readme update #3
Conversation
README.md
Outdated
| @@ -66,8 +66,19 @@ samples = RunACSampler(num_of_samples, params_AC); | |||
| The sampling function starts by generating the number of required samples and then runs OPF for each of the samples and filter those with feasible OPF solutions. Importing power grid data, grid modifications and solving OPF are all done within [PowerModels.jl](https://github.com/lanl-ansi/PowerModels.jl) framework. Since some of the generated samples might not be feasible, the sample generation continues in an iterative manner until the required number of samples with feasible solution is met. Currently, if more than 60\% of the samples lead to infeasible OPF in the first iteration, the algorithm returns an error to indicate the fact that the choice of parameters might not be suitable for the used grid. | |||
|
|
|||
| #### Output Structure | |||
| The output data `samples` is an array of dictionaries where each element of array has the corresponding sample parameter values and the OPF solution. For each input parameter type, the order of array of data is based on the sorted dictionary of the original data in the `base_model`. For examples, `samples[i]["rate_a"]` is a vector containing the thermal rating branch values in sample i and the order is the same order of the keys in `sort(base_model["branch"])`. Or `samples[i]["qd"]` is a vector containing the load reactive power values in sample i and the order is the same order of the keys in `sort(base_model["load"])`. | |||
| The output data `samples` is an array of dictionaries where each element of array has the corresponding sample parameter values and the OPF solution. For each input parameter type, the order of array of data is based on the sorted dictionary of the original data in the `base_model`. For examples, `samples[i]["rate_a"]` is a vector containing the thermal rating branch values in sample i and the order is the same order of the keys in `sort(base_model["branch"])`. Or `samples[i]["qd"]` is a vector containing the load reactive power values in sample i and the order is the same order of the keys in `sort(base_model["load"])`. | |||
There was a problem hiding this comment.
sample parameter values and the OPF solution
I thought the solution was not being included because it is (relatively) cheap to calculate? Also, the files are called input_*.
There was a problem hiding this comment.
Or is this referring to the output of OPFSampler.jl directly, and not the data we are providing.
There was a problem hiding this comment.
In which case, this needs a clarifying comment, as there will be a discrepancy there.
There was a problem hiding this comment.
oh thanks, good catch. Actually to be precise, you are right. This is the output of Sampler and not the data we are providing so I should add a comment in the data we are providing that we have removed the OPF_output from the output to lower the size.
There was a problem hiding this comment.
Please check if the changes in the new commit clarify things.
There was a problem hiding this comment.
@AlexRobson I reorganized things a bit and added more description on the original PowerModels data and the mapping between our samples and that data. Can you please take a look to see if it is more clear?
Codecov Report
@@ Coverage Diff @@
## master #3 +/- ##
=======================================
Coverage 86.14% 86.14%
=======================================
Files 4 4
Lines 166 166
=======================================
Hits 143 143
Misses 23 23Continue to review full report at Codecov.
|
Co-Authored-By: Alex Robson <[email protected]>
No description provided.