Bjorn’s Corner: The challenges of Hydrogen. Part 5. The Hydrogen tank.

By Bjorn Fehrm

August 21, 2020, ©. Leeham News: In our series on hydrogen as an energy store for airliners we start the design discussion of a hydrogen-fueled airliner by understanding the onboard storage of hydrogen better.

While there is present knowledge from for instance the space launcher industry, the storage demands for launchers are hours rather than days. Several implementations of longer storage aeronautical tanks have been done, among others by NASA/Boeing for high flying UAVs.

Airbus and the Russian aircraft industry were also active with research during the 1990s and Tupolev built a test aircraft that included a complete hydrogen fuel system (Figure 1).

Figure 1. The Tu-155 Hydrogen research aircraft. Source: Tupolev.

How to store hydrogen efficiently on airliners?

We will discuss the tank design for liquid hydrogen (LH2) in this Corner. It’s the area that forces a major change compared to the airliners as we make them today, and the subject has enough variables that we focus the aircraft fuel system in the next Corners.

Hydrogen is the most common chemical substance in the universe. It melts from solid material to liquid at 14K (-259°C) and boils to gas at 20K (-253°C). It means, unless we want to store it under extreme pressure (which is not weight efficient for an aircraft), we need to store the LH2 between 14K and 20K in our tanks.

For cars, buses, trucks, and forklifts it’s commonly stored in gas form at ambient temperature in pressure tanks of 350 or 700 bars (35 or 70 MPa, 5 kpsi or 10 kpsi).

As discussed, this is too volume inefficient for airliners and we, therefore, store hydrogen in its liquid phase as LH2. This means we need to figure out how we keep hydrogen below 20K in the tanks. You do it by insulating the tanks extremely well.

The common designs are an inner liner that can sustain the characteristics of LH2 (among others no permeable leakage and no embrittlement), followed by an insulation layer to keep the fluid at the below 20K it has when filled from the fueling truck.

This is not possible for all parts of the tank over a longer period. At parts where the temperature exceeds 20K, hydrogen will boil off and there will be a top layer in the tank of hydrogen gas, Figure 2.

Figure 2. Typical liquid hydrogen tank. Source: Fullcryo.

A certain loss of hydrogen is accepted, the boil-off factor, which is less than a percent per day for well-insulated tanks. The boil-off gas also replaces any LH2 that has been consumed from the tank through B.

The gas is not allowed to build high pressure in the tank, above a fraction of a bar a pressure relief valve will vent the hydrogen to the surroundings or to for instance a hydrogen fuel cell system that replaces the APU in the aircraft (A in Figure 2).

To keep the fuel cell feed when the boil-off is not sufficient, a path from the liquid side with a liquid-to-gas heat exchanger is included at C.

The inner tank (the liner) is typically aluminum, with a foam style insulation or a vacuum chamber on the outside. Designs that use a vacuum layer have an insulation layer on top. The vacuum functions according to the Dewar or thermos principle to reduce heat transfer to the LH2 area. The additional insulation is a safety measure in case the vacuum is lost. Without the insulation layer, the boil-off in the tank would be strong, creating a potentially dangerous outflow of gas.

As hydrogen gas is a non-toxic gas it can be vented to the atmosphere for any excess or leaked gas. Schemes where the boil-off gas is re-liquified by cooling, such as the storage tanks for missile launcher sites, are appearing for stationary longer-term storage, Figure 2.

(We use the word Cryogenic a lot in this series, it also appears in Figure 2. Here its meaning from Wikipedia: In physics, cryogenics is the production and behavior of materials at very low temperatures.)

Figure 2. Traditional and future LH2 storage at launcher sites. Source: NASA.

It has not found its way to airborne tanks yet, the system complexity and weight doesn’t pay off for missions of up to 15 hours.

There are several reasons why efficient tanks have spherical or cylindrical or near-spherical/cylindrical shape:

  • The key characteristic of a tank is its low heat leakage, which depends on the external surface area to contained volume. This ratio is minimal for a sphere. The next practical and efficient shape is a cylinder.
  • The tank shall endure pressure variations from internal hydrogen gas and the external atmosphere as the tanks are normally placed outside the pressurized area of the aircraft. An aircraft cycles from 1 bar to 0.2 bar (14.5 to 2.9 psi) external pressure and 20-30°C to -65°C several times a day, creating a material fatigue problem. Round shapes handle such pressure variations best.

The insulation of the tank is designed to give the desired heat transfer for the storage of a day or so. For longer parking of the aircraft, the LH2 is returned to the airport tanks via the tanking trucks.

In the next Corner, we look at tank placement and compromises in tank shape.

35 Comments on “Bjorn’s Corner: The challenges of Hydrogen. Part 5. The Hydrogen tank.

  1. Perfectly clear. We see on cars like Toyota Mirai Composite overwrapped pressure vessels for compressed hydrogen that allow a highter pressure in the tanks, ideally over jet Engine burner pressure. With Engines/APU running consuming hydrogen makes the fuel pump design and its shaft power requirement easier to meet. It will be interesting to see what new certification requirements will come with LH2 powered Aircrafts.

    • As Bjorn mentioned, automotive tanks are high pressure and do not store LH2, which can exist in liquid form only below the critical point of 13 atm and 33K. Aircraft tanks probably will operate at conditions well below that in order to get the needed liquid storage density, and for safety.

      This is as opposed to the Mirai tank which is charged to 70 atm and ambient temperature 20C, and stores only 5 kg of hydrogen, for use as a gas in a fuel cell at 4 atm. The high pressure sealed tank is needed to retain fuel during overnight periods. Bjorn indicated aircraft LH2 must instead be offloaded overnight to prevent loss.

      With an LH2 pump operating before the vaporizing heat exchanger, the shaft work can be small. The expansion ratio for hydrogen is 750 to 1, although less at fuel injection pressures. The evaporative tank losses can be either combusted in a separate pre-burner (as you mentioned earlier) or compressed to mix with the main burner gas.

      • “This is as opposed to the Mirai tank which is charged to 70 atm and ambient temperature 20C, and stores only 5 kg of hydrogen,”

        The tanks store hydrogen at 70 MPa (10,000 psi).
        … or just short of _700_ atm. ( state of the art upper end for mobile tanks. But: 90kg structure for 5kg fuel)
        Burst pressure of those vessels should be better than twice that.

        • Yes, thank you Uwe. it’s 700 atm in the case of those H2 storage tanks.

          The filling and discharge of those tanks is a science in itself. The pressure and temperature state path must be controlled as the tank both fills and drains, to maximally utilize the storage capacity. Fully computerized process with sensors in the tank. Vehicle inhibit interlock so the driver doesn’t pull away during fill.

        • “Burst pressure of those vessels should be better than twice that.”

          According to CS25, this is correct for liquid filled contents. However, since LH2 may boil off (clogged vent event has to be considered), I wonder wether requirement for gaz may apply; it requires a burst pressure from 4 to 10 times the nominal one (as far as I remember, please correct me if wrong).
          Hydraulic accumulators have to meet gaz requirement for example.

          • The automotive H2 gas tanks are designed for maximum working conditions of 900 atm and 85C, with a 2.25 safety factor. So Uwe is correct in his claim of handling double the pressure. The nominal tank full pressure is 700 atm, but the filling pressure may range up to 900.

            For LH2, more than 15 atm would result in a supercritical fluid, and then you need a cryo-compressed tank, so the pressure maximum would be quire high, 250 atm or more before safety factor. It’s difficult to build that kind of cryo-tank and they are heavy, so for aircraft use they may rely on multiple venting and pressure relief safety schemes instead.

  2. It would help to have the pressure data ref in PSI for the US. I’ve worked with atmospheres so I can convert but not a lot have. No feel at all for Mpa which is not used here at all, even for bottling Jack Danial’s.

    For those not familiar Atmo ref is 14.7 psi, 10 Atmo’s is a bit over 140 psi.

    By today standards you will need 3 of everything.

    Pump and heater systems, safety relief.

    Reefer systems are hard enough for a cold freezer at -20F,

  3. Sounds like aircraft shapes are prefect for LH2 cylindrical storage. The floor in a double bubble is only there to support the passenger deck and secondary tube cross stiffening.
    That leaves the principle problem remaining of where to put passengers and baggage in a design thats begun from scratch for LH2.
    Come back A380 3 decker we still need you.

    • The A380H is a fun concept. It needs deep surgery but there might be money around for new dry wings in carbon fiber. New RR Ultrafan hydrogen burning engines. The top deck full length Al-Li LH2 tank and pax on mid deck only, limited MTOW so one set of Main Landing gears is enough making space for cargo. Maybe as a one-off technology demonstrator with French goverment money to Dassault to make them ready to make huge carbon wings for Airbus once UK is totally outside EU? The Hydrogen burning Advance RR engines can be certified and assembled in Dahlewitz?

      • With all those excess A380s, how about a Super Super Fire Fighting Tanker?

        • The ned them in California now, however cheaper to convert the A340-500’s sitting around…

  4. One important aspect of onboard storage of LH2 is the structural efficiency or gravimetric density of the tanks.
    Gaseous storage of H2 typically gets 5 to 10% structural efficiency. Meaning, a fully charged tank weighing 100 kgf, carries only 5 to 10 kgf of compressed H2. This is unacceptable for aviation use.
    Liquid storage of H2 shall achieve 35-50% structural efficiency. Much better than GH2, but still a critical factor.

  5. Wing mounted pods would be a good way to storage plus reduce loads on the wings.

    • Typically, insulated hydrogen tanks would be about the cross section and size of an engine. Making them long and smaller cross section will introduce COG issues. So possibly better stored in the fuselage. You can then introduce super slim wings. IMO

      • Yes, if you design dry wings in CFRP you can optimize in a new way and having the LH2 tanks as an extension of the fuselage makes for low drag design. You might need trim tanks up in front in the space between nose landing gear and the front cargo hold making integration and pipe routing harder. The final optimization of tank pressure for the H2 is not done yet (if it should be Pamb or higher pressure), note as the APU/Engines are started ideally they will consume all boiled off H2 gas.

  6. Does anyone know what happens when a tank of H2 at minus 240 deg C is ruptured? is the boil-off graceful or explosive? The answer will ceratinly affect what the regulators feel about certifing it for passenger use.

    • If the tank is vented, there is no pressure explosion but a spill of cryogenic fluid. The fluid will freeze whatever it touches while simultaneously flashing into vapor, at an expansion ratio of 750 to 1. The heat of vaporization would be quite large, so absence of a source of that heat would reduce the rate of vapor production over time. Obviously a fire could supply the needed heat, in a snowball effect.

      If the spill environment is sealed or not adequately vented, there could be a secondary pressure rupture or explosion. If the spill area is top-vented, hydrogen vapor is much lighter than air so will rapidly vent and diffuse upwards.

      If there is an ignition source in a non-vented spill area formerly occupied with air, and the mixture passes through the flammability limits, there could be a detonation explosion. In a top-vented spill area with ignition, the more likely outcome is rapid flame combustion diffusing upwards.

      My guess would be that any hydrogen storage or piping would exist within isolated vented containment at lower than cabin pressure, perhaps with nitrogen inerting and a relief valve. This would prevent any combustion from taking place due to a spill or leak, and safely vent the hydrogen to the outside and above the aircraft.

      What happens in a crash is not predictable, but again my guess is that a dedicated aft fuselage section with strong bulkhead separating passengers, is the safest storage method. It offers the best protection for passengers, as the main risk would be inertia of the LH2 carrying it forward in a crash, either around or through the bulkhead.

      A rupture contained by the bulkhead could safely spill, and then either vent or combust skyward. Perhaps that fuselage section would have a cryogenic liner that is designed to fail at the top rear in the event of a crash. That could contain and redirect the LH2 momentum rearwards, and allow safe vaporization. Kind of like an airbag that blows out in the rear.

      All speculation of course, this would all need to be investigated and tested. But it’s a good question as safety will have to be the same or better than current standards.

  7. Transport Canada will do flight testing of the 737 MAX next week. First they will visit Boeing to fly the simulator, then the test aircraft will be flown to Vancouver for evaluation by Canadian pilots. Part of the recertification process for Canada.

    The EU has lifted travel restrictions for Canada, so perhaps if Canada reciprocates, there would be an opportunity to repeat this process for EASA.

    • Pilots are not covered by travel restrictions or quarantines and every country has various exemptions for ‘critical people’

      • The EASA team wasn’t able to travel to the US for the FAA flight tests, they claimed their entry was denied, so the exemptions did not work for that case.

        Perhaps they could do a similar circumvention through Canada, if the authorities would permit it. The US and Canada have a mutual essential travel agreement.

        So the issue would be if Canada would allow the team from Europe, and if the US would apply Canadian rules to them once there. Possibly not, though. It was just a thought to speed things along. EASA will have to test at some point.

        • It seems totally implausible EASA would be denied entry into US, though Pebble Mine is now off approval as someones boy goes fishing there, anything is possible.

          And all Boeing needs to do is take the software to Europe (or send it) load it on any 737MAX anyone is willing to offer up and then have EASA do their tests.

          Or Boeing can load software on a US 737MAX and send it over.

          There is the wiring work that still needs to be done for allowed service and I assume EASA will go with FAA on that one.

          • Boeing currently has the one 737-7 aircraft instrumented for certification, as that model was the next certification step before the grounding. So that aircraft would have to go to Europe with a Boeing support team. Currently that’s not possible, but perhaps in time.

            They will likely have a 737-10 model instrumented in the future, so that would be another eventual option, after certification by the FAA.

          • They will continue trying to sideline EASA as has been done in the past.
            No entry for EASA personnel is happenstance/enemy action.

          • It may be more accurate to say that the administration denied the entry request from the FAA/Boeing/EASA. Likely because of Trump’s feelings of unfairness towards the EU, which have caused a lot of other difficulties as well. Hopefully much of that will be reversed in the coming months.

            Boeing/FAA have nothing to be gained from excluding or sidelining EASA. Without EASA approval, there would be a split in world regulations regarding the MAX. Everything has been done to include EASA and avoid that outcome. No one wants to see that happen. The FAA has been right to pursue that strategy, and it appears that EASA has cooperated fully.

            The entry denial is another hurdle to be overcome, but it will be overcome in time, as all the others have been,

          • Why is Transport Canada allowed NOW and not earlier?
            Obviously there was a deal made so FAA allowed EASA and Transport Canada to have free requirements instead of flight testing.

            EASA wasn’t allowed to enter when tennis players were allowed, right. Seems to be very strange that FAA/Trump allow flight testing now.

            STS has two modes, fast and slow, fast maybe for stall prevention. In the meantime Boeing could have altered STS and made it more aggressive.

            FAA is Trump and they don’t want to do a good job. There might still be many parts which were self-certified by Boeing with undue pressure and jedi mind tricking and FAA are not touching them. Best example is the “novel” use of the stabilizer.

            Norwegian is on court and suddenly they have partly Chinese owners. It should be easy for them to flight test a MAX without MCAS. Norway is part of EASA, Norwegian might get instruments from EASA.

            There was a report in a UK magazine in January that the 777X has a software like the MAX has, but I couldn’t read it. Now Clark wants to have it confirmed that the 777X has no new embedded software than the 777.

            If I were a regulator I would not accept a MAX-7 for flight testing. With the crashes and Boeing’s shown behavior I would not accept anything from Boeing. Boeing still kept their jedi mind tricking in February 2020, modus operandi for 787 certifications.
            If I were a regulator I would support flight testing through Norwegian, just to record some flight data without MCAS. These data would likely show a different behavior which then would question everything FAA/Boeing did since the groundings.

            Obviously there is a need for MCAS, otherwise it wouldn’t be there, to make the flight behavior more linear which is stall prevention. Then as Philip said long ago the stabilizer needs redundant systems which obviously are not there. That’s why FAA is not touching the “novel” use mentioned by JATR.

            Other regulators might have commented on FAA’s certification process and won’t let the MAX fly. Not all countries might have signed an agreement which allows the FAA to certify worldwide. Especially when the FAA act criminal too. Seems TrumpFAA is losing another battle.

            Seems no MAX certification this year and not next year too.

            Boeing should start to clean the house and the FAA too.

            It is proven now that worldwide certification agreements don’t work.

          • I can see the 737-7MAX test bird (its going to Canada in a shuffle thing to do the tests).

            All I can find is nebulous statements about why EASA is not in US.

            Also vague statement that Boeing needs to do more work but does not say what.

            Boeing can fly the 737-7MAX to Europe as well.

            Maybe they are hoping the administration will be gone soon so they don’t have to worry about a Tweet to boycott?

            That means come Nov 3 (or so) Boeing is free to act.

          • “”And all Boeing needs to do is take the software to Europe””

            The new software is not needed.
            The first thing to do would be to flight test without MCAS, to see why MCAS is needed. That would result in the systems which are needed and software only might not be enough.

            But the political agreement might say that Europe would allow FAA to certify for Europe. EASA can’t easily do their own flight testing. Trump is playing this card, but not all countries might have signed this political agreement. And same as Trump did in the past when he cancelled agreements, other countries might cancel this political agreement too and won’t accept FAA certifications anymore 🙂

            We all know what was standard business, self-certifying with undue pressure, some jedi mind tricking, some secrets and most of these standard business results are still there, FAA is not correcting them.
            MAX production stopped in December 2019, 3 days after EASA didn’t accept Boeing’s self-certifications anymore. And here we are now, over 17 months after the groundings and nothing has changed, in February 2020 Boeing still used undue pressure for 787 self-certifications, another Frankenstein plane.

        • Leon, these are just further unsubstantiated rants against official conclusions that don’t support your views. You are not going to accept the evidence the FAA found and documented in their report, in conjunction with multiple investigations and regulators from around the world. Just as you won’t accept certification when it occurs.

          That evidence addresses all the questions and issues that were raised, by all parties, including all of the points you continue to highlight. This is all in the report. We will see in the commentary if other issues are raised, or if others disagree with the findings. But these questions are now mostly resolved and settled, pending the commentary and final approval.

          I know that won’t matter to you, but for the benefit of others reading here, it’s important to distinguish opinion from fact. You have a dissenting opinion and that’s fine, but the factual findings of the testing and analysis have aligned with the other side.

          EASA and Transport Canada have their own requirements for changes and certification, and will continue to pursue them, but neither Boeing nor the FAA have objected, so those will be implemented too. Most other regulators have said they will follow the majority consensus. China remains silent so their response is uncertain, but cooperation may be unlikely in the face of multiple tensions.

  8. Hi Rob,

    I invented a new buoyancy system concept back in 2010 for Hybrid Airship transportation but only recently got interest in the ideas from both Government and private parties.

    I am cognicent with some of the issues you describe ibn the article but I wonder if you would be kind enough to help me with one potential issue. Could hydrogen be recycled from its liquid phase to gas and back to liquid… and again back to gas and so on? Is the technology to allow for this mature or even available is really my question I think?

    • The phase transition from liquid to gas and back again is a matter of energy transfer in & out. The heat of vaporization must be supplied to drive the liquid into vapor. Then it must be withdrawn again to condense the vapor back into a liquid.

      If it’s to be a cyclic process, then you need both a source of heat and a refrigeration system. For hydrogen, the liquid temperature is so low (20 K) that a conventional heat pump refrigeration system would be difficult to construct & impractical.

      Typically for liquefaction, the hydrogen gas is first compressed and cooled by ambient air to room temperature, then further cooled to about 80 K by liquid nitrogen. Then expanded through a nozzle or orifice to reach the liquefaction point. This process trades volume increase for temperature decrease. The product is a two-phase mixture from which liquid is drained off, and vapor is recycled back into the cooling stage. That process is irreversible so no energy recovery is possible.

      The source of heat for liquid hydrogen vaporization could be natural surroundings, since the liquid temperature is so low. But you’d need a large transfer surface area or thermal mass to provide the amount of heat needed.

      Another factor can be contamination. Atmospheric gases will liquify and/or solidify before hydrogen, and those components can gum up the works. So there needs to be filtering or screening of those components if they are present in the hydrogen.

      Anyway, hope that helps. It would be challenging to make this into an efficient cyclical process. It can be done in a moderately small package if the goal is only to recover evaporative losses from a vented LH2 tank. But to cycle a large amount of hydrogen repetitively, would require a large liquification facility, and a large heat source.

    • Why not having a compressor and Composite overwrapped pressure tanks like AGA Genie? It weights some but you size the balloons for certain max lift and as you need less lift you run the compressors to fill the pressure tank with hydrogen from the balloon. Atlas Copco and similar makes suitable ones that could run on DC from fuel cells powered by the hydrogen you have on-board.

  9. Rob

    Many thanks for your email reply and this will be included in my proposal as background information! If you’re OK with that.

    • That’s fine with me. I hope you are successful in your proposal.

      In thinking about this more overnight, I realized that the conversion rate may be more important than the total amount of conversion. If the rate can be small but can happen over a longer time, that could still handle a larger amount of hydrogen. The rate is what would determine size, weight, and power requirements of the components.

      For a buoyancy application, the effect of a small rate might be amplified by using/storing the nitrogen gas needed for the hydrogen liquefaction plant. This could be kept in a diaphragm-separated compartment of the hydrogen envelope, such that the volume of hydrogen could be varied by adding nitrogen and removing hydrogen at the same time, This would keep the overall pressure relatively low but would multiply the effect of lowering buoyancy. The process would be reversed to increase buoyancy.

      You’d need dual small liquefaction plants for both nitrogen and hydrogen, but they could share some of the components to save weight. Liquifying hydrogen would produce nitrogen boil-off to be used in the envelope compartment, and vice-versa. You’d need liquid storage too, for both elements. But LN2 could be used to chill and lower losses of LH2. The advantage is that although hydrogen must be conserved, replacement of nitrogen losses could be extracted from the atmosphere.

      The power requirement would still be a challenge to carry with the vehicle over a long time. Perhaps solar could help out, if the weight can be kept down. A fuel cell or hydrogen fueled engine could be the other options. Or a conventional engine with added weight of fuel.

      Not sure how this would all balance out for size, weight, buoyancy control, lift capacity and endurance. But it’s an interesting thought. From the history of airships, control and stability in weather is one of the biggest problems. But there have been improvements in those areas in recent years, judging by testing of new designs.

      • Thank you for your suggestion as it all helps to form the basis of my proposal!

        Once again thanks for the ideas.

  10. If H2 is below about 40K you can cool it by throttling. I have seen large LH2 tanks that take a small amount of the gas off of the top, pressurize it in a small compressor, re-cool it with the bulk liquid, and then expand it to re-liquefy it.
    There is also the issue of pressure rating. Some of our LH2 tanks were rated 60psi (4atm) and some 200psi (14atm). These were all vacuum/pearlite insulated, and the higher pressure ones had SS inner tanks and very low boil off rates.

Leave a Reply

Your email address will not be published. Required fields are marked *