search

PyPSA / technology-data

Compiles assumptions on energy system technologies (e.g. costs and efficiencies) for various years.
https://technology-data.readthedocs.io
GNU General Public License v3.0
50 stars 37 forks source link

electrolysis investment costs #109

Closed lisazeyen closed 2 months ago

lisazeyen commented 8 months ago

Investment cost assumptions of electrolysis from DEA

Investment cost assumptions of DEA are very optimistic, especially for the near-term. Cost could be adjusted to a smaller plant size (1 MW), and the upper uncertainty bound from DEA.

fvdborre commented 8 months ago

In a recent BCG report the cost projections of H2 is increased because of a more realistic cost estimation based on "complex 'real-life' electrolyzer systems incl. BOP,a many components with limited cost reduction potential".

image https://media-publications.bcg.com/Turning-the-European-Green-H2-Dream-into-Reality.pdf

Furthermore, a Dutch study finds also much higher projected investment costs. image Source: https://energy.nl/wp-content/uploads/tno-2022-p10111_detzweeda_projections-of-electrolyzer-investment-cost-reduction-through-learning-curve-analysis.pdf

p-glaum commented 8 months ago

image from https://zenodo.org/records/10101328. Reference cost for a 500MW unit

euronion commented 8 months ago

Mind this refers to PEM electrolysis, while we are usually (and in the DEA data) assuming Alkaline

martavp commented 8 months ago

They also take higher cost assumptions in the H2 network paper from DTU using Balmorel, which is one of the reasons (not the only one!) that makes SMR-CC competitive in their results.

cost_Kountouris

I added this here for completeness but I'm not fully sure about taking DEA assumptions for 1MW.

fvdborre commented 8 months ago

In a Dutch rapport used to determine the height of the financial support ( SDE++) they assume an investment support of 1800€/kW.

image https://www.pbl.nl/sites/default/files/downloads/pbl-2021-conceptadvies-sde-plus-plus-2022-waterstofproductie-via-elektrolyse-4392.pdf

Many countries have a hydrogen strategy or a hydrogen support system. I assume the figures used in the context of support schemes should be reliable sources of information for the current costs. I believe figures provided by technology suppliers tend to underestimate the costs.

lisazeyen commented 7 months ago

My suggestion

Checked sources (including the above mentioned)

BCG report

TNO report 2022 (Amsterdam, Netherlands)

From the Odenweller paper plot created from me based on input excel

electrolyser_data_odenweller

Dutch report

today 1800 Eur/kW

IEA Global Hydrogen Review 2023

IEA world energy outlook 2023

p.305

iea_world_outlook

Comparison investment cost in SI for the learning paper)

electrolysis_SI_learningpaper

DEA for small 1 MW plant p.128

fvdborre commented 7 months ago

The IEA places a significant emphasis on announced electrolyzer plant projects, yet many of these announcements were made before the necessary policies were in place to support such initiatives. Consequently, these announcements do not result from policy support but rather act as drivers for policy development. The primary objective behind unveiling large electrolyzer plant projects is not necessarily their construction but rather the influence they wield on policy direction. Consequently, it is advisable to disregard the announcements and focus solely on projects with a Final Investment Decision (FID). This approach is crucial for maintaining accuracy in IEA's cost projections. image

When extrapolating historical values for installed electrolyzer plant capacity, a limited dataset yields the following trend. While four data points may seem scant, it provides a more reliable foundation than relying on announced projects. image image

Assuming a long-term goal of building 1000 GW electrolyzer capacity (sufficient to meet current hydrogen demand), a logistic curve emerges with the following parameters:

image

This gradual deployment based on a natural growth makes more sense than the IEA assumption for two primary reasons. Firstly, it ensures that electrolysers are acquired at a more cost-effective rate. Secondly, the investment aligns with a period when electricity costs are lower due to the increased integration of renewables into the grid.

image

martavp commented 7 months ago

For completeness, the DEA is currently updating their cost assumption for electrolyzers. They have asked for feedback by Dec. 4 so the updated costs will be released soon.

For 100 MW AEC, the new capital cost assumptions are: 1200 EUR/kW (2020), 875 EUR/kW (2025), 575 EUR/kW (2030), 300 EUR/kW (2050)

These are slightly higher than your suggestion @lisazeyen, but could have the benefit of reading data from the same database.

assume IEA cost assumptions today (1700 USD/kW), 2025 (850 US/kW), 2030 (680 USD/kW) (from IEA Global Hydrogen Review 2023 and 2050 (mean 400 USD/kW) from IEA world energy outlook 2023

I'm also fine with just sticking to your suggestion @lisazeyen

euronion commented 7 months ago

Great to hear DEA is updating the numbers, thanks for letting us know @martavp ! I'd favour us sticking with the DEA numbers if they become available soon.

Thanks @fvdborre for your detailed analysis! In addition to the purpose of announcements you describe, I also see them as a means to increase production capacities and trigger economies of scale (in an industry where until recently production was more artisanal than industrialised). An S-curve approach with an assumed learning rate does help for a harmonized framework (see your comment in the other issue), does require assumptions on target capacities, growth rate, learning rate and a starting point (investment costs today) - all of which I'd not feel comfortable in assuming, so we'd need to stick with IEA numbers again anyways.

fvdborre commented 7 months ago

The issue at hand arises from discrepancies between the figures provided by the IEA and the communication of the Hydrogen Science Coalition, found at https://h2sciencecoalition.com/ . Given this inconsistency, it is evident that at least one of the two parties must be in error.

One notable discrepancy pertains to the Compound Annual Growth Rate (CAGR) implied by the IEA's figures. A global installed capacity of 175 GW in 2030 implies a staggering 106% CAGR, a figure significantly higher than the historical electrolyzer CAGR. This projection appears unrealistically high, especially when compared to the CAGR of renewable technologies in Flanders.

image

Upon further reflection, it occurred to me that the technology behind alkaline electrolysis is not novel; it is rooted in the chloralkali process. This process is well-established, with an annual production of 100 Mt of chlorine, 83% of which utilizes membrane technology. It is crucial to note that the purported cost reductions do not accurately reflect the advancements and experience already achieved with membrane chloralkali technology.

I apologize for my critical assessment. In my experience, hydrogen is frequently employed as a delay tactic, deflecting attention from more impactful and effective measures.

euronion commented 7 months ago

Critical assessments are good as long as they are productive, no need to apologize.

As for IEA or the H2 sci coalition - they are likely both wrong. Which is okay, we have to see them in their specific contexts. A worthwhile read which looks at different growth rates is this paper by Odenweller et al (2022): https://www.nature.com/articles/s41560-022-01097-4

The technologies are similar but not identical.

Anyway, we are heading offtopic. How about you have a look at the revised numbers by the DEA as soon as they are out and we halt the discussion for now?

martavp commented 5 months ago

New cost assumptions for electrolyzers from DEA just published

https://ens.dk/sites/ens.dk/files/Analyser/technology_data_for_renewable_fuels.pdf https://ens.dk/sites/ens.dk/files/Analyser/data_sheets_for_renewable_fuels.xlsx

For 100 MW AEC, the new capital cost assumptions are: 1200 EUR/kW (2020), 875 EUR/kW (2025), 550 EUR/kW (2030), 300 EUR/kW (2050)