Fix Monte Carlo random sampling for VRE siting availability by patrickbrown4 · Pull Request #249 · ReEDS-Model/ReEDS

patrickbrown4 · 2026-10-07T00:13:51Z

Summary

This PR partially addresses #41 by fixing the Monte Carlo sampling functionality for VRE siting availability.

Latin hypercube sampling for VRE siting availability is still broken; we can decide whether to fix that here or in a followup PR.

Technical details

Implementation notes

The original sampling approach for siting was designed around the old region/class profiles; now that we use site-level profiles, that approach would be much more time- and memory-intensive.
We used to reweight the profiles based on the capacity in the sampled supply curve. Profiles do vary at the site level (see Fix Monte Carlo sampling for demand and VRE availability #41 for some background discussion and plots), but not very much.
So now we read a set of site-level profiles for the most permissive siting scenario included in the specified Monte Carlo distribution. For example, if the distribution includes both limited and reference access, the reference-access site profiles will be used; if it includes reference and open, the open-access profiles will be used.
- There are sites in limited that are not in reference/open, and sites in reference that are not in open. So for Monte Carlo sampling, we ignore site mismatches between the profiles and the supply curves (printing messages to the user)

Additional changes

writecapdat.py: I was having trouble following the structure while debugging, so reorganized to put the globals at the top and renamed a few variables (sorry)
- The substantive changes ended up just being to assign_class(); I was getting float classes (instead of integers) without this change, which were messing up downstream processing.
mcs_sampler.py: We no longer need the exogenous or prescribed capacity processing, so those parts are removed along with the recf profile processing.
mcs_distributions_default.yaml:
- Updated AEO switches from 2025 to 2026
- Tweaked a few scenario definitions to match what was used in the conference paper (we could still do more here to unify the naming and clarify the structure and options)
I wanted to check the offshore-zones functionality (MARICTNYNJPAOH_Offshore) since I touched the supply curves but it was already broken. So I included a fix for it in this PR.

Issues resolved

#41 (partially)

Known incompatibilities

I'm still getting an error for the MonteCarlo_LHS (latin hypercube) case when siting is included; I'm not sure why it's behaving differently from MonteCarlo_Random. I think the fix to random sampling is valuable enough to get in on its own, but if the LHS case ends up being a quick/easy fix, we can include it here too.

TODO before merge

Double check that the supply-curve sampling results make sense. Are all the site capacities somewhere between the capacities in the scenarios that define the distribution?
Debug full-USA MC run (error in missing_class_resource.csv)

Validation, testing, and comparison report(s)

Changes to USA_defaults are within the noise: results-v20261009_mainK0_USA_defaults,v20261009_mcK0_USA_defaults.pptx

The MonteCarlo_Random case now works with VRE siting included in the distribution.

Here's a full-US example with 10 samples and MCS_dist_groups set to load_flat_country.siting_lowmid_country.gasprice_flat.upv.onswind.ofswind.battery.nuclear.geo:
results-1,2,3,4,5,6,7,8,9,10.pptx

20261019 update

Still minimal change to USA_defaults: results-v20261017_mainK0_USA_defaults,v20261017_mcK0_USA_defaults.pptx
And USA_decarb:
results-v20261017_mainK0_USA_decarb,v20261017_mcK0_USA_decarb.pptx
I ran the full cases_test.csv, and all except the following are working:
- MonteCarlo_LHS (Failing because I added siting to the list of distributions. If you'd rather keep it working with a smaller set, we can remove siting.)
- Pacific_rep15: infeasible, but infeasibility model is infeasible, so not sure it's real
- Simple: Fixed by Fix cases_small.csv and merge it with cases_test.csv #222

Checklist for author

Details to double-check

Charge code provided to reviewers
Included comparison reports for appropriate test cases
Code formatting standardized
Reusable functions used where possible instead of copy/pasted code

General information to guide review

Zero impact on results of default case
No large data file(s) added/modified
No substantive impact on runtime for full-US reference case
No substantive impact on folder size for full-US reference case
No change to process flow (runreeds.py, reeds/core/solve/solve.py)
No change to code organization
No change to package requirements (environment.yml or Project.toml)

Did you use LLM tools (chatbot or copilot) in the preparation of this PR? If so, describe how

No

…n process_unitdata

…s and settings

…esent

…y one site in a class

…apital_adder_per_mw fixed for supply curve files in mcs_sampler.py

…rocess_unitdata.py)

patrickbrown4 · 2026-10-19T16:31:35Z

+        if (tech == 'wind-ofs') and int(sw.GSw_OffshoreZones):
+            gdf_joined['r'] = gdf_joined['ba']
+        else:
+            gdf_joined['r'] = gdf_joined['r_nearest']


This is the only substantive change to this script (to get offshore zones working again); the rest were just to help me follow things while debugging

patrickbrown4 · 2026-10-19T16:32:35Z

            'copy_files',
-            'process_unitdata',
            'mcs_sampler',
+            'process_unitdata',


Need to run after the sampling since it assigns existing capacity to sites

patrickbrown4 · 2026-10-19T16:33:24Z

 ,Default Value,Pacific,USA_defaults,Mid_Case,USA_decarb,github_Pacific,github_Everything,github_MA_county_CC,Pacific_CC,Pacific_weks,Pacific_full_year,Interday_storage,Pacific_2020,Pacific_rep15,WY_county,WECC_county,PJM_county_CC,NYVT_mixed,OR_water,MonteCarlo_Random,MonteCarlo_LHS,Everything,Simple,USA_fast,USA_faster,MultiMetricRA,Pacific_DR,Pacific_MGA,Pacific_MGA_RV,Pacific_LoadSite95,MARICTNYNJPAOH_Offshore,R2P
 ignore,1,0,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
-GSw_Region,cendiv/Pacific,,country/USA,country/USA,country/USA,,st/ID.WY.NE.IA.IL,st/MA,,,,,,,st/WY,interconnect/western,transreg/PJM,st/NY.VT,st/OR,st/NE.NY.PA,st/NE.NY.PA,st/ID.WY.NE.IA.IL,st/KS,country/USA,country/USA,st/NY.NJ,,,,,st/MA.RI.CT.NY.NJ.PA.OH,
+GSw_Region,cendiv/Pacific,,country/USA,country/USA,country/USA,,st/ID.WY.NE.IA.IL,st/MA,,,,,,,st/WY,interconnect/western,transreg/PJM,st/NY.VT,st/OR,,,st/ID.WY.NE.IA.IL,st/KS,country/USA,country/USA,st/NY.NJ,,,,,st/MA.RI.CT.NY.NJ.PA.OH,


Changed from NE.NY.PA to Pacific to avoid the islanded-zones warning added in #243

patrickbrown4 · 2026-10-19T16:33:57Z

 GSw_HourlyClusterAlgorithm,,,,,,,,,,,,,,,,,,,,user,user,,,,,,,,,,,
 MCS_runs,,,,,,,,,,,,,,,,,,,,2,2,,,,,,,,,,,
-MCS_dist_groups,,,,,,,,,,,,,,,,,,,,tech.hydro.nuclear.gas.coal.load_country,upv_tri.nuclear_tri.ng_fuel_price_tri.load_country_unif,,,,,,,,,,,
+MCS_dist_groups,,,,,,,,,,,,,,,,,,,,load_flat_state.gasprice_flat.siting_flat_state.upv.onswind.ofswind.battery.nuclear.geo,load_country_unif.gasprice_tri.siting_country_discrete_flat.upv_tri.nuclear_tri,,,,,,,,,,,


I added siting to the LHS case, but it's not yet supported. @bsergi let me know if you'd rather leave it out until we fix it for LHS.

patrickbrown4 · 2026-10-19T16:35:07Z

-            case=os.path.dirname(os.path.normpath(inputs_case)),
-            agg=agg,
-        ).reset_index()
+        df = reeds.io.get_supplycurve(full_path, case=inputs_case)


Now we just get the initial supply curve and assemble the full curve (i.e. add the interconnection costs) only once downstream in writesupplycurves.py, rather than assembling it twice (which complicates the MC sampling)

patrickbrown4 · 2026-10-19T16:35:48Z

We could spend more time unifying the naming of the distributions in this file; let me know if you think we should do that here

patrickbrown4 · 2026-10-19T16:37:55Z

Some of the changes here are just formatting. The main substantive change is that I removed the exogenous/prescribed sampling (since it's now handled downstream in process_unitdata.py) and only sample the capacity from the supply curve, inheriting the cf and profiles from the most permissive siting scenario included in the distribution.

patrickbrown4 · 2026-10-19T16:40:44Z

Sorry, a lot of these changes were just to help understand the organization while I was debugging.

patrickbrown4 · 2026-10-19T16:44:10Z

-        pd.read_csv(os.path.join(inputs_case, 'classification_geothermal.csv'))
-        .query(f"access_case == '{sw.GSw_SitingGeo}'")
+
+def get_class_cf_min(


I think this set of 3 functions is the main substantive change. While testing different zone sets, I was getting issues for cases where there was only a single site in a zone/tech_class combo. Now instead of defining min and max for each class, we just define min, and take the first match.

patrickbrown4 added 12 commits September 4, 2026 13:16

mcs_sampler.py: remove exog_cap and prescribed_builds

c2e11c3

mcs_sampler.py: fix some indentation and indexing

b7c7ae8

move mcs_sampler before process_unitdata since we use supply curves i…

ed241a2

…n process_unitdata

Merge branch 'main' into pb/mc

79f333d

Merge branch 'main' into pb/mc

8c05a10

Merge branch 'main' into pb/mc

76df0c1

mcs_sampler.py: drop recf sampling and weighting

d86316b

writecapdat.py: avoid float class numbers in cap_exog

cbdf8a3

recf.py: allow for missing sites; for MC, use only the most-open profile

23df9a4

add siting back to MonteCarlo_Random test case

c54435e

add siting to MonteCarlo_LHS

4730cd5

mcs_distributions_default.yaml: update AEO scenarios; tweak some name…

7216908

…s and settings

patrickbrown4 requested review from bsergi and removed request for bsergi October 7, 2026 00:13

github-actions Bot added data_changes model_changes docs switches labels Oct 7, 2026

writecapdat.py: fix get_class_cf_bounds() when not all classes are pr…

81878eb

…esent

patrickbrown4 marked this pull request as draft October 7, 2026 20:09

patrickbrown4 added 3 commits October 7, 2026 16:18

writecapdat.py: simplify class assignment and fix situations with onl…

d22bf6b

…y one site in a class

writecapdat.create_exog_rsc(): fix upv tech name

26953f0

Merge branch 'main' into pb/mc

b0bde73

patrickbrown4 marked this pull request as ready for review October 7, 2026 23:23

patrickbrown4 marked this pull request as draft October 7, 2026 23:39

patrickbrown4 added 5 commits October 8, 2026 13:02

only write original supply curves from copy_files.py; keep cf/class/c…

0bf9f13

…apital_adder_per_mw fixed for supply curve files in mcs_sampler.py

inputs.parse_cases(): allow MCS check to be skipped

0fd65ae

Merge branch 'main' into pb/mc

b004682

write assembled supply curves for r2x

a7068c5

Merge branch 'main' into pb/mc

7530134

add supplycurve_init* files to futurefiles.csv

77ae3ea

patrickbrown4 marked this pull request as ready for review October 9, 2026 15:57

patrickbrown4 requested a review from bsergi October 9, 2026 16:13

patrickbrown4 added 3 commits October 16, 2026 10:29

Merge branch 'main' into pb/mc

e82daa1

reeds.io.get_supplycurve(): use offshore sitemap for wind-ofs

6c1f97d

fix offshore zones (use sitemap assignment in get_supplycurve() and p…

0c55e25

…rocess_unitdata.py)

patrickbrown4 commented Oct 19, 2026

View reviewed changes

Navigation Menu

Repository navigation

Fix Monte Carlo random sampling for VRE siting availability - #249

Fix Monte Carlo random sampling for VRE siting availability#249
patrickbrown4 wants to merge 25 commits into
mainfrom
pb/mc

patrickbrown4 commented Oct 7, 2026 •

edited

Loading

Uh oh!

patrickbrown4 Oct 19, 2026

Uh oh!

patrickbrown4 Oct 19, 2026

Uh oh!

patrickbrown4 Oct 19, 2026

Uh oh!

patrickbrown4 Oct 19, 2026

Uh oh!

patrickbrown4 Oct 19, 2026

Uh oh!

patrickbrown4 Oct 19, 2026

Uh oh!

patrickbrown4 Oct 19, 2026

Uh oh!

patrickbrown4 Oct 19, 2026

Uh oh!

patrickbrown4 Oct 19, 2026

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

patrickbrown4 commented Oct 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Technical details

Implementation notes

Additional changes

Issues resolved

Known incompatibilities

TODO before merge

Validation, testing, and comparison report(s)

20261019 update

Checklist for author

Details to double-check

General information to guide review

Did you use LLM tools (chatbot or copilot) in the preparation of this PR? If so, describe how

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

patrickbrown4 commented Oct 7, 2026 •

edited

Loading