Talk:Green hydrogen

Summarize Timeline Fact Check

In this edit by @HydrogenEagle added 13 bogus sources. The DOIs do not exist and the titles do not appear in Google Scholar. No authors are listed. I removed these sources but left the content for now. Hopefully correct sources will be added. Johnjbarton (talk) 16:21, 23 June 2025 (UTC)

Thanks for catching that and thanks to Ita140188 for reverting the bogus content. False sources are absolutely unacceptable. I will give HydrogenEagle a warning. Clayoquot (talk | contribs) 20:04, 23 June 2025 (UTC)

I don't think the solution of 'Green hydrogen' page to be reverted to June 2023 after being hardly edited after two years; mistakes in the references can happen; but instead of deleting the entire edit (which has taken precious time to do), you may notify me to fix the errors and allow sufficient time for that. HydrogenEagle (talk) 03:26, 24 June 2025 (UTC)

No I don't think edits based on fantasy references can be relied on even if we ^{[citation needed]} bomb them. Greglocock (talk) 03:59, 24 June 2025 (UTC)

This excuse seems disingenuous to me:

• "mistakes in the references can happen"

All 13 sources were looked "ok" with reasonable titles and DOI numbers. All 13 were bogus, seemingly invented by combining titles and numbers without any actual sources. One or two misspellings or numbers would be a "mistake in a reference". This is something else. I did not revert the content but it seems to me that if the sources are fabricated the content is also invented. Johnjbarton (talk) 15:20, 24 June 2025 (UTC)

I reverted the content, and I think this should be automatic in these cases. I don't have the authority for this, but in my opinion I would also block the offending user, since these additions are extremely dangerous for Wikipedia, introducing content that seems correct and well-written while being inaccurate or even completely invented and without sources. Ita140188 (talk) 06:46, 25 June 2025 (UTC)

That seems harsh but I agree. It should be a standard part of a user's induction (such as it is ) that they agree that will not use LLMs to generate references. Greglocock (talk) 23:41, 25 June 2025 (UTC)

I will propose a topic block for the editor if this happens again. However, editors should not be asked to agree to any limitations on tools they choose to use. Not only are these limitations unenforceable but they are not aligned with any Wikipedia goal. Edits are the only things we can check and we should only take action against bogus content or sources in edits. Johnjbarton (talk) 00:54, 26 June 2025 (UTC)

@Johnjbarton So there is encouraging use of citationbot, and sources like Worldcat to verify citations being added. I use AI a lot to format citations but it almost always messes up isbn, oclc and doi. So the editor needs to verify citations before they commit them. Duncnbiscuit (talk) 01:10, 26 June 2025 (UTC)

Yes editors are always responsible for every single edit. Wikipedia relies on a lot of software based analysis tools and the AI tools will just be another one after people understand their limitations. Johnjbarton (talk) 01:16, 26 June 2025 (UTC)

Independently of what tool has been used, adding content with references that do not exist should be considered vandalism in my opinion. Same as adding content referenced to real sources that do not support at all what's said in the text. In summary, faking references is very bad. Ita140188 (talk) 06:41, 26 June 2025 (UTC)

And "it's not me, it's AI hallucinating" is not a viable excuse. As mentioned before, editors are entirely responsible for their edits. Ita140188 (talk) 06:43, 26 June 2025 (UTC)

Summarize Timeline Top Qs Fact Check

In my opinion, current prices are intrinsically inappropriate news. Economic analysis of long term trends or fundamental shifts would be encyclopedic, but the currency values are not. Other opinions? Johnjbarton (talk) 18:08, 16 July 2025 (UTC)

Tricky one. The literature is full of boosterism, the reality is that green hydrogen projects are being cancelled at a huge rate - something like 99% of Australia's green hydrogen projects have been cancelled (or variations thereof). A comparison of predicted prices vs actual price trends would be useful, unboosterish, and of course insufficiently woke to get published. Greglocock (talk) 03:13, 17 July 2025 (UTC)

https://hydrogen-central.com/green-hydrogen-prices-will-remain-stubbornly-high-for-decades/ is at least a start Greglocock (talk) 03:37, 17 July 2025 (UTC)

@Johnjbarton any content that quickly goes stale isn't a good choice for an article. as pointed out in the WP:MOS use of words like current is discouraged. the usual work around being 'as of date', which pins the assertion to a historical point in time. time series data trends and dynamically updating content using wikidata templates might work to address content going stale if there is a suitable source. For example, population data for a suburb article where the data is drawn from the local government stats body.

I also often strip out future focused content, for example, infrastructure planned by one government but cancelled by its successor. Again, you could focus on the historic decision point and add the contradictory decision point when it has happened. Duncnbiscuit (talk) 09:32, 17 July 2025 (UTC)

Cost is a major aspect of green hydrogen and there are plenty of scholarly, non-news sources sources about it, e.g. . I've been working on a rewrite of our H2 cost sections but with my kids out of school for the summer I'm afraid I'm making slow progress. If anyone wants to take a stab at improving this stuff, please do. Clayoquot (talk | contribs) 20:57, 17 July 2025 (UTC)

Thanks, btw your links need "[link space text]" or bare link.

Johnjbarton (talk) 01:59, 18 July 2025 (UTC)

There seems to be a consensus to avoid content that quickly go stale, including announcements about projects that haven't been built. I think some main points to get across are:

1) There is a large cost gap between green and gray hydrogen. I put in some currency numbers to convey what the gap is like. I wouldn't mind expressing this more simply, but a single number like "approximately 3x more expensive" doesn't seem neutral when the gap is sometimes described as "six times more expensive".

2) The cost gap between green and blue hydrogen should be added somewhere.

3) Why green H2 costs more, e.g. electrolysers and electricity.

4) The cost of green H2 is expected to fall due to economies of scale and technology improvement. We can leave out the thorny question of how much costs will fall. Clayoquot (talk | contribs) 23:32, 25 July 2025 (UTC)

Point four is up for grabs, the plants are so expensive that they need 24/7 power, hence energy costs are high Greglocock (talk) 19:54, 26 July 2025 (UTC)

Is there a source that says green H2 prices are not expected to come down? Clayoquot (talk | contribs) 03:31, 27 July 2025 (UTC)

I doubt it the MSM are mostly just quoting press releases and even I doubt anyone would launch a full production process that is more expensive than the current lab scale demonstrations . However the need for base load power is a cost that the promoters have only recently revealed, and will increase the cost . Greglocock (talk) 07:14, 27 July 2025 (UTC)

Point four is a mainstream point of view which can be attributed to the IEA. Predictions are opinions, not facts, so I'm fine with explicit in-text attribution. If there's a point of view that says costs won't come down, we could give due weight to that point of view but obviously it would need to be sourced. I looked and could not find this point of view in any reliable source. Even the link you posted above (https://hydrogen-central.com/green-hydrogen-prices-will-remain-stubbornly-high-for-decades/) says production costs will come down - just not by a huge amount. BTW full-scale green H2 plants are already in operation. Clayoquot (talk | contribs) 00:39, 10 August 2025 (UTC)

"BTW full-scale green H2 plants are already in operation." feeling lazy, what's your definition of full scale (I'd arbitrarily say 200 tonnes/day), and which ones? Greglocock (talk) 03:02, 10 August 2025 (UTC)

I used the term full-scale to mean facilities that produce green H2 for commercial use, as opposed to facilities that exist for experimental/demonstration purposes. The fact that green H2 production is measured by the IEA means that such facilities exist. Clayoquot (talk | contribs) 15:42, 31 August 2025 (UTC)

Maybe answer the question? Greglocock (talk) 21:14, 31 August 2025 (UTC)

I decided it wouldn't be appropriate to do so because Wikipedia is not a discussion forum. Clayoquot (talk | contribs) 21:26, 31 August 2025 (UTC)

It occurred to me that the entire "Projects" section could be replaced by an up-to-date map of operational green hydrogen projects, and a sentence or two saying "As of 2024, x green hydrogen facilities were in operation in y countries".

The IEA has something close, however it would show all electrolytic hydrogen facilities, only around half of which are known to use renewable energy exclusively.

A map could be generated from the IEA's H2 production database. The database is an Excel spreadsheet that you would need to filter on Status="Operational" and "Technology_electricity=Dedicated renewable".

What do people think of doing this? (Getting a volunteer who has the skills to do it is another matter.) Clayoquot (talk | contribs) 02:59, 1 September 2025 (UTC)

That would be great. Unfortunately I do not have time to do it myself at the moment Ita140188 (talk) 05:35, 1 September 2025 (UTC)

China's claims for green hydrogen seem vastly at odds with the IEA database. I propose to use IEA figures rather than propaganda Greglocock (talk) 03:43, 2 September 2025 (UTC)

Can you point to specific numbers/references? Ita140188 (talk) 01:18, 3 September 2025 (UTC)

Most egregious is Inner Mongolia, claimed 66.9 kt/y, IEA 12.0. https://www.blackridgeresearch.com/blog/latest-list-of-largest-biggest-green-hydrogen-h2-projects-plants-in-peoples-republic-of-china-prc . One explanation is that the promoters may be including all phases of a project, and the IEA only counts what is actually working. There's also a disparity between claimed GH2 output and power supply, a 100 MW power supply 24/7 could only electrolyse 16 kt/y, real capacity factors for onshore wind and solar are maybe 0.3 and 0.2 at best. Greglocock (talk) 01:45, 3 September 2025 (UTC)

It is not clear to me if the stated capacity is the electrolyzer capacity or the nameplate capacity of the associated wind/solar plant. In general we should avoid using these blog posts as references as they are very imprecise (eg.: The Longyuan Power Green Hydrogen Project is a 600 MW project located in Alxa Left Banner of Alxa League, in the Inner Mongolia Autonomous Region in China. In addition to 600 MW, it is powered by 400 MW wind power and 200 MW solar PV to produce clean energy. what does a "600 MW project" mean? Is that the sum of solar and wind peak capacities? Then why do they say "in addition"? What is the peak capacity for hydrogen production?) Ita140188 (talk) 03:32, 3 September 2025 (UTC)

More believable numbers for Lancaster in California say 650 MW solar for 22 kt/y of hydrogen, giving a capacity factor of 0.21 Greglocock (talk) 01:56, 3 September 2025 (UTC)