Simple SCOPF model leveraging MadNLP two-stage solver

Project.toml

@@ -17,6 +17,15 @@ MadNLP = "0.10"
 MadNLPGPU = "0.10"
 julia = "1.11"
 
+# The two-stage SCOPF (scopf_twostage_model) needs MadNLP's SchurComplementKKTSystem
+# support for design-only constraints + the deterministic GPU Schur assembly, which live
+# on the `schur-design-constraints` branch (not yet released). Pin both MadNLP and the
+# MadNLPGPU subpackage to that branch so `Pkg.test()` (local and CI) resolves it instead
+# of the registry release. Remove once these land in a tagged MadNLP/MadNLPGPU.
+[sources]
+MadNLP = {url = "https://github.com/MadNLP/MadNLP.jl", rev = "schur-design-constraints"}
+MadNLPGPU = {url = "https://github.com/MadNLP/MadNLP.jl", rev = "schur-design-constraints", subdir = "lib/MadNLPGPU"}
+
 [extras]
 CUDA = "052768ef-5323-5732-b1bb-66c8b64840ba"
 CUDSS = "45b445bb-4962-46a0-9369-b4df9d0f772e"

README.md

@@ -59,6 +59,65 @@ model, cons, vars, lengths, sc_data_array = goc3_model(
 result = madnlp(model; tol=1e-4)
 ```
 
+### Generic N-1 security-constrained OPF
+
+Independent of the GOC3 formulation, `scopf_model` builds a lightweight, hard-constrained
+N-1 SCOPF. Each contingency is a generator or branch (line) outage; the base case plus all
+contingencies are solved together as one AC OPF, with bounded corrective generator redispatch
+in the post-contingency scenarios. Line outages are applied by zeroing the branch admittance
+(so its flows are forced to zero), and MATPOWER `rateA = 0` ("unlimited") is mapped to a large
+finite rating.
+
+```julia
+using ExaModelsPower, MadNLP, DelimitedFiles
+
+# One outage per entry; (type = :gen, idx = g) is also supported.
+idxs = vec(readdlm("data/case118.Ctgs", Int))         # 1-based branch indices
+contingencies = [(type = :branch, idx = l) for l in idxs]
+
+model, vars, cons = scopf_model("case118.m", contingencies; form = :polar)
+result = madnlp(model; tol = 1e-4)
+
+# On a GPU backend, solve the condensed KKT system with cuDSS:
+using MadNLPGPU, CUDA, CUDSS
+model, vars, cons = scopf_model("case118.m", contingencies; backend = CUDABackend())
+result = madnlp(model;
+    kkt_system    = MadNLP.SparseCondensedKKTSystem,
+    linear_solver = MadNLPGPU.CUDSSSolver,
+    tol = 1e-4)
+```
+
+#### Two-stage / Schur form
+
+`scopf_twostage_model` builds the same problem on a `TwoStageExaCore` (base case = first stage,
+each contingency = a scenario) so it can be solved with MadNLP's `SchurComplementKKTSystem`,
+which factorizes the per-contingency blocks in parallel. It returns an extra `post_solve_info`
+NamedTuple with the Schur dimensions:
+
+```julia
+model, vars, cons, info = scopf_twostage_model("case118.m", contingencies; backend = CUDABackend())
+result = madnlp(model;
+    kkt_system    = SchurComplementKKTSystem,
+    linear_solver = MadNLPGPU.CUDSSSolver,
+    kkt_options   = Dict(:schur_ns => info.ns, :schur_nv => info.nv,
+                         :schur_nd => info.nd, :schur_nc => info.nc))
+```
+
+> **Note.** The base-case physics are first-stage *design* constraints. The current
+> `SchurComplementKKTSystem` does not yet support design-only constraint rows (solving asserts
+> `m == ns*nc`, off by exactly `info.nc_design`). The model is structurally correct; the Schur
+> solve is gated on upstream support for design-only constraints.
+
+A runnable example covering both formulations (`--mode single`/`twostage`/`compare`), both
+coordinate systems (`--form polar`/`rect`), and a choice of case (`--case case9`/`case118`) is in
+[`examples/scopf.jl`](examples/scopf.jl). Options are ArgParse CLI flags:
+
+```bash
+julia --project=./examples examples/scopf.jl --help
+julia --project=./examples examples/scopf.jl --case case9 --mode twostage --form rect
+julia --project=./examples examples/scopf.jl --gpu --inertia free
+```
+
 ### Multi-period optimal power flow
 ```julia
 model, vars, cons = mpopf_model(

data/case118.Ctgs

@@ -0,0 +1,176 @@
+2
+3
+4
+5
+6
+8
+10
+11
+12
+13
+14
+15
+16
+17
+18
+19
+20
+21
+22
+23
+24
+25
+26
+27
+28
+29
+30
+31
+32
+33
+34
+35
+36
+37
+38
+39
+40
+41
+42
+43
+44
+45
+46
+47
+48
+49
+50
+51
+52
+53
+54
+55
+56
+57
+58
+59
+60
+61
+62
+63
+64
+65
+66
+67
+68
+69
+70
+71
+72
+73
+74
+75
+76
+77
+78
+79
+80
+81
+82
+83
+84
+85
+86
+87
+88
+89
+90
+91
+92
+93
+94
+95
+96
+97
+98
+99
+100
+101
+102
+103
+104
+105
+106
+107
+108
+109
+110
+111
+112
+114
+115
+116
+117
+118
+119
+120
+121
+122
+123
+124
+125
+126
+127
+128
+129
+130
+131
+132
+135
+136
+137
+138
+139
+140
+141
+142
+143
+144
+145
+146
+147
+148
+149
+150
+151
+152
+153
+154
+155
+156
+157
+158
+159
+160
+161
+162
+163
+164
+165
+166
+167
+168
+169
+170
+171
+172
+173
+174
+175
+178
+179
+180
+181
+182
+185
+186