Since PR #150 overpowering is not working as expected

[jmcollin78]

In PR #150 , we change the way the power is configured.
This lead to a regression in overpowering calculation.

Expected:
The overpowering calculation should calculate the max(device_power / nb_heater, device_power * on_percent) in VTherm overpowering calculation,

So that when on_percent is 100% all the power should be available, or only one heater power (in average).

[adi90x]

Hello,
This may lead to an unsecured situation 🫤 where we fired up a heater that use more than the average.
Don't we want to secure by checking for full power available ?
Regards