Bjorn’s Corner: The challenges of hydrogen. Part 26. Auxiliary power

February 26, 2021, ©. Leeham News: Last week we discussed auxiliary power generation for a hydrogen aircraft and found that a fuel cell system had many attractions.

However, it’s more challenging to develop than a hydrogen-converted APU, and we were asked to work through this case as well.

Figure 1. The principal parts of a single-aisle APU. Source: United Technologies.

Hydrogen APU as auxiliary power?

The carbon-fueled APU is primarily used on the ground when the engines are not running. It supplies the aircraft with bleed air through the load compressor, Figure 1, and electricity through the AC generator. Electric-driven pumps then provide hydraulic power.

The installation is in the rear tailcone as the unit is noisy and requires air intakes and a muffled exhaust pipe, Figure 2.

Figure 2. An A320 APU installation. Source: Airbus.

The APU is a simple gas turbine with one centrifugal compressor stage, giving at best a pressure ratio of 8:1. The shaft power out efficiency is around 20%-25%. The efficiency of the bleed air from the load compressor is then below 20% as a load compressor is about 80%-85% efficient.

Figure 3 shows our ballpark power and weight table for our 165 seater for the hydrogen APU or fuel cell auxiliary power. We now assume a classical aircraft architecture with engine bleed air and auxiliary gearboxes with alternators, hydraulic pumps, and air/electric starters. Wing de-ice is bleed air.

Figure 3. An APU versus Fuel Cell auxiliary power comparison. Source: Leeham Co.

Once again, it’s not an exact comparison, just a check where we compare “grosso modo.”

We can see in Figure 2 the APU has a rather elaborate installation in the aircraft’s tail. But the fuel cells at the wing roots are no easy installation either. They are pressure, temperature, and humidity sensitive, so we most likely must provide a pressurized and temperated compartment with a regulated humidity level. We assume the two installations are similar in terms of weight and space requirements.

The fuel cell alternative is 65kg heavier, but it wins the efficiency race with a margin. If we assume a fuel cell efficiency of 45%, it’s more than double the efficiency of the bleed and electrical power delivery of the APU which both are below 20% (we use a 90% efficiency for the APU AC generator). Over the lifetime of the aircraft, this will count, especially with a higher LH2 price than today’s fuel prices.

As before, the fuel cell produces H20 and O2 as output that we can use as onboard water and emergency oxygen.

We see the fuel cell is still the more attractive alternative, but it will be costlier in development. Converting an APU to hydrogen fuel is straightforward. It was successfully done in the Airbus Cryoplane project.

51 Comments on “Bjorn’s Corner: The challenges of hydrogen. Part 26. Auxiliary power

  1. Thanks Bjorn. Good to see the breakout and relative efficiencies. I’d expect to see the conventional APU increase in size with the more-electric aircraft, transitioning to fuel cells with the LH2 aircraft.

    • on that point: how does the size of the 787-9 APU compare to the A350-900 APU (closest flying example of a modern conventional vs more electric architecture)?

      the PW APS5000 on the 787 produces 450kva

      the Honeywell HGT1700 on the A350 produces 1700 HP, which according to google is about 1695kva equivalent (of course it has to generate a lot of bleed air so it wouldn’t be producing that much electricity)

      • The HGT1500 original A350 APU provides 150KW electrical. Honeywell’s literature says the HGT1700 is also 150 KW electrical.

        • the APU HP numbers seem to be ground pressure referenced.
          Electric power ( and bleed?) must be available at cruise altitudes.

          APS5000 787 APU produces 2 time 225kW
          HGT1700 A350 produces 150kW plus hyd. ?plus bleed?

          both at 40k+Feet.?
          Powerlapse over altitude should be 5:1 ?
          hmm, spec available power should be reduced for cruise altitude.

      • The output from an APU is shaft power (the 1700hp which is 1268kW). This shaft power then transforms into bleed air in the load compressor, which is harder to quantify as to how much of this power it consumes (you need thermodynamic calculations that I do with GasTurb). The shaft power is also used to drive a generator which is the kVA. You can deduce the kVA from the shaft power by dividing it first with the efficiency of the generator which is about 0.9.

        So 500kW of the 1268kw goes to electric, and this is the only power offtake of that APU. What about the remaining 768kW? The APU can deliver this electric power all the way to 43,100ft, where the thin air reduces the APUs shaft power output. So it’s a margin for this demanding design point.

        The A350 HGT1700 is also a 1700hp shaft power APU, ie 1268kW shaft power out. It’s also specified to 43,100ft. Remember the A350 APU must produce bleed air for cabin ECS and engine start in addition to the electric 150kVA, which is a less efficient form than the 787 all-electric APU power delivery. We can, therefore, make a rough calculation that the HGT1700 delivers something like 150+270=420kW of joint power to the A350, the 80kW difference to the 787 APU is losses in the bleed air system, including its distribution, and then the 768kW is margin for high altitude operation.

        For the APUs it’s just different ways of delivering power to the aircraft systems (= energy when you sum it over time).

        • Think you have to add the 787 cabin compressors efficiency for the cabin air comparison withthe A350 APU load compressor provided cabin air.
          Both use Air cycle machines before air enters the cabin.

    • This article, though not outright wrong, gives the wrong impression. It’s a wonderful mix of semi-facts and quotes from experts who don’t know what they talk about (as I have pointed out over the last three years, none of these have put pen to paper and actually checked if stuff works, they just assume, following others that they assume know something).

      Every aircraft OEM studies all alternatives, and in their communication, they make sure to include the words that the community perceives as the positive ones. They are not lying by that, just now telling everything.

      The reality is that full electric doesn’t work beyond air taxis (<50nm range), hybrid makes little sense (complication for no gains) and hydrogen works up to single aisle. Then it’s SAF, Sustainable fuels. I have covered all this since 2017 in Corners, search for ePlane in our search box.

      • @ Bjorn
        I’m aware that the article is full of hot air, and that it mentions concepts that are not practicable — but, as you say/imply, the press triggers on certain catchphrases. One could even ask oneself if the press is being deliberately fed such catchphrases by OEMs as a way of projecting a “green image”. I only included the link because it’s an indication of the “buzz on the wire”.

        Personally, I attach greatest merit to use of truly sustainable biofuels, such as algal fuels. Hydrogen comes next on my list, and battery power is right down toward the bottom. I am absolutely not a fan of battery propulsion — the production of Li for batteries is a filthy technology, and then there’s the huge EOL/recycling problem after a number of years…not to mention the thorny issues of limited range and unacceptable recharging time. But the aircraft industry is going to have to contend with yups who are calling for electric aircraft — without having a clue about the subject — until they tire of the issue and move onto another “fad of the day”.

  2. IIRC from your article last week, the intent was to avoid taking power from the main engines, and have the “auxiliary power” be the main power supply for on-board systems. In this case the fuel cells or APU would run continuously on the ground and in the air, and only shut down when the aircraft is parked.
    I don’t know how that affects fuel cell performance over time, but I’m pretty sure that the APU and its moving parts will degrade over time. And more importantly, it will require a lot of expensive maintenance.

    Also, this hot piece of heavy machinery would be right next to the LH2 tanks (assuming it stays in the tail for the reasons you mention). Would that not present some thermal challenges, and some safety questions (H2 leak into APU bay, APU blow out…) ?

  3. The 65kg difference is equivalent to 15mn of operation for a 250kg/h.
    If the fuel cell is twice as efficient, the weight breakeven is reached after 30mn of operation (if the fuel weight is comparable, eg LH2 + container vs jet fuel)

    • Yes, thanks.

      I will spend the next Corner to sum this up and make yearly and lifetime projections of the different system architectures.

      • For cost comparisons I think it’s important to include the very sharp scale up in production of Fuel Cells that’s happening right now. In the time frame we’re talking about here, 2035, projections are that world wide productions of fuel cells will be in the millions of units annually. From single digit thousands in 2019 and double digit thousands this year, maybe triple digit thousands at end of 2021 (annualized capacity in Dec 2021, it’s a big maybe with Covid). Cost is expected to drop accordingly. This isn’t just Toyota, GM/Honda & Hyundai, but also more specialist companies like Ballard, Plug Power and Cummins. The Heavy Duty Specialty guys are finding ways to cut cost dramatically as well. So I’m not sure the assumption that implementation of a Fuel Cell will be costly is all that applicable in 2035, although it is certainly applicable now. Anyways, lots of uncertainty here, and difficult to project accurately.

  4. With a recuperator, turbotech-aero.com -backed by Safran) claims a much better efficiency, almost doubled for a small 55-90kW turbine. Maybe it could be scaled up to APU-sizes – but efficiency improvements would not be as good. 300kW recuperated microturbines are 33% efficient, but can reach 42% experimentally.

    GA Diesel engines are over 40% efficient and can reach over 1kW/kg. They are reliable enough, Continental claims one shutdown per 164’000 hours.

    • Recuperation really comes into play with LH2-based turbines (due to the LH2 to H2 process sucking up a lot of heat). You will hear a lot about this going forward from the engine industry. This is the potential of hydrogen gas turbine engines beating today’s carbon-fueled ones for efficiency.

    • It will be awhile before a fuel cell APU is certified and ready to be installed in aircraft. Even then, as Bjorn pointed out, it will be for the single-aisle market and smaller. In the meantime, as Marc pointed out, there is room for development and improvement of the conventional APU. I suspect the reasons were as they gave them, loss of immediate market potential. Significant barriers to enter that market right now with a new company.

        • The APU was linked to the NMA and it was going to cost money to duplicate what was out there and then you have to force it on the Airlines.

          More bad management of jealousy of ROI that Boeing can’t get because they put their efforts into pie in the sky and not building good aircraft.

          Huge losses over stupid moves.

          • Force it ?
            The APU is not a customer choice anymore , [Allied Signal/Honeywell on NG/Max], Boeing doesnt have engine choices for 737 and 777X either.

          • Agreed, but when I say force that means forcing a unknown and new product into the mix.

            Whole different parts source and failure modes and not a competitive cross offering.

            I guy gasoline and diesel from a known good source and do not risk the crap you can get at the fly by night stations.

          • The JV is to expand BGS into “uncharted” territories.

            FG: Boeing-Safran APU joint venture Initium frozen due to ‘cancellation’ of NMA

            Seems Boeing’s next new jet is further and further away than many expect.

  5. Are the turbines injecting LH2 or gasified H2 into the turbine. Have been thinking about it and made some calculation and according to my calculations you can only run a turbine using gasified H2 under a small power band without it becoming very complicated. Or is an APU small enough that the heat sink for the LH2 is not a problem. Fuel cells solve this problem because the efficiency difference between a large FC and a dozen small FC is small so instead of one changing the output of big one FC you change the number of running small FC

    ps. I think that this is the reason that the Airbus LH2 NB is a hybrid. Having a pure large gaseous H2 burning turbine with LH2 storage is just to complicated

    • The answer to an Airbus Hybrid is in Bjorns comments above
      “It’s a wonderful mix of semi-facts and quotes from experts who don’t know what they talk about “

      • I’m saying you can’t run an A320 size aircraft with only a turbine powered by LH2 because you get a problem with gasifying the huge amount of LH2 for a large power range. I have read his article. It didn’t look at the gasifying problem.

        • “The Turbofans can be optimized for delivering propulsive power, where the LH2 is supplied to the combustor injectors at the appropriate temperature via a heat exchanger in the engine pylon.”
          and
          ‘The major changes are in the combustor. It needs a different design to burn the gaseous hydrogen which enters at about 200K (-73°C). To take the LH2 from 20K liquid to 200K gas just before entering the engine a heat exchanger is used.
          https://leehamnews.com/2020/09/18/bjorns-corner-the-challenges-of-hydrogen-part-9-hydrogen-gas-turbines/
          It appears to be a change from 20K to 200K

          Max power is only needed at takeoff and then only for a few minutes. I dont think the huge power range you think exists in practice. Its not like military fast jets who slam the throttle from idle to full power in seconds over their flight envelope

          • 200K means the heat exchanger would only deliver 95K H2 gas if you double the load if you let the energy transfer be constant. Obvious you will get more energy out of the heat exchanger with more LH2 and a higher temperature gradient but not more than double.

        • nothing forces you to inject gaseous H2. LH2 works.
          look at any rocket engine ( H2/O2 fueled.)
          another solution is what the people at Reaction Engines do.

          • From memory Reaction Engines is very H2 rich and flying at very high aka cold heights. Neither is possible with a simple jet turbine powered aircraft.

            Rockets are only started ones. (in reality a few times but not like aircraft engines) and the injector is cooled by just blowing unburned H2 out to cool it to operating temperature. I doubt that that is possible with a busy airport.

            But it is more that running a standard turbine with 200K H2 gas is likely a simple solution that does not require a lot of change. 30k LH2 on the other hand is IMHO not a simple job of changing the fuel piping from kerosene to pressurized LH2. With 30K you get weird phase transitions, super conductive materials and materials that overall become brittle.

          • “flying at very high aka cold heights. Neither is possible with a simple jet turbine powered aircraft.”

            The change in temp from LH2 to gas gets it to 200K or -73C.
            What are the ‘very cold temperatures ‘ at altitude you are talking about ?
            At 30,000 ft outside air is -45C or so, and 0.28 atm, still ‘hotter’

            Then there is the engine compressed air as it enters the burners, much higher pressure as we can get overall pressure ratio in a modern turbine over 40 for say a PW1000 type engine. The compression raises the temperature as well. Maybe some type of intercooler stage for the high thrust takeoffs to improve efficiency as well?
            I know practically nothing about jet engines yet the issue you raise , seem at a basic level, to be a minor one.

          • Reaction Engines is a single stage to orbit design. What i remembered (wrongly) was that they cooled the air to liquid oxygen. H2 has about the same molar heat capacity as N2 and O2. LH2 has, if you include the energy needed to gasify an (obvious bullshit) temp of -10K. Maximum H2 contration of H2 so everything burns (2 H2+O2–> 2 H2O) is 2 parts for every 5 parts of air. 5*200K +2*(-10K)/7=140K. That is way above the temperature O2 condensates.
            Solution. Use way more LH2 than you burn.

            It seems their design is just cooling the compressed air to hot but not incredible hot temperatures but they still use more LH2 than they burn. This is (maybe) a good solution to go hypersonic but why not go to synthetic hydrocarbons with subsonic aircraft. There is likely no penalty in electric power needed between the fuel for a plane that burns all their hydrocarbons and a plane that uses only part of their H2.

            Difference between a kerosene turbine and a gaseous H2 turbine is that kerosene is liquid. Its density wont change that much with temperature of pressure. So a simple fuel injector that simply measures the volume of kerosene that is injected works great. H2 is a gas. density is very dependent on temperature and pressure. LH2, is super cold and if you include the vaporizing energy that it has a virtual super cooled gas temperature of about -10K. To get to 200K the intercooler needs to deliver a lot of heat. This can be done with a massive amount of not so much warm than 200K or a small mass of really hot. If you want to heat more LH2 in the case of massive low heat than the end result is almost 200 H2, but this is an airplane and mass is the enemy so it is more LH2 with a small amount of real hot. You can solve this a bit by running the intercooler pump faster but the work fluid will be colder so you still end up with colder H2 but with more H2 in the intercooler. Goes pressure even up in that case? How much volume should the injector pump so it is the same mass as injected LH2. How fast should the intecrooler pump run. Those are also not linear and likely to create oscillations.
            Changing the flow of LH2 isn’t that hard i suspect but for large changes it will be interesting

    • How about going back to H2 ICE + hybrid? BMW had built a hundred or so hydrogen powered ICE cars.

      • You mean piston engines? Knowledge problem. All the people who designed the aircraft engines are long death.

        • This is about Science, abstracted knowledge and not about trade secrets that can “die out”.

          • True, but the theoretical better solutions have much more practical experiences.

  6. It makes me wonder if it might not make sense to completely eliminate the APU and only have a large Lithium battery capable of running air conditioning, full flight system actuators and even electric ground taxiing. It could recharge from the main engines but also ground power.

    • APU is less of a short term energy supply source than
      a safety item. ( one reason why modern APU systems can be started at cruise altitude. That was not a given in olden times.)

    • Batteries are 200 times heavier for delivery of the same power*time to the aircraft during ground ops than hydrogen, it’s not a viable solution.

      • After the liquid hydrogen reaches a certain tank pressure it would need to vent. Obviously rather than venting to the atmosphere it would power the fuel cells (unless collection facilities were connected) , so there is no point to my idea on an LH2 aircraft apart from the weight issues you pointed out.

        I was thinking that if the ATR72 has no APU then the a new generation of certified Lithium batteries might provide the power level needed to replace some of the functions of an APU such as air conditioning, backup power for hydraulics so that main engine could be shutdown more often. I estimated a battery at current Tesla levels would provide the same power as an APU of the same weight for 20 minutes.

        • You could have checked your thesis, I found this immediately:

          The Tesla 85 kWh battery pack weighs 1,200 lb (540 kg) and contains 7,104 lithium-ion battery cells in 16 modules wired in series. (It has an energy density of 0.157 kWh/kg).

          The APU weighs 140kg and delivers way more energy than 85kWh. It delivers 400kW for a normal ground stop of 30 minutes = 200kWh. If the stop takes longer, no problem, just leave the APU on.

          Forget about batteries and airliners other than for firing up the aircraft and for short-term emergency purposes.

Leave a Reply

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