Yesterday around noon my solar production dropped to exactly zero for exactly seven hours, then came back as if nothing has happened. (It’s not a Sense problem, because the old Solarcity monitor that reports directly from the inverter showed exactly the same thing).
The inverter (and entire system) is from 2008 (14 years!), and I wonder if it might be near EOL. Looks again good today, but will keep an eye on it. Has anyone seen anything similar?
Never happened in the past (Sense installed in Nov 2018). Nobody messed with the panel according to my security system.
For illustration:
Sense data on the left, Tesla (originally Solarcity) on the right. Both show the same gap.
(EDIT: not sure why the posts also shows the tag “wishlist”. I did not select that and even if I edit and remove that tag, it keeps showing up…)
Wow ! If both Tesla app and Sense show the same thing, plus Sense keeps on reporting Total Usage, it’s most certainly an inverter outage of some kind - can blame it on network or power outages affecting reporting. Maybe the inverter switched off due to some internal protection mode.
Next time that happens, can you take a look at the voltage that Sense measures from the grid and take a screenshot and post it here?
It might be that someone in your neighborhood got solar installed and the 2 of you are now pushing the voltage up to a level where the inverter shuts down. When demand goes up, PV generation goes down aka grid voltage goes back to normal, inverter switches back on.
The following link is from Australia but the concept is the same everywhere:
I have seen this happen to someone who lived near a solar field in California.
He opened a case with the local utility. They exchanged or did a different setup of a transformer and he never had the problem again.
Another possibility is that modern inverters can be configured to back down generating power with change of the grid frequency.
What brand/model of inverter do you currently have?
I have homeassistant connected to sense and thanks to @kevin1 I now have this graphic of the grid voltage during solar production:
@altenbach it looks like you have a great use-case for common.sense.com
I assume with at least one additional neighboring Sense on the same transformer that the solar gain issue (thanks @dannyterhaar) could be confirmed.
If only flywheels were easier.
In the absence of an empty EV battery, dump the energy into water heating or somesuch?
Zooming out from the individual over-production issue I arrive at an image of a “self-replicating” residential solar panel micro-factory going high gear at noon. And the conflicting image of a very complex spider-like gridbot that connects and reconnects everything, spinning conductors in a highly organic manner.
On a more real world level I suppose the question is whether this issue is solvable with existing hardware (i.e. a software fix) and if not (or when not) how to intelligently pre-empt the hardware modifications to avoid failure. Possibly a great Sense Lab within the common.sense.com framework.
Thanks for all the input and suggestions. Some interesting points there!
However, I think I fixed the problem, though. (But will continue to watch it, of course).
I spend some time to check all the connections and it appeared that one of the solar phases was not fully tightened at the circuit breaker (was able to give it another quarter+ turn!).
After tightening that screw, system has been working flawlessly!