ABOUT THE AUTHOR: Al Bowers is a Senior Aerodynamicist
for NASA at the Dryden Flight Research Center located on Edwards AFB in
the Mojave desert of California. His presentation reflects his thoughts
based on NASA's work and does not attempt to interpret or represent what
airlines or aircraft producers think or are going to do.
(Note: The actual slides range in size from
49k to 104k, average is about 60k)
Blended Wing Body (BWB) is being considered as the next generation commercial
airliner. The trend is towards larger aircraft that can carry more
people, economically while reducing the number of operations from
airports. He noted that recent surveys have identified about 60%
of the delays are due to the number of aircraft saturating the airspace,
as anyone who has been delayed can attest, the ramps and runways of airports.
This movement of more people on fewer aircraft has been defined by NASA
as “The Lure of Large Aircraft”. There are a lot of other infrastructure
problems that also need resolving like terminal congestion, parking facilities
and, adequate loading gates.
There is a very competitive
large aircraft market as illustrated by the AirBus decision to produce
the A3XX that could carry about 650 people on two decks. The intra-Asian
market is another area that can utilize high density loading. They
are already doing it with Boeing Super 747s rigged for full economy seating
to haul 550 people over the short distances between cities. The trade
off is less fuel, but it isn’t needed for the short runs. This is
going to be a problem for the Chinese in about 10 years as they become
more affluent and want to travel throughout their country.
Another aspect of large
aircraft design is the ability to adapt it to the all cargo market.
Al didn’t hasn’t really seen the full logic behind the idea yet, but NASA
is pursuing it. With used Boeing 747s available at relatively low
prices, along with other smaller aircraft that are readily available, the
market for a new large cargo hauler may not be as great as expected by
NASA. However, the military gets interested it design and helps defray
some of the startup costs, then the picture for the commercial markets
Gavin asked the question
about whether or not the design would allow for doing quick conversions
between people and cargo moving to get better airframe utilization.
Al noted that due to internal structure and layout of this particular design,
it would be very difficult to do the conversion on a daily basis.
The interior design includes a lot of chordwise bulkheads to form several
different passenger compartments across the span of the center section.
There would be more passageways to negotiate with seat pallets getting
them to the doors and removing the overhead compartments.
Al went on to say that
you really have to start thinking differently when it comes to unconventional
configurations, but there are potentials for breakthroughs. Eventually,
someone will take the bold step to do the development work on these designs
and then sell them to the air traveling public. This is one of the
biggest questions that there isn’t a good answer for right now. People
have been used to the “tube with wings” concept for almost 90 years and
it will take some doing to get them into an unconventional one.
(ed. – Some years
ago there was a proposal for multi-blade, external fans for aircraft like
the MD-80, but surveys found the public wouldn’t fly on them because they
had “propellers”. But maybe the introduction of aircraft like the
B-2 and some of the next generation fighters currently starting qualification
testing will turn the tide toward flying wing acceptance by the public.)
So where are these potentials.
The biggest kicker is to take the body of the airplane and morf it with
the wing, then you get a body that produces lift merging with
the spanloader idea. You can’t take it to the point of a true flying
wing due to the added wing area at the outboard ends creating too much
drag. So you end up with a blended wing body that looks like the
one below. The lift to drag ratio can be increased from something
like the 747’s 17 to the a range in the mid 20’s for the BWB. This
savings in drag translates into substantial economic and environmental
benefits. This particular model would be expected to use 20-25% less
fuel, require 10-15% less weight (or conversely allow for more paying payload)
and result in 10-15% lower direct operating costs.
This was all started by a
design study in 1989 by Dr. Dennis Bushnell, Chief Scientist at NASA Langley.
He foresaw the need for a commercial aircraft that could carry 800 passengers
over 7000nm and a speed of .85 Mach. This was the result of that
design study which was originally McDonnell Douglas’.
One of the more interesting
facets of this design was the position of the engine inlets. Since
they are right down on the wing surface, they are ingesting the boundary
layer so any airflow sucked into the engines can be ignored as drag.
This gives a huge increase in the L/D due to the decrease in drag.
There are also a lot of control surfaces on this version, however, the
larger inner surface has been eliminated in follow-on designs. As
part of what Al was talking about earlier, notice the 290’ span that won’t
fit into the current passenger terminal infrastructure. This makes
this configuration non-viable as a solution to the high density passenger
Al then moved from the outside features to the inside layout of the airframe.
The diagram shows how thisapplies the spanloader concept by having the
weight out where the lift was being produced. The passenger compartment
goes out into the wing structure area which is obviously different that
a conventional fuselage. Outside of the passenger area are the main
fuel tanks which also run out into the wings, further moving weight out
to the lifting areas. This is entirely different than the point loads
of the fuselage arrangement.
In an overlay comparison of the BWB to the 747, you can graphically see
why there is a problem with this particular BWB design. You can park
747s side-by-side at current passenger terminal gates, but the BWB’s 290’
span makes this impossible. Both Boeing and McDonnell Douglas looked
into folding the wings like aircraft carrier jets, but determined that
the public would not like to fly on an airplane that looked broke.
Another idea was to caster the wheels so the aircraft could come into the
gate area slightly sideways, but this means higher weight in the landing
Staying on the inside, Al put up a slide of a full scale mockup of a section
of the passenger compartment. One of the first questions everyone
asks is where are the windows. In this design there are no real passenger
windows, but each seat will have a multi-functional LCD screen on the seat
in front of them. A selector will allow the passenger to select from
a number of views, including looking to the rear and straight down.
The other obvious thing in
the pictures are the really heavy structural walls between the compartments.
Al now went on to answer Ralph Wilcox’s question about how hard is it to
pressurize a square box versus a cylinder. The heavy walls are one
of the ways and due this extra weight they also cut into the ultimate potential
gains Al talked about in the first part of his presentation. However,
he also commented that it is expected enough gains will be made on the
aerodynamic side to offset the extra structural weight. Gavin asked
about putting a series of round section within the wing to carry the pressurization
loads. Al commented that this was looked at, but in the final analysis
it was determined that weight wise it is better with the current design
parameters. He did note there are some fatigue questions that still
need to be worked out before there is any commitment to building something
like the BWB.
Al moved along to the direct operating cost analysis between a 747, a new
conventional design like theAirbus 3XX, and the BWB at the year 2015.
The numbers all show the BWB makes gains in the areas of operating costs,
fuel efficiency, gross weight and nitrous oxide emissions. This last
item is of great concern to NASA since they have been linked to the green
house gases. Here there was a 17% expected gain for the BWB predicated
on the fact there are no major breakthroughs in engine design during this
period. Some of the gains will come from a combination of many little
improvements over the entire airframe versus one or two major improvements.
Gavin ask Al whether
or not the airlines would be behind these types of changes in aircraft
design. Al commented that in his opinion changes in the environmental
laws will probably create the need for such aircraft to meet things like
emission standards. If airport and airspace congestion rules are
changed, the aircraft will have to change and the airlines will go along
because they will have too.
The next slide was a comparison of the benefits and challenges. The
benefits include: lower operating costs; lower production costs; reduced
airport/airspace congestion; lower fares; reduced environmental impact
and; improved safety. Operating costs he had already covered.
Lower production costs come from not have as many tight bends so the manufacturing
costs go down. Although the number of aircraft at terminals won’t
go down, they will be moving more passengers with each departure which
will impact congestion by preventing its escalation. It is felt this
design concept is at least as safe, and possibly saver, than a convention
The challenges included:
structures and materials; aero-structural integration; aerodynamics; controls;
propulsion-airframe integration; systems integration and; infrastructure.
Structures is back to the pressurization issues and the integration issue
revolves around making the structure clean enough to work aerodynamically
and achieve the savings potential.
Aerodynamics is a separate
issue from the aerostructures. Imagine that this design has an elliptical
span load associated with it, so that is the minimum induced drag for this
vehicle. Then think about the lift coefficient that needs to be produced
for this type of wing. Since the span load is chord dependent the
center body section with its wide chord had no problem meeting the requirements.
But as you move out towards the tips you reach a pinch point where the
chord narrows sharply. The problem is going to tip stalling due to
the high taper ratio and the loading out there. This is a problem
for the aerodynamicist since the aircraft must takeoff and land.
This means you need to generate high lift coefficients, which puts you
close to the stall, which is also close to the departure. Of course
the last thing you want is a passenger aircraft with bad departure characteristics,
so how do you get the lift coefficients at the pinch point to avoid these
problems or at least degrade elegantly so you don’t lose control of the
Ralph asked the
question about boundary layer control at that point on the wing.
Al commented that it had been looked at and there was still a problem even
using vortex generators. Boeing went to slats on the outboard section
since this would generate a lot of lift on this portion of the wing.
The disadvantage of this system is that slots and slats have really bad
hysterisous effects, so once stalled it might be hard to get back.
NASA is still looking into this area.
On systems integration,
Al noted that this area is becoming more and more complicated. This
is the digital fly-by-wire systems so you can control the way in which
the aircraft reacts to the control inputs. This particular design
has a nose slice just before reaching the stall, so some method is needed
prevent in inadvertent departure. The digital controls with its accelerometers
and other sensors feeding back information, the control surfaces in the
affected area can be deployed upward to decrease the span loading and move
it inboard. This will prevent the airplane from departing, but it
is so sensitive that any external changes can have major effects on the
Gavin asked a question
about what types of construction techniques would be used for this aircraft.
Al said it was planned to be built by bending tin and used the 747 as an
example. If you look at the outer wing panels on the 747 and compare
them to the same panels on the BWB you find they are very similar.
Since the BWB was originally a McDonnell Douglas design, the outer wing
sections were based on the DC-12 which was never produced because of the
buyout by Boeing. This then became a good starting point so the center
body construction problems became the focal point of further development.
Al moved on to the really big issues of structure and aero-structural integration;
non-cylindrical pressure vessel. How do you pressurize something
that doesn’t look like a tube or a sphere. Initial thoughts were
to use conventional metallic structures, but more recently thoughts have
been turning towards composites like graphite stitched epoxy resins.
They are questioning whether this would help with the pressure structure
problems and perhaps also save some weight.
There is another issue
with joints between the various panels. One of the things NASA does
with their test aircraft is go through a ground vibration test. Hopefully
this predicts what the structural modes are in the wing. The is a
mass suspended by a fairly rigid beam structure which will vibrate at a
particular frequency and a guess is made as to what it will be based on
the existing structure.
With metal airframes
there is an I-beam with a plate on the top, the skin, that is riveted in.
It turns out that due to the factors of give, flex and friction the actual
frequency actually, when tested, comes out lower than the prediction.
This goes back to the fly-by-wire system where the pilot can make a jerk
input to the stick which would give an almost perfect square wave input
to the system. The system looks at it as a change to the angle of
attack. In most airplanes the change would occur gracefully with
some overshoot and then stabilize out, which is the short period frequency.
If this frequency is the same frequency as the structural wing bending
the aircraft will catastrophically fail. The pilot can’t be told
not to make these types of control inputs, especially if they are fighting
an aircraft in turbulence while landing.
Now we bring in the
composite structure. Some composites joints are glued together and
other are not, so in some cases there are butt joints where the load transfers
are harder to calculate. In tension and compression there is pretty
good data, but not in the bending. Apparently the joints don’t handle
the stresses that same way in each direction so this makes the calculation
much more difficult. At this point in time there just isn’t a lot
of experience on how to handle these types of joints on airplanes.
This is due to the load having to transfer from one skin, through the flange
or other connecting structure, to the other skin. Since the cloth
fibers are not running continuously along the known stress line, the calculations
become much more complex.
Another major issue
that will need to be worked in the future, but is not a top priority at
this point, is the outer surface “bulging” that will occur as the aircraft
is pressurized. These bulges will form in-between each of the main
structural bulkheads forming the passenger compartments. Obviously
this will deform the elegant cruise airfoil shape that is being planned,
so it has to be taken into consideration in the design. When doing
this with composites it becomes even more difficult due to the lack of
experience in this area.
Another set of challenges that up early on were just the aerodynamics.
This is still looking at an 800 passenger, big BWB aircraft with three
engines sucking in air over the upper surfaces boundary layer. After
running computation fluid dynamics to predict what the flow would be, they
found two areas of reverse flow. One was where the lift coefficient
was very high, right at the sharp break in the trailing edge and right
over the control surfaces. The other problem is putting people inside
the airfoil which now has to be thicker than people are tall. This
gives you a very thick, transonic airfoil which sets up new challenges
to overcome the various shock waves and resulting wave drag. All
these are bad things that need to be addressed and viable solutions found
is another issue that needs to be looked at. When you put all the
control surfaces in close together in one location fore and aft, the whole
aircraft is affected by turbulence at one time. So given the same
level of technology in control feedback systems as currently used on conventional
aircraft with control surfaces spread out over the wing and tail surfaces,
the ride quality of the BWB, or a flying wing, will be worse. New
feedback technologies need to be developed for the improving ride quality
without letting things get so sloppy that the pilot can’t control the airplane.
Along with the
digital fly-by-wire comes the all electric subsystems; everything is electronic
onboard the aircraft. This is a very common in the military in aircraft
like the Air Force’s F-16, but is just now starting to become more prevalent
in commercial aircraft. Doug Fronius asked if this meant no hydraulics
and Al commented that there were still hydraulic actuators being controlled
electrically at the site of the control surface. Doug noted that
the next generation of military aircraft coming along will truly be all-electric
with no hydraulic subsystems. Although there are some supporting
subsystems on commercial aircraft at the present time, both Doug and Al
indicated all-electric main systems were probably a long way off.
asked about the thickness of the boundary layer. Al said he wasn’t
sure about the numbers, but did estimate it could be a couple of feet thick
at the rear of the centerbody section of the wing. The chord at this
point is close to 160’, so Al wouldn’t be surprised if the it reached these
Ralph Wilcox asked about what the wing
loading would be for this type of wing. Al commented that when you
go to a flying wing you need to bring the numbers back down from what you
would have for a conventional aircraft. So instead of having the
105-120 numbers for things like the Boeing 7X7 series, you need something
like 95-105. This is easier to do on the BWB since you have so much
more wing area.
Gavin Slater asked
about the problems associated with boundary layer ingestion (BLI) on the
three engines at the back of the aircraft. Engines don’t like to
see a lot of distortion at the compressor face and, this will become even
worse at high angles of attack. The boundary layer will get much
thicker and it presents many problems in designing ducts or making changes
to the engines themselves. (ed. – This problem has led to a different
engine configuration as shown below. The engines have been moved
up on pylons to get them out of the boundary layer flow at all times.)
Al moved on to the infrastructure problems noted earlier. The ICAO
wants to stay with the 80m (262 ft.) wingspan separation between terminal
gates, so this becomes an intractable problem. The 800 passenger
version was also a double-decker which presented another set of problems
with the existing gate structures. The two decks also raised concerns
over passenger safety in a crash situation where the upper deck could collapse
onto the lower one. The last issue was how to you handle 800 passengers
for several airplanes at a time in terminals not designed to handle that
volume of humanity.
SO, this all had the engineers at Langley pulling their hair out.
The upper left segment of the slide shows a 1% spin tunnel model at Langley.
They tried a couple of different models and the one that is going to be
flight tested showed high yaw rates. There also appears to be an
auto-rotation tumble mode that was observed in the tunnel tests.
The lower left segment shows a model in the acoustics chamber being tested
of radiation’s from things like the radio antennas. The other pictures
on the slide show some of the various models that were put through further
tunnel testing leading up to changes in the configuration that will be
Al then laid some background on the changes that were being made for the
upcoming flight test model. The top picture below are of the models
Ilan Kroo and his graduate students at Stanford University built and flew
as proof of concept vehicles. The top one is a 6’ R/C model flown
with fairly stable static margins. There were two versions, one with
gas power for longer flights and one with electric power (cleaner and quieter).
The lower picture below is the 17’ version with a true closed loop control
system using a MacIntosh laptop computer as the processor for this system.
This was a twin engine, gas powered model with multiple control surfaces,
each one controlled by an electric actuator designed and built by the students.
Ilan also designed the internal instrumentation systems.
The large number
of control surfaces presented their own unique troubles. Since any
one control deflecting upward will cause the aircraft pitch up the question
became one of how best to control the mixture of movements. They
had to determine whether there were any pitfalls like airflow separation
as the controls moved in differing amounts. Some of these questions
are being answered through wind tunnel tests.
Andy asked Al
what caused the phenomenon on the model where the control surfaces appeared
to be flopping around during ground taxi. Al indicated the controls
were responding to the inputs from onboard sensors being activated by bumps
in the taxiway. Due to the low airspeed associated with taxiing,
the control deflections to correct the sensed conditions were large and
therefore drove the surfaces to their stops. The electric actuators
are very fast, so it appears the surfaces are just really loose on the
hinges. Al noted this occurs on current fly-by-wire aircraft in the
military as they taxi, you just have to look closely to see it.
As Al continued,
he noted one of the big questions revolves around something they call “behavior
quantification”; how does the aircraft behave, what does it do. They
have problems with stall, the aeroelastic properties cause dive problems
during the pull out, and a mach buffet with an “ugly” tuck.
Then there is the issue of engine out performance since it doesn’t have
real strong directional stability. If you loose an engine, what is
the Vmca really going to be? The last issue is stability margin in
terms of pitch, lateral and overall directional control.
Up to this point
everything that he had been talking about revolved around the 800 passenger
version that was studied under a NASA contract that ended in 1996.
At that time the Douglas division of McDonnell/Douglas
picked up the study and continue it on their own money. These studies
resulted in design you see below, which was frozen sometime late last year.
The Douglas division of Boeing then continued the project and have further
refined the design, which no longer looks like this drawing. However,
this is the design version NASA is continuing with for wind tunnel testing
and construction of the remotely piloted vehicle (RPV). It is 14.2%
scale with a 35’ wing span. The flight control systems are being
designed at the Dryden facility, but the actual construction and installation
of the boxes is being done at Langley. Boeing is also providing engineering
support for the vehicle.
One of the first
things you notice on this design is the engines having been raised out
of the boundary layer flow. This was the fastest and easiest way
to solve the inlet and compressor problems trying to deal with the turbulent
flow off the rear of the wing, especially in high angle of attack flight.
Since this was going to be one of the hardest problems to solve and may
not have been economically feasible, the simpler approach was taken.
to see in the drawing, the control surfaces are spread out differently.
There are now 16 control surfaces instead of 22, but it still have the
split elevons in the last four outboard elevons, with the outermost two
being ganged together. It is all electric with no hydraulics on board,
using control actuators taken from an air-launched munitions program.
These actuators are designed to work for 90 seconds, yet the test vehicle
is being designed to last for 100 hours, so there was another problem than
needed to be worked out. It turned out that the actuators were way
over designed for the munitions application and probably would last much
longer in the less demanding environment of the BWB. The model was
also being built in such a manner that the actuators could be easily removed
if they showed signs of failure.
The next problem
they had to overcome is which digital controller would drive which control
surfaces. To maintain a minimum level of redundancy, each controller
must drive actuators at different parts of the wing so the failure of one
controller will not bring the aircraft down. A little later there
was a discussion between Al and Doug about the software NASA is using so
the controllers know which one takes the lead and how the entire system
determines when the lead controller is actually have a problem and needs
to be relieved by a secondary unit. This got a little deep on the
technology side, so not much more will be covered here.
Power will be
provided by three Williams turbojet engines from target drones that are
designed to operate for about 10 hours. The same issues came up here
as with the actuators, plus there was no low idle setting (below 90 lbs.)
on the fuel control (the Navy didn’t need low idle for in-flight launch
so it was not designed in by Williams). The spool up time was also
not acceptable since the engine would take 33 seconds from idle to the
full thrust of 240 lbs. After long discussion with Williams and a
lot of money, NASA got their supply of engines with a 35 lb. idle setting
and a much faster response time from idle to full power. Since the
test vehicle only needs about 190-200 lbs. per engine, a detent was placed
on the pilot’s throttle console to limit thrust at 200 lbs. This
will make the engines last well beyond the 10 hours. The pilot will
have the ability to remove the detent in the event of an engine failure
and more power is needed from the remaining engines.
The wing does
have slats on the leading edge, however, they do not have actuators so
are locked in a fixed position depending on the type of test to be conducted.
The original plan called for a ballistic recover chute to save the model,
but that has been removed and replaced by triple redundancy in the control
software (remember spreading out the control surfaces between controllers)
and dual redundancy on hardware. The various antennas will be integrated
into the structure to try and keep the surface as clean as possible.
There will be a spin recovery chute since plans call for spin testing of
this model, along with attempts to make it tumble.
Al mentioned that
the weights shown in the slide were dynamically scaled to the commercial,
full size version. He gave the example of scaling something down
by one half, which results in a piece of hardware that has one quarter
of the area, one eighth the volume and one eighth the mass (dynamic scaling).
The other scaling factor is moment of inertia. NASA has a very specific
set of weights, moments of inertia, area and size targets for this project.
The 2,700 lbs. represents the scaled maximum takeoff weight of the full
size aircraft. On the other hand, the 2,300 lbs. minimum (empty)
weight is well above the truly scaled 1,300 lbs. This is due to the
fact you can’t always economically scale down things like actuators and
other types of hardware, computer systems and other instrumentation.
They are not satisfied with the high weight and are trying to get it down
below 2,000 lbs., but they are not sure this will happen. The construction
is all carbon fiber and Al gave as an example of weight savings the winglet’s
vertical surfaces. These have a design weight of 7 oz. including
all hinges, horns, ribs, etc., and will have to withstand a max speed of
140 knots. Based on what they know now it doesn’t look like they
will make that weight, but at least it will be the minimum they can make
One other thing
he wanted to cover on this slide was the relationship between the 14.2%
scaling factor and the 35’ wing span. If you do the math they don’t
come out to the 289’ of the model he was talking about at the beginning
the presentation. That’s because this model is for the smaller, 450
passenger, single-deck version that Boeing is working towards. This
is still bigger than a current B-747, so many of the earlier noted problems
regarding infrastructure still apply.
The next slide shows the molds being built at the Langley facility.
The pictures were taken in June and work has now progressed to the point
of laying cloth into the molds. However, all of the structural analysis
has not be completed so they are not sure how many layers of carbon fiber
will be needed. So the engineers are in the catch-up mode trying
to get the numbers put together so the technicians can get back to building.
The 3% wind tunnel model was nearing completion and it was expected it
would go to the tunnel in late September. This model is correct configuration
for the skins that are waiting for the final numbers.
The lower right
portion of the slide shows the general layout of the test model with fuel
tanks and elevons located in and one the center section. The wings
will be separate which is mainly so the model can be shipped more easily
with the plan being to use a large motorhome that doesn’t have any interior
(just a big box on wheels). Alternate shipping plans include using
Air Force C-17 training missions and, as a last resort, paying to ship
it on a Super Guppy.
What NASA is doing
here is looking at the flight characteristics of the BWB class of vehicles.
Although the test model doesn’t reflect the latest developments in design
changes (as a result of freezing the design for the models purposes), all
the test data will be extrapolated to the latest versions using data obtained
over the life of the project.
The next slide was a listing of the goals for the ELP (Envelop Limits Program)
research, which is where all this is leading too. Things like what
happens when you stall it, spin it, tumble it and can it be recovered aerodynamically.
Can these modes be prevented aerodynamically using the controls surfaces
on the aircraft and, if you can’t what will it take in terms of the system
point Al took a few minutes to explain a little more about dynamic scaling.
If everything goes as planned, the model’s scaling will exactly duplicate
the aerodynamic forces are work in all phases of flight. So the numbers
they get from a wing tip’s helix angle as it falls off in a stall will
be exactly what the full scale aircraft would see under the same angles
of attack. There are a couple of gotcha’s here, one being mach number
since the model will be nowhere near that of the full size aircraft.
For the low speed end this shouldn’t be too much of a problem, but they
can’t do anything about the differences in Reynolds numbers between the
For those of you
who are concerned with how your government spends its money, Al noted one
of the things they are trying to do keep this program under cost control
and set an example for future projects. It is a partnership venture
between the design engineers at Langley, the flight research group at Dryden
and Boeing to do everything right the first time around.
At this point Al went into his summary. The BWB offers potential
for substantial economical and environmental benefits. The BWB is
a high-risk, high-payoff conceptual platform and it is felt NASA’s involvement
is appropriate. NASA is now committed to investigating the low-speed
stability and control attributes of this confirguration.