add unit commitment for links

PyPSA / pypsa-ariadne

High resolution, sector-coupled model of the German Energy System

https://ariadneprojekt.de/en/model-documentation-pypsa/

MIT License

6 stars 4 forks source link

add unit commitment for links #76

Closed JulianGeis closed 1 week ago

JulianGeis commented 2 months ago

Coal is operated too flexible in PyPSA-DE.

This PR adds unit commitment on different links including [OCGT, CCGT, nuclear, coal, lignite, oil, urban central solid biomass CHP]

lindnemi commented 2 months ago

if UC is only active for coal, how much compute time does this PR add?

JulianGeis commented 2 months ago

Solving time in s for the networks 2020-2045 are (6171 + 8275 + 13474 + 6970 + 5280 + 6249) with UC and (2539 + 4299 + 8559 + 4490 + 4429 + 5494) without UC. So in general it adds about 55% on top of the solving time, which is about 5h.

lindnemi commented 2 months ago

quite a lot for such a minor part of the model :/ Could you post the primary_energy_detailed.png from both runs here? Would be nice to see what's the impact on the outputs.

And could you please benchmark what happens when only start up costs are set?

lindnemi commented 2 months ago

(mainly in terms of solving time, but also outputs)

JulianGeis commented 2 months ago

Effects for UC on coal and lignite in DE in 2020:

coal

reduction of electricity production from coal by 4 % from 40.25 to 38.71 TWh
reduction in variance of elec production from coal from 6 - 12 %

lignite

increase of electricity production by lignite by 1 % from 170.29 to 171.55 TWh

plots

you do net see any difference in the plots

lindnemi commented 2 months ago

@nworbmot

UC does not seem to influence the results significantly. So maybe we should not use it after all? Would you have other ideas for how to reduce coal production in 2020?

lindnemi commented 1 week ago

validation for 2020 is now fine without UC, so i close this for now