Solark 15K - Daily disconnect from grid (duke energy)

[jarred125]

After installing our SolArk 15K it became apparently we have a frequent issue here with our grid supply. At least once a day the SolArk will disconnect from the grid b/c the voltage and frequency drop below acceptable thresholds. It has happened three times today, typically the grid frequency dips from 59.9/60Hz to around 55Hz (voltage drops around 60V) when this happens. We have enough battery and the SolArk switches to this, but if we happen to be charging our car at the time we have to quickly stop the charge as the inverter won't supply the full charging amperage the car will demand.

Does anyone else see this with your grid connection, or am I just special and need to call Duke to ask them what is going on?

 

[Supervstech]

He'll, I would CERTAINLY report this to duke...
55! Wow, anything over 1 hz is crazy...

 

[Supervstech]

Where do you live?

 

[jarred125]

Where do you live?

Outside of Cincinnati, OH on the east side.

 

[jarred125]

Just put a service request in, it seems they take this pretty seriously

 

[#40Fan]

Have you confirmed these readings with a meter or are you getting the info from the Sol-Ark screen?

 

[jarred125]

Have you confirmed these readings with a meter or are you getting the info from the Sol-Ark screen?

The Solark only, it’s literally a split second and random.

 

[DIYrich]

I had a similar problem during the summer. Best I can tell, a momentary glitch in one of the 120v lines, and the systems disconnects from the grid. A minute or two later, it reconnects. It happened a few times a week (sometimes twice within a few minutes). 5 minute logs are worthless to diagnose the problem. I'm thinking of getting Solar Assistant to get second by second data of when the glitch occurs. Maybe a nearby power problem that sent enough noise through the line to trigger the disconnect.

I don't think it was a frequency issue (other than 0hz at 0v). Hard to image a "glitch" in the frequency unless the transformer is failing (somehow skipping cycles). It takes a lot to move the frequency of the grid.

 

[DIYrich]

, but if we happen to be charging our car at the time we have to quickly stop the charge as the inverter won't supply the full charging amperage the car will demand.

Definitely want to put that circuit on something that will automatically shed the load when switching to battery.

 

[Partimewages]

I had a similar problem during the summer. Best I can tell, a momentary glitch in one of the 120v lines, and the systems disconnects from the grid. A minute or two later, it reconnects. It happened a few times a week (sometimes twice within a few minutes). 5 minute logs are worthless to diagnose the problem. I'm thinking of getting Solar Assistant to get second by second data of when the glitch occurs. Maybe a nearby power problem that sent enough noise through the line to trigger the disconnect.

I don't think it was a frequency issue (other than 0hz at 0v). Hard to image a "glitch" in the frequency unless the transformer is failing (somehow skipping cycles). It takes a lot to move the frequency of the grid.

My computer UPS does this from time to time. We aren't losing power but maybe dropping below the voltage parameters built in.

 

[reardencode]

Which grid mode is your 15k in? IIRC the UL1741SB included increased requirements around inverter ride-through to prevent nuisance disconnects (and in the process protect the grid).

 

[DIYrich]

Which grid mode is your 15k in? IIRC the UL1741SB included increased requirements around inverter ride-through to prevent nuisance disconnects (and in the process protect the grid).

I'll have to check which grid mode I am in. Ride through is to protect the grid, not to prevent nuisance disconnects (UL could care less about that). In most of the instances, we were drawing from the grid, not supplying it. So, a disconnect when voltage drops, actually helps the grid, not hurt it. I'd rather not try to ride through a 15% voltage drop.

 

[reardencode]

You can write the priority in either direction. Preventing bunches of inverters from going offline during _high or low_ voltage/frequency events will both protect the grid and reduce nuisance disconnects.

 

[jarred125]

Definitely want to put that circuit on something that will automatically shed the load when switching to battery.

Resurrecting this thread a bit to post what solution I found for this issue.

Duke Energy didn't take it seriously and just abandoned the problem calling it a "solar issue".

I poll the 15K every 3 seconds with Home Assistant + HACS + Solark integration + Tesla wall connector integration. When it sees that the grid goes to 0W (above -1 and below 1) and the Tesla wall connector shows the L1 voltage above 50V (sits at 0 unless charging and then ramps to 120ish), it will stop the wall connector from charging.

 

[DIYrich]

Seems like a quality of power problem. Give Duke Energy a copy of the logs. Also send a copy to the Utility Regulators.

How many amps is your Charger? Is it low enough to put on the Smart Load?

 

[jarred125]

Seems like a quality of power problem. Give Duke Energy a copy of the logs. Also send a copy to the Utility Regulators.

How many amps is your Charger? Is it low enough to put on the Smart Load?

The wall charger is on a 60A circuit but is max 48A due to the Model 3 onboard charger.