

nominal_full_pack_energy: 13542,
energy_charged: 10655678,
energy_discharged: 9379650,

The 2nd PW is a few months younger because the original one has to be replaced due to failed cooler pump.

The PW are located in the basement.

 

[gsdubs]

The word "aggregate" makes sense in your interpretation, it's the word "throughput" that doesn't make sense to me. But I was thinking of the battery cells, if I think of the inverter, then it's true that the energy must pass through the inverter twice, once on input, then again on output - so I think you're right, and it's in that context that Tesla uses "aggregate throughput"....

My opinion: the 37.8MWh of aggregate throughput warranty limitation for PW refers to output only, not the "round trip" input + output.

The 37.8MWh value is obtained by assuming a user fully discharges and charges a PW once a day, every day, for 10 years. For reference, see this explainer on how manufacturers estimate aggregate throughput over a battery's life and fill in the values for PW:

13.5 Nominal capacity in kWh (Cn)
70% Capacity retention at end of 10-year warranty period (nu)
90% Round trip efficiency
100% Depth Of Discharge (DOD)
3650 Number of Cycles at end of service life, assuming 10-year warranty period
37695 total throughput in kWh over 10-year warranty period (C) [Add in 2-3 days for Leap Years and you get to 37.7MWh, pretty much dead on the warranty limitation]

Tesla is likely covering for the scenario where someone cycles their PW's from 100% to 0% more than 1x/day, although I would see no economic benefit of someone doing so.

Bottom line: feel free to ride the Powerwall every day even with grid charge, grid export, VPP, load shift....as long as it goes from 100% to 0% and back to 100% only once per day, you are warranted to get 10 years out of it.

For those of us who aim for a slightly longer life, my hunch is that using the PW "gently" both reduces the number of full cycles, and the DOD, which should correlate to longer life. My original PW2 was installed on 4/5/21 but it died on 11/16/22. Replacement was installed on 11/30/22 and I noted an initial full pack energy of 14.811kWh. Today I have a reading of 14.449kWh full pack energy, which is 97.6% of original, but still 107% of the 13.5kWh rating. Looks like I lose about 280 Wh/year in capacity, or 1.9% per year relative to initial full pack energy.

 

[offandonagain]

My opinion: the 37.8MWh of aggregate throughput warranty limitation for PW refers to output only, not the "round trip" input + output.

Agreed. Two other data points in support of this:
1) The warranty doc in the footnote says "measured at the battery AC output".
2) I called Powerwall support and they confirmed only discharge is counted.

 

[gsdubs]

Agreed. Two other data points in support of this:
1) The warranty doc in the footnote says "measured at the battery AC output".
2) I called Powerwall support and they confirmed only discharge is counted.

If only discharge is counted, shouldn't lifetime "cycles" be energy_discharged / 13500? This automatically includes the PW's actual round trip efficiency.

I see other folks using an average of charged and discharged, but not sure why that is more meaningful

Not a huge difference but it gets into the modeling of when a system hits 37.8MWh throughput limit for warranty

 

[offandonagain]

If only discharge is counted, shouldn't lifetime "cycles" be energy_discharged / 13500? This automatically includes the PW's actual round trip efficiency.

I see other folks using an average of charged and discharged, but not sure why that is more meaningful

Not a huge difference but it gets into the modeling of when a system hits 37.8MWh throughput limit for warranty

I think the number of cycles is higher than that, because as the battery degrades the discharges get shallower. Based on your formula above, it's energy_discharged / (13500 * average_degradation_rate[0.85] * round_trip_efficiency[0.9])

Regardless, it's simpler to just use the lifetime discharged energy: there's no ambiguity (which is important in the context of a warranty) and the Powerwalls track this metric internally.

 

[Matt-FL]

If you are using the Powerwalls as solar consumption/backup only, the throughput limit does not apply. It is only for those using them outside of those parameters, perhaps VPP or other.

 

[gsdubs]

I think the number of cycles is higher than that, because as the battery degrades the discharges get shallower. Based on your formula above, it's energy_discharged / (13500 * average_degradation_rate[0.85] * round_trip_efficiency[0.9])

Regardless, it's simpler to just use the lifetime discharged energy: there's no ambiguity (which is important in the context of a warranty) and the Powerwalls track this metric internally.

I agree - the ground truth is lifetime discharged energy

I was getting fancy and trying to use the formula to predict when my system will hit the 37MWh limit based on current and historical performance...need to think about that some more

 

[gsdubs]

If you are using the Powerwalls as solar consumption/backup only, the throughput limit does not apply. It is only for those using them outside of those parameters, perhaps VPP or other.

I'm using it for VPP, occasional grid charging, and I grid-exported a couple times....so I'm planning on Tesla invoking the 37.8MWh cap on me...

 

[CWhite]

I have 2 Powerwalls, one is only a year old the other was installed in 2017. I only use it for backup so it rarely discharges except for the 10 times a year the utility buys power from us. They are mounted outside. The Netzero app showed no degradation, pleasant surprise if correct.

 

[Vostok]

Are any of you doing energy arbitrage, and can you provide some idea of what you think your ROI will be, along with your Peak and Off-Peak rates? I'm trying to decide whether I should get a system to try this out, but since my home isn't a candidate for solar, I'd have to be battery only and use arbitrage to try to make it pay for itself / pay for my other energy use... Thanks!

Well I’m one of those data tragics who has logged their PW2 data since installation - I now have nearly half a million rows of data.

So I have done that calculation exactly - what has been the value of the energy discharged by the PW2, based on how much and when the discharging occurred. In my case there is also foregone FIT (when solar is charging the battery).

FITs are rapidly trending towards zero in Australia though, as rooftop solar penetration approaches 50%, there is simply no-one to sell the energy to. So loss of FIT is rapidly becoming irrelevant in that calculation.

My grid rates (AUD) are currently 62.9c/kWh peak, 33.2c/kWh shoulder, and 19.2c/kW offpeak. So there is benefit to be had in grid charging at offpeak rates to use that power at peak times the following day if solar production is predicted to be poor and won’t fill the battery.

Anyways, based on that, my PW2 on its own has a payback time of 12.3 years. Not economic based on typical ROI expectations. And not a surprise to me, I bought it because I wanted the technology, not because it was going to be a rip-roaring investment.

However my solar array on its own has a payback period of 5.6 years, and blended with the PW2 the overall payback is 7.7 years. Which is acceptable.