Powerwall efficiency 86%....

[sorka]

Since October, I've charged 3270 kWh into my 3 poweralls and used 2811 kWh. This is about 86% efficient. Is this what others are seeing? The fans have rarely been on as I haven't used high current loads yet (that's coming with AC this summer).

I suspect my efficiency will drop quite a bit this summer as the powerwalls will expend a lot more energy to stay cool.

 

[Will792]

Since October, I've charged 3270 kWh into my 3 poweralls and used 2811 kWh. This is about 86% efficient. Is this what others are seeing? The fans have rarely been on as I haven't used high current loads yet (that's coming with AC this summer).

I suspect my efficiency will drop quite a bit this summer as the powerwalls will expend a lot more energy to stay cool.

Where do you get these aggregates?

 

[holeydonut]

I'm pretty close to you in terms of ratio. According to the Tesla app in 2021, I've used 1,653 kWh from the PW's, and I've sent in 1,901 kWh. Assuming I have a residual of 19.8 kWh sitting in the Powerwalls right now, that means 88% efficiency.

Edit: How do you know when the fans come on? Do you mean the wub wub wub sound? I hear that all the time when these are charging and discharging.

 

[sorka]

Where do you get these aggregates?

These are pulled from rest API day by day:

https://owner-api.teslamotors.com/api/1/energy_sites/<site id>/calendar_history?kind=power&end_date=<date/time>

via a script that feeds my dashboard.

 

[sorka]

I'm pretty close to you in terms of ratio. According to the Tesla app in 2021, I've used 1,653 kWh from the PW's, and I've sent in 1,901 kWh. Assuming I have a residual of 19.8 kWh sitting in the Powerwalls right now, that means 88% efficiency.

I'm getting 86.95% with your numbers. Where is 88 coming from? So yea, still really close.

 

[wjgjr]

The problem with calculating PW efficiency is that it combines at least two factors - steady-state usage (basic operations+ heating/cooling) and round-trip efficiency losses.

I keep my battery in backup-only mode, which means that I can go months without any exports. What I have found is that, for me, the daily energy use for my two PWs during that time seems to be between 0.55 and 0.6 kWh (in a basement.) If I were to look at my all-time numbers (excluding the initial charge) I would get an overall efficiency of about 70% (mostly because I did run off-grid before PTO or it would be even lower.) But, if I adjust for that estimate of the steady-state use, I am a bit over 90%.

So, my point is that one has to make sure the comparisons take into account the location of the PWs as well as the operating mode.

 

[holeydonut]

I'm getting 86.95% with your numbers. Where is 88 coming from? So yea, still really close.

I did (1,653 + 20) / 1,901 ... but maybe that's the wrong way to do the calc?

 

[sorka]

I'm pretty close to you in terms of ratio. According to the Tesla app in 2021, I've used 1,653 kWh from the PW's, and I've sent in 1,901 kWh. Assuming I have a residual of 19.8 kWh sitting in the Powerwalls right now, that means 88% efficiency.

Edit: How do you know when the fans come on? Do you mean the wub wub wub sound? I hear that all the time when these are charging and discharging.

Just realized you said you used the app. I hadn't thought about trying that. So when I use the app for any period greater than 1 day, it only shows "to powerwall", not from powerwall.

Also interesting is that the lifetime from powerwall in the app is showing 2854 but rather than than the 2811 I got from the API but I've checked and am not missing any 5 minute data chunks all the way back to the day the system was activated. Strange.

 

[sorka]

I did (1,653 + 20) / 1,901 ... but maybe that's the wrong way to do the calc?

I was using (1901/1653)*100

 

[sorka]

The problem with calculating PW efficiency is that it combines at least two factors - steady-state usage (basic operations+ heating/cooling) and round-trip efficiency losses.

I keep my battery in backup-only mode, which means that I can go months without any exports. What I have found is that, for me, the daily energy use for my two PWs during that time seems to be between 0.55 and 0.6 kWh (in a basement.) If I were to look at my all-time numbers (excluding the initial charge) I would get an overall efficiency of about 70% (mostly because I did run off-grid before PTO or it would be even lower.) But, if I adjust for that estimate of the steady-state use, I am a bit over 90%.

So, my point is that one has to make sure the comparisons take into account the location of the PWs as well as the operating mode.

Yea, if you're just sitting in backup mode, discharge both chemical and just running the onboard electronics will be a much higher percentage of your usage. I'm offsetting 100% of all my peak usage with the powerwalls so basically whatever goes in comes out within the same day cycle.

 

[Redhill_qik]

Each Powerwall has a battery leakage loss of about 0.175-0.250kWh per day just sitting there, which isn't consider in the 90% efficiency rating.

 

[sorka]

Just realized you said you used the app. I hadn't thought about trying that. So when I use the app for any period greater than 1 day, it only shows "to powerwall", not from powerwall.

Also interesting is that the lifetime from powerwall in the app is showing 2854 but rather than than the 2811 I got from the API but I've checked and am not missing any 5 minute data chunks all the way back to the day the system was activated. Strange.

The only hour I'm missing is the jump forward in daylight savings on 3/14 but that's not really missing.

 

[gpez]

Each Powerwall has a battery leakage loss of about 0.175-0.250kWh per day just sitting there, which isn't consider in the 90% efficiency rating.

This is exactly correct. I'm still not sure why the battery leakage isn't configurable to "top up" from the grid or solar, seems like powering the components of the Powerwall should just use grid power...

In the app I see 0.3kwh/day go from solar to the Powerwall even though mine is and has always been configured in backup mode. That's close to 1kwh/day for 3x Powerwalls or about 210kwh for October - April. 2811 / 3060 = 91.86%

 

[holeydonut]

I was using (1901/1653)*100

Oh ok, I was adding the residual ~20 kWh sitting in my batteries right now. Presumably that's part of the generation that hasn't been sent back to the house yet.

 

[sorka]

This is exactly correct. I'm still not sure why the battery leakage isn't configurable to "top up" from the grid or solar, seems like powering the components of the Powerwall should just use grid power...

In the app I see 0.3kwh/day go from solar to the Powerwall even though mine is and has always been configured in backup mode. That's close to 1kwh/day for 3x Powerwalls or about 210kwh for October - April. 2811 / 3060 = 91.86%

Note getting the same thing:

From October 19th to yesterday is 192 days at 0.3*192 = 57.6 kwh

2811 / (3270 - 57.6) * 100 = 87.5%.

EDIT: Ahh. 3 powerwalls

 

[sorka]

Oh ok, I was adding the residual ~20 kWh sitting in my batteries right now. Presumably that's part of the generation that hasn't been sent back to the house yet.

That will throw it off a little. My powerwalls came 30% charged and at midnight which is what the rest API pull goes to each day, I'm at about 50% so it's a really insignificant. I didn't raise this issue until I had at least 6 months of data.

 

[sorka]

Note getting the same thing:

From October 19th to yesterday is 192 days at 0.3*192 = 57.6 kwh

2811 / (3270 - 57.6) * 100 = 87.5%.

Whoops. Just realized 3 powerwalls!

2811 / (3270 - 57.6*3) * 100 = your numbers

 

[holeydonut]

Hmmm ok so for someone who uses their Powerwalls in self-powered (or related) mode, it looks like the round trip efficiency is 86 to 90%...

I guess this gives some benefit that the power company gives "100%" efficient round trip energy for what passes through their meter. Although I still think the PG&E EV2-A time of use differential is far worse than a 12% round trip loss in the Powerwalls differential.

 

[sorka]

Hmmm ok so for someone who uses their Powerwalls in self-powered (or related) mode, it looks like the round trip efficiency is 86 to 88%... not the 90% Tesla advertises.

I guess this gives some benefit that the power company gives "100%" efficient round trip energy for what passes through their meter. Although I still think the PG&E EV2-A time of use differential is far worse than a 12% round trip loss in the Powerwalls differential.

90% if you don't include onboard operational costs. But yes, the reality is if you're going to use powerwalls at all, it will be anywhere from 87% to possibly much much less if you're using it for backup only.

 

[wjgjr]

90% if you don't include onboard operational costs. But yes, the reality is if you're going to use powerwalls at all, it will be anywhere from 87% to possibly much much less if you're using it for backup only.

That is why I would describe the technical efficiency as excluding the normal operational overhead. I wouldn't describe a PW in backup-only mode as less efficient, just less used. And, like so many other connected devices these days, there is a power cost just for having them (which, it was pointed out, also includes the battery loss.) It also avoids potentially interpreting the data as suggesting that days with higher PW exports are more efficient when it may just be that they are better masking the fixed loss.

Certainly from a monetary/value standpoint it makes sense to take everything into account. And, though not huge, the energy costs do add up - roughly $15/year/PW out here just to keep it in backup only, and that is with lower-priced (~$0.14/kWh) energy out here.