Can we Design Hydrogen-Fuelled Aircraft?

S H Salter, Engineering and Electronics, University of Edinburgh.EH9 3JL.

The collection of temperature measurements by David Travis following the 3-day grounding of all US civilian flights after 9/11 showed the astonishing effect of jet exhaust on the environment. If burning hydrocarbon fuel in the stratosphereever becomesa criminal offence, the aviation industry will have an interesting problem. A possible solution is the use of hydrogen as a fuel. Is this technically possible?

The Airbus 380 carries 250 tonnes of fuel with a total calorific value of about 1013 joules. Fuel is stowed in wing tanks but this would be a volume of about one eighth of the fuselage. The calorific value per unit mass of hydrogen is about 3.5 times that of jet fuel and so the weight of hydrogen for the same range would be only about 70 tonnes. Unfortunately the ratio of density of jet fuel to un-pressurized hydrogen is about 9000, so the design problem is how to reduce the volume ratio by about 2500. If we compress hydrogen to reduce its volume by a factor of, say, 100 we still have a fuel volume of 25 times the liquid fuel one or 3.2 times the fuselage volume. The cube root of 3.2 is 1.47 so by increasing all three fuselage dimensions by this factor we could have an aircraft with enough volume for all fuel in the fuselage but no passenger space. An increase by a factor of about 1.6 in both diameter and fuselage length would give enough volume for passengers provided they did not feel unhappy about being close to so much hydrogen.

The immediate reaction against the proposal will be triggered by embedded folk memories of the Hindenburg. Any use of hydrogen will need careful public relations. The Hindenburg survival rate was 64%, much better than crashes of modern conventional aircraft. Deaths were caused by jumping not burning. People who stayed aboard until the wreck reached the ground were unharmed. It is likely that the fire started in the fabric dope rather than the hydrogen. Because spilt hydrogen moves rapidly upwards there is much less risk than from a liquid fuel or heavier-than-air gases like butane or propane which regularly cause devastating explosions in boats and buildings. Furthermore the heat radiated by the invisible hydrogen flame is much lower than that from carbon particles in hydrocarbon flames. We can argue that hydrogen is actually safer than petrol and hydrocarbon gases.

We can spend the 180 tonne fuel weight-saving on gas storage bottles in the form of a low-permeability skin surrounded by wound carbon fibres. A helical winding of aluminium sheet with a low diffusion coefficient for hydrogen looks good. It can be made with the linear equivalent of spot welding. The axial stress in a thin-wall tube under pressure is only half the hoop stress, so we can use the gas tubes as fuselage strength-members. Once the fuselage bending moments are known, we can choose the wrap angle of the windings to give the right balance of directional strength. One structure might be a bundle of nine tubes in an hexagonal array with six full of hydrogen and three containing passengers. A cross section is sketched in the figure. Other configurations are being studied.

The smooth stress paths of the gas bottles would be badly disrupted by the conventional design of landing gear. Can we get rid of it? The requirements for processing the variable energy flows from renewable-energy sources have led to the development of new high-pressure oil machines using digital rather than analogue control of machine displacement. These machines have very high conversion efficiencies and very easy interfaces to computers (see ) . The extremely accurate control of very large energy flows allows many new applications. One of these involves replacing the landing gear of large passenger aircraft with a ground vehicle. Please suspend disbelief until you have considered the following facts:

  1. The landing gear of the A380 weighs 20 tonnes, say, 200 passengers. This weight is carried round the world for many hours and then used for only a few minutes on each flight.
  1. The landing gear occupies a substantial volume of the internal space. The volume restriction limits the travel of the landing gear and so increases acceleration forces.
  1. The requirement for openings compromises the structural integrity of the fuselage and adds weight, even more passengers.
  1. Landing gear must perform with very high reliability despite the weight penalty and extreme temperature cycling.
  1. The full weight of the aircraft must be passed to the ground through highly stressed points.
  1. Gas turbines are very inefficient for moving aircraft at slow speeds.
  1. On the A380 the shape of the landing gear doors and opening spoils the aerodynamic fairness.
  1. There is a severe design conflict between tyre weight, tyre life and braking performance.

An alternative might be to provide the function of the landing gear by a special-purpose ground vehicle. It would of course have to have VERY reliable links to the aircraft ground approach electronics so as to be in exactly the right place and moving with the right velocity underneath an aircraft on final approach. However there would be no weight, volume or temperature compromises.

The contact between the landing vehicle and the aircraft would be provided by a nest of large air-filled tubes like very large, very soft V-block, running the full length of the fuselage. This would spread the weight evenly into the aircraft skin. The tube surfaces could have vacuum suckers, like an octopus, which could apply shear forces evenly to the aircraft skin. The bags could be on a frame which could have hydraulic actuators to give a much longer travel than the legs of the landing gear. Tilting this frame would remove the need for the angling of the rear underside of the fuselage required to prevent ground contact at V-Rotate. This would further reduce drag during flight. The absence of fuselage penetrations could allow safe water landings for emergency. Runways can have parallel lakes presenting a much lower fire hazard if fuel is spilt. The impact loading on the runway would be much reduced and it might even be possible to revert to grass runways with several parallel operations from any wind direction.

The ground vehicles could use Diesel engines with much higher efficiency at taxi speed than gas turbines. They could have higher acceleration during take off and higher deceleration during landing. The hydraulic transmission would also allow regenerative braking, so the kinetic energy from one landing could be used for the next take-off. All-wheel steering and the option of direct side movement would allow much better use of ground space. The ground vehicle could have many more tyres, which need have no weight or volume compromise to achieve high braking. It could have an air-knife to dry runway surfaces and remove snow. There would be plenty of time to inspect and exchange landing vehicles and they would be in use for a much higher fraction of the time. The landing vehicles could gently lower aircraft on to passive supports at each loading pier and be used for other movements while aircraft were being boarded or serviced.

The volume of most aircraft wings is much below that of the fuselage and so there is not a strong reason to use gas tubes as structural wing members. However they would offer a way to offset the extra drag of the larger frontal cross-section. From the original work by Prandtl, it has long been known that sucking air from the upper surface of an aerofoil section will reduce the drag by an amount which far offsets the power needed for a suction pump. Schlichting in figure 14.9 of Boundary Layer Theorygives a graph showing a factor of more than two. An objection to suction on wings, where the outer skin is a structural member, is that perforations and slits cause stress concentrations. This should not apply to wing spars made as gas tubes supporting an unstressed skin.

It is important that using fuel does not move the centre of gravity of the aircraft. This happens automatically with fuel stowed in wing tanks. If large quantities of fuel are to be stored in the fuselage it will be necessary to have the centre of pressure of the wings close to the centre ofgravity of the fuselage-engine combination. The choice of a ground-based landing vehicle suggests high wings and engine placement above the wing. In theory at least, this will give some advantage from higher air-velocity over the upper wing surface and lower noise transmission to ground level. It is much easier to service and inspect equipment if you do not have to reach above your head. Cranes lifting an engine upwards are much more convenient than forklift trucks working from below. While some change in the architecture of maintenance hangers would be required, high engines accessed from above would by no means be unwelcome to ground crew.

Gas tubes may not be ideal for connections to a low-chord wing and so the longer attachment line of a delta wing, such as used in the Vulcan and Concord and many fighter designs, should be investigated. Suction may be able to offset some of the disadvantages of the delta wing as applied to civilian aircraft provided always that they can land safely after a failure of the suction system. A delta wing with a deep thickness and a leading edge made from very strong but transparent material, perhaps poly carbonate, might even allow passengers to sit in the wing enjoying a splendid view if their vertigo allows.

The range of the A 380 is 15000 kilometres. While this may have been chosen for passenger convenience with the properties of present fuels, it is larger than necessary for trans-Atlantic flights and could allow a further volume reduction. The San Francisco to Sydney distance is only 12000 km and stops in mid Pacific could be very attractive.

Before we waste time on radical new aircraft designs and ground-based landing systems, it is necessary to confirm that burning hydrogen in gas turbines at high altitudes will be a chemically appropriate solution. If we burn hydrogen in ambient air there will be no release of carbon dioxide but there will still be the formation of nitrogen–oxygen compounds collectively known as NOXes. If these are cooled very rapidly, as in the adiabatic expansion of an internal combustion engine, they can be ‘frozen’ at the high-temperature equilibrium state with lots of very nasty acids. The lower combustion pressure and slightly slower cooling of a jet exhaust should be less severe but we want to quantify the severity of the problem. There may even be problems from ice crystals formed from the exhaust. I have asked colleagues at the National Centre for Atmospheric Research at BoulderColorado for an opinion.

There is one engine design in which the combustion products cool slowly enough for almost all the NOX production to revert to ambient values. This is the Stirling engine originating from 1815 but abandoned because of the absence of materials with good thermal conductivity and high hot strength. Much better materials are now available. By an extraordinary coincidence, the digital hydraulic systems needed for the speed and accuracy of the ground-based landing gear can also radically change the design of Stirling engines by using hydraulics to replace the crank and connecting rods of the conventionalStirling engine. A Stirling-engined aircraft would probably have to use a ducted fan or propeller propulsion but these could still allow civilian aviation to continue in a NOX-sensitive world.

The best way to do experiments on high-altitude engine-chemistry might be from a balloon. Do we know anyone with an interest in this area?