Open dlcole3 opened 5 months ago
I think this is a bug in the indexing.
I looked into this issue and it seems that we default to interpreting int indexes the the bus location and not the number. if you pass the bus object then it works fine. We could make this error more informative.
this should work
using PowerNetworkMatrices, PowerSystems, PowerSystemCaseBuilder
rts_sys = build_system(PSITestSystems, "test_RTS_GMLC_sys")
ptdf = PTDF(rts_sys)
bus = get_component(Bus, rts_sys, "Cobb")
vector = ptdf[:, bus]
I found that the PTDF matrix does not recognize the lookup dictionary when trying to access the vector of line data corresponding to a given bus. For example, in the code below, I get an error on the last line because the number for the bus "Cobb" (321) is greater than that dimension of the matrix (73). This is using Julia 1.9 with PSY v3.3.0, PNM v0.10.0, and PSCB v1.2.5.
This error came up because I was using the code from PowerSimulations.jl here to compute the LMPs of a Sienna simulation. This was fairly easy to work around (just mapping the values in
ptdf.axes[1]
to the order they appear) but wasn't sure if this was expected behavior. Wondering if it would make sense to haveptdf.axes[1]
be the bus names instead of the bus numbers to avoid this issue. For example, if I instead callptdf["A1", :]
, I do get the expected behavior where it gives me the column corresponding to Line "A1".Also, as a related note, it was a little confusing to me that the PTDF matrix expects the indexes to be passed as column then row, rather than row then column (e.g.,
ptdf["A1", 101]
rather thanptdf[101, "A1"]
) since the dimensions inptdf.axes
are listed with the bus number as dimension 1 and line names as dimension 2. Just an observation as a first time user. Overall though, thank you for making and supporting the Sienna suite. It has been very helpful.