"Ghost"-Demand drop to 0, causing battery to lower discharge when demand is actually still the same about solarflow-control

[reinhard-brandstaedter]

What you are running into here is a issue with too quick updates to the SF limit. Notice the 1-2second interval where it triggers a new limit? I have made this observation as well and in combination with too quick new values from the smartmeter (seems also every 1s interval). This leads to false predictive values. Kind of a vicious circle that I'm investigating.

We have already built a "rate-limit" for the SF limit switching, to accommodate for it's slow reaction time (it takes 30-60s until a limit is actually applied on the SF side).
I think that little dip and jump in the orange phase actually triggered that behavior.

Note that we trying to address this by rather limiting via inverter where possible (faster) and will try the hub limit only in cases where not possible otherwise. So it's important to have a working limiting on the inverter.

[mavo]

As this seems also somewhat relates to my other issue or is very close and we touched upon this in the other issue, I'll close this one for now.