search

cpmodel / FTT_StandAlone

Future Technology Transformation models
GNU General Public License v3.0
9 stars 1 forks source link

Investigate differences FORTRAN and Standalone FTT:P #124

Open Femkemilene opened 2 months ago

Femkemilene commented 2 months ago

Ian and me recently discovered a mistake in the CSC input files, which we've corrected. After correction:

In the 2022_Baseline_update FORTRAN version, the 2050 global solar generation is 38 PWh. This is only 13 PWh in the Standalone main version. Given that neither have had the new data update, the values should be much closer together.

jp-camecon commented 2 months ago

After a quick look, I think a good place to look at is calculation of MEWI. Particularly end of life replacements. This is done differently in the Fortran version @pv-camecon @rhayward1 would know more about the particulars as this came in as part of iter R&D but I can clearly see that MEWI does not line up even in history in both models and will have implications for learning via cumulative capacity additions (MEWW)

Femkemilene commented 1 month ago

Another difference: MGAM for solar PV is changed in the idiom file from 2019.. This should be changed to the excel sheet before the baseline_2022_update is made into the Main Fortran branch. I understand the need to stop too much solar coming in, but tweaking gamma values post-hoc isn't a great way. Instead, one can set the lead time to a higher value to mimic grid constraints.

For now, I've change the gamma values to align in SA (new branch: https://github.com/cpmodel/FTT_StandAlone/tree/explore_differences_fttp).

The carbon tax values are different between the two model versions. I believe @ldxib2 has some script to convert between the two versions, right? The FORTRAN version seem to be the updated version of the carbon tax of SA.

The pre-2018 major differences in METC could be explained by a bug in the fuel tax/subsidy variable, which is now fixed. This means we now have 23 PWh of solar, vs 38 PWh in the FORTRAN version.

Femkemilene commented 1 month ago

Further key difference is MSSP & MLSP (per-technology storage costs). This is roughly 50% higher in the Standalone version, which I believe explains the remaining pre-2018 differences in METC. @pv-camecon: do I remember correctly that something changed with this variable due to exchange rates / inflation recently?

Femkemilene commented 1 month ago

Update: there is now a PR for the MEWI differences :).

Femkemilene commented 1 month ago

Almost there: we're now at 31 PWh for the SA and removing some odd regulation on onshore and offshore wind in the FORTRAN version at 35.5 PWh there. Shall I commit these removed regulation to the Master @pv-camecon?