search

lanl-ansi / PowerModelsRestoration.jl

A PowerModels Extension for Optimization of Power Network Restoration
https://lanl-ansi.github.io/PowerModelsRestoration.jl/stable/
Other
22 stars 9 forks source link

Non-convex AC-MLD #43

Closed Dx-Wang closed 4 years ago

Dx-Wang commented 4 years ago

I guess the run_ac_mld_uc solves the relaxed SOC-MLD formulation in your paper, right? Is there a way to solve the non-convex AC-MLD formulation with this package? Or is there another package for AC-MLD? Thanks!

ccoffrin commented 4 years ago

There are the examples of running the model with different formulations, https://github.com/lanl-ansi/PowerModelsRestoration.jl/blob/master/test/mld_uc.jl

The specifics of each formulation name can be found here, https://lanl-ansi.github.io/PowerModels.jl/stable/formulation-details/

The function run_ac_mld_uc is a heuristics approach that combines the SOC and AC MLD formulations to quickly find feasible solutions to the non-convex AC-MLD problem. The implementation can be reviewed here, https://github.com/lanl-ansi/PowerModelsRestoration.jl/blob/master/src/util/ac-mld-uc.jl

I can also recomend having a look at the

Dx-Wang commented 4 years ago

Thanks a lot for your prompt response.

So the results I get from run_ac_mld_uc should be a solvable power flow case, right?

ccoffrin commented 4 years ago

Yes, as long as the termination status is reasonable the variable assignments should AC-PF feasible. However, note that some components may be removed as part of the solution process, so you need to update the component status as indicated in the result's solution.

Dx-Wang commented 4 years ago

Thanks for pointing it out. Is calling update_data!(model, result["solution"]); sufficient for such updates?

ccoffrin commented 4 years ago

Yup!

Dx-Wang commented 4 years ago

Got it. Thanks a lot!