Add linear branch flow model and branch flow storage problem type #676
Conversation
…for AbstractWRModel
Codecov Report
@@ Coverage Diff @@
## master #676 +/- ##
==========================================
+ Coverage 92.94% 92.98% +0.04%
==========================================
Files 42 42
Lines 7014 7070 +56
==========================================
+ Hits 6519 6574 +55
- Misses 495 496 +1 Continue to review full report at Codecov.
|
There was a problem hiding this comment.
Thanks for the contribution! Overall looks very good. I have a few minor requests about naming and best use of the type hierarchy.
Also, please add a note to the change log and add yourself to the list of contributors in the readme.
On a side note, can you tell me a little bit about your interest in this model? Do you have a specific application in mind or do you just like to have it for posterity sake?
src/core/types.jl
Outdated
| @@ -28,6 +28,10 @@ abstract type AbstractConicModel <: AbstractPowerModel end | |||
| "for branch flow models" | |||
| abstract type AbstractBFModel <: AbstractPowerModel end | |||
|
|
|||
|
|
|||
| "for variants of branch flow models that target LP solvers" | |||
| abstract type AbstractBFLPModel <: AbstractBFModel end | |||
There was a problem hiding this comment.
I think the name AbstractBFAModel for Branch Flow Approximation, would be more consistent with the other model naming conventions.
src/core/types.jl
Outdated
| } | ||
| ``` | ||
| """ | ||
| mutable struct LPBFPowerModel <: AbstractBFLPModel @pm_fields end |
There was a problem hiding this comment.
The model name should match the abstract type name, so BFAPowerModel following the previous comment.
src/form/shared.jl
Outdated
| @@ -343,3 +343,28 @@ function constraint_current_limit(pm::AbstractWModels, n::Int, f_idx, c_rating_a | |||
| JuMP.@constraint(pm.model, p_to^2 + q_to^2 <= w_to*c_rating_a^2) | |||
| end | |||
|
|
|||
| "" | |||
| function constraint_storage_loss(pm::AbstractWModels, n::Int, i, bus, r, x, p_loss, q_loss; conductors=[1]) | |||
There was a problem hiding this comment.
I think the more general type we want here is AbstractWConvexModels. Note that AbstractWModels includes AbstractACTModel which is a nonconvex model that does not require a current variable. This constraint can be limited similarly to the ACPPowerModel case.
src/form/wr.jl
Outdated
| @@ -337,7 +337,7 @@ end | |||
|
|
|||
|
|
|||
| "do nothing by default but some formulations require this" | |||
| function variable_current_storage(pm::AbstractWRModel; nw::Int=pm.cnw, bounded::Bool=true, report::Bool=true) | |||
| function variable_current_storage(pm::AbstractWModels; nw::Int=pm.cnw, bounded::Bool=true, report::Bool=true) | |||
There was a problem hiding this comment.
Same as above, AbstractWConvexModels. Also now as a shared variable definition, it can move to form/shared.jl.
test/opf.jl
Outdated
| @test isapprox(result["objective"], 11277.9; atol = 1e0) | ||
| end | ||
| # @testset "24-bus rts case" begin | ||
| # result = run_opf_bf("../test/data/matpower/case24.m", LPBFPowerModel, ipopt_solver) |
There was a problem hiding this comment.
Curious, what is the issue with this case?
There was a problem hiding this comment.
This case is infeasible for this formulations. I haven't figured out why.
|
The intention with the linear branch flow approximation is to use it for operation and sizing of storages in LV/MV grids. I need a computationally efficient implementation for problems with many timesteps. All your comments have been addressed and the code should be ready for a new review. Thanks for your feedback! |
|
BTW, if you continue to work on your fork you will want to sync and start from the master as I squash merged this PR. |
|
Thanks! No need for tagging it right away, you can do it when you find it convenient. |
This PR adds the linearized branch flow model formulation for OPF and PF including tests. A storage problem for branch flow has also been added (
build_mn_opf_bf_strg) and tested for both the new linearized branch flow and existing branch flow.build_mn_opf_bf_strgis the branch flow equivalent of bus injection model problembuild_mn_opf_strg