Ya-but the rabbit, courtesy of Google Gemini
TL&DR Summary: sure, you can re-use an existing gas pipeline to carry some quantity of hydrogen at some pressure, and do so safely. How much pressure and under what circumstances and operating conditions, and how much energy can be delivered via that repurposed pipe, and whether or not this makes any economic or environmental sense in a decarbonized future- these are all questions that don’t have such a clear answer. The devil’s in the details, and in some cases, he’s definitely not hiding there in complex matters- he’s plainly visible from a distance, in simple and well known issues. The result is an object lesson in how someone can ask an engineer a “yes/no” question, and receive the answer “Yes, but…”, and then do the very human thing which is to hear only the word “yes” and more or less ignore the rest.
Why Do I Care?
I’m the principal author of a peer reviewed research review paper on the re-use of existing fossil natural gas infrastructure for pure hydrogen and hydrogen/fossil gas blends. The paper is open access, and was published in Energy Science and Engineering, which is a high impact scientific journal.
https://scijournals.onlinelibrary.wiley.com/doi/10.1002/ese3.1861
My co-authors in that paper include several prominent scientists working for the Environmental Defense Fund in the US and Europe, and two of my co-founders of the Hydrogen Science Coalition, all of whom are qualified scientists and engineers in their own right. Our paper looked at every aspect of the fossil gas production, transmission, distribution and end-use, and evaluated the feasibility and issues which come along with hydrogen-methane blends and with pure hydrogen, given that this option is being offered by the fossil natural gas industry as a beneficial re-use of their existing infrastructure. Our paper sailed through peer review, in part because the major issues contained therein had already been extensively peer reviewed here on LinkedIn, which has in fact greater reach to people who, in my opinion, have the most practical knowledge about these issues: engineers working in the private sector, as I still am and have been for over 30 years- people who, by and large, rarely have occasion to read the academic press.
https://www.linkedin.com/pulse/hydrogen-replace-natural-gas-numbers-paul-martin
To briefly summarize the results of our review, the main problems are in the gas transmission network- the long, high pressure pipelines which carry gas between countries and regions. The distribution network could perhaps be re-used for pure hydrogen without damage to that network’s components, but every end use device on the network would need replacement or substantial retrofit first, and numerous other concerns such as hydrogen leakage and permeation leading to effective GHG emissions and a greater risk of fires and explosions, would need to be addressed. The transmission pipes, on the other hand, were never designed to handle hydrogen, nor were their associated compressors, valves and the like. The significant differences in properties between hydrogen and a typical natural gas which is mostly methane, make the transmission pipes vulnerable to damage, and the existing compressors fundamentally unsuitable for the new duty.
While our paper deliberately steers clear of the economics of using green hydrogen as a fossil gas replacement, when one does such an evaluation, it’s clear that the notion of hydrogen as a gas replacement is fraught, even in a dreamy future time when green hydrogen is very, very much cheaper than it is today. Why? The cost per tonne of CO2 emissions averted is just ridiculously high even if one assumed that the gas network itself didn’t need modification at all. Any risks or costs we might undertake in the process of repurposing the gas network as part of a green hydrogen based GHG emission mitigation scheme are therefore without a valuable destination, and hence not worth the cost or risk involved. Here’s my basic arithmetic proving that point:
https://www.linkedin.com/pulse/why-hydrogen-blending-gas-network-bollocks-paul-martin-i5sdc
Needless to say, some people in the gas transmission and distribution industry took issue with our conclusions. One entity, the Dutch gas network operator Gasunie, an entity wholly owned by the State of the Netherlands, was particularly vocal in criticism of our work. Gasunie had concluded in a public report titled “Hyway27”, that the transition of their existing transmission assets to carry pure hydrogen in the future was basically a “solved matter”, so they dismissed our paper publicly as “old news”.
https://www.gasunie.nl/en/expertise/hydrogen/hyway-27
A read of Gasunie’s Hyway27 final report is interesting. It seems to me to be an object lesson to engineers in relation to how their work can be taken out of context by a motivated client.
At Section 4, the Hyway27 report reads as follows:
“This confirms that the wall thicknesses of existing pipelines, which are determined by the diameter of the pipeline, as well as the rated pressures and steel quality are adequate for hydrogen transmission as well at a similar rated pressure (Bilfinger Tebodin , 2019).”
The Hyway27 report goes on to imply in its summary that all that’s required to repurpose certain Dutch gas transmission pipelines for pure hydrogen is to clean the lines, replace some valves, and make a small adjustment in operating pressure.
Well that’s clear enough then, isn’t it? Myself and my co-authors must have just gotten this whole issue wrong, because we didn’t bother to do a good enough search for references when writing our review!
Bilfinger Tebodin 2019: “Research Technical Aspects of Hydrogen – In Existing Pipelines for the Energy Transition”
When you look into it, you find that the Bilfinger Tebodin report referenced in support of that rather strident, conclusive statement, is not publicly available- or at least by means that my Google-fu was able to obtain. Repeated requests to Gasunie for copies of that report, so its conclusions and qualifications could be interrogated, fell on deaf ears. In other words, Gasunie’s position was more or less that “We know it’s all OK, on the basis of information that we’ve received but aren’t willing to share with you. Trust us!”
But fortunately, I have about 30,000 followers here on LinkedIn, and a resourceful follower who shall remain anonymous, was able to obtain a Dutch .pdf of that report. With some effort the report was translated to Word and then to English via Word’s auto-translate feature, as my Dutch is limited to a few choice cusswords. However, because the report isn’t public, I’m not able to share the whole text with you to read it yourself. If you really want to read it and draw your own conclusions, sadly you’ll likely need to make a Freedom of Information request to the Dutch Ministry of Infrastructure and Water Management, for whom the report was written by Bilfinger Tebodin, a large and well respected engineering consultancy.
Here are a few quotes from the report, with my thoughts:
“· The design factors that have been applied over the years for high-pressure natural gas pipelines are in line with the design factors used for new hydrogen pipelines to be built. This means that the wall thicknesses used for the existing pipelines, corresponding to the relevant pipe diameters, design pressures and steel qualities, are suitable for the use of hydrogen at a comparable design pressure.”
When you read that bullet point in isolation, you might conclude (wrongly!) like Gasunie seems to have done, that the pipelines are in fact suitable for re-use “at similar pressure” with hydrogen. But a knowledgeable reader understands that wall thickness checks for hoop stress are just one of several things that need to be satisfied for pipeline safety.
“· The damage mechanism that deserves extra attention for hydrogen applications under natural gas design conditions is fatigue crack growth. For smaller pipes (≤ DN400) with a lower steel grade (Re~ 245 N/mm2), no excessive crack growth is expected, even for larger pressure changes (∆p~ 30% of the design pressure). These small existing pipelines do not need to be subjected to an extensive quantitative analysis if they are to be used for hydrogen gas applications.
For larger pipes (> DN400) with a higher steel grade (Re ≥ 415 N/mm2), smaller pressure changes ((∆p ≤ 10% of the design pressure) are also not expected to result in excessive crack growth. When larger pressure changes are expected, there is a real chance of fatigue crack growth (PM’s emphasis). In this case, a quantitative analysis will have to be performed, which may result in limitations in business operations in the form of lower operating pressures and/or pressure changes.”
For reference, DN400 (16″) pipe is pretty small as gas pipelines go, and Re (yield strength) of 245 N/mm2 (36 ksi) is mild steel– yeah, these lines are FINE with hydrogen! While mild, low alloy carbon steels are commonly used in chemical plant piping for hydrogen at ambient temperatures, people don’t generally make long distance fossil gas high pressure transmission pipelines out of mild steel, and haven’t for a long, long while.
Real pipelines are going to be > DN400 and > 415 N/mm2 (60 ksi) yield for the most part. And the issues with the acceleration of fatigue cracking and the loss of fracture toughness on exposure to molecular hydrogen is well documented and discussed thoroughly in our paper, and even more thoroughly in my article here in relation to extensive DVGW testing done on pipeline steels to examine this risk.
https://www.linkedin.com/pulse/german-gas-pipelines-fundamentally-suitable-carrying-hydrogen-martin
The softest, lowest yield strength gas transmission pipeline steels used in the modern era are generally made of something like API 5L grade X42, which is also commonly used in hydrogen pipelines. X42 has a yield strength of 42 ksi, i.e. somewhere between these figures, i.e. between mild steel at 36 ksi and a typical 60 ksi pipeline steel. For the higher yield strength steels, the rules based portion of ASME B31.12, the code used for the design of hydrogen pipelines, would apply a “material performance factor Hf” of between 0.874 and 0.776, i.e. de-rating the design pressure of a pipe of given wall thickness to between 87.4% and 77.6% of what it would be allowed to have if it were carrying natural gas. If the design pressure of the line is increased (above the 66 bar design pressure which seems to have been assumed in the case of the Gasunie lines), that Hf factor drops further- all the way down to 0.606 for 207 bar design pressure (gas pipelines with MAWP in excess of 150 bar exist in many places in the world). And for harder steels- X80 for instance with its 80 ksi yield strength, ASME B31.12 requires an even further de-rating via an even lower Hf factor. There are also exposure factors which differ between B31.8 for gas pipelines and B31.12 for hydrogen pipelines, which, combined in the worst case with the Hf factor for high pressure high yield strength pipe material, can reduce the safe allowable working pressure of a gas pipeline to as little as 1/3 the design pressure it previously had for carrying fossil gas.
The consequences of reducing design pressure are reduced operating pressure, which drops the useful capacity of the line to deliver energy to its customers, or reduces the energy efficiency by greatly increasing the energy consumption of compressors used to push the gas through the line at a now un-economic velocity. The latter option is not just a waste of energy, albeit a fairly small amount of energy relative to the amount being delivered, it also represents extra cost for the installation of larger, more powerful, and also more frequent, compressor stations along the pipeline.
https://www.linkedin.com/pulse/reduced-pipeline-pressure-eats-energy-paul-martin-8n3zc
You’ll also note that they say that for larger lines of higher yield strength steels, pressure fluctuations need to be kept within 10%…and at 10% pressure fluctuation maximum, such a pipe would have near zero “line pack” i.e. it could store no meaningful amount of hydrogen useful to keep downstream users operating if supplies were interrupted. Keeping the line within these limits would basically drop the capacity of the line dramatically, because any major user would be unable to turn on or off suddenly (when a turbine tripped for instance) without inducing an excessive pressure fluctuation unless they were drawing a very minor fraction of the gas flowing in that pipeline.
The report goes on to say:
“Steels with a yield strength Re ≥ 415 N/mm2, tensile strength Rm ≥ 800 N/mm2 and/or a hardness of 22 HRC or 250HB are prone to fatigue crack growth in hydrogen gas applications.”
They further say,
“. In particular, the DN600 and DN900 pipes show a rapid growth of the depth of the axial crack in the WBZ of the longitudinal seam, for pressure changes of 20 to 30% of the design pressure. In the worst case, fatigue crack growth rates exceed the acceptance criterion of 0.01 mm/cycle (see Figure 5) by a factor of 25.”
They recommend that the pipelines be inspected for fatigue cracks, a pressure fluctuation threshold be set to limit fatigue (in the conclusions they suggest a pressure fluctuation limit of 10%), and then the line should be tested for 1 -2 years with frequent examination to observe fatigue crack growth. To me, that sounds like an experiment with potentially significant consequences to the public safety, so I’d want to be sure, were I a Dutch citizen, that such an experiment was first given very careful scrutiny by a qualified and disinterested 3rd party whose sole concern was the public safety and wellbeing. That would go doubly for examination of the experiment and the interpretation of its results.
Comments of George Verberg, former CEO of Gasunie
An article in the Dutch science magazine De Ingenieur, about our work, came to the attention of George Verberg, the former CEO of Gasunie for 12 years. George was initially convinced by the Hyway27 report’s support for the repurposing of existing gas networks, but our study made George review the previous work with somewhat more skepticism. I encourage you to read (translating if necessary) his comments here:
https://www.linkedin.com/pulse/dag-li-genoten-die-belangstelling-hebben-voor-h2-en-het-verberg-kdhfe
A few important points that George makes are supported by our work and bear repeating here:
1) He mentions that the gas network initial working pressure is 66 bar, which would be reduced to 50 bar “initially with hydrogen”. That is almost exactly consistent with the Hf= 0.776 de-rating per B31.12, i.e. assuming that no other design factors need adjustment. The required pressure de-rating would be greater in systems outside the Netherlands, where design pressures are higher and stronger, harder pipe materials are used, and where differing design factors are required between the two applicable codes.
2) He goes on to mention the effect of the greater velocity required to deliver the same amount of heat energy to customers in the form of hydrogen relative to natural gas, combined with the pressure de-rating, would result in considerably reduced practical capacity for the lines. And that if pressure fluctuations are required to be kept to within 10% to reduce the risk of fatigue cracking, it’s quite clear that the effective line pack would disappear, complicating pipeline operation.
3) He mentions that the Hyway27 report is inconsistent with regard to discussing the compressors. Compressors are long delivery, expensive pieces of equipment, and it’s clear from the Hyway27 report and from numerous other references that replacement, not retrofitting, will be required. Strangely, the cost of replacing compressors is not accounted for. Furthermore, although it seems evident that Hyway27 would propose to deliver similar energy flows in the pipes (i.e. at 3x the velocity, to account for hydrogen’s lower heat content per unit volume), the fact that the compressors would necessarily be 3x as large and powerful has not been considered. In fact when one designs piping, one generally selects a pipesize based on an economic velocity, which optimizes the cost of pipe materials and installation against the cost of compressors and energy to run them. One doesn’t generally try to shove more gas through a line that is too small by installing bigger compressors.
4) Hyway27 does conclude that valves need replacement, but George points out that the valve frequency is assumed to be one every 32 km instead of the present one every 7-10 km in the network. The reduced frequency of isolation valves is in fact the exact opposite to what might be expected from a hazard analysis examining the relative hazards between hydrogen and fossil gas.
George calls for the public release of the Bilfinger Tebodin report, which I fully support.
Lessons to be Learned
Bilfinger Tebodin wrote a carefully worded and, on first review, technically accurate report. When asked the question, “can the gas network be re-used for pure hydrogen?”, they answered as good engineers should do- not with a yes or no, but with a list of conditions and provisos under which a conclusion of any kind could be drawn, specifically in relation to the particular system they were asked to review. Unfortunately, we engineers all eventually learn during our careers that our opinions, even when expressed carefully with all the necessary qualifications, can be taken out of context and used to draw broader conclusions than we’d intended. That seems to me to have been quite clearly the case here. The good engineers answered “Yes, but…” and Gasunie stopped listening after hearing the only thing they wanted to hear- “Yes”.
It’s also very important to examine the references used to support claims – especially when claims are surprising or are of potentially significant public importance. Readers of Gasunie’s Hyway27 final report may have been satisfied to see a report by a qualified engineering company given as evidence of their statements, but unless that report itself is publicly available, its conclusions and provisos can’t be interrogated- and hence the conclusion must remain suspect.
Finally, industry reports- especially when the industry in question is examining more or less whether they will have a value proposition of any kind in a decarbonized future- need to be taken with an appropriate measure of salt. My suggestion is that the appropriate measure is rather more like an excavator bucketful than it is to a pinch!
Disclaimer: this article has been written by a human, not a fluffy “ya-but” rabbit- and humans are known to make mistakes from time to time. Show me where I’ve gone wrong, with good references (public ones that I can interrogate) and I will be happy to edit my work to correct it.
If, however, your principal objection to my work is that it dumps a pile of rabbit-pellets on your pet idea, i.e. your dreams of a future selling hydrogen rather than fossil methane door to door- then you are encouraged to take it up with my employer, Spitfire Research Inc., who will be very happy to tell you to hop off and write your own article.