Wrong bounds when batteries have different bounds

[llucax]

Here's the bug:

frequenz-sdk-python/src/frequenz/sdk/microgrid/_power_distributing/_distribution_algorithm/_battery_distribution_algorithm.py

Lines 126 to 136 in bcc902e

	# To satisfy the largest exclusion bounds in the set we need to
	# provide the power defined by the largest bounds multiplied by the
	# number of batteries in the set.
	power_exclusion_upper_bound = Power.from_watts(
	max(bounds.exclusion_upper for bounds in battery_metrics).as_watts()
	* len(battery_metrics)
	)
	power_exclusion_lower_bound = Power.from_watts(
	min(bounds.exclusion_lower for bounds in battery_metrics).as_watts()
	* len(battery_metrics)
	)

If one battery has exclusion bounds, it is applying those to all the batteries. So the total exclusion range becomes much bigger. For example, if only one battery had exclusion bounds, the rest of them should be usable normally.

[llucax]

There is another bug related to bounds in battery pools:

• Battery pool bounds are not calculated correctly #1180

[Marenz]

I don't think this is a bug.
There is no other way to leave the exclusion zone of battery A without also applying the same power to battery B as they are on the same inverter.

[llucax]

This also looks INVALID to me. @shsms

We would still need to investigate why we got bounds quite different from what the user requested in the power manager, but this doesn't look like a good lead.