[asystemoffields]

Closes #247

Changes proposed in this Pull Request

The CO2 liquefaction cost assumptions came from a 2004 Mitsubishi Heavy Industries/IEA ship-transport report and the investment cost was about five orders of magnitude too low: 12.89 EUR2004 per (tCO2/h), i.e. about 21 EUR/(tCO2/h) in the 2025-EUR outputs, where the Danish Energy Agency gives 1.33 million EUR2020 per (tCO2/h). As suggested in #247, this PR substitutes the rows in inputs/manual_input.csv with the values from the DEA "Technology Data for Carbon Capture, Transport and Storage" datasheet 411 Liquefaction biogas (datasheets updated February 2024; the sheet was added on 2023-02-11; cost data in 2020 EUR):

parameter	before (MHI/IEA 2004)	after (DEA 411)
investment	12.89 EUR2004/(tCO2/h) → 21.28 EUR2025	1.33 million EUR2020/(tCO2/h) → 1,672,255 EUR2025
FOM	5 %/year	4 %/year (note H: "Fixed O&M is taken as 4% of CAPEX"; the sheet lists 0.05 mill EUR/(tCO2/h)/yr, 3.8 % after rounding)
lifetime	25 years (guesstimate)	20 years (technical lifetime)
electricity-input	0.123 MWh/tCO2	0.137 (2020), 0.133 (2025), 0.129 (2030), 0.125 (2040), 0.122 (2050) MWh/tCO2: compression + drying + liquefaction
heat-input	0.0067 MWh_th/tCO2 (drying)	0 (drying is electric in the DEA sheet, 2 kWh/t, included above)
carbondioxide-input	1 t/t	unchanged

The DEA entry is a 2 tCO2/h compression-and-liquefaction plant for a biogas retrofit (the minimum cost-effective size per note A; lower/upper estimates 1.06/1.6 million EUR/(tCO2/h)); it includes the compression train, pre-coolers, drier, refrigeration and installation, but not gas cleaning (note G). It is the only liquefaction entry in the catalogue, so large terminal-scale plants are not represented; happy to adjust if you prefer a different scaling.

Outputs were regenerated with snakemake -c1 --forceall compile_cost_assumptions compile_cost_assumptions_usa (pandas 2.3, as pinned); on master the same command reproduces the committed outputs exactly, and with this change only the CO2 liquefaction rows differ (EU and US files). I kept heat-input as an explicit 0 rather than dropping the row so that downstream lookups keep working, and kept the unit string EUR/t_CO2/h.

Not done here: adopting the Feb-2024 CCTS workbook itself (the bundled one is the 2021 version). Its DAC sheets were renamed and split, so that update would also change the DAC numbers and the sheet mapping; it would make sense as a separate PR.

Checklist

• Code changes are sufficiently documented; i.e. new functions contain docstrings and further explanations may be given in doc. (data only; sources and notes are in the csv)
• Data source for new technologies is clearly stated.
• Newly introduced dependencies are added to environment.yaml (if applicable). (none)
• A note for the release notes doc/release_notes.rst of the upcoming release is included.
• I consent to the release of this PR's code under the GPLv3 license.

(Prepared with Claude Code (Claude Fable 5.1); the values were cross-checked against DEA datasheet 411 during preparation, and the PR was reviewed by the author before submission.)

🤖 Generated with Claude Code

[euronion]

Thanks for raising this PR!

Two things:

1. Looking at the MHI numbers, I think this is simply a mistake in converting the numbers. 80 MUSD/t/a for a 20 kt/d plant should be 96 kUSD/t/h , not what we have at the moment.
2. Your numbers are as you write for a different scale, different application.

Would you be fine to adjust this PR?

a) Rename your technology selection from the DEA to something like "CO2 liquefaction small scale" and keep your values.
b) Fix the original "CO2 liquefaction" value with the correct conversion from USD/t/d to EUR/t/h?

[euronion]

Suggested change

* Replace the CO2 liquefaction cost assumptions (investment, FOM, lifetime, electricity and heat input) from a 2004 Mitsubishi Heavy Industries/IEA report with the Danish Energy Agency's Technology Data for Carbon Capture, Transport and Storage (datasheet 411, 2020 EUR). The investment cost was about 5 orders of magnitude too low (about 21 EUR/(tCO2/h) instead of about 1.7 million EUR/(tCO2/h) in 2025 EUR). (https://github.com/PyPSA/technology-data/pull/289)

* Update the 'CO2 liquefaction' assumptions with the Danish Energy Agency's Technology Data for Carbon Capture, Transport and Storage. Investment costs increase by orders of magnitude (https://github.com/PyPSA/technology-data/pull/289).

[euronion]

We like to keep the release notes concise and without specific numbers.

[euronion]

Add the direct link instead https://ens.dk/media/5729/download (more permanent)

Please avoid "" in the format, it confuses some editors for csv formats. Use single quotes ' or no quotes instead.

[asystemoffields]

Thanks for the review, both points make sense. Done in fcbd101:

• CO2 liquefaction keeps the original MHI/IEA 2004 entry with the conversion corrected. The report gives a construction cost of 80,000,000 US$ for a design rate of 20,000 t/d (p. 11), i.e. 833 t/h, so 96,000 US$ per (t/h); the old 12.89 had divided the cost by the annual throughput (20,000 t/d × 365 d × 85 % running rate). I converted at the ECB 2004 average rate (1.2439 US$/EUR) to 77,177 EUR(2004)/(t/h), which the pipeline inflates to about 127 kEUR(2025)/(t/h). I did the currency step in the CSV because add_manual_input does not run clean_up_units: the four NPC rows with USD/(tCO2/h) come out of the pipeline still labelled USD, only inflated. That looks like a separate issue; happy to open one if you agree.
• CO2 liquefaction small scale is the DEA datasheet 411 entry (2 t/h biogas retrofit), with the direct link https://ens.dk/media/5729/download and no escaped double quotes in the CSV (inner quotes are now single quotes).
• Release note shortened to your suggestion's style, without numbers.

Outputs regenerated (only the two liquefaction technologies change); pytest test: 52 passed.

(Comment, like the PR, prepared with Claude Code on the account owner's behalf.)