Pace settings for fair distribution on CT002 breaks saturation detection

[jkrijtenburg]

What's the issue?

Pace settings for fair distribution on CT002 breaks saturation detection.
When 1 battery (e.g. the Venus E) is full, with default settings, the saturation of this consumer is reported to be 0.0%. The other, not so full , battery does have an increasing saturation, preventing the load previously taken on by the Venus E to be transferred to the Venus A.
For me, the result was that last night, after one battery was empty, the remaining battery only provided half of the power needed, resulting in energy import and unused battery contents.

Just now (middle of the day, 1 battery full), as a test, I overruled the pace settings, by setting the PACE_BASE_STEP to 0. After restart of Astrameter, the battery that wasn't full immediately took as much as it could of the excess power from my solar panels.

Your setup

• Home assistant app
• Venus E3 and Venus A
• Home Wizard

Configuration

Use default Pace settings (i.e. 30 and 100 for base and max)

Error messages or logs

No logs. (I wanted to create a log on debug level, but currently someone is using the oven so I can no longer recreate the scenario).

[tomquist]

Thanks for the report. I can't fully reproduce this in simulation but I'll take a look once you're able to provide logs.

[jkrijtenburg]

Ok thanks. Producing a log might take a few days, since I need one full or empty battery and consistent but significant 'demand'. What do you need, Astrameter logs on debug level, or something else?

[jkrijtenburg]

@tomquist this is from this afternoon.

Logs.txt
I had around 700 remaining excess solar power that should have been added to the Venus A. But Astrameter seemed to keep trying to balance on the Venus E with my manually selected pace base step of 15, but with the failure, the saturation of my venus A in MQTT insights was increased to 100%, blocking the battery to charge more then already going.
After this I removed the pace max step and set the pace base step to 0. the result was that the saturation of my venus E grew to 100% and that the Venus A took on the full load of excess solar power.
Of that success after disabling pace I have no logs. I hope this is enough. If not then I would have to recreate the scenario deliberately somewhere in the next few days.

[jkrijtenburg]

@tomquist
Today reproducing the issue succeeded partially. Registering of saturation on the wrong battery did not occur this time, however.
This time, there was no saturation detection at all, as long as pacing was enabled.
I tested 4 scenarios:

1. unchanged, with pacing disabled,
2. Pacing enabled with base step 15 and max step 200
3. Pacing on defaults (no config values)
4. back to pacing disabled.

I have screenshots of mqtt data of all 4 scenarios.
I have logging for the start of scenario 2, 3 and 4 (which is back to 1).
log.txt
screenshots of battery statistics.docx