A LO2/kerosene SSTO rocket design, w/o AOL
burnside
A LO2/Kerosene SSTO Rocket Design (long)
Mitchell Burnside Clapp
Pioneer Rocketplane
(view with a fixed pitch font such as courier or monaco)
Abstract
The NASA Access to Space LO2/hydrogen single stage to orbit
rocket was examined, and the configuration reaccomplished
with LO2/kerosene as the propellants. Four major changes
were made in assumptions. First, the aerodynamic
configuration was changed from a wing with winglets to a
swept wing with vertical tail. The delta-V for ascent
was as a result recalculated, yielding a lower value due to
different values for drag and gravity losses. The engines
were changed to LO2/kerosene burning NK-33 engines, which
have a much lower Isp than SSME-type engines used in the
access to space study, but also have a much higher
thrust-to-weight ratio. The orbital maneuvering system on
the Access to Space Vehicle was replaced with a pump-fed
system based on the D-58 engine used for that purpose now
on Proton stage 4 and Buran. Finally, the wing of the
vehicle was allowed to be wet with fuel, which is a
reasonable practice with kerosene but more controversial
with oxygen or hydrogen. Additionally, in order to reduce
the technology development needed, the unit weights of the
tankage were allowed to increase by 17 percent.
After the design was closed and all the weights
recalculated, the empty weight of the LO2/kerosene vehicle
was 35.6% lighter than its hydrogen fuelled counterpart.
Introduction
NASA completed a study in 1993 called Access to Space, the
purpose of which was to consider what sort of vehicle
should be operated to meet civil space needs in the future.
The study had three teams to evaluate three different broad
categories of options. The Option 3 team eventually settled
on a configuration called the SSTO/R. This vehicle was a
LO2/hydrogen vertical takeoff horizontal landing rocket.
The mission of the Access to Space vehicle was to place a
25,000 pound payload in a 220 n.mi. orbit inclined at 51.6
degrees. The vehicle had a gross liftoff weight of about
2.35 million pounds. The thrust at liftoff was 2.95 million
pounds, for a takeoff thrust to weight ratio of 1.2. The
empty weight of the vehicle was 222,582 pounds, and the
propellant mass fraction (defined here as
[GLOW-empty]/GLOW) was 90.5%.
Main power for this vehicle was provided by seven SSME
derivative engines, with the nozzle expansion ratio reduced
to 50. This resulted in an Isp reduction from 454 to 447.3
seconds. Each engine weighed 6,790 lbs, for an engine sea
level thrust to weight ratio of 62.
Aerodynamically the vehicle was fairly squat, with a
fineness ratio (length:diameter) of 5. The overall length
of the vehicle was 173 feet and its diameter was 34.6 feet.
It had a single main wing (dry of all propellants) of about
4,200 square feet total area, augmented by winglets for
directional control at reentry. The landing wing loading
was about 60 lb/ft2. The oxygen tank was in the nose
section. The payload was mounted transversely between the
oxygen and hydrogen tanks, and was 15 feet in diameter and
30 feet long.
This design exercise was among the most thorough ever
conducted of a single stage to orbit LO2/LH2 VTHL rocket.
It was probably the single greatest factor in convincing
the space agency that single stage to orbit flight was
feasible and practical, to borrow from the title of Ivan
Bekey's paper of the same name.
A LO2/kerosene alternative
A number of people have been asserting for some time that
higher propellant mass fractions available from dense
propellants may make single stage to orbit possible with
those propellants also. The historical examples of the
extraordinary mass fractions of the Titan II first stage,
the Atlas, and the Saturn first stage are all persuasive.
Further, denser propellants lead to higher engine thrust to
weight ratios, for perfectly understandable hydraulic
reasons.
It has not usually been observed that higher density also
leads to significant reductions in required delta-v.
There are two major reasons that this is so. First, the
reduction in volume leads to a smaller frontal area and
lower drag losses. The second, and more significant, reason
is that the gravity losses are also reduced. This is because
the mass of the vehicle declines more rapidly from its
initial value. The gravity losses are proportional to the
mass of the vehicle at any given time, and hence the
vehicle reaches its limit acceleration speed faster.
NASA itself has implicitly recognized this effect. When the
Access to Space Option 3 team examined tripropellant
vehicles, the delta-v to orbit derived from their work was
29,127 ft/sec, for precisely the reasons described in the
previous paragraph. This compares to a delta-v of 30,146
ft/s for the hydrogen-only baseline, as reported in a
briefing by David Anderson of NASA MSFC dated 6 October
1993. To be clear, these delta-v numbers include the back
pressure losses, so that no "trajectory averaged Isp"
number is used. They did not, however, report any results
for kerosene-only configurations.
To come to a more thorough understanding of the issues
involved in SSTO design, I have used the same methodology
as the Access to Space team to develop compatible numbers
for a LO2/kerosene SSTO. There are four major changes in
basic assumption between the two approaches, which I will
identify and justify here:
1: The ascent delta-v for the LO2/kerosene vehicle is
29,100 ft/sec, rather than 29,970 ft/sec. The reason for
this is argued above, but I ran POST to verify this value,
just to be sure. The target orbit is the same: 220 n.mi.
circular at 51.6 degrees inclination. The detailed weights
I have for the NASA vehicle are based on a delta-v of
29,970 ft/sec rather than the 30,146 ft/sec reported in
Anderson's work, but I prefer to use the values more
favourable to the hydrogen case to be conservative. The
optimum value of thrust to weight ratio turns out to be
slightly less than the hydrogen vehicle: 1.15 instead of
1.20.
2: The aerodynamic configuration is that of Boeing's RASV.
Without arguing whether this is optimal, the fineness ratio
of 8.27 and large wing lead to a much more airplane-like
layout, better glide and crossrange performance, and
reduced risk. The single vertical tail is simpler and safer
than winglets as well. Extensive analysis has justified the
reentry characterisitics of this aircraft. The wing is
assumed to be wet with the kerosene fuel, as is common on
most aircraft. The fuel is also present in the wing
carry-through box. The payload is carried over the wing
box, and the oxidizer tank is over the wing. This avoids
the need for an intertank, which in the NASA Access to
Space design is nearly 6,600 pounds.
3. The main propulsion system is the NK-33. The engine has
a sea level thrust of 339,416 lbs, a weight of 2,725 lbs
with gimbal, and a vacuum Isp of 331 seconds. Furthermore,
it requires a kerosene inlet pressure of only 2 psi
absolute, which dramatically reduces the pressure required
in the wing tank. It also operates with a LO2 pressure at
the inlet of only 32 psi. The comparable values for the
SSME are about 50 psi for both propellants. This will have
a substantial effect on the pressurization system weight.
4. The OMS weight is based on the D-58 engine. This engine
is used for the Buran OMS system and the Proton stage 4. As
heavy as it is the Isp is an impressive 354 seconds. NASA's
vehicle used a pressure fed OMS, which is a sensible design
choice if you're stuck with hydrogen and you wish to
minimize the number of fluids aboard the vehicle. But
because both oxygen and kerosene are space-storable, there
is no reason to burden the design with a heavy pressure fed
system.
Using the same methodology for calculating masses, and
accepting the subsystems masses as given in the Access to
Space vehicle, a redesign with oxygen and kerosene was
accomplished. The results appear in Table 1.
Table 1: Access to Space vehicle and LO2/kerosene
alternative
Name O2/H2 LO2/RP
Wing 11,465 11,893 lb
Tail 1,577 1,636 lb
Body 64,748 33,741 lb
Fuel tank 30,668 - lb
Oxygen tank 13,273 17,271 lb
Basic Structure 14,610 10,274 lb
Secondary Structure 6,197 6,197 lb
Thermal Protection 31,098 21,238 lb
Undercarriage, aux. sys 7,548 5,097 lb
Propulsion, Main 63,634 36,426 lb
Propulsion, RCS 3,627 1,234 lb
Propulsion, OMS 2,280 823 lb
Prime Power 2,339 2,339 lb
Power conversion & dist. 5,830 5,830 lb
Control Surface Actuation 1,549 1,549 lb
Avionics 1,314 1,314 lb
Environmental Control 2,457 2,457 lb
Margin 23,116 16,105 lb
Empty Weight 222,582 141,682 lb
Payload 25,000 25,000 lb
Residual Fluids 2,264 1,911 lb
OMS and RCS 1,614 1,261 lb
Subsystems 650 650 lb
Reserves 7,215 8,895 lb
Ascent 5,699 7,587 lb
OMS 679 541 lb
RCS 837 767 lb
Inflight losses 13,254 17,445 lb
Ascent Residuals 10,984 15,175 lb
Fuel Cell Reactants 1,612 1,612 lb
Evaporator water supply 658 658 lb
Propellant, main 2,054,612 3,034,972 lb
Fuel 293,604 843,048 lb
Oxygen 1,761,008 2,191,924 lb
Propellant, RCS 2,814 2,556 lb
Orbital 2,051 1,756 lb
Entry 763 800 lb
Propellant, OMS 19,357 15,452 lb
GLOW 2,347,098 3,246,156 lb
Inserted Weight 292,486 211,185 lb
Pre-OMS weight 271,482 186,152 lb
Pre-entry Weight 252,125 170,700 lb
Landed Weight 251,362 169,900 lb
Empty weight 222,582 141,682 lb
Sea Level Thrust 2,816,518 3,733,080 lb
Percent margin 11.6% 12.8%
Assumed Isp(vac) 447.3 331.0 s
Ascent Delta-V 29,970 29,100 ft/s
OMS delta-V 1,065 987 ft/s
RCS delta-V 108 107 ft/s
Deorbit Delta-V 44 53 ft/s
Reserves 0.28% 0.25% lb/lb
Residuals 0.53% 0.50% lb/lb
Wing Parameter 4.56% 7.00% lb/lb
TPS parameter 12.37% 12.50% lb/lb
Undercarriage parameter 3.00% 3.00% lb/lb
Wing Reference Area 4,189 5,528 ft2
Density of fuel 4.4 50.5 lb/ft3
Density of oxygen 71.2 71.2 lb/ft3
Volume of fuel 66,276 16,694 ft3
Volume of oxygen 24,733 30,785 ft3
Fuel tank parameter 0.42 - lb/ft3
Oxygen tank parameter 0.48 0.56 lb/ft3
Some discussion of the results and justification is in
order.
The wing is about 40 percent heavier as a percentage of
landed weight than for the hydrogen fueled baseline. When
considered as a tank, it is about 60 percent heavier for
the volume of fuel it encloses. Its weight per exposed area
is about the same and the wing loading is half at landing.
No benefit is taken explicitly for the lack of a
requirement for kerosene tank cryogenic insulation.
The tail is assumed to have the same proportion of wing
weight for both cases. This is conservative for the
kerosene wehicle because its single vertical tail is
structurally more efficient.
The body of the kerosene vehicle has three components. The
oxidizer tank has an increased unit weight of about 17
percent. This is done in order to avoid the need for
aluminum-lithium, which was assumed in the Access to Space
vehicle. The basic structure group is unchanged, except
that the intertank is deleted and the thrust structure is
increased in proportion to the change in thrust level.
The secondary structure group is mostly payload support
related, and was not changed.
The thermal protection group is in both cases about 12.5%
of the entry weight. This works out to 1.107 lbs/ft2 of
wetted area for the kerosene vehicle, which is common to
many SSTO designs.
The undercarriage group is 3% of landed weight for both
vehicles. There is no benefit taken for reductions in gear
loads for the kerosene vehicle due to lower landing speed
and lower glide angle at landing.
The main propulsion group includes engines, base mounted
heat shield, and pressurization/feed weights. The engines
are far lighter for their thrust than SSME derivatives. The
pressurization weights are reduced in proportion to the
pressurized volume for the kerosene vehicle. No benefit is
taken for reduced tank pressure.
Here is as good a place as any to point out the erroneous
assertion that increased hydrostatic pressure is going to
lead to increased tankage weights. There is no requirement
for a particular ullage pressure except for the need to
keep the propellants liquid. It is the pressure at the base
of the fluid column rather than the top of the column that
is of engineering interest. The column of fluid exerts a
hydrostatic load on the base of the tank, but this load
does not typically exceed the much more adverse requirement
for engine inlet pressurization. For the kerosene vehicle,
the hydrostatic load at the base of the oxygen tank is 49
psi, which is compatible with the pressures normally seen
in oxygen tanks for rocket use. The load declines after
launch because the weight goes down faster than the
acceleration goes up.
The bottom line here is that dense propellants may require
you to alter a tank's pressurization schedule, but not to
overdesign the entire tank. Structures are sized by loads
and tankage for rockets is sized principally by volume, and if
the vehicle is small, by minimum gauge considerations.
This is not completely true for wet wings, however, as
discussed previously. In this particular example, there is
no need for high pressure in the wing tank either, because
of the low inlet pressure required by the NK-33.
The OMS group is the only other major change, as discussed
above. The reliable D-58 engine has been performing space
starts for decades and will serve well here. The
acceleration available from the OMS is about 0.12 g, which
is standard.
All the other weights are pushed straight across for the
most part. A brief inspection suggests that this is very
conservative. Control surface actuation requirements are
certainly less, electrical power requirements less, much
better fuel cells available than the phosporic acid type
assumed here, and reduced need for environmental control.
Nonetheless, rather than dispute any of these values it is
easier simply to accept them.
The margin is applied to all weight items at 15% execpt for
the engine group at 7.5%. The justification for this is that
the main and OMS engine weights are known to high accuracy.
The vehicle has an overall length of 1955 inches, and a
diameter of 236.4 inches. The wing has a leading edge sweep
of 55.5 degrees and a trailing edge sweep of -4.5 degrees.
Its reference area is 5,632 square feet, of which 3,992
square feet is exposed. The wing encloses 16,694 ft3 of
fuel, with a further 5% ullage. The carry-through is also
wet with fuel. The wing span is 1293 inches, and the taper
ratio is 0.13.
The payload bay has a maximum width and height of 15 feet.
It sits on top of the wing carry through box. The thrust
structure from the engines passes through and around the
payload bay to the forward LO2 tank. The payload bay is 30
feet in length. It has a pair of doors, the aft edge of
which is just forward of the vertical tail leading edge.
The engine section encloses 11 NK-33 engines, with a 4 - 3
- 4 layout. The engines are each 12.5 feet long, and
additional structure and subsystems take up another 6.5
feet.
The oxygen tank comprises the forward fuselage, which
encloses 30,785 ft3 of oxygen, with a further 5% ullage.
The length of the tank is about 100 feet. The ventral
surface of the tank is moderately flattened as it moves
aft, to fair smoothly with the wing lower surface. This
flattening reduces its length by about 5% with respect to a
strictly cylindrical layout. The aft edge of the oxygen tank
is about even with the forward payload bay bulkhead. A
compartment of about 13.9 feet provides room for some
subsystems and a potential cockpit in future versions.
Conclusion
The methods of the NASA Access to Space study were used to
design a single stage to orbit vehicle using existing
LO2/kerosene engines. An inspection of the final results
shows that the vehicle weighs about 36.5% less than its
hydrogen counterpart, with reductions in required
technology level and off the shelf engines. The center of
mass of the vehicle is about 61% of body length rather than
68% for the Access to Space vehicle, which should improve
control during reentry. The landing safety is considerably
improved by lower landing speed and better glide ratio.
Structural margins are greater overall. The vehicle
designed here appears to be superior in every respect:
smaller, lighter, lower required technology, improved
safety, and almost certainly lower development and
operations cost.
David L. Burkhead
burn...@bix.com (burnside) wrote:
:>A LO2/Kerosene SSTO Rocket Design (long)
[ 8< Lots of good stuff >8 ]
:>Conclusion
:>The methods of the NASA Access to Space study were used to
:>design a single stage to orbit vehicle using existing
:>LO2/kerosene engines. An inspection of the final results
:>shows that the vehicle weighs about 36.5% less than its
:>hydrogen counterpart, with reductions in required
:>technology level and off the shelf engines. The center of
:>mass of the vehicle is about 61% of body length rather than
:>68% for the Access to Space vehicle, which should improve
:>control during reentry. The landing safety is considerably
:>improved by lower landing speed and better glide ratio.
:>Structural margins are greater overall. The vehicle
:>designed here appears to be superior in every respect:
:>smaller, lighter, lower required technology, improved
:>safety, and almost certainly lower development and
:>operations cost.
Why am I not surprised?
David L. Burkhead "If I had eight hours to cut down
dav...@dax.cc.uakron.edu a tree, I'd spend seven sharpening
FAX: 330-253-4490 my axe." Attributed to Abraham
SpaceCub Lincoln
http://GoZips.uakron.edu/~david8
JHOLL4
>A LO2/Kerosene SSTO Rocket Design (long)
>Mitchell Burnside Clapp
>Pioneer Rocketplane
Nice job, Mitchell. I've only had time to skim through it
briefly, but I have two questions.
First, why are you eating the weight of a separate OMS engine
instead of reusing the NK-33's for this? I'm guessing that the
NK-33 is not deep-throttable, but I don't have any detailed
engineering information on this engine.
Second, with 11 NK-33's providing takeoff thrust, have you
looked at engine-out scenarios at all? At first glance it would
seem likely that the craft can make orbit after a single complete
engine failure, but I wondered if you have done any detailed analysis
of this.
I'll probably have a little more input after I have time to
digest the details of the paper.
--Cathy Mancus <man...@vnet.ibm.com>
Ken Myrtle
burnside wrote:
snip
> It has not usually been observed that higher density also
> leads to significant reductions in required delta-v.
> There are two major reasons that this is so. First, the
> reduction in volume leads to a smaller frontal area and
> lower drag losses. The second, and more significant, reason
> is that the gravity losses are also reduced. This is because
> the mass of the vehicle declines more rapidly from its
> initial value. The gravity losses are proportional to the
> mass of the vehicle at any given time, and hence the
> vehicle reaches its limit acceleration speed faster.
>
> NASA itself has implicitly recognized this effect. When the
> Access to Space Option 3 team examined tripropellant
> vehicles, the delta-v to orbit derived from their work was
> 29,127 ft/sec, for precisely the reasons described in the
> previous paragraph. This compares to a delta-v of 30,146
> ft/s for the hydrogen-only baseline, as reported in a
> briefing by David Anderson of NASA MSFC dated 6 October
> 1993. To be clear, these delta-v numbers include the back
> pressure losses, so that no "trajectory averaged Isp"
> number is used. They did not, however, report any results
> for kerosene-only configurations.
>
snip
Nice paper
Could you tell us what the limiting acceleration was for each
of the designs? I assume that the higher the limiting acceleration
the better the LO2/kerosene rocket will look. Do you have any numbers
on how delta-v depends on the limiting acceleration?
Ken Myrtle myr...@sfu.ca
MbClapp
The minimum throttling level I have heard claimed for the NK-33 is 25%,
which strains even my credulity. Using the NK-33 for OMS insertion would
involve first of all an untested restart of the engine and secondly a 0.65
g impulse, which is a little hard to control accurately. The D-58 engine
is really better suited to this purpose.
With 11 engines (perhaps not the optimal number, but that's how it worked
out in the design), you can lose one engine at takeoff and still make a
stable orbit, though not necessarity the one you had in mind. After about
thirty seconds you can lose a second engine.
The engines were not throttled during ascent. They were just shut down one
by one. With a large engine count the difference is only about 50 ft/sec
of delta-V.
The limit acceleration was set for both vehicles to 3 g. Personally I
don't see a problem with 5, but having seen many NASA studies baseline 3 g
I wanted to compare apples to apples rather than apples to zebras or
Buicks.
The more I think about this rocket the more fun I think it would be to
build. Unless I am really wrong about the weights, which of course is
possible, this would be a straighforward effort. You could even fly from
existing shuttle pads like the idle one at Vandenburg AFB SLC 6.
What would a flight cost?
TPS maintenance should run something like 25$/ft^2/flight, for a total of
$500,000
Each engine should take about $450,000 from flight to flight, including
overhauls and depreciation, so engine maintenance should run around $4.95M
Propellants are cheap. $77/ton for LO2 and $1/gallon for kerosene gives
about $250,000/flight.
Maintaining the airframe itself and all its pieces internally runs
possibly another $1M/flight. It's pretty simple.
The subtotal so far is $6.7 million. I've probably forgotten some stuff,
so let's call it $10M even, and throw on a 100% profit margin for a total
of $20M per flight.
That means charging about $800/pound and making money. Of course those
pounds are ISS orbit pounds. Due east to a low orbit I haven't calculated
the payload, but I'd guess it's on the order of 32,000 lbs. So the price
is more like $625/pound.
But there is a lot of room for improvement in the NK-33 flight and
reusability characteristics. There are also things we could do to reduce
the costs of TPS maintenance and inspection, like deleting the stuff
altogether on much of the upper surface by using higher temperature resins
for the carbon fibre composite, and automating the inspection procedure.
If we did as well as the SR-71 in terms of ops costs as a multiple of
propellant costs (i.e. 10X), we would be talking about a per-flight cost
of $2.5M. If we did as well as an airliner, a flight would be $0.75M. A
fighter squadron is in between somewhere in terms of the operations cost
multiplier.
So taking the SR-71 case, a flight to ISS' orbit could have a direct cost
of about $100/pound. That's pretty good, in anyone's book (although it
still means that my ticket would cost 18 grand. I really need to start
that diet!).
Mitchell Burnsid Clapp
Pioneer Rocketplane
Phil Fraering
Mr. Burnside Clapp,
(hmm, could I call you Mitchell? It would end the endless worrying
I have about whether I'm supposed to be using a hyphen or not)...
Anyway, with regards to the current vehicle, isn't there a company that
already has exclusive rights to the NK-33?
How does the engine used on the Zenit stack up as a replacement?
Phil Fraering The above is the opinion of neither my internet
p...@acadian.net service provider nor my employer.
wrk: 318/2699112
home:318/2619649 I'm currently having net problems; if you want
to make sure I'll see it, or expect a response,
email it.
Jan Vorbrueggen
mbc...@aol.com (MbClapp) writes:
> The minimum throttling level I have heard claimed for the NK-33 is 25%,
> which strains even my credulity.
Hmmm. Didn't Stennis run some deep-throttling tests on SSMEs last year,
successfully - down to 25%, IIRC? If even the SSME can do it...or is that
easier for a high-pressure engine such as the SSME than for low pressure ones?
Jan
Rich Kolker
Phil Fraering wrote:
>
> Mr. Burnside Clapp,
>
> (hmm, could I call you Mitchell? It would end the endless worrying
> I have about whether I'm supposed to be using a hyphen or not)...
>
No hyphen, just like on the T-shirts.
++rich
Cathy Mancus
In <19970205154...@ladder01.news.aol.com>, mbc...@aol.com (MbClapp) writes:
>The minimum throttling level I have heard claimed for the NK-33 is 25%,
>involve first of all an untested restart of the engine and secondly a 0.65
>g impulse, which is a little hard to control accurately. The D-58 engine
>is really better suited to this purpose.
OK, I'll buy that. I didn't realize the NK-33 was not rated for
in-flight restart. I really should track down more technical
info on the available engines...
>The engines were not throttled during ascent. They were just shut down one
>by one. With a large engine count the difference is only about 50 ft/sec
>of delta-V.
>Each engine should take about $450,000 from flight to flight, including
>overhauls and depreciation, so engine maintenance should run around $4.95M
Are you assuming the NK-33's run at 100% power at launch? If so,
it's an interesting question how much the lifetime might be extended
(read: lower maintenance costs per flight) might be extended by
tacking on an extra engine and running them all at about 92% power.
>The subtotal so far is $6.7 million. I've probably forgotten some stuff,
>so let's call it $10M even, and throw on a 100% profit margin for a total
>of $20M per flight.
>That means charging about $800/pound and making money.
I don't see any design cost amortization in there. I would expect
it to at least be comparable to ops cost, if not much greater. So
you aren't *really* making money. Did you run a costing model
to see what this configuration was likely to cost to design,
build, and validate? I'm curious.
--Cathy Mancus <man...@vnet.ibm.com>