Professional Documents
Culture Documents
Vol. , No. ,
Buoyancy control is identified as a significant problem in airships, past and present. An existing proposal to
enhance the control of buoyancy in lighter-than-air aircraft is reviewed, and a new method of buoyancy control (the
inflatable vacuum chamber) is proposed. The differences between the existing proposal and new proposal are
highlighted, including the new possibility of an airship that lands vertically onto the ground without a mooring mast, a
ground handling party, or a runway. Theory of lightweight stiff inflatable structures is discussed and a specific design
and construction method is proposed. The design is optimized and confirmed elastically stable; design specifications
are given. A typical operation cycle (including liftoff, flight, and touchdown) is detailed. An area of potential
theoretical advancement is identified.
I.
Introduction
II.
BARTON
Fig. 1 The USS Los Angeles after being struck by a tailwind while
moored on 25 August 1927.
III.
Fig. 2
Inflatable-Vacuum-Chamber Concept
BARTON
We average this for all orientations of the fibers to find the average
elastic energy per unit volume of fiber material:
Z Z
1
u sin d d
4 0
2
x y 2y y z 2z z x
E x
10
15
10
15
10
15
u
this problem, but as the design load further lightens, it becomes nearly
impossible to maintain the ratio of stiffness to strength, hence the
difficulty in designing lightweight vacuum containers as compressive
structures. This is the motivation for the idealized inflatable solid
just described. Its design load can be reduced arbitrarily while
maintaining its strength-to-stiffness ratio. Thus, it can then be shown
that, if this idealized spherical shell is sufficiently thick, the linear
mass density of the fibers is sufficiently small, and the space between
the fibers is sufficiently pressurized; the structure remains elastically
stable when vacuum is introduced into the central space. Let us
further consider this pressurized network of fibers.
In the idealized case, a pressurized matrix of tension fibers can be
described mechanically as a continuous solid having a specified
strength and elasticity. We will here show that, if the tension fibers are
isotropic, the idealized bulk solid has isotropic elasticity and
Poisson's ratio of one-fourth. We will also show that the elastic
modulus per mass and single-axis strength per mass of this idealized
continuous material is one-sixth that of the fiber material.
We imagine a dense isotropic network of thin elastic tensile fibers
having all possible orientations:
2
3
cos
a 4 sin cos 5
(2)
sin sin
(4)
(5)
(6)
Fig. 4
BARTON
Fig. 5
IV.
Proposed Structure
V.
VI.
Fig. 6
BARTON
Fig. 8
mass per unit volume of vacuum that has both elastic stability and
positive tension in all membranes. As shown in the Appendix the
mass of gas the structure contains when vacuum is not present is
directly proportional to the total membrane mass. Thus, the weight of
the pressurizing gas is not considered when optimizing the structure.
The optimized structure is found to have an outside radius that is
1.50 times its inner radius, 44 lobes, and a pressure of 3.08 absolute
atmospheres in its wall. The failure mode of the structure is shown in
Fig. 8. Tables 1 and 2 give the specification of the specific design.
VII.
System Operation
Parameter
Value
General design specifications
BARTON
Table 2
Weight components
Fabric weight
Extra helium required to pressurize
Helium removed when evacuated
Total weight added to airship when fully charged
Reduction in payload assuming discharge on takeoff
Total buoyancy control range
Weight of air displaced when central space is evacuated
Weight of air undisplaced when central space is vented
Weight of air immediately undisplaced on venting
Weight of air displaced when wall is vented
Weight of air immediately displaced when wall is vented
Table 3
Initial state
2.998 tons
1.042 tons
0.276 tons
3.764 tons
2.998 tons
7.552 tons
2000 kg
2000 kg
1429 kg
7552 kg
4531 kg
Pressure Vacuum
First
Second
C
C
1.43 t
0
C
C
1.43 t 4.53 t
C
C
3.10 t
0
C
D
4.53 t
0
Delayed
weight
change
0.57 t
2.45 t
2.45 t
3.02 t
Final state
Pressure Vacuum
C
D
D
D
D
D
D
D
C charged, C discharged.
Note: Delayed weight change occurs as adiabatic heating/cooling
dissipates.
wall of the IVC remain open allowing the reduction in ambient
pressure to further evacuate all parts of the IVC, thus buoyancy
remains balanced.
Once an altitude of 10,000 ft is reached, valves to the IVC are shut,
locking-in the density of the IVC system. Because of the fixed
volume of the IVC the airship will tend to stay at 10,000 ft. Avariation
of 1000 ft in altitude will create a change in buoyancy of about
0.13 tons that will tend to return the airship to 10,000 ft. This
oscillation of the airship's altitude will have an approximate period
of 17 min.
For the next 10 h or so, the airship flies its course of approximately
1000 km during which it consumes almost all of the 5.55 tons of fuel
added. Simultaneously, a helium pump (thought to weigh about half a
ton) removes helium from the interior of the IVC and from the
envelope of the airship and forces it into the wall of the IVC to create a
near vacuum in the interior of the IVC and a pressure of close to
28.0 psi gauge in its wall. This net compacting of the helium results in
a reduction of buoyancy of close to 5.55 tons, precisely balancing the
weight of lost fuel. The net buoyancy remains at 0.00 tons, and the
fuel tank is close to empty.
As the helium pump continues to increase the gauge pressure in the
wall of the IVC beyond close to 28.0 psi (where close to corresponds
to the small weight of fuel remaining), the altitude for aerostatic
equilibrium (which was 10,000 ft) begins to decrease, and the airship
begins to descend. As the density and pressure of the ambient
atmosphere increases, the buoyancy of the IVC system begins to rise
and (in absence of continued pumping) the gauge pressure in the wall
of the IVC would begin to drop. Thus, pumping continues to surpass
changes in ambient pressure and bring the gauge pressure in the wall
of the IVC to close to 30.6 psi as the airship approaches the ground.
Net buoyancy remains at 0.00 tons.
As touchdown approaches the pilot prepares to vent the vacuum in
the central space of the IVC. When first contact with the ground is
made an auxiliary air ballonet in the central space of the IVC is vented
rapidly decreasing the net buoyancy by 1.43 tons. Net buoyancy is
now 1.43 tons and, thus, traction with the ground is secured. During
the next 15 min, as frictional heat dissipates to atmosphere, an
addition 0.57 tons of weight are added to the ship increasing traction
with the ground. For rapid return to the air, refueling and superheating
of the lifting gas occur simultaneously. The auxiliary air ballonet is
gradually emptied by allowing helium into the central compartment
of the vacuum tank.
Note that in this proposed system, the IVC vacuum space is finally
filled with helium and not air. Note that the buoyancy of the ship is the
same in either condition. In both cases, some gas already in the ship is
moved into the IVC to make room for more air to flow into the airship.
The addition of new air to the airship is what actually changes the
weight of the ship. Additionally, moving helium instead of air into
the IVC has the advantage of increasing the pressure height of the
ship whenever vacuum is vented. This might be useful in the event
that the IVC system is used in-flight for disaster avoidance. A similar
argument shows the advantage of pressurizing the wall of the IVC
with helium. Furthermore it is useful for the central space of the IVC
and the wall of the IVC to contain the same gas, so that only one pump
is required to manipulate the gas. Alternately, placing the vacuum
system into an air atmosphere (and filling it with air) has the
advantage of being able to rapidly exhaust gas to the exterior of the
ship (which reduces the retention of thermally disturbed gas on
takeoff, eliminates the need for overpressurization protection in a
nonrigid airship, and simplifies structural considerations in a rigid 2
airship (see Sec. IX). It has the disadvantage of permanently reducing
pressure height and/or payload capacity of the airship.
VIII.
BARTON
276 and 55 MJ, respectively. At four times the volume and one-fourth
evacuation these become 148 and 24 MJ, respectively. Assuming
adiabatic compression in the pump and isothermal compression in the
tank, the energy E required to pressurize or vacuumize a volume V to
a pressure P in an atmosphere of pressure P0 and inverse relative heat
capacity at constant pressure (2/7 for ideal diatomic gasses) is given
in Eqs. (7) and (8), respectively:
1
P V
PV
P0 V
P
Epressure 0
(7)
1
1 P0
Evacuum
(8)
IX.
When the airship lands the central space of the vacuum tank must
be vented to gain traction with the Earth. There are several ways this
can be accomplished. A first option is to vent directly with atmospheric air resulting in direct weight gain. This first option has the
disadvantage of reducing payload capacity or pressure height of the
airship. The airship can reach full pressure height only if 100% of its
gas volume is occupied by helium. The presence of any air volumes
that cannot be exhausted results in a loss of payload capacity or
pressure height. A second option (which resolves this first issue) is
to vent with helium from the lifting-gas volume causing the volume of
the lifting gas to decrease and, thereby, causing the air ballonet to
increase in volume creating the increase in weight indirectly. With
this second option, one finds that the air ballonet blower is generally
unable to keep up with the massive volume of air that must be moved
to prevent implosion of a nonrigid airship. Even in a rigid airship, it
X.
BARTON
Utank
5
NkT tank
2
(9)
P0 V atm NkT 0
(11)
5
U0 NkT 0
2
(12)
In the final state, its internal energy must increase by the amount of
atmospheric work:
(13)
P
v2
P
0
0
2
where
NkB
CP
2
5
2
7
(14)
where A is the area of the orifice through which the tank is vented. The
average thermal energy in the tank during and immediately after
venting is equal to the sum of the thermal energy and work energy that
has entered the tank during venting (see Sec. X):
1
1
1 NkT tank
1 NkT 0 P0 V
1
T0
NkT 0 T tank
1
This heating of the gas (which comes from the work energy that was
required to create the vacuum initially) results in the tank filling to
only 3/5 and 5/7 of its total mass capacity for monatomic and
diatomic gases, respectively. The tank fills to full capacity only once
this frictional heat has dissipated. In a medium-sized airship this will
take about 15 min. The pressure in the tank is given by
5
V atm V tank
7
Here, P and are the pressure and density in the flow, respectively,
and P0 and 0 are the stagnation/ambient pressure and density. The
adiabatic condition requires
1
P
0
P0
(10)
This same gas in the initial state also obeys ideal gas properties:
Utank U0 P0 V atm
XI.
When the vacuum tank is vented some time is required for the gas
to move into the central space. Once the valve is opened gas accelerates toward the valve opening expanding adiabatically according to
Bernoulli's equation for compressible fluids:
Ptank P0
V T tank
P1
V0 T0
V 0 1 1P1 P0
t
0
1
V 1
dV 0 p
A
V_
ss
20
P1
1
P0
P0
Thus, when the tank volume is 1632 m3 and fully evacuated (not
oversized) and fitted with a valve having an orifice area of 0.25 m2 (a
valve weighing approximately 200 kg), the fill time for helium and air
is 11.3 and 42.9 s, respectively. (Note that fill time for oversized tanks
with a valve of the same weight is less.) Helium and air will fill the
tank to three-fifths and five-sevenths full mass capacity, the mass
defect being related to the frictional heat of filling the tank. When the
medium is air, one might consider exhausting this heated air to
atmosphere to replace it with air of ambient temperature to acquire the
BARTON
XII.
The vacuum tank can fail in two primary ways: leakage and
explosion. First, the vacuum tanks gas impermeable skin can fail,
leaking helium into the lifting-gas volume of the airship. In the case of
slow leaks, the helium pump can replace the leaking helium with no
effect on the airship. In the case of faster leaks, the vacuum volume (if
charged) can be used to temporarily take up some of the leaking
helium thus temporarily preventing the buoyancy of the ship from
changing. As pressure continues to drop in the wall of the vacuum
tank the vacuum must be released to prevent geometric instability of
the vacuum tank structure. Finally, when pumping and venting of
vacuum can no longer keep pace with the leak, the extra helium will
begin to accumulate in the lifting-gas volume forcing air off the ship.
This loss of air will result in an uncontrollable increase in buoyancy.
A discharge of lifting gas may then be required to prevent a breach of
pressure height and safely land the ship.
If the leak is fast or results from a violent tearing of the vacuum
tank's fabric, the available pathways for exhausting air may be insufficient to prevent overpressurization of the airship and the resulting
structural damage. In anticipation of such difficulties, one could
potentially design the airship to fail in a very specific way when overpressurized, so that pressure is quickly relieved, damage is confined
to a single area, and the remainder of the lifting gas is retained. This is
the subject of future research.
After the vacuum tank system has failed it can no longer provide
the buoyancy control function required to land the ship. An emergency landing will be secured by other means (loss of aerodynamic
lift [runway landing], mooring, ground handling party, vectored
thrust, jettison of lifting gas, etc.).
XIII.
XIV.
Conclusions
0
2
Z
composite volume
6
tr6
4
Pgas
7
07
5 dV
(A1)
Note that the tension in the composite is shown to occur only in the x
direction. This is clearly not true but also does not affect the result,
because the trace is invariant under rotations of the coordinate
system.
Because the stress tensor at the boundary of the volume is
everywhere zero, and because the divergence of the stress tensor is
zero (momentum conservation), it is true that all volume ij dV 0ij .
Also, because the trace of the integral is the integral of the trace, it is
true that
Z
Z
tr ij dV tr
ij dV tr0ij 0 (A2)
all volume
all volume
3Pgas V gas
thus mcomposite
composite working load
3Pgas
composite
gas composite working load
(A3)
Acknowledgments
The author would like to acknowledge Jack Sams for bringing
together the right people to generate this idea, Ian Winger for
developing techniques for constructing inflatable structure from
fabrics, Helena Safron for editing this document and illuminating
errors in the author's logic, Edwin Mowforth for sharing his lifetime
of experience with lighter-than-air aircraft, and Frank Flaherty and
Robin Winton for useful discussion. The author would also like to
acknowledge David Van Winkle for making possible the trip to
Denver to present this work, Juan Ordonez for administering the
continued fabrication research, and the anonymous reviewer that has
helped bring the manuscript into its current form.
Reference
[1] Mowforth, E., Improvements, Airship Technology, edited by Khoury,
G. A., and Gillett, J. D., Cambridge Aerospace Series 10, Cambridge
Univ. Press, New York, 1999, pp. 359384.
10
BARTON
BARTON
11
Queries
1. AU: Please review the revised proof carefully to ensure your corrections have been inserted properly and to your satisfaction.
2. AU: As per journal style, section citations in the text denoted by their respective label numbers only.
3. AU: Please check and confirm the references and their respective citations in the text.