Allow variable ON and standby power values for more accurate long-duration energy estimation in the fixed strategy.

[gieljnssns]

Hi, first of all thanks for the great work on Powercalc — it's an incredibly useful integration.

I have a question (or perhaps a feature request): Would it be possible to introduce the ability to define variable or dynamic values for power_on and standby_power?

The current fixed value approach works well for short or intermittent usage, but I’ve noticed that for devices like electric heaters, which can stay ON for several hours, this leads to less accurate estimations. Since the integration only applies a fixed power_on value during the entire ON duration, the total energy consumption is calculated only after the full ON period ends. This means that if, for example, a heater runs for 6 hours, the energy use is only correctly accounted for at the end of that period — and it assumes the same constant usage the entire time.

In reality power draw fluctuates a little bit.

Would this be technically feasible, or is there another way to achieve this kind of dynamic behavior using the current capabilities of Powercalc?

Thanks again for your work and time!

Something like this?

powercalc:
  sensors:
    - entity_id: switch.test
      fixed:
        power: 4-6
      standby_power: 0.4-0.6

[bramstroker]

I don't like to overcomplicate standby power more.
There is no way for powercalc any way to know when it is consuming 4 or 6 watt. So even when I build something which fluctuates the power randomly it will still be somewhere around 5W average.
So I don't see any advantage in something like this.

What do you mean with This means that if, for example, a heater runs for 6 hours, the energy use is only correctly accounted for at the end of that period?
It should update multiple times within this 6 hours. By default every 10 minutes, it should force a state update for power sensor on this interval and update energy sensor accordingly.

[gieljnssns]

This is an example of a group.

On the second image you can see that there isn't any energy consumption every hour for Media, there are only green bars in the hours there was any change idle/playing

Now I have made a template sensor to add to the Media group

    ###  Random watt sensor  ##
    - name: "Random watts"
      unique_id: 27dc873d-2129-4ad3-96b5-be4310206b54
      state: >
        {% set options = [-1.0, -0.5, 0.0, 0.5, 1.0] -%}
        {% set t = now().hour * 60 + now().minute -%}
        {% set index = t % options | length -%}
        {{ options[index] }}
      unit_of_measurement: "W"
      device_class: power
      icon: mdi:variable

This is an image whit my random watts sensor included in the group.

Before and after random watts

[bramstroker]

How do you have this setting in the group?

And could you show history of the energy sensor of Media?

[gieljnssns]

Force calculate energy is off

[gieljnssns]

There has always been a real power sensor in that group, but I had recently removed it.
After that, I noticed that the calculation was no longer correct.
I’m going to enable force calculate energy now.
Thank you for your response.

[bramstroker]

It should also work without this setting.
Enabling this setting will force create a riemann sum sensor with integrates to a kWh from the power sensor.
But the default setting it will just sum the underlying energy sensors. As long as the underlying member sensor increase their energy state the group sensor should also update.
So I would suggest to have a look into one of the member sensors which should contribute to the standby power. Look into that state history, does it slowly increase when in standby? Of does it stay the same?

[gieljnssns]

I have the energy sensors of those members disabled

[bramstroker]

Ah, that would be the reason. You should not disable these as powercalc needs these in the group energy calculation.
You could make them hidden, but disabling actually disables the energy sensors completely so they don't calculate.

An alternative would be to enable force calculate as previously suggested. But when you have some real energy sensors in the group this is not preferrable as the riemann sum is an approximation and will never be as perfect as real energy sensors.