How are your Powerwalls exporting at 12 kW? Do you have the newer update that allows for more than 5 kW continuous per Powerwall?
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
New Powerwall Advanced Options [Toggles for charging from and discharging to grid from powerwalls]
- Thread starter dailo
- Start date
How are your Powerwalls exporting at 12 kW? Do you have the newer update that allows for more than 5 kW continuous per Powerwall?
solar + PW = 12KW seems to be what's happening; as solar output drops the PWs increase but i don't think i've ever seen more than 10KW from the powerwalls.
what's funny about this is that 12KW seems to be some magic (hardcoded?) number. in theory i have 200A service and 12KW is 50A @ 240V so the system is far below the max (which is OK with me, but still seems strange). my inverter can produce 5KW peak so 15KW would be the theoretical max and 12KW is close enough.
what's funny about this is that 12KW seems to be some magic (hardcoded?) number. in theory i have 200A service and 12KW is 50A @ 240V so the system is far below the max (which is OK with me, but still seems strange). my inverter can produce 5KW peak so 15KW would be the theoretical max and 12KW is close enough.
My max export of solar + PW is 14.4kW which is 60A @ 240, which makes sense since there is only an 80A breaker for my system to the main panel.
I noticed the same behavior and see there is a new alert on the system SW_EOC which makes me think TOU mode now does some sort of calibrating. If I leave it in TOU mode I don’t see it but switching modes will cause it to happen sometimes.
duh, you are absolutely right about this. i need to go see what size breaker is between the PW and main panel. probably similar.
what is SW_EOC? is this something visible in the API?
i'm just hoping the thing becomes more responsive to EE/ES changes again because if it takes 20min to respond i'm going to blow way past the SOC i am targeting for the remainder of the day.
It’s something I see from the alerts API in pypowerwall. I assume it is something related to software, no clue what it stands for though. Just know whenever I see that alert my PW wont do anything when I toggle between EE and Solar.
Astro,
It may be better to let it dump more during peak. I am on PG&E's EV2-A rate with off peak 12m to 3pm, partial peak 3-4pm, peak from 4-9pm, and back to partial peak from 9pm-12m. Export Everything dumps down to reserve during peak, but tapers the rate so as run the house as well till the end of peak. Since the sun is down by 9, the house then imports partial peak from 9 on. My first thought was to avoid that import, just like before export everything was available. But last month, Wayne pointed out (here: New Powerwall Advanced Options [Toggles for charging from and discharging to grid from powerwalls] ) that it was better to export at peak and re-import at partial peak. The differential during summer EV2-A is roughly 56¢ vs 45¢ per kWh, i.e. 11¢ more credit per evening kWh if you export and re-import. Strange, I know.
If you still want to not re-import, you might try editing your pricing periods. I did this the other day, and PW covered the evening house use, and did a couple periods of export as wall, brining it down to my 20% reserve setting right at midnight. Most of the export was done early on, during actual peak times. So this strategy might work for you.
Another approach to preserve some charge for the evening hours might be to adjust the reserve setting.
I originally thought I'd need to use a script, but the built in logic is doing the right thing for me.
My single PW system exports at 3.5 kW. This is the rated max output of my inverters, and shows up as a setting in my site-info read out. I have just added more solar and am asking Tesla to increase this setting on mine. There is an input for this in the installation wizard, but changes there do not seem to propagate to the site-info or PW behavior.
My hunch is that this configuration setting is to prevent more power export than the utility approved in one's permission to operate. One of the things the utility does in the interconnection review is to check the local distribution capacity to ensure it can handle the export.
for me, with the fog around here, in the summer my solar production is often gone by 5-6pm. i am also on EV2-A.
maybe my PWs were actually misbehaving because i never saw a taper down. here is what happened when i let it run without switching modes:
the reserve was set to 30% on that day. you can see that the PW just blasted out everything for a sum total between 12KW and 10KW until the reserve was hit. that graph is what led me to start micromanaging the EE vs SO setting so that i could make it to the end of the day with no further consumption.
this is what happened on the next day. i changed modes but didn't leave enough in reserve and ended up importing around 10:15PM.
you can see some other weirdness there where the PW stopped charging in the morning, then around 2PM reached 100% but 20 mins later it changed its mind and decided the SOC was < 100% and charged some more. i have actually seen that a lot when in EE mode, but never saw it when SO was the only mode that existed. in fact the same thing happened on the 13th above. the divot in the SOC just occurred with no PW discharge.
what's interesting is that i'm now on day #2 of 5KW grid output. the first time i switched on EE, i saw a 5KW cap on day 1, and then every day after that i saw a 12KW cap. maybe that was actually a bug in the firmware. as mentioned i think i got a new version the other night. and yeah, as it turns out 5KW is the size of my solar inverter (non-tesla).
i agree that re-importing starting at 9PM is a better play especially on NEM1. i'm just going to leave the thing in Export Everything mode for a few days and see what it does, then adjust the reserve accordingly to just make it to 9PM.
I noticed that this week my export limit came back. It coincides roughly with getting approved for the VPP, so I wonder if setting the export limit is part of the preparation for the VPP. For some reason, the export limit that is now set is slightly different from the one that I had originally. It's 7.8 kW instead of 7.4 kW.
Without the export limit, the Powerwalls just seem to dump at maximum output (9.8 kW or 4.9 kW per Powerwall) during peak. With the export limit, they just discharge enough to stay at that limit for as long during the peak as they can.
You can see the export limit by checking https://<powerwall ip>/api/site_info - the value is labelled min_site_meter_power_kW
Without the export limit, the Powerwalls just seem to dump at maximum output (9.8 kW or 4.9 kW per Powerwall) during peak. With the export limit, they just discharge enough to stay at that limit for as long during the peak as they can.
You can see the export limit by checking https://<powerwall ip>/api/site_info - the value is labelled min_site_meter_power_kW
Last edited:
Mine was changed also. min_site_meter_power_kW was -3.5, now is -8.04822021484375. Precision down to the femto-Watt, strangely.
I also recently was enabled for VPP, but also requested an increase because of newly added solar panels. Tesla has also been playing with the availability of the new features on folk's PWs, so there is no telling why these changes are happening, what with Tesla being so non-forthcoming on what/why/when them make changes.
I have not yet seen how this affects PW behavior, but it has exported to the grid more than 3.5 a couple times in the last couple days. So I think that min_site parameter is what it uses to pick the export rate. But it appears to include solar export in it's calculation, adjusting PW discharge rate to make the total export max match the setting.
SW
Last edited:
Strange. Mine also sometimes slows the charge rate as it approaches 100%. I figure it may ba a thermal consideration. Also my SOC drifts some when there is no flow, perhaps also due to temperature changes. That top-off yours did looks similar to that.
But, those graphs are very nice. I was considering sending a screen shot of the app, but it may take 4 images to tell the story, while yours shows all four flows on one image. What do you use to create those?
SW
aesculus
Still Trying to Figure This All Out
Mine is still set to 10 as my VPP application was rejected. Supposedly the numbers I entered didn't jive so I am back in the pending pool.
Anyway mine has exceeded 10 kW on a number of occasions when (solar - house load) was added to the 10 kW (i.e. 14.5 kW).
Mine can slow too. The most EE max periods I have seen in peak period was 2 full and 2 that diminished. Not sure what the first diminished set that followed the first max rate discharge was or how it was determined. And then there was another max discharge event followed by a tapering discharge that got the PWs to my reserve at the time the peak period ended. I suspect if I had set a lower reserve I might have seen three of these sets.
Astro,
Here is what mine did when I adjust the peak period to run from 4 to midnight.
These are the PW and grid flows. It was set to Export Everything, Time Based.
PW charged in the AM. Shown in the lower frames is that it then let solar run the house and export the excess. Then when peak kicked in at 4 pm, PW took over the house while solar continued (unshorn here) to export. around 6, PW decided to dump it's estimated excess to the grid, after which it dropped back the discharge rate to cover the house till very nearly midnight. Two little details are of interest: In the grid flow pic, you can see that PW dumped a bit more twice as midnite approached, but then just before midnight it hit my 20% reserve, so the house imported for the last few minutes, so little that it is not visible in the house pic.
So it appeared to do what you were trying to accomplish, but I did not need to intervene and change setting to make it do this.
Mine did act quite differently for the first few days, so perhaps it was re-learning my typical solar and house use so it could make estimates and discharge accordingly.
The house and solar complete the story. (Your combined chart is hipper!)
I expect yours will settle down and with the rates set normally, work as you wish. By the way, even though I was on NEM1, I set the sell price 3¢ lower to account for NEM2 Non-Bypassable-Charges. Sadly, I had to switch to NEM2 to register my new solar addition so that PG&E's estimated solar would not limit my export. EE would put me over, so I would lose the benefit..
I've been playing with other settings since. Today it is in Self Consumption, but the cloudy sky is not going to get back to 100%, so I'm adding Grid Charging, but had to up the reserve to make it charge. This, by the way, it is doing at 1.7kW, who knows why.
Fun with electrons!
SW
I forgot to mention that I also had weird pauses in charging for a couple of weeks, but they stopped about 8 days ago. Here's an example from two weeks ago. Note that at the time I had no export limit:
(this graph produced using graphana/influxdb/telegraf, a modified version of this: GitHub - mihailescu2m/powerwall_monitor: Monitoring for the Tesla powerwall)
it's just PVOutput - all the battery data goes into the extended parameters. it's not the greatest since PVOutput doesn't really understand storage at all; the extended parameters are all just advisory data. this is why i had to create another system that just watches the meter (via an Eagle) and spoofs solar generation when energy is flowing to the grid.
there's a python uploader that can talk to the local tesla API and push to PVOutput, which would give you the graphs shown in my screenshots.
Thank you. That looks interesting...
This also looks interesting.
For now I think I may just stick with the provided data logging: Tesla app, Enphase w/ consumption monitoring, TED5000 on the Model Y Wall connector circuit (which is outside of my PW's purview.
We've had clouds all day here in Oakland, so I'm guessing your EE trials have been hampered a bit. I've been playing with Self Powered to see if it would act like an Import Nothing mode. Sadly, the overcast meant I had to grid charge to get to 100% SOC, so the pretty no-import screen shot (which I want for an NEM3 conversation) will have to wait for better weather.
Grid charging was a bit strange. First I had to set the reserve to 100% to get it to grid charge, and then, when set to self-powered, it charged at 1.7 kW. When I set it to Time-Based, it started grid charging at 5 kW. I was expecting 3.3 kW, which is what the PW spec sheet says it does when grid charging, and what it has done when Storm Watch was active. So both 1.7 and 5 kW strike me as odd. Yet more mysteries from the Mind of PW.
(An off topic quip: Do we need an artificial psychologist to analyze the behavior of an artificial intelligence? Not that PW is an AI, but FSD beta behavior in the cars sure raises some questions...)
SW
indeed the weather has sucked these last couple of days and we are all set for another foggy 4th of july.
so yeah, haven't learned much.
i am reluctant to use grid charging (outside of storm watch) because of the ITC. but i did try it one cloudy day and saw similar behavior as you have described - charging at around 6KW in time-based mode.
i guess we are just seeing either bugs in the logic or the heuristics doing funny things...
so yeah, haven't learned much.
i am reluctant to use grid charging (outside of storm watch) because of the ITC. but i did try it one cloudy day and saw similar behavior as you have described - charging at around 6KW in time-based mode.
i guess we are just seeing either bugs in the logic or the heuristics doing funny things...
