[WIP] [docs] Complex number AC Optimal Power Flow tutorial #3382
Conversation
Codecov ReportPatch and project coverage have no change.
Additional details and impacted files@@ Coverage Diff @@
## master #3382 +/- ##
=======================================
Coverage 98.06% 98.07%
=======================================
Files 34 34
Lines 4922 4928 +6
=======================================
+ Hits 4827 4833 +6
Misses 95 95 ☔ View full report in Codecov by Sentry. |
| # following question: what is the most cost-effective operation of electricity | ||
| # generators while meeting constraints on the safe limits of network components? | ||
|
|
||
| # We will use the 9 _bus_ network test case `case9mod` to explore this problem |
There was a problem hiding this comment.
In there, it seems "case9mod" is some standard name but it's only where from. Is it coming from some benchmarking suite ?
| # Even before solving, we can substitute the lower bound on each generator's real power range | ||
| # (all 10, as it turns out in this case) | ||
|
|
||
| objval_basic_lb = value(lower_bound, objective_function(model)); |
There was a problem hiding this comment.
It's unclear why this is a lower bound since you have quadratic terms
I think your choice of putting it in conic is good. |
|
Thanks @blegat for the review. Could you see if the latest commit addresses your comments? |
| # The techniques of *convex relaxations* can also be used to improve on our | ||
| # current best lower bound: | ||
|
|
||
| ## println("Objective value (better lower bound): $(objval_better_lb)") |
There was a problem hiding this comment.
Yes. Still to do.
|
I made a few formatting changes. Perhaps this would be better in the application section? https://github.com/jump-dev/JuMP.jl/tree/master/docs/src/tutorials/applications |
|
|
||
| # The purpose of this tutorial is to highlight JuMP's ability to directly | ||
| # formulate problems involving complex-valued decision variables and complex | ||
| # matrix cones such as the [`HermitianPSDCone`](@ref) object. |
There was a problem hiding this comment.
TODO: add an SDP relaxation so this sentence is true.
|
Yes, that seems reasonable since it cuts across a few areas and is fairly technical in places. |
|
|
||
| Y = | ||
| A * SparseArrays.spdiagm(1 ./ z) * A' + | ||
| SparseArrays.spdiagm(LinearAlgebra.diag(A * SparseArrays.spdiagm(y_sh) * A')) |
There was a problem hiding this comment.
@odow I agree with this formatting but the computer says no?
There was a problem hiding this comment.
Hmm. I think the formatter must be changing its mind. I didn't change this. Just ran the formatter locally.
There was a problem hiding this comment.
Is there a formatting file I'm not seeing by just running
JuliaFormatter.format_file("./docs/src/tutorials/conic/optimal_power_flow.jl", verbose=true)
from the ./JuMP/ dir?
There was a problem hiding this comment.
No that should work. Maybe we have different versions of JuliaFormatter. Otherwise break it up a bit so there aren't so many nested calls?
There was a problem hiding this comment.
I think that would help readability anyway.
|
@jd-foster, I've sent you an invite to join This will give you write permission to push to the |
|
Thoughts on taking Hypatia as a dependency in the docs environment? It worked well for the hardest SDP example here while SCS seems to struggle, at least on the default settings. |
|
Potentially Clarabel? I used it in #3385 for higher precision. |
|
Closing in favor of #3395 |
Still plenty to add and fix here, but the first pass at #3363 .
It currently is a nonlinear AC-OPF formulation, but the idea is to add to it the original intent, SDP relaxation of AC-OPF using
HermitianPSDCone. Not sure whether it should be underconicornonlinear...