The length of result['solution']['gen'] did not match the inputted pm_data['gen]

[yasirroni]

I'm testing PowerModels on PGLib. In running pglib_opf_case2746wop_k.m, the length of input is pm_data['gen] is 514, while the output is result['solution']['gen']. Can you enlighten me the meaning of this? Thank you.

[yasirroni]

Related to #928, debug and understanding the behavior of pm_data = _matpower_to_powermodels!(mp_data).

[ccoffrin]

I cannot follow your question. Please review the package docs and elaborate on what you think may be going wrong.

[yasirroni]

Same data, why solve_opf result has fewer generator.

using InfrastructureModels
using PowerModels
using Ipopt
using JuMP

case = "pglib-opf/pglib_opf_case2746wop_k.m"
nlp_solver = JuMP.optimizer_with_attributes(Ipopt.Optimizer, "tol"=>1e-6, "print_level"=>0)

result_ac = solve_opf(case, ACPPowerModel, nlp_solver)

result_ac["solution"]["gen"]
# Dict{String, Any} with 431 entries:
#   "306" => Dict{String, Any}("qg"=>2.05902e-7, "pg"=>0.15)
#   "1"   => Dict{String, Any}("qg"=>-0.0961256, "pg"=>1.4)
#   "54"  => Dict{String, Any}("qg"=>0.601058, "pg"=>1.2)
#   "101" => Dict{String, Any}("qg"=>0.151089, "pg"=>0.4)
#   "371" => Dict{String, Any}("qg"=>0.0749997, "pg"=>0.5)
#   "464" => Dict{String, Any}("qg"=>0.000998251, "pg"=>0.001)
#   "475" => Dict{String, Any}("qg"=>0.0, "pg"=>0.012)
#   "447" => Dict{String, Any}("qg"=>0.0, "pg"=>0.008)
#   "335" => Dict{String, Any}("qg"=>5.1374e-8, "pg"=>0.25)
#   "362" => Dict{String, Any}("qg"=>0.00999991, "pg"=>0.53376)
#   "505" => Dict{String, Any}("qg"=>0.00761, "pg"=>0.19344)
#   "491" => Dict{String, Any}("qg"=>0.01699, "pg"=>0.058)
#   "299" => Dict{String, Any}("qg"=>0.32169, "pg"=>0.8)
#   "168" => Dict{String, Any}("qg"=>1.06376e-7, "pg"=>0.09)
#   "159" => Dict{String, Any}("qg"=>-0.199888, "pg"=>1.35)
#   "403" => Dict{String, Any}("qg"=>0.0, "pg"=>0.03)
#   "228" => Dict{String, Any}("qg"=>1.76333e-7, "pg"=>0.06999)
#   "332" => Dict{String, Any}("qg"=>3.76483e-8, "pg"=>0.4)
#   "270" => Dict{String, Any}("qg"=>0.697152, "pg"=>1.4)
#   ⋮     => ⋮

pm_data = open(case) do io
    mp_data = PowerModels._parse_matpower_string(read(io, String))
    pm_data = PowerModels._matpower_to_powermodels!(mp_data)
end

pm_data["gen"]
# Dict{String, Any} with 514 entries:
#   "306" => Dict{String, Any}("vg"=>1.035, "mbase"=>25.6, "source_id"=>Any["gen"…
#   "407" => Dict{String, Any}("vg"=>1.035, "mbase"=>3.6, "source_id"=>Any["gen",…
#   "1"   => Dict{String, Any}("vg"=>1.03, "mbase"=>233.2, "source_id"=>Any["gen"…
#   "54"  => Dict{String, Any}("vg"=>1.03, "mbase"=>150.0, "source_id"=>Any["gen"…
#   "101" => Dict{String, Any}("vg"=>1.035, "mbase"=>57.5, "source_id"=>Any["gen"…
#   "371" => Dict{String, Any}("vg"=>1.035, "mbase"=>55.5, "source_id"=>Any["gen"…
#   "41"  => Dict{String, Any}("vg"=>1.03, "mbase"=>150.0, "source_id"=>Any["gen"…
#   "464" => Dict{String, Any}("vg"=>1.035, "mbase"=>0.1, "source_id"=>Any["gen",…
#   "65"  => Dict{String, Any}("vg"=>1.0, "mbase"=>252.4, "source_id"=>Any["gen",…
#   "475" => Dict{String, Any}("vg"=>1.035, "mbase"=>1.2, "source_id"=>Any["gen",…
#   "447" => Dict{String, Any}("vg"=>1.035, "mbase"=>0.8, "source_id"=>Any["gen",…
#   "335" => Dict{String, Any}("vg"=>1.035, "mbase"=>30.2, "source_id"=>Any["gen"…
#   "362" => Dict{String, Any}("vg"=>1.035, "mbase"=>53.4, "source_id"=>Any["gen"…
#   "505" => Dict{String, Any}("vg"=>1.035, "mbase"=>19.4, "source_id"=>Any["gen"…
#   "491" => Dict{String, Any}("vg"=>1.035, "mbase"=>6.0, "source_id"=>Any["gen",…
#   "326" => Dict{String, Any}("vg"=>1.035, "mbase"=>52.2, "source_id"=>Any["gen"…
#   "299" => Dict{String, Any}("vg"=>1.035, "mbase"=>129.3, "source_id"=>Any["gen…
#   "168" => Dict{String, Any}("vg"=>1.035, "mbase"=>9.2, "source_id"=>Any["gen",…
#   "159" => Dict{String, Any}("vg"=>1.035, "mbase"=>255.0, "source_id"=>Any["gen…
#   ⋮     => ⋮

[yasirroni]

I will assume that missing value in result_ac["solution"]["gen"] means the generator is not available and the value of qg and pg is 0. If that is not what intended, please let me know. I'm on my way writing a benchmark of various solver out there.

[ccoffrin]

That is the correct. Any component that is deactivated via the status field will not appear in the solution data. You may be interested in the basic-data feature, where this kind of situation does not occur, https://lanl-ansi.github.io/PowerModels.jl/stable/basic-data-utilities/

If you are interested in benchmarks you may find this effort interesting, https://github.com/lanl-ansi/rosetta-opf and https://youtu.be/tvBNQcuU-hY

I am going to close this issue now because the software is working as expected.