w-flo / eu-emission-factors Goto Github PK

Hey, figured it would be best to create a new issue here to discuss this particular topic :)

I've been reading through the Rust code, but since it's not my speciality (in any way 😅), I'm struggling a bit with understanding how your manual matching works for the particular case where there are multiple names (separated by |) for both generation and emission.

How do you know what generation unit belongs to which emissions entry?

As an example, I am looking at Karlshamnsverket in Sweden. Your CSV row looks like this:

But the 2 emission entries have very different values:

And the entries from ENTSO-E:

According to https://en.wikipedia.org/wiki/Karlshamn_Power_Station I would expect both should map to the10-563-005832-2004:138 (and not the gas turbine thing), but if you already have an approach of how to map these it would be great :)

I'm not sure if this affects other power plants, but at least Knapsack 1 and 2 seem to be missing some generation data.

For example, Knapsack 2 generated some electricity on 2023-12-06 according to energy-charts, but it's 0 according to ENTSO-E. The yearly sum doesn't seem to add up for Knapsack 1, too: Unit 1 only has ~70% of energy-charts total generation, Unit 2 only has 40% of energy-charts total generation. And even then, emissions seem way too high for Knapsack 2 if the plant was really shut down between February and October.

Dear @w-flo ,

thanks a lot for your great work. Unfortunately it has not been integrated so far. I'd like to discuss with you how we can get these well-calculated emission factors to a broader audience and if possible inside the EM repo.
Would you be interested in such a discussion?

Best
Siggi

As noted in electricitymaps/electricitymaps-contrib#5417 (comment), some plants should be able to pass the plausibility check even when they have an emission factor < 300, because they are not actually (100%) using the fuel they claim to use.

These might be powered by a combination of biomass and fossil fuels:

• Rya KVV
• Amercentrale

This needs some kind of manual override, probably in manual_matches.csv.

Eurostat hasn't updated this dataset yet for 2023: https://ec.europa.eu/eurostat/de/web/products-datasets/product?code=nrg_chdd_a

So the current emission factors for combined heat and power plants in the 2023 output are probably even worse than they are for previous years.