Implausibly High Capacity of H₂/Gas-Fueled Generators in Results (KN2045_Bal_v4, 3H resolution)

[mohsenmansouri]

I faced an issue while running the model with the KN2045_Bal_v4 scenario at a 3-hour time resolution. In the output for both 2040 and 2045, the result shows a total installed capacity of over 250 GW for H₂/gas-fueled generators in 2040 and 2045, despite these generators contributing only a small share of total energy generation.

I would appreciate any guidance on whether:

• this is a known behavior in this scenario/resolution combination,

• there are parameter settings or constraints I should review (e.g., availability, fuel cost, or emission limits),

• or if this could stem from input assumptions in the scenario itself.

Thanks!

[lindnemi]

With #81 the workflow for public and TU Berlin internal users is now the same. We never saw the issue you describe in our internal runs. The public runs should now use the same data. Could you please re-run your model with the updated workflow and confirm that the issue persits?

[mohsenmansouri]

Thanks for the feedback.

Today I cloned the repository again and ran the model from scratch using the following commands:

snakemake build_scenarios -f
snakemake ariadne_all

When using resolution_sector: 365H, the total installed capacity for H₂ and gas-fueled generators was around 80 GW, which seems plausible.

However, when I changed the resolution to 3H, the total installed capacity for H₂ and gas increased significantly to 296 GW. I've shared a snapshot of the output nodal_capacities.csv file showing this result.

cluster			27	27	27
opt			none	none	none
planning_horizon			2030	2040	2045
Link	DE	CCGT	10469.94655	8590.7931	7149.63793
Link	DE	OCGT	7207.11502	56.85366	56.85366
Link	DE	urban central gas CHP	28356.59431	28356.59431	28356.59431
Link	DE0 0	H2 OCGT	0.15998	12816.12384	12816.124
Link	DE0 1	H2 OCGT	0.16791	45516.20769	45516.20787
Link	DE0 2	H2 OCGT	0.16801	90709.63548	90709.63567
Link	DE0 3	H2 OCGT	0.16802	37539.52348	37539.5237
Link	DE0 3	H2 retrofit OCGT		1729.48921	1729.48921
Link	DE0 4	H2 OCGT	0.17123	32041.45444	32041.4546
Link	DE0 5	H2 OCGT	0.1687	9803.88503	9803.8852
Link	DE0 6	H2 OCGT	0.17593	21811.45481	21811.45496
Link	DE0 6	H2 retrofit OCGT		5399.70585	5399.70585
Link	DE0 7	H2 OCGT	0.16066	3719.00306	3719.00324

[lindnemi]

Thanks, that looks really strange, i will take a look

[lindnemi]

These csvs are based on the n.statistics.optimal_capacity() function, which returns the capacity from bus0 to bus1. For these technologies this is the amount of gas or hydrogen that they can consume. This is much bigger than their electrical/heat output capacity. Please take a look at

ariadne_comparison/Capacity-Electricity-Gas.png or at the corresponding variable in ariadne/exported_variables.xlsx, which should contain the correct electrical capacities.

[lindnemi]

@mohsenmansouri is this issue solved?

[mohsenmansouri]

Yes, sorry for late reply. Thanks.