Better specify PowerFlowData parametric type aliases

[GabrielKS]

This:

is

• deeply unsettling and
• problematic for multiple dispatch.

[GabrielKS]

I'm not fully convinced that these pseudo-subtypes should be based on the types of the matrices they contain; that feels like an implementation detail, even though it's a pretty fundamental one. One can at least in principle imagine a pair of PowerFlowDatas with the same types of matrices except you want to solve them differently. Maybe the PowerFlowData contains the PowerFlowEvaluationModel it was constructed with, is parameterized on that, and that's what the aliases key onto (or at that point it's readable enough that we don't need the aliases)? Or maybe one is required to pass a PowerFlowEvaluationModel to solve_powerflow! along with the PowerFlowData? Guess this overlaps with #38.

[GabrielKS]

For now I'll just strictly specify vPTDFPowerFlowData as has been done for all the other aliases. I still think this might warrant rethinking more deeply later.