PowerModels Parser Overzealously Updates Bus Type

[GabrielKS]

PowerSystems allows generators to be connected to PQ buses. So does the PSS/E data format, as far as I can tell. However, when PowerSystems parses such a thing from a PSS/E file, it converts the bus to PV:

PowerSystems.jl/src/parsers/pm_io/data.jl

Line 1698 in c932c0b

@info "active generators found at bus $(bus["bus_i"]), updating to bus type from $(bus["bus_type"]) to 2" maxlog =

Though probably intended as a data cleaning measure (quoth @claytonpbarrows), this strikes me as a bug as it leads to a Sienna system that does not correspond to the valid PSS/E system to the extent possible.

Note that the opposite case, which we also implement — converting a PV bus to PQ if there are no generators connected — is less problematic: the PSS/E spec states that PV buses must have at least one corresponding generator record, so PV buses without generators are invalid and we are free to do whatever we want with them.

I commented out the offending lines of data.jl and the PSY tests still passed, but I'd love to hear from others on the best thing to do here. Here's a minimal PSS/E RAW file that reproduces the issue (bus 1434):

0, 100.0, 33, 0, 1, 60.0
basic_2024
PSS/E 33.3 RAW via PowerFlows.jl, 2024-08-13T15:36:56.175
1434, 'B1434_S', 20.0, 1, 1, 10, , 1.0131474402609133, 21.436317553216863, 1.1, 0.9
4242, 'B4242_ND', 20.0, 3, 2, 2, , 1.02, 0.0, 1.1, 0.9
0
0
0
1434, 'OP', 520.0000000000001, 0.0, 3600.0, -3600.0, 1.0131474402609133, , 638.0, , , , , , 1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Q

[jd-lara]

@GabrielKS see this issue for reference on this topic lanl-ansi/PowerModels.jl#797