batpred icon indicating copy to clipboard operation
batpred copied to clipboard

Predbat charging battery via "HoldChrg" during peak times

Open stevedundee2 opened this issue 2 months ago • 6 comments

Describe the bug Predbat has charged my battery for short periods twice this evening, during the daytime peak rate.

I think could be because I made bread last evening. Does predbat think I'll do the same this evening and is making sure my battery won't be drained? If so, I think it would be better to just wait and see instead of being proactive.

Expected behaviour Predbat should not choose to charge at peak rates.

Predbat version 1.2.9 / 8.25.3

Environment details

  • Givenergy
  • Standard HAOS
  • Intelligent Octopus Go
  • Predbat is configured to consider the past 7 days but for yesterday to have twice the waiting of the other days, i.e. 2,1,1,1,1,1,1

Screenshots

Log file Clicking Predbat debug yaml file doesn't seem to do anything

stevedundee2 avatar Oct 03 '25 18:10 stevedundee2

unlikely be able to do anything with this without the logfile or debug file, plan, etc, but if you have yesterday's house load set to twice be weighted at twice the other days then that's probably what's a contributory factor

gcoan avatar Oct 04 '25 09:10 gcoan

Yes sorry I know it was a bit sparse! I was rushing to capture it while it was still happening but it stopped before I could get the plan. Would my apps.yaml help? I have it configured to prioritise self-sufficiency over money.

Also this (not debug) log file?

But I perfectly understand if this is not a good use of your time - perhaps if someone else reports the same symptom?

apps.yaml

predbat.log

stevedundee2 avatar Oct 04 '25 10:10 stevedundee2

I'm not sure if mine is exactly the same issue or not, but also seen charges happening at peak rate the last few days (on IOG). This morning it was charging on hold charge, and now it plans to have a charge (though battery already at 100%). Debug/log files attached if useful (but only from now showing the planned charge, not what happened earlier).

predbat_debug.yaml

predbat.log

Edit - it seems to have recalculated and isn't planning a charge any more - not sure if my debug got the before or after state, sorry.

ablack89 avatar Oct 06 '25 11:10 ablack89

Ok, back to having a planned charge at 1pm (at peak) rate now.

predbat_debug (1).yaml

predbat (1).log

ablack89 avatar Oct 06 '25 11:10 ablack89

Similar to the above, I've noticed that HOLD CHARGE at % will often charge the battery way past the expected maximum.

The HOLD CHARGE value sent to the AIO was 55%, however the AIO was already at 58% when this command was sent.

Image

Not sure if this causes the AIO to start charging indefinitely.

I'm expecting more of this situation where my battery won't likely make it from 05:30 to 00:30, and a HOLD CHARGE command would be issued, which then causes massive imports that Predbat doesn't seem to recognise as imports. I believe the only reason this would change from HOLD CHARGE to DEMAND is when an EXPORT is determined later in the plan due to the inoppurtune IMPORT at the HOLD CHARGE command.

lewiswatt avatar Oct 19 '25 12:10 lewiswatt

Have seen similar behaviour tonight - predbat was showing 'hold charge' for less than the current charge % from 2330-000, but was setting a charge slot in the inverter at this time (shown in app and portal) for 100%, leading to the battery force charging.

On agile and expecting negative prices in a few hours (and force charge planned then as expected), plenty of battery to last until then so definitely not expected behaviour to charge earlier.

tim-b2 avatar Oct 25 '25 23:10 tim-b2

Believe to be fixed in the current release, at least for the case ablack89 posted

springfall2008 avatar Nov 22 '25 18:11 springfall2008

Thanks!

I've not noticed it happening for a while now. A couple of days it has let the battery go down to its minimum in the evening. The house then just used what grid energy it needed and no more, as I would expect it to when the current price is peak rate.

stevedundee2 avatar Nov 23 '25 09:11 stevedundee2

Please post a debug yaml if you can catch it in the act so to speak

springfall2008 avatar Nov 23 '25 10:11 springfall2008