Changed SubgridContainer to represent galvanically seperated grids

[staudtMarius]

Resolves #1226
Resolves #1310

[danielfeismann]

Almost there :)

[danielfeismann]

LGTM

[danielfeismann]

One thing we missed here is clarifying whether we consider the transformer to be part of the 'SubGridContainer', and if so, whether it is part of the lower or upper grid.

Excluding the transformer would make voltages within the container very simple and clear. However, this would mean that not all grid elements would be within the 'SubGridContainer'. I therefore tend to think that the connecting transformer should belong to the grid with the lower voltage level (but not the connecting node).

What do you think @sebastian-peter and @staudtMarius?

Once we agreed on a way we should update PSDM Docs and adapt OsmoGrid, which revealed this.

[sebastian-peter]

@danielfeismann I think traditionally, we have (almost) handled it in the exact way you described. The transformer is supposed to be part of the lower sub grid. The upper connecting node of the transformer (which I think you meant there) is part of the sub grid as well, though, because otherwise we would have a connecting element (the transformer) for which one node (the upper) does not exist within the container.

Check out ContainerUtils#filterForSubnet for reference (filtering for transformer with lower node, but also adding the upper node):

PowerSystemDataModel/src/main/java/edu/ie3/datamodel/utils/ContainerUtils.java

Lines 313 to 319 in c93fcf4

	Set<Transformer2WInput> transformer2w =
	input.getTransformer2Ws().stream()
	.filter(transformer -> transformer.getNodeB().getSubnet() == subnet)
	.collect(Collectors.toSet());
	/* Add the higher voltage node to the set of nodes */
	nodes.addAll(
	transformer2w.stream().map(Transformer2WInput::getNodeA).collect(Collectors.toSet()));

In my opinion, this approach is sufficiently well reasoned and should stay this way, however, there should definitely be documentation on this.