Energy Dashboard shows odd entry on graph after midnight

Configuration
1

I can see that just after midnight, I get a spike but I know this must be due to how it is resetting the counter. But I’m not sure how I can dertermine which sensor I’m meant to add to sort this out.

IMG_0203
IMG_02032160×1620 281 KB

Should I be using the sensors that are the ‘total’ or the daily one?

2

Please share the history graphs for the energy sensors you are using. Use the History panel, not the entity pop-up card. This ensures you will be looking at state data, not statistics (that can hide some glitches).

3

Hi.

Here are the picture requested.

Here are the sensors that it shows, I can’t find the one ‘Consumed Solar’.

Sensors
Sensors437×208 12.3 KB

I have the daily and Total for the Export. The energy dashboard is currently configured with the ‘Total’ sensor.

GivTCP Energy Export Energy Today kWh
GivTCP Energy Export Energy Today kWh602×248 19.4 KB

GivTCP Energy Export Energy Total kWh
GivTCP Energy Export Energy Total kWh602×246 21.2 KB

And I have the Electricity Import, again the is using the ‘Total’.

Smart Meter Electricity Energy Import
Smart Meter Electricity Energy Import602×266 24.1 KB

Currently my Battery Storage doesn’t have anything in the settings, but I had used

GivTCP Energy Battery Discharge Energy Today kWh
GivTCP Energy Battery Discharge Energy Today kWh602×225 23.8 KB

I can see that it would appear that the drop to 0, occurs a few seconds after midnight.

Should I be using the total sensor, rather than Daily?

4

I used to think so. I saw a lot of issues like this where the value did not reset exactly at midnight and thought it was the cause but user @karwosts told me in a post once (that I now can’t find) that the energy dashboard should be able to handle this. Maybe they have some advice.

5

Statistics collection (and therefore energy dashboard, which relies on statistics) is unaffected by sensors that reset to 0 when:

  • the sensor is state_class: total_increasing
    or
  • the sensor is state_class: total, and the change to zero is accompanied with an update of the last_reset attribute.

In either of those cases, the state changing to 0 will not affect the long term statistics accumulation.

6

Hi All,

Basically, I think I’ve fixed it.

This morning I had this showing.

IMG_0205
IMG_02052160×1620 307 KB

Last night I swapped most of my sensors to total, but missed the export. I changed that this morning and the spike after midnight has now gone.

IMG_0206
IMG_02062160×1620 287 KB

It helped using my PC so I could have a couple of pages open and compare the sensors I had set in the energy dashboard and the history page so I could view the sensor graphs better.

I’m not sure, if the energy help Webpage, that is linked from the energy dashboard config page, says anywhere that it’s best to use total sensors rather than the daily ones. I assume my daily ones where setting to zero just after midnight as they where being controlled by the GivTCP integration.

Regards