Update interzonal transmission costs/distances and use directly calculated values by patrickbrown4 · Pull Request #95 · ReEDS-Model/ReEDS

patrickbrown4 · 2026-05-15T23:13:22Z

Summary

In the same vein as 2003, this PR:

Adapts the interzonal transmission costs, distances, and planned/optional additions for spatial flexibility as laid out in Reformat spatial inputs pipeline for more flexibility #16
Uses the latest reV methods and assumptions to calculate the interzonal least-cost paths and costs

Technical details

Implementation notes

Least-cost paths

There are a few updates to the interzonal least-cost path procedure:

The interzonal transmission costs have been out of sync with interconnection costs since 1836. This PR brings them in sync, so both are now using costs from the MISO 2025 Transmission Cost Estimation Guide (parent, description, data)
- A new pair of data files from MISO, inputs/transmission/conductor_{ac or dc}.csv, provide lookup tables for converting from the assumed kV to MW capacity
We used to assume 500 kV costs for all interzonal greenfield lines, regardless of location or scale (so tiny counties with low-voltage lines got to use 500 kV costs, which are cheaper on a USD/MW basis than lower-voltage lines). Here, we instead use the maximum AC voltage currently connecting each pair of zones, with a 138 kV floor. So interfaces that are crossed by higher-voltage lines get lower USD/MW expansion than interfaces crossed by lower-voltage lines.

Spatial flexibility

The old future-capacity scenarios (which were specific to z134) have been removed. The scenarios used in NTP have been maintained (planned_lines-NTP_{MT or P2P}.csv), but now specify individual from/to endpoints
Similarly, planned additions are now specified as individual endpoints
- TransWest Express and SunZia have been added; see the descriptions in inputs/transmission/README.md
- The offshore zone backbones are in inputs/transmission/newlinks_offshore_backbone.csv
- The offshore zone radial connections are in inputs/zones/{GSw_ZoneSet}/newlinks_offshore_radial.csv
  - These need to be specified for each GSw_ZoneSet because the user might want to control which zone they make landfall in, particularly at county resolution
  - Offshore zones and county resolution are now compatible (they were not before)
We keep the old single-bin transmission upgrade supply curve (TSC) file for greenfield 500 kV with GSw_ZoneSet=z134 as an example of the required file format when using TSC, but will eventually remove it when we add a full supply curve
- The greenfield costs are put into this format automatically in transmission.py

Additional changes

Features that we no longer need with the new format have been removed:
- inputs/shapefiles/transmission_endpoints (now in inputs/zones/{GSw_ZoneSet}/zonehash.csv)
- Transmission aggregation functionality in copy_files.py
- aggregate_regions.py (Disaggregate and re-aggregate inputs with regional scope of legacy (134) zones #102 removed everything except for transmission aggregation from aggregate_regions.py; given the removal of transmission aggregation here, we no longer need the script)
- Zone aggregation functionality in reeds.io.get_zonemap()
transmission.py has been reorganized to make it importable
min_co2_spurline_miles (= 20): moved from hard-coded in transmission.py to scalars.csv

Switches added/removed/changed

GSw_TranScen: Allowable options changed to none (default, meaning only interfaces with existing capacity can be expanded), NTP_MT, or NTP_P2P

Issues resolved

Validation, testing, and comparison report(s)

I ran the full cases_test.csv suite and everything worked except for WECC_county, which failed in 2015, the same year in which it fails on the main branch.

Here's some background info and plots of the new costs: 20260515 - transmission costs distances.pptx

Compare reports are available in the same deck.

There's more LCC and VSC in the default solution since we've added existing/planned lines that were originally left out
The big line from Chicago to Michigan now goes to Indiana since we only include directly connected AC interfaces
I made sure the WECC_county case runs through 2015

Checklist for author

Details to double-check

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

…ogging style

…ines-{}.csv

…; only define co2_routes for r,rr pairs with nonzero distance

…thout an interfaces list

…ap() (no longer need to modify zone map in aggregate_regions.py)

wesleyjcole · 2026-05-27T14:42:08Z

-Network reinforcement costs are approximated by tracing a path along existing transmission lines from each wind/solar POI to each zone "center" within the same state;
-the zone center is usually taken as the largest population center in the model zone but is sometimes (for zones without large urban centers) assigned to a high-voltage substation within the zone.[^ref35]
-A cost for each reinforcement route is calculated using the cost surface described above, with capital expenditure (CAPEX) costs multiplied by 50% to approximate the lower cost for reconductoring compared to greenfield transmission construction.
+Network reinforcement costs are approximated by tracing a path along existing transmission lines from each wind/solar POI to a nearby urban center.


Can you quantify "nearby" and "urban center"? I can't remember the specifics, but this seems like a good place to document those.

Added some lines to describe the current approach.

wesleyjcole · 2026-05-27T14:50:04Z

@@ -0,0 +1,9 @@
+voltage_kv,kcmil,conductor_type,conductor_quantity,amp


What is kcmil?

It's a silly U.S. unit but it's the standard measurement of wire gauge: https://en.wikipedia.org/wiki/Circular_mil

wesleyjcole · 2026-05-27T14:54:36Z

-Cross Sound Cable,330,-72.90222222,41.28666667,-72.8675,40.95916667
-Neptune Cable,660,-73.55111111,40.76055556,-74.35308333,40.47371667
-Trans Bay Cable,400,-121.89666667,38.03083333,-122.38583333,37.75472222
+name,MW,year_online,trtype,certain,from_lon,from_lat,to_lon,to_lat


Is year_online the number of years until it is online? If so, it might be worth updating the name, as I was expecting something like "2028" for the value rather than "0".

Added a description to inputs/transmission/README.md and changed these to their actual historical online year

wesleyjcole · 2026-05-27T14:55:36Z

@@ -0,0 +1,3 @@
+name,MW,year_online,trtype,certain,from_lon,from_lat,to_lon,to_lat
+TransWestExpress,3000,2032,VSC,1,-107.176147,41.747242,-112.590781,39.541825


This one has year_online as what I would expect, so I think I'm not understanding this column correcting, or what a zero value for this column means.

wesleyjcole · 2026-05-27T17:43:36Z

+    ### TEMPORARY 20260402: Drop county interfaces with no distance/cost
+    if (level == 'r') and (sw.GSw_RegionResolution in ['county', 'mixed']):
+        transmission_line_fom = get_transmission_fom(case, interface_params)
+        indices = ['r', 'rr', 'trtype']
+        drop = (
+            dfout
+            .merge(transmission_line_fom.reset_index(), on=indices, how='left')
+        )
+        drop = list(drop.loc[drop.USDperMWyear.isnull(), indices].itertuples(index=False))
+        dfout = dfout.set_index(indices).drop(drop).reset_index()


Is this what is making WECC county runs fail (mentioned in your PR text)? Or does this make it not fail?

Yunzhi is nearly done with the WECC county bug fix, so it would be nice to not have it immediately broken again if that can be avoided, even if that meant a temporary patched solution until new reV data is available.

No, it's a bit different; this is to keep GSw_ZoneSet = z3109 (county-level, which includes counties with no transmission links to neighbors) working until we finish reformatting the spatial pipeline, which will allow use of GSw_ZoneSet = z2972 (which aggregates those counties into their neighbors and avoids making islanded zones).

The missing SunZia line at county resolution is caught in check_inputs.py. But I'll add it before merging.

The SunZia cost/distance has been added, so the WECC_county case now runs through 2015, the same as on main.

…Detail = 1

…ME.md

…tances

wesleyjcole

Thanks for the updated explanations and documentation.

kodiobika

Thanks for doing this! LGTM

kodiobika · 2026-06-04T16:17:49Z



-def get_hvdc_lines():
+def get_hvdc_lines(filename='hvdc_existing.csv'):


No action required here, but this got me wondering - is there a well-defined difference between reeds/inputs.py and reeds/io.py? My instinct would've been to put something like this in the latter but I'm not actually sure which, if any, makes more sense in this case

My very rough notion has been to use inputs.py for things that are "further up" in the pipeline (closer to the direct inputs/ data) or that require more processing, and io.py for interacting with intermediate files from the {case}/inputs_case folder. I'm sure I haven't been super consistent though (inputs.py is newer and I didn't port everything to it from io.py that would fit those guidelines.) I was also worried about io.py becoming too huge if we use it as a catch-all for everything, but I'm not sure how important that is.

Maybe we could do a standalone cleanup/organizational PR after we're done with all the spatial stuff? We'll need some more layers of organization anyway as we start moving b_inputs.gms into Python.

kodiobika · 2026-06-04T16:29:28Z



-def get_dfmap(case=None, levels=None, exclude_water_areas=False):
+def get_dfmap(case=None, levels=None, exclude_water_areas=True):


Just checking - the fact that all the instances throughout the code where get_dfmap is currently being used and exclude_water_areas isn't specified will now exclude water areas (where they were previously including them) is intentional right?

get_dfmap() until now has used the 134-zone shapefile for everything except county zones, and that shapefile excludes water areas. So I think excluding them by default is most consistent, but I didn't actually go through and check all the areas where water areas might matter.

Do you remember offhand the places where including water areas is intended and important?

Ah got it, thanks that SGTM. No instances come to mind (and from a cursory scan it looks like there are none) so I think we're good

kodiobika · 2026-06-04T16:59:46Z

@@ -0,0 +1,9313 @@
+start,end,polarity,voltage,cost_MUSD,length_miles


nit: It'd be nice for consistency to specify the voltage units too, but no need if it's too much of a hassle at this point

Yeah fair; I kept these because they're the column names for inputs to and outputs from the reV Routing (reVRt) model. I've added notes on the units to inputs/transmission/README.md; think that's enough for now?

Makes sense, yup that SGTM

kodiobika · 2026-06-04T17:00:37Z

@@ -0,0 +1,499 @@
+name,from_lon,from_lat,to_lon,to_lat,polarity,voltage,cost_MUSD,length_miles


nit: Ditto on voltage units ('voltage' -> 'voltage_kv')

kodiobika · 2026-06-04T17:15:12Z

+    interface_params['geometry'] = interface_params.apply(_make_line, axis=1)
+    interface_params = gpd.GeoDataFrame(interface_params, crs='EPSG:4326')
+    interface_params['straight_miles'] = (
+        interface_params.geometry.to_crs('EPSG:5070').length


Just curious, why 'EPSG:5070' here instead of the usual 'ESRI:102008'?

They're both equal-area projections, but EPSG:5070 is zoomed in more on the contiguous US, while ESRI:102008 is for all of North America (there are examples of the two at https://gis.stackexchange.com/a/457667). That means that within the CONUS, there's a bit less distortion in distances and areas using EPSG:5070 than ESRI:102008. (Eventually it'd be nice to change everything to EPSG:5070 and make it controlled by a global switch.)

…arams (avoids NTP_MT/NTP_P2P being used in place of SunZia)

patrickbrown4 added 30 commits April 15, 2026 10:52

copy branch from .gov repo

a536d1b

Merge branch 'main' into pb/transcost

c679d84

move scale_column() to reeds.output_calc to avoid propagating bokeh l…

b06f904

…ogging style

remove ba/county planned tranmission files and replace with planned_l…

0a3679f

…ines-{}.csv

transmission: propagate name changes

33920d0

transmission.py: move firm import limits to get_firm_import_limits()

760f3c8

transmission.py: use new data structures

620d6e2

add first newlinks_offshore_{}.csv files

46c0e37

add newlinks_offshore_radial.csv for other zone definitions

9a103c2

functify transmission.py

2ab5e98

remove transmission_endpoints shapefile

17df830

transmission.py: downselect to modeled regions; add leading *

0d3de04

move min_co2_spurline_miles to scalars; use DC distance for pipelines…

81c51b9

…; only define co2_routes for r,rr pairs with nonzero distance

add interfaces arg to inputs.get_interface_data() to allow reading wi…

c04e64a

…thout an interfaces list

define transmission_line_fom for both directions

5bb7d0e

include offshore links in transmission.py

e568e25

fix VA_W zonehash

d9410e4

validate_zoneset(): check for distances for all interfaces

04233e4

remove unused transmission aggregation options

3fd3d92

reeds.inputs.parse_regions(): convert to usual case and **kwarg inputs

44ce609

move get_zones() from reeds.inputs to reeds.io and simplify get_zonem…

e48d4b2

…ap() (no longer need to modify zone map in aggregate_regions.py)

add SunZia

3235e40

transmission.py: check for missing costs upfront; cleanup

f590f43

transmission_cost_distance.csv: add missing B2B connections

6d4f037

transmission.py: reorganize functions

a962c0f

add individual-line transmission costs & distances

c091f25

transmission.py: add missing docstrings

8adb7dc

Merge branch 'main' into pb/transcost

8b83e0b

retail rates: transmission_distance.csv -> transmission_miles.csv

b2b7cf7

transmission.py: don't add artificial p19-p20 link

ede6efa

patrickbrown4 added 2 commits May 22, 2026 12:18

Merge branch 'main' into pb/transcost

21971fb

Merge branch 'main' into pb/transcost

6603bad

patrickbrown4 mentioned this pull request May 22, 2026

Disaggregate and re-aggregate inputs with regional scope of legacy (134) zones #102

Merged

13 tasks

fix OR_water

cb02783

patrickbrown4 requested review from kodiobika and wesleyjcole May 22, 2026 21:28

patrickbrown4 added 2 commits May 26, 2026 09:53

fix reedsplots.plot_diff() for transmission

28503e7

speed up runstatus.py by avoiding reeds import

91e5f5d

wesleyjcole reviewed May 27, 2026

View reviewed changes

patrickbrown4 added 8 commits May 27, 2026 13:45

Merge branch 'main' into pb/transcost

91acfcf

runreeds.check_compatibility(): allow county resolution with GSw_CO2_…

e4c32cd

…Detail = 1

rename hvdc files for clarity; describe contents in transmission/READ…

8d3f490

…ME.md

Merge branch 'main' into pb/transcost

e2a16a6

remove aggregate_regions.py

3f1efa4

add transmission cost+distance for SunZia and z70

f34d31c

update docs: describe urban centers; describe how to update costs/dis…

a905f83

…tances

Merge branch 'main' into pb/transcost

bd19514

patrickbrown4 requested a review from wesleyjcole June 1, 2026 20:37

wesleyjcole approved these changes Jun 3, 2026

View reviewed changes

Merge branch 'main' into pb/transcost

2be92d1

kodiobika approved these changes Jun 4, 2026

View reviewed changes

patrickbrown4 added 6 commits June 4, 2026 12:33

cleanup from review

1283792

more cleanup; add missing files to inputs/transmission/README.md

a4d19b7

stop referring to n-0/n-1 capacity (it's all n-1)

1894171

transmission.py: only include lines from GSw_TransScen in interface_p…

5ae5dbc

…arams (avoids NTP_MT/NTP_P2P being used in place of SunZia)

more cleanup

64b728b

prasplots.plot_interface_flows(): fix DC delimiter

5d79a9d

patrickbrown4 merged commit 62e6948 into main Jun 5, 2026
10 checks passed

patrickbrown4 deleted the pb/transcost branch June 5, 2026 16:20

patrickbrown4 mentioned this pull request Jun 5, 2026

Calculate greenfield transmission costs for all zone geometries directly #37

Closed

33 tasks

		@@ -0,0 +1,9 @@
		voltage_kv,kcmil,conductor_type,conductor_quantity,amp

		@@ -0,0 +1,3 @@
		name,MW,year_online,trtype,certain,from_lon,from_lat,to_lon,to_lat
		TransWestExpress,3000,2032,VSC,1,-107.176147,41.747242,-112.590781,39.541825



		def get_hvdc_lines():
		def get_hvdc_lines(filename='hvdc_existing.csv'):



		def get_dfmap(case=None, levels=None, exclude_water_areas=False):
		def get_dfmap(case=None, levels=None, exclude_water_areas=True):

		@@ -0,0 +1,9313 @@
		start,end,polarity,voltage,cost_MUSD,length_miles

		@@ -0,0 +1,499 @@
		name,from_lon,from_lat,to_lon,to_lat,polarity,voltage,cost_MUSD,length_miles

Conversation

patrickbrown4 commented May 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Technical details

Implementation notes

Least-cost paths

Spatial flexibility

Additional changes

Switches added/removed/changed

Issues resolved

Validation, testing, and comparison report(s)

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!

Uh oh!

Uh oh!

Uh oh!

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!

Uh oh!

wesleyjcole left a comment

Choose a reason for hiding this comment

Uh oh!

kodiobika left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

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!

Uh oh!

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!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

patrickbrown4 commented May 15, 2026 •

edited

Loading