December 25, 2023
By Michael Barnard

New Hydrogen Pipeline Vs HVDC Study Less Wrong, More Clearly Shows Hydrogen Uneconomic

ChatGPT & DALL-E generated panoramic image of a rusting hydrogen pipeline that abruptly ends
 in mid-air, set against a vast, open landscape.

For the past few years there’s been a sub-genre of studies trying to pretend that making hydrogen in one place and using pipelines to move it to other places is better than just moving the electrons using HVDC transmission. It’s the Oxford Institute for Energy Studies’ turn to get it wrong, although they are less wrong than preceding efforts.

A year or so ago I looked at a couple of the first papers, and some things leapt out at me. The papers were 2018’s Relative costs of transporting electrical and chemical energy by Saadi et al and 2021’s Cost of long-distance energy transmission by different carriers by DeSantis et al.

As I noted at the time, they had mistakes in common, and glances at other studies found the same ones. The first problem was that they assumed that there was a very large source of green hydrogen molecules in a single location suitable for putting into a 1,000 km pipeline. This isn’t the case unless you make hydrogen from natural gas at a gas field. All of the small number of kilometers of current hydrogen pipelines have a centralized natural gas steam reformation system and use a few hundred kilometers of pipelines to get the natural gas to major industrial consumers, mostly oil refineries. That’s true in Germany and it’s true in Alberta.

They assume hydrogen that’s dirt cheap to manufacture, $2-$4 per kilogram per the 2021 study. Because green hydrogen takes 50 to 60 kWh of electricity per kilogram, depending on balance of plant, to manufacture, dehydrate and compress, and because a 28 component electrolysis plant is a high capital cost, this would require $.02 electricity 24/7/365. The high capacity factor requires firmed electricity, which means lots of wind and solar farms spread over a large area, some storage and lots of transmission to the electrolysis plant, all of which means that the electricity won’t be $0.02 per kWh.

What this means is that the assessments ignore a bunch of infrastructure required for the electrolysis facility and hence a bunch of costs. Firmed electricity will cost closer to $0.10 per kWh in the real world for decades, outside of some places with massive legacy hydro dams like northern Norway, Quebec and British Columbia, where $0.5-$0.06 per kWh can be found. That means hydrogen that’s at minimum 2.5 times more expensive, $5 to $10 per kilogram, just to manufacture. The earlier study even used $1 per kilogram hydrogen, which would require free electricity and someone else providing a gift of a free electrolysis plant with no strings or costs attached.

The papers assume that the hydrogen and electricity just magically appear at the end of the pipeline and transmission, instead of the hydrogen first consuming a good fraction of the energy in the electricity. Then they assume, incorrectly, that HVDC lines are and must be small, while hydrogen pipelines are big.

Even with all of this, at $1 per kilogram in one paper, the hydrogen is twice the cost per unit of energy as natural gas.

Of course, electricity gets into the high efficiency distribution grid, while at the end of transmission, hydrogen gets much more expensive due to distribution costs. For context, dirt cheap hydrogen made from natural gas costs $1-$2 to manufacture, yet costs €15 to €25 at pumps in Europe and $36 at most pumps in California right now. Costs go up radically as soon as you leave pipelines.

Then both papers ignore that hydrogen as an energy carrier has to be used at the other end, and molecules are almost always less efficient than using electricity. Fuel cells are 50% efficient on average at turning hydrogen into electricity for motors while electricity used directly or through batteries is much more efficient.

Burning hydrogen for heat under 200° Celsius competes directly with heat pumps, which are three times as efficient on average, and in industrial settings usually more efficient. There is zero room for residential or commercial heating as a result.

The vast majority of industrial heating applications above 200° are electrifiable directly, typically with efficiency gains over using burnable fuels. Even at energy parity, the energy will cost multiples because so much electricity is lost between wind turbine and heat.

This is exergy. The hydrogen pipeline studies ignore it.

Then there was this year’s DNV study, Specification of a European Offshore Hydrogen Backbone, in which four molecules-for-energy analysts in the molecules-for-energy side of the firm worked on a report bought and paid for by the European association of pipelines, which really needs molecules for energy to be a thing or it and its members cease to exist. Yup, no conflicts of interest there.

I assessed it in an initial and followup article a few months ago. It was slightly better than the 2018 and 2021 study, in that it included more of the balance of plant for the hydrogen. Did it correct the mistakes above? Only in that hydrogen got slightly more expensive as they included more real world requirements, but it was still asserting that green hydrogen manufactured offshore would be delivered to the end of the transmission line for €3.21 per kilogram in 2050 in the best possible case, and that this was cheaper than any transmission of electrons.

Still missing balance of plant. Low balling offshore infrastructure and operational costs. Using HVAC transmission instead of HVDC transmission to get electricity to shore for onshore electrolysis. Giving HVDC transmission very high transmission losses compared to hydrogen pipeline efficiencies. They make pipeline operational costs much lower than HVDC operational costs, although HVDC has no moving parts while pipelines have lots of moving parts. They make offshore wind capacity factors very high for hydrogen electrolysis. They make onshore solar capacity factors very low.

With all of these thumbs on the scale for hydrogen, it’s still ten times more expensive than liquid natural gas, the most expensive form of electricity any country imports today. This is not the basis for an energy economy, but the basis for economic disaster. Do they make that comparison? No, no they don’t.

Once again, this is just at the end of hydrogen transmission but before costs multiply for distribution.

The entire report was structured to meet the need of the clients to pretend that manufacturing molecules of hydrogen offshore at wind farms and then constructing pipelines all the way to major demand centers was the most cost effective model, and the DNV analysts contorted numbers and the space time continuum until the client’s needs were satisfied.

And so, into this context comes the Oxford Institute for Energy Studies and their November 2023 report Hydrogen pipelines vs. HVDC lines: Should we transfer green molecules or electrons?

They do a couple of things right, which is worth noting. The first is that they start with the same amount of electricity at the beginning of the pipeline and HVDC, 9,600 MWh. They then apply most electrolysis facility energy losses before putting hydrogen into the pipeline, and apply the much smaller efficiency losses to electricity before it gets into transmission. This is good. This is much closer to an apples to apples comparison.

They calculate the energy delivered at the end of the pipeline or transmission to be 1,152-5,712 MWh for the pipeline and 7,872-8,832 MWh at the end of the transmission line. Sharp eyes will note that’s an awful lot more energy at the end of the transmission line. Being optimistic for the pipeline and conservative for the transmission and doing a simple average of both pairs of numbers finds 3,432 MWh for the pipeline, or 34% of the green electricity delivered, and 8,352 MWh for transmission, or 87%.

That means the hydrogen pathway delivers only 40% of the energy in the electrons pathway. That means that all else being exactly equal, the energy will cost 2.5 times what the electricity costs. They do some more work to quantify the costs of the up front electrolysis and then asserted higher costs of HVDC transmission than pipelines, but that’s a bit of a wash.

They burden transmission a bit with a greater requirement to bury it than they use for pipelines, which is odd as pipelines tend to be buried for great lengths of them as well. That’s a bit of a wash.

But okay, only 40% of the energy delivered from the wind farms because they actually did some work on electrolysis, desalination and compression. That’s at least closer to an apples to apples comparison.

But then they fall over. First off, they note that getting firmed electricity in sufficient quantities to an offshore electrolysis facility will take a lot more wind farms, but then don’t factor the cost of building all of those extra wind farms and power cables into the equation. Like the earlier reports, they kind of pretend that all of the electricity just magically gets to the offshore electrolysis platform and the pipeline starts there.

And their electrolysis facility is pretty light, missing as a major component the hydrogen dehumidifier that removes water vapor from it before transmission. As there are a lot of components, they are glossing over the balance of plant a bit. At least they have a balance of plant though, which as I’ve been been pointing around deeply flawed International Council of Clean Transportation reports related to hydrogen is usually missing from assessments.

But then at the other end of the pipeline, they fail on the exergy problem again. They treat the energy delivered as molecules and the energy delivered as electrons as being equal, when they aren’t. Fuel cells are 50% efficient on average at turning hydrogen into electricity, requiring 5 times as much electricity at the beginning of the journey instead of using electricity directly or through batteries. For heat under 200° Celsius, 7.5 times as much electricity starting the journey. For heat over 200°, where there is virtually always an electric option and it’s usually more efficient, even assuming equal heat energy in hydrogen and electricity means 2.5 times as much electricity at the beginning of the journey.

They do note that there is no market for hydrogen for energy today, yet use units of energy for hydrogen almost exclusively. They cite ranges of demand for hydrogen, but never question the expectation of massive amounts more hydrogen or alternatives for the energy storage use case they express as a value proposition of pipelines. They note that there are vastly more HVDC transmission lines in operation, approved and under construction and new hydrogen pipelines are barely at the design stage and have no private industry stepping up to build them, and then assume that hydrogen pipelines will be built.

And, once again, hydrogen distribution costs are ignored.

The data in their report makes it clear that there are virtually no places where putting electrons into electrolyzers to make hydrogen to put into pipelines makes the slightest fiscal sense compared to transmitting the electrons, yet they conclude that the technologies are complementary and both will be used in large amounts. They don’t question the entire premise of hydrogen as a carrier of energy despite the almost complete lack of its use for that today.

Considering these technologies as standalone competitors belies their complementary nature. In the emerging energy landscape, they will be integral components of a complex system.

They don’t have the courage to state what the data clearly shows. No one is going to build hydrogen pipelines for energy. What’s going to happen is that HVDC will be used to carry electrons everywhere, and where there’s an industrial demand for hydrogen as a feedstock, it will be manufactured at point of use, just like 85% of hydrogen consumed today.