Accuracy factors for Utility Meter

[robb1 (Robert Brown)]

I have read many of the posts here around the Energy Dashboard and the Utility Meter but have not seen my issue raised. For me the Utility Meter is reporting data a factor of 10 too high.
I am confident of the power (W) figure from my current sensor as shown in this chart. It uses an update interval of 60 sec from platform: atm90e32

It shows that my energy consumption (area under the curve) should be under 600 Wh for any one hour running time. Instead the Utility meter gives a number that is too high by a factor of 10.

Here is my config.yaml

utility_meter:
  liv_aircond_energy:
    source: sensor.esp32_6channel_6c_ct2_watts_2
    cycle: hourly

and here is my customize.yaml

sensor.liv_aircond_energy:
  unit_of_measurement: Wh
  friendly_name: LivingRm AirCond Energy
  device_class: energy
  state_class: total_increasing  
  last_reset: '1970-01-01T00:00:00+00:00'

Is there a Utility Meter parameter for setting the integration interval to account for the steep spike when the air cond starts?

The other concern is that the Energy Dashboard reports kWh rather than Wh. Is this a bug that it ignores the entity config or have I missed something?

What is the best way around these issues? Should I use the Riemann Integration instead? Use a template sensor to apply a divide by 10 multiplier?

[tom_l]

The utility meter does not take power sensors. It takes energy sensors. It is not an integrator. All it does is change an ever increasing total to one that resets at your chosen cycle.

If you don’t have an energy sensor then yes you need to use the Riemann Sum helper to convert your power sensor to energy before using that in a utility meter. Be sure to use method: left in the Riemann Sum config to minimise approximation errors.

[robb1 (Robert Brown)]

Thanks for that clarification and prompt reply. On the basis of this I have made a suggested wording addition to the docs for the Utility Meter.
I will post back in a few hours with some new data.

[tom_l]

It is not specifically for energy sensors. It can be any sensor type with an always increasing total, like water or gas consumption.

[robb1 (Robert Brown)]

Thanks @tom_l It’s looking good in the Energy Dashboard now.