Aerospike engines (long)
Jeff Greason
Ben Muniz (586-3578) ; /home/auspex_d0/usr2/yqq7121 <yqq...@rs0200.rdyne.rockwell.com> wrote:
>In article <3k03l8$a...@agate.berkeley.edu>,
>Afam <af...@uclink.berkeley.edu> wrote:
>>
>>What is an aerospike engine? how does it work and how is it different
>>from other engine designs?
>
>Where can I get more information?
>
>There are many references to aerospikes in the open technical literature:
>one that I've seen noted but haven't had a chance to get is: Ballard, R. O.,
>"The Aerospike/Aeroplug Engine: A Technology Development Summary", Sverdrup
>Technology Inc., MSFC Group, Contract NAS8-37814, 1991.
>
>
>Follow-up to sci.space.tech.
Since the subject has come up again, I thought I'd re-post the
aerospike bibliography I put together after I asked a similar question
about a year ago...
This is a summary of the responses I got about technical references
on aerospikes. Thanks to all who responded.
I was looking for fairly introductory aerospike references, but
somewhat beyond the "popular science" level. I didn't get that, but
I did get an excellent set of references at an advanced level. Thanks
to the two individuals working in the field who sent me a very detailed
bibliography (summarized below);
David Garza (University of Texas)
dga...@sparc2000.utsi.edu
Michael Fick (Technical University of Munich)
mf...@asterix.lrt.mw.tu-muenchen.de
A common thread in all the responses was agreement that aerospikes are
not a readily acessible subject in the modern literature; they all had
some horror stories to share about the difficulty in tracking down
information. Much of it is in government and company reports which
are difficult to obtain.
Introduction:
-------------
"Aerospikes" and "Plug Nozzles" (the terms are often used interchangeably),
confine the expansion of exhaust gases to a narrow region around the
perimeter of the rocket nozzle. It is this "expansion" region in which
the thermal energy of the rocket combustion products is converted to
kinetic energy, producing thrust.
The primary interest in aerospikes is that, due to the details of the
supersonic expansion profile of the jet, they can be "altitude compensating",
optimizing their exhaust profile for different ambient pressures, improving
Isp on a through-atmosphere mission.
One note: according to "Rocket Propulsion Elements" (cited below), these
terms are actually not interchangeable. A "plug nozzle" confines the
exhaust by placing a physical plug in the nozzle, while a true "aerospike"
engine uses a jet of cooler gases in the center of the nozzle to achieve
the same effect. But I'm not at all sure this distinction is clear in
the literature.
Rocketdyne/Pratt & Whitney/GE are claimed to have done some tests on aerospike
engines, but few detailed performance references could be obtained. In
particular, one respondent said that Rocketdyne was ready to do a flight
engine for SSME except that the contract RFP specified conventional nozzles
so Rocketdyne wouldn't have an unfair advantage. Rocketdyne's data is
proprietary.
References:
-----------
Introductory:
-------------
Sutton, George P., "Rocket Propulsion Elements", 6th edition
(my ISBN was out of date),
"In the section on nozzles there's a few paragraphs introducing the
concept of plug nozzles/aerospikes. Just enough to introduce the
idea; but a nice place to start since you probably own this book
anyway (or should)"
General:
--------
Graham, A.R., "The Plug Nozzle Handbook"; General Electric Co., August,
1968, Contract #:NAS9-3748
"... has a restriction "Government Agencies Only" on it ... Early,
'50's-'60's, research was done mostly by GE, and this is supposed to be
the only textbook/handbook on plug nozzles with lots of design info, but
it won't have anything on the Rocketdyne work."
Ballard, R.O., "The Aerospike/Aeroplug Engine"; Sverdrup Technology, 1991,
Contract #:NAS8-37814
"the most up to date summary of info ..."
Sutor, A.T., "Rocket Engine Nozzle Compendium"; Rocketdyne Div., 1985,
Contract #:F33657-82-C-0346
"... compiled for the Air Force, and may also have a restriction
on distribution ... summarizes Rocketdyne's experience with aerospikes,
and includes a full page plot of slipstream effects."
Wasko, Robert A., "Performance of Annular Plug and Expansion-Deflection
Nozzles Including External Flow Effects at Transonic Mach Numbers";
Lewis Research Center, April, 1968, NASA TN D-4462
"... a 32 page report on cold-flow test of full-length and truncated
plug, as well as an expansion-deflection nozzle. Testing was done in
still air, as well as Mach numbers from about .5 to 2.0, and the truncated
plugs and E-D were tested both with and without base flow"
O'Brien, C.J., "Unconventional Nozzle Tradeoff Study"; Aerojet Liquid Rocket
Company, July, 1979, NTIS: N79-28224, Contract #: NAS3-20109
"... a long report on a proposal to use an aerospike for a Space Tug, and
it includes a fairly long section on plug nozzle background. 300+ pages"
Huang, D.H., "Aerospike Engine Technology Demonstration for Space Propulsion";
Rocketdyne Div., 1974, Contract #: F04(611)-67-C-0116
"describes the development of a 25,000 lb flight-weight engine at
Rocketdyne, but doesn't have much on engine performance once it was
built."
Sergeant, R.J., "An Experimental Hot Rocket Model Investigation of a Plug
Cluster Nozzle Propulsion System, Part I: Base Thermal and Pressure
Environment For a Module Chamber Pressure of 300 psia and Simulated
Altitudes to 150,000 Feet"; Cornell Aeronautical Lab,
CAL Ho. HM-2045-Y-5(I), September 1967
On exact design procedures for the plug:
----------------------------------------
G.V.R. Rao, "Exhaust Nozzle Contour for Optimum Thrust"; Jet Propulsion,
June 1958, p 377-382
G.V.R. Rao, "Spike Nozzle Contour for Optimum Thrust"; (Rocketdyne),
Ballistic Missile and Space Technology, Vol.2, Pergamon Press, New York,
1961
Lee, C.C., "FORTRAN Programs for Plug Nozzle Design"; TN R-41; NASA CR-51300,
March 1963
Lee, C.C., "Computation of Plug Nozzle Contours by the Rao Optimum Thrust
Method"; Brown Engineering TN R61; NASA CR-51301, July 1963
On integration issues with specific vehicles (an issue with aerospikes):
------------------------------------------------------------------------
Fanciullo, T.J, D.C. Judd, C.J. O'Brien, "Operationally Effective
Nozzles for SSTO Vehicles"; Aerojet Propulsion Div., GenCorp Aerojet,
1992, Contract #: SDIO84-90-C-0030,
<Deals with Delta Clipper>
Immich, H., R.C. Parsley, "Plug Engine Systems for Future Launch Vehicle
Applications"; MBB Deutsche Aerospace and UTC Pratt & Whitney, June 1993,
AIAA 93-2560, AIAA/SAE/ASME/ASEE 29th Joint Propulsion Conference and
Exhibit, June 28-30, 1993, Monterey, CA
<Deals with General Dynamics proposal competing with Delta Clipper>
Gielda, Thomas P., T.M. Walter, Ramesh K. Agarwal, "Single Stage Rocket
Performance Prediction and Test"; McDonnel Douglas, March, 1992,
AIAA 92-1386, Contract #: SDIO84-91-C-0029, AIAA Space Programs and
Technologies Conference, Huntsville, AL, March 24-27. 1992
<Deals with Delta Clipper>
Heald, Dan A., "Plug Nozzle Propulsion System"; General Dynamics, February,
19922, NTIS: N93-10013
<Deals with General Dynamics proposal competing with Delta Clipper>
Fanciullo, T.J, D.C. Judd, C.J. O'Brien, "Critical Engine System Design
Characteristics for SSTO Vehicles," Aerojet, February, 1992,
NTIS: N93-10014, Contract #: SDIO84-90-C-0030
<Deals with Delta Clipper>
"Both this and the above are in 1992 JANNAF Propulsion Meeting, Vol 1,
NTIS: N93-10001"
Heald, Dan A., Kessler, Thomas L., "Single Stage to Orbit Vertical Takeoff
and Landing Concept Challenges," General Dynamics, Oct, 1991, IAF 91-205,
42nd IAF International Astronautical Congress, Montreal, Oct 5-11, 1991
Disclaimer: While I am an Intel employee, all opinions expressed are my own,
and do not reflect the position of Intel, NETCOM, or Zippy the Pinhead.
============================================================================
Jeff Greason "We choose to go to the Moon in this decade,
<gre...@ptdcs2.intel.com> and do the other things, not because they
<gre...@ix.netcom.com> are easy, but because they are hard." -- JFK
Kevin W. Ryan
engines: it's been a while, but I think I can shed _some_ light on this.
Feel free to correct me if I'm totally bogus on this.
Normal bell nozzles are built to a length where the expanding gasses
are at ambient pressure at the end of the bell. If the bell is too long,
the gasses tend to separate from the bell (not good - lots of turbulence,
etc.), while if it is too short you lose energy. Different stages of
current boosters have different (relatively) length bells as a result,
with upper stage engines being longer. Some of the SSTO orbit designs have
adjustable length bell engines to overcome this difficulty.
Plug or aerospike engines have a tapered cone in the center, combustion
occurring in a ring around it. The point stretches out quite a distance,
shaped much like the space between two bell nozzles side by side. The
outside containment is ambient pressure and gas momentum. This design
adjusts directly to ambient pressure, while still providing good
performance. In addition, the large diameter engine is a good match to the
back end of a reasonably sized booster.
Pure plug engines have a _long_ cone for the greatest efficiency, which
tends to be unwieldy and heavy. Aerospike engines have a short plug, with
a flat end, and use gas injection at the end to effectively fill in the
rest of the cone.
Linear spike engines have a stretched wedge with combustion occurring
on either side - a decent fit to the back edge of a lifting body design.
Hope this is helpful!
--
kwr
Internet: kr...@access.digex.com
Henry Spencer
> Normal bell nozzles are built to a length where the expanding gasses
>are at ambient pressure at the end of the bell. If the bell is too long,
>the gasses tend to separate from the bell (not good - lots of turbulence,
Actually, most bell nozzles for first stages are designed with some
overexpansion at sea level, because flow separation is a problem only
when overexpansion is severe.
The problem with flow separation is not so much turbulence and such, but
the fact that the separation is seldom *symmetrical*. The result is that
the jet emerges along one side of the nozzle or the other, rather than
on the centerline, and unpredictable large side forces result. Apart from
the unwanted thrust vectoring, the stresses may be too much for engine
attachments or gimbal actuators.
For the J-2 engine, which was badly overexpanded at sea level because it
was an upper-stage engine, it was a practical necessity to be able to test
at sea level. This was done in one of two ways. Most tests could be done
adequately by just locking the nozzle in place with a big external brace,
so the gimbal actuators wouldn't have to take the side loads. In some
cases it was necessary to insert a water-cooled separation collar into
the nozzle, to force symmetrical separation.
There's been a tendency in recent engines to design the "core engine" to
be (at most) slightly overexpanded at sea level, with a separate nozzle
extension that accomplishes the rest of the expansion. That way you can
test the core in the open air without worrying about flow separation.
People have played with the idea of trying to deliberately trip
symmetrical flow separation in some way that could be used in flight.
(The J-2 nozzle insert was strictly for ground testing -- trying to eject
it with the engine running would have been, uh, exciting.) Nobody's
pursued it seriously, though.
--
There is a difference between | Henry Spencer
cynicism and skepticism. | he...@zoo.toronto.edu
Kevin W. Ryan
Spencer) wrote:
>In article <kryan-25039...@dcc00195.slip.digex.net>
kr...@access.digex.com (Kevin W. Ryan) writes:
>> Normal bell nozzles are built to a length where the expanding gasses
>>are at ambient pressure at the end of the bell. If the bell is too long,
>>the gasses tend to separate from the bell (not good - lots of turbulence,
>>etc.)...
>
>Actually, most bell nozzles for first stages are designed with some
>overexpansion at sea level, because flow separation is a problem only
>when overexpansion is severe.
>
>The problem with flow separation is not so much turbulence and such, but
>the fact that the separation is seldom *symmetrical*. The result is that
>the jet emerges along one side of the nozzle or the other, rather than
>on the centerline, and unpredictable large side forces result. Apart from
>the unwanted thrust vectoring, the stresses may be too much for engine
>attachments or gimbal actuators.
> ...
>People have played with the idea of trying to deliberately trip
>symmetrical flow separation in some way that could be used in flight.
>(The J-2 nozzle insert was strictly for ground testing -- trying to eject
>it with the engine running would have been, uh, exciting.) Nobody's
>pursued it seriously, though.
I seem to remember that the Sprint nuclear interceptor missile (part of
the Safeguard ballistic missle defense design, never deployed) used freon
injection into the bell of its solid fuel motor to vector the exhaust and
steer the missile. It required _serious_ vectoring with minimal
mechanicals, as it was a short range 100G bird. Was this design (if anyone
knows) overexpanded to make this vectoring easier?
Just curiousity... I always thought that it was an interesting design.
--
kwr
Internet: kr...@access.digex.com
Jim Glass ; JF ; GLASS ; x586-0375 ; (W) ; 634-000
|> In regards to the design of and difference between plug and aerospike
|> engines: it's been a while, but I think I can shed _some_ light on this.
|> Feel free to correct me if I'm totally bogus on this.
|>
|> Normal bell nozzles are built to a length where the expanding gasses
|> are at ambient pressure at the end of the bell. If the bell is too long,
|> the gasses tend to separate from the bell (not good - lots of turbulence,
|> etc.), while if it is too short you lose energy. Different stages of
|> current boosters have different (relatively) length bells as a result,
|> with upper stage engines being longer. Some of the SSTO orbit designs have
|> adjustable length bell engines to overcome this difficulty.
|>
So I guess engines made for operation in vacuum (space engines) must be infinitely
long, eh?
Actually, NO engines with bell nozzles are built to a length where the exhaust
products are at atmospheric pressure.
Typically, bell nozzles are set at a "percent length" (percent of a 15-degree
half-angle cone) of around 80%. The area ratio is set by an optimization
procedure. Even booster engine nozzles are typically overexpanded (exhaust
at less than atmospheric). Separation does not occur until the exhaust pressure
is well below atmospheric. For example, the SSME, at an area ratio of 77,
has an exhaust pressure of around 2-3 psia.
<snip>
The shapes of US bells are calculated by using M-O-C type aero codes. Strangely,
Soviet (Russian) bells are subtly different in shape, with a more conical figure.
Why I have no idea. Since method-of-characteristics solutions OUGHT to give
the same results for anybody, a visitor from Mars would expect them to all
look more or less the same.
Maybe they know something we don't. In fact, I'm sure of it.
Jim Glass
*********************************************************************************
* gl...@rdyne.rockwell.com * "The Earth is the cradle of the Mind, *
* Jim Glass * But you cannnot stay in the cradle forever." *
* Member of Technical Staff * - Konstantin E. Tsiolkovsky *
* Systems Architecture *************************************************
* Rocketdyne Division * *
* Rockwell International * The opinions expressed here *
* Mail Stop IB-59 * are my own and do not reflect *
* 6633 Canoga Avenue * those of my employer! *
* P.O. Box 7922 * *
* Canoga Park, CA 91309-7922 * (But then, you knew that!) *
* (818) 586-0375 * *
*****************************************************************
Ben Muniz (586-3578) ; /home/auspex_d0/usr2/yqq7121
Kevin W. Ryan <kr...@access.digex.com> wrote:
> In regards to the design of and difference between plug and aerospike
>engines: it's been a while, but I think I can shed _some_ light on this.
>Feel free to correct me if I'm totally bogus on this.
You are correct. It seems that I forgot to include sci.space.tech when I
posted the following a while back on sci.space.policy. For those of
you who've read it already, please excude my duplicate post.
--------------- cut here ---------------
A while back (pre sci.space split) I posted in regard to the following:
In article <1992Dec3.1...@ke4zv.uucp> ga...@ke4zv.UUCP (Gary
Coffman) writes:
>tested. Later it intends to use aerospike engine designs that have
>*never* been tested, even on the ground. It will be difficult for
On Date: Fri, 1 Jan 93 17:02:56 PST, Brian Stuart Thorn <BrianT@cup.
portal.com> writes:
> Much has been said recently about aerospike engines, specifically
> in regard to potential use on the DC-1. I've never heard of them
> before, can someone give me (and anyone else in the dark) a brief
> description of what an Aerospike Engine is?
I've consolidated parts of my replys and expanded others. Maybe someone
would like to edit this for the FAQ?:
The following information comes from an article "Nozzle Design" by R. A.
O'Leary and James E. Bech in the Spring 1992 (No. 8) issue of Rocketdyne's
*Threshold* magazine (call (818) 586-2380/2771 or write to Rockwell Aerospace/
Rocketdyne Division, 6633 Canoga Avenue, Mail Code AB57, Canoga Park, CA, USA,
91304-7922 to get a copy).
What is an aerospike?
Briefly, an spike (or "plug") engine uses an exhaust nozzle that can be thought
of as a conventional bell shaped nozzle turned inside-out. The aerospike
nozzle is a truncated version of an ideal spike. Here is a cross-sectional
view:
Bell Ideal Spike Aerospike
xxxxx x------------x x------------x
| | \ / \ /
\ / \ / \ /
/ \ \ / \______/
/ \ \ /
/ \ \ /
\/
(x indicates the combustion location).
For a better diagram, see <Rocket Propulsion Elements> by George P. Sutton,
6th edition, J. Wiley & Sons, 1992, p. 70.
In a bell nozzle combustion gases flow through a constriction (throat) and
then the expansion away from the centerline is contained by the diverging
walls of the nozzle up to the exit plane. Bells nozzles are a point design
with optimum performance at one specific ambient pressure (i.e., altitude).
Careful design is needed to achieve desired high altitude performance while
avoiding flow separation at the walls of the nozzle near the exit when
operating at low altitudes (launch), which can lead to loss of performance and
possible structural failure of the nozzle due to dynamic loads [flow
separation is responsible for the large nozzle motion on the SSMEs during
startup transient - watch closely during next launch]. Therefore a compromise
altitude must be used for the design point of a bell nozzle.
In an spike nozzle the opposite takes place - the gas flow is directed
radially inward from an annulus at some diameter away from the centerline.
This flow is directly exposed to ambient pressure and its expansion is thus
directly coupled to the external environment (continuous altitude
compensation with no moving parts). Thus, a very high area ratio nozzle (high
vacuum performance) can also operate efficiently and safely at sea-level.
Truncating the ideal spike to save weight results in a wake at the base which
has some performance loss. This can be offset by pumping secondary flow (about
1% of primary flow) into the base region to elongate the wake which then forms
an aerodynamic countour similar to the truncated structure (hence the name
"aerospike").
Have aerospike nozzles been been tested?
The article "Nozzle Design" states "During the 1960's, Rocketdyne tested
numerous aerospike engines, ranging in size from subscale, cold-flow models
to this 250,000-pound-thrust oxygen/hydrogen shown at a test stand in Nevada
(picture of engine firing). The low altitude performance advantage of
the aerospike over conventional bell nozzle is clearly seen". The curve
presented looked like this:
1.0 .----.----.----.----.----.----.----.----.----.----.
| + T +T T + +/ +/ +/ +/ T +/
| T + /
: + /
| /
0.9 | /
: / Area Ratio = 75:1
Nozzle | / nozzle length = 25% equivalent
Efficiency | / conical
: /
0.8 | / / Bell Nozzle
| / + Aerospike (predicted)
: / T Aerospike (test data)
| /
|
0.7 :----.--/-.----.----.----.----.----.----.----.----.
| | | | | | | |
10 100 200 400 800 2000 4000 10000
Pressure Ratio: Pc/Pa
Various propellants and both conical (1-D) and axial (2-D) models were
been tested. I have heard from several sources that Rocketdyne's
*original* proposal for the Space Shuttle Main Engines used an aerospike
design based on these tests. At the California Space Development Council's
"Making Spaceflight Affordable" conference held in San Diego in February 1992,
Vern Larson from Rocketdyne gave a presentation on the aerospike test program.
Also, I've heard reports that the Germans experimented with them during WWII,
but I have not seen documentation to confirm this.
Why aren't they used?
At that same conference, I asked Max Hunter ("father" of the Delta rocket
and a major player in the SSTO field) why it seemed that an aerospike was
not baselined for the DC-X or the proposed DC-Y and DC-1. He replied
that there was concern regarding the lack of *flight-test data* (he
acknowledged that there was plenty of ground test data), in particular
for the transonic regime. However, the Rocketdyne article states ". . .
from Mach 1 to about Mach 3, *wind tunnel tests* (emphasis mine) indicate
a drop in nozzle efficiency due to the slipstream turning into the nozzle
region . . . Nevertheless, the interval of time that is spent in this
adverse flight regime is short for typical flight trajectories, and
overall performance of the aerospike nozzle remains well above that of a
conventional bell-type nozzle". Note that wind tunnel test results combined
with CFD simulations are usually sufficient for preliminary design of
experimental aircraft flight vehicles.
Where can I get more information?
There are many references to aerospikes in the open technical literature:
one that I've seen noted but haven't had a chance to get is: Ballard, R. O.,
"The Aerospike/Aeroplug Engine: A Technology Development Summary", Sverdrup
Technology Inc., MSFC Group, Contract NAS8-37814, 1991.
Ben Muniz (ISU '94) munizb%rwtms2...@beach.rockwell.com w(818)586-3578
International Space Station Alpha: Structural Loads and Dynamics
Views/Commitments expressed do not represent Rockwell Aerospace/Rocketdyne
Jeff Greason
Jim Glass ; (W) ; 634-000 <yqg...@sunshine.rockwell.com> wrote:
>The shapes of US bells are calculated by using M-O-C type aero codes.
>Strangely, Soviet (Russian) bells are subtly different in shape, with a more
>conical figure. Why I have no idea. Since method-of-characteristics
>solutions OUGHT to give the same results for anybody, a visitor from Mars
>would expect them to all look more or less the same.
>
>Maybe they know something we don't. In fact, I'm sure of it.
>
>Jim Glass
Maybe I'm missing something obvious here; since a purely conical
nozzle represents a well-known simplification of the bell to
save weight and fabrication complexity at the expense of some
performance, why is it surprising if someone else chooses an
intermediate point between bells and cones?
Jim Glass ; JF ; GLASS ; x586-0375 ; (W) ; 634-000
|> In article <1995Mar31.1...@nb.rockwell.com>,
|> Jim Glass ; (W) ; 634-000 <yqg...@sunshine.rockwell.com> wrote:
|> >The shapes of US bells are calculated by using M-O-C type aero codes.
|> >Strangely, Soviet (Russian) bells are subtly different in shape, with a more
|> >conical figure. Why I have no idea. Since method-of-characteristics
|> >solutions OUGHT to give the same results for anybody, a visitor from Mars
|> >would expect them to all look more or less the same.
|> >
|> >Maybe they know something we don't. In fact, I'm sure of it.
|> >
|> >Jim Glass
|>
|> Maybe I'm missing something obvious here; since a purely conical
|> nozzle represents a well-known simplification of the bell to
|> save weight and fabrication complexity at the expense of some
|> performance, why is it surprising if someone else chooses an
|> intermediate point between bells and cones?
|>
|> and do not reflect the position of Intel, NETCOM, or Zippy the Pinhead.
|> ============================================================================
|> Jeff Greason "We choose to go to the Moon in this decade,
|> <gre...@ptdcs2.intel.com> and do the other things, not because they
|> <gre...@ix.netcom.com> are easy, but because they are hard." -- JFK
|>
It's suprising because we (and presumably they) try to optimize our nozzles
to minimize losses. In particular, you don't want shocks messing up your
flowfield in the nozzle. If you and I each design a nozzle, using M-O-C,
for the same inlet conditions and gas properties, and we are each smart,
and make no errors, the nozzle contours 'ought' to come out very close
in shape. They are dictated by the M-O-C solution, right?
Well, Russian nozzles look "funny", "strange" to western eyes. They have
a different contour than ours. Have we been designing nozzles wrongly for
40 years? Have they? What is going on here? *I* dunno; I'd love to find out.
If your claim is that the Russians are compromising performance to save weight,
I have two responses:
(1) In my experience, their engines are typically HEAVIER than ours for a given
thrust (spurious marketing claims to the contrary notwithstanding);
(2) My preliminary examination of the various efficiency terms in U.S. and
Russian engines has led me to conclude that we are superior combustion engineers
and that they are superior aerodynamicists(!). In other words, it appears to me,
using limited data, that the Russians take large losses in the chamber and then
get much of the loss back in the nozzle; WE have very efficient combustors and then
throw more away in the nozzle. This is my *IMPRESSION*; if it is correct, then
the funny-looking Russian nozzles are actually better than our "optimum" designs(!)...