Energy dashboard shows wrong power consumption of individual device

Energy
1

Hi everyone.
I’m running into a strange issue with my energy dashboard.

Here is what I’m trying to accomplish.
I am using shelly devices to control lights, roller-shutters and the outdoor power-outlets. I do not need to monitor the consumption of every single device individually, but I’d rather monitor the 3 different types as a group (lights, shutters, outlets). Therefore I created helper-groups, to combine the energy consumption sensors of the single devices into a ‘total’ sensor.
As an example, this is the configuration of the power-outlets group-sensor:

image
image581×957 32.9 KB

The data of the sensor looks valid. So in this example, the outlets consumed ~0.4kWh today (between 06:00-08:00h).
image

However, if I add that sensor as an individual device to the energy dashboard, the usage is completely misrepresented. The graph shows a usage of 3.19kWh between 04:00-05:00h and no further consumption thereafter.
image

The roller-shutters have the same problem. Sensor shows a consumption of 0.03kWh until now for today, while the energy dashboard shows 13.78kWh.
For whatever reason the lights group is working, even though the configuration is the same on all sensor-groups.

I’m at a loss of where to start looking for the error… Any advise?

2

I suggest waiting til tomorrow…

I have similar behavior from the solar panels if the system hasn’t read the solar energy out for a while… then it puts the total since the last update in.

3

Thanks for the advise, but it’s been already 3 days since I set up the sensor groups and added them to the dashboard.

1 Like
4

Grouping energy sensors rarely works well, there is too many subtleties around resets for it to work reliably.

I see your history graph omits the 4-5am period. What happened then?

5

I just screenshoted that timeframe because nothing should have happened before 06:00h since those outlets only power the christmas-lights, which trigger at 06.00h.
But I’m not at home right now, so I can’t check the history…

6

I fixed it using a workaround…

I now use a sensor-group to sum up all the current power sensors and then use a integral-sensor to generate the statistics, which seems to work for now…