Matpower conversion

[rb004f]

This pull request changes the default input format to follow the matpower conventions

@bluejuniper Take a look at the matpower based input files. I removed redundant/unused data fields that were in the json files. If you see something we should preserve, I will add them back in

@ccoffrin If you see something about the matpower input files that looks strange, let me know. But we should be moved away from the json input files now.

[ccoffrin]

In the data files would you consider something like (note curly braces),

%column_names%  hi_bus lo_bus gmd_br_hi gmd_br_lo gmd_k gmd_br_series gmd_br_common baseMVA type config
mpc.branch_gmd = {
	1	3	1	-1	1.793	-1	-1	100	'xf'	'gwye-delta'
	1	2	-1	-1	0	-1	-1	100	'line'	'none'
	2	4	3	-1	1.793	-1	-1	100	'xf'	'gwye-delta'
};

Same goes for gmd_bus. This is basically a short coming of using matricies to specify component data, all values in matricies should have a common type. Cell arrays are the easy way to get around this in InfrastructureModels.

Then I think we can avoid the need for, lanl-ansi/PowerModels.jl#305

[ccoffrin]

Only one minor suggestion. If the datasets you are using can be made open access (we should discuss the legal details), then I would highly recomend we propose them as an official IEEE PES GMD test set as part of PGLib, https://github.com/power-grid-lib/

[ccoffrin]

Consider touching up the output data writer to clean up the -29.99999 values.

[rb004f]

Agreed. This was a more elegant way of handling the data. I have updated the data sets accordingly. And, yes, this reduces the need for lanl-ansi/PowerModels.jl#305