restrict definition of q37_feedstockShares

mellamoSimon commented 7 months ago

Purpose of this PR

restrict the definition of q37_feedstockShares to relevant sets.

Type of change

[x] Bug fix

Checklist:

[x] My code follows the coding etiquette
[x] I performed a self-review of my own code
[x] I explained my changes within the PR, particularly in hard-to-understand areas
[x] I checked that the in-code documentation is up-to-date
[x] I adjusted the reporting in remind2 where it was needed
[x] I adjusted forbiddenColumnNames in readCheckScenarioConfig.R in case the PR leads to deprecated switches
[x] All automated model tests pass (FAIL 0 in the output of make test)
[x] The changelog CHANGELOG.md has been updated correctly
Test runs are here: /p/tmp/simonlei/this-is-remind/remind/output/ and here /p/tmp/pehl/Remind/output/SSP2EU-NPi_2024-03-06_11.31.54 with a working input gdx.

mellamoSimon commented 7 months ago

test runs are running in the above mentioned folder

0UmfHxcvx5J7JoaOhFSs5mncnisTJJ6q commented 7 months ago

@mellamoSimon, do https://github.com/mellamoSimon/remind/pull/8. What is the purpose of q37_FossilFeedstock_Base?

mellamoSimon commented 7 months ago

What is the purpose of q37_FossilFeedstock_Base?

that was set up by @fschreyer to prevent the model from using biomass or synfuels in baseline runs. However, I think that now that might be unnecessary because we improved emissions accounting (including incineration) and we are not supposed to use baseline anymore...

mellamoSimon commented 7 months ago

I'm halting this merge, cause the tests failed for NPi

0UmfHxcvx5J7JoaOhFSs5mncnisTJJ6q commented 7 months ago

I'm halting this merge, cause the tests failed for NPi

:point_right: https://github.com/mellamoSimon/remind/pull/8

mellamoSimon commented 7 months ago

I strated new tests in the same folder

0UmfHxcvx5J7JoaOhFSs5mncnisTJJ6q commented 7 months ago

The current REMIND input gdxes are thoroughly broken, so I started a second test with the previous input gdx (c1256714220e99250f791cc35e61ae51d3bdc7da) at /p/tmp/pehl/Remind/output/SSP2EU-NPi_2024-03-06_11.31.54.

mellamoSimon commented 7 months ago

Quick update as Michaja is out of the office for the rest of this week. Below is the summary for this run:

The current REMIND input gdxes are thoroughly broken, so I started a second test with the previous input gdx (c1256714220e99250f791cc35e61ae51d3bdc7da) at /p/tmp/pehl/Remind/output/SSP2EU-NPi_2024-03-06_11.31.54

So I'm not sure if I should merge yet. Maybe a new calibration will fix things but I don't know about the workflow in this case: should we re-calibrate and test first and then merge? Thanks!

REMIND run finished!

Model summary:
  gams_runtime is 12.9 hours.
  full.gms exists, so the REMIND GAMS code was generated.
  full.log and full.lst exist, so GAMS did run.
! abort.gdx exists, a file containing the latest data at the point GAMS aborted execution.
  non_optimal.gdx exists, because iteration 69 did not find a locally optimal solution. modelstat: 6 (Intermediate Infeasible)
  fulldata.gdx exists, because iteration 70 was successful. modelstat: 2 (Locally Optimal)
  Modelstat after 70 iterations: 2 (Locally Optimal)
  full.log states: *** Status: Execution error(s)

Infeasibilities extracted from full.lst with nashstat -F:
itr      region      solvestat               modelstat                   resusd      objval           
  6      CHA        NORMAL COMPLETION         LOCALLY INFEASIBLE      544.922     64.0247890145572   F
... 3 infeasibilities omitted, show all with nashstat -a ...
 69      MEA     TERMINATED BY SOLVER    INTERMEDIATE INFEASIBLE     1038.762    -37173.3285338495   F
If infeasibilities appear some iterations before GAMS failed, check nashstat -a carefully.
The error that stopped GAMS is probably not the actual reason to fail.

Failed markets in fulldata.gdx:
 - peoil: 2005, 2010, 2015, 2020, 2025, 2030, 2035, 2040, 2045, 2050 ...
 - pegas: 2005, 2010, 2015, 2020, 2025, 2030, 2035, 2040, 2045, 2050 ...
 - pecoal: 2005, 2010, 2015, 2020, 2025, 2030, 2035, 2040, 2045, 2050 ...
 - good: 2005, 2010, 2015, 2020, 2025, 2030, 2040, 2045, 2050, 2055 ...
No failed tax convergence in fulldata.gdx.

Renato-Rodrigues commented 7 months ago

The change in this PR fixes a clear bug that is currently active in the REMIND code. I would say that if the changes do not cause pre-triangular issues, and if in the runs all industry equations are limited to the right dimensions, we will be in a better shape to figure out which other issues are holding back the REMIND convergence with this merged than not. So I would say, go ahead and merge the changes here, as long as the points I raised above have no issue.

mellamoSimon commented 7 months ago

[ FAIL 0 | WARN 0 | SKIP 6 | PASS 80 ] :)

remindmodel / remind