RESTRICTED

REPORTS cOlLEC 10
JUN ,2 5 1971;
STANfORD UNIVfR lTV
ENGlNEERING WtR RV

POWER PLANTS /

A REPORT PREPARED FOR THE AAF

SCIENTIFIC ADVISORY GROUP

By
FRANK L. W ATTENDORF
Engineering Division, Air Materiel Command
Wright Field, Dayton, Ohio \} N I VfRs,
9;-'\} l'J..
H. S. TSIEN 4..0
Guggenheim Laboratory <
California Institute oj TechnoloKY::.
If)
POL DUWEZ
Jet Propulsion Laboratory
Califronia Institute oj Technology

Plibl/sheJ May, 1946

by
HEADQUARTERS ~ MATERIEL COMMAND
PUBLICATIONS BRANCH. INTELLtifENcET.2, • WRIGHT FIELD, DAYTON, OHIO

RESTRICTED
 The AAF Scientific Advisory Group was activated late
in 1944 by General of the Army H. H. Arnold. He se·
cured the services of Dr. Theodore von Karman, reo
nowned scientist and consultant in aeronautics, who
agreed to organize and direct the group.
Dr. von Karman gathered about him a group of Ameri-
can scientists from every field of research having a
bearing on air power. These men then analyzed im-
portant developments in the basic sciences, both here
and abroad, and attempted to evaluate the effects of their
application to air power.
This volume is one of a group of reports made to the
Army Air Forces by the Scientific Advisory Group.

lhIs docum.nt contains information affecting the natlonal d.f.nse of the Unit.d StCltes within
the m.aning af the apionag. Act, 50 U. S. C., 31 Clnd 32. as amend.d. Its transmission ar the
r ..... latlon of its contents in any manner to an unauthorlz.d p.rson is prohibit.d by law.
 AAF SCIENTIFIC ADVISORY GROUP

Dr. Th. von Karman

Director

Colonel F. E. Glantzberg Dr. H. L. Dryden

Deputy Director, Military Deputy Director, Scientific

Lt Col G. T. McHugh, Executive

Capt C. H. Jackson, Jr., Secretary

CONSULTANTS

Dr. C. W. Bray Dr. A. J. Stosick

Dr. L. A. DuBridge Dr. W. J. Sweeney
Dr. Pol Duwez Dr. H. S. Tsien
Dr. G. Gamow Dr. G. E. Valley
Dr. I. A: Getting Dr. F. L. Wattendorf
Dr. L. P. Hammett Dr. F. Zwicky
Dr. W. S. Hunter Dr. V. K. Zworykin
Dr. I. P. Krick Colonel D. N. Yates
Dr. D. P. MacDougall Colonel W. R. Lovelace II
Dr. G. A. Morton Lt Col A. P. Gagge
Dr. N. M. Newmark Lt Col F. W. Williams
Dr. W. H. Pickering Major T. F. Walkowicz
Dr. E. M. Purcell Capt C. N. Hasert
Dr. G. B. Schubauer Mr. M. Alperin
Dr. W. R. Sears Mr. I. L. Ashkenas
Mr. G. S. Schairer

LAYOUT & ILLUSTRATION

Capt M. Miller
Capt T. E. Daley

Ii
 TABLE OF CONTENTS

Page
Part I
GAS 11JRBINE PROPULSION
Introduction .................. ............................................ . 1
Reciprocating Engine .•...•••••............••.•.•.•...•..••••....••..•... 1
Com pound Engines ...••..•..•••............•...•.•.......••..••..•.•.... 2
Free-Piston Engine ......................................................................................... . 3
Gas Turbine ......... ................................................................................................ . 4
Turbojet. . ....... . . . .............. . ....... ..................... ...... . .... ............................. .............. 8
Recommendations for Future Research. . . • . • . • . . . . . • • . . . . . • . • . . . . . . . . . . .. 13
Tables ................................................................• 18
Illustrations ................•.•.•...........•.....................•...... 21
Part II
EXPERIMENTAL AND THEORETICAL PERFORMANCE OF AERO PULSE
ENGINES
The Present Status of Aeropulse. . . • . . . . . . • . . . . • . . . • . . . . . . . • . • . • . . . • . . . . .. 41
The Specific Fuel Consumption ..........•.......•.....•.••••.•......... 41
Frequency of Pulsation and Thrust ..........................•.......... 47
Possible Improvement of the Existing Form of Aeropuls4:. . . . • . . . . . . . . . . . .. 51
Valveless Aeropulse .................... '. . . . . . . . . . • . . . . . . . . . . . . . . . • . . . . .. 52
Improvement in Specific Fuel Consumption ........................•••.. 52
Increase in Thrust Coefficient. . . • • . • . . . . . . . . . . • . . • . . . . . . . . . . . . . . . . . . . .. 53
Concluding Remarks ...•........•....•.................•....•.......... , 53
Appendices. • . . • . . . . . . . . . • . • . . . . . . . . . . . . . . . . • • . . . . . . . . . . . • . . • • • • . . . . . . .. 54
Part III
PERFORMANCE OF RAMJETS AND THEIR DESIGN PROBLEMS
Introduction. . . . . . . . . • . . . . . . . . . . . . . • . . • . . . . . . . . . . . . • . . . . . . . . . . . . . . . . . • .. 61
Basis of the Theoretical Analysis. . . • . • . . . . . . . . . . . . . . . . . . . . . . • . . . . . • . .. . .. 63
Calculated Performance. . . . . • . . . . . . . . . • . . . . . . . • . . . . . • . . • . . • . • • • • . . . . . • • .. 66
Weight of a Ramjet. . . . . . • . . • . . . . . . . . . . • . . . . . . . • . . . . . . . • . • . • . • • • . . . . . . •• 70
Boosting of the Ramjet at Low Forward Speeds •••.......•..•............. 70
Research and Development on the Ramjet. . . . . . . •. .. . . . .•. • .•. . ... . . . . . •. 70
Ramjet Combustion as a Problem in Fluid Mechanics. . . . . • . . . . . . . • . . . . . . .. 71
References. . . . . . . .. .. . . . . .. . . . . • . • . • . . . . . . . . .. . . .. • . . . .. . . . .. . . . . .. • .. .. 74
Appendix. . . . . . . ... . . . . . . . . . . . . . . • . • . . . . . . . . . . . . . . . . . . . . . . . . . . . . • • • • • • . .• 75

Part IV
FUTURE TRENDS IN THE DESIGN AND DEVELOPMENT OF SOLID AND
LIQUID FUEL ROCKETS
Types of Rockets and Their Present Applications. . . . . . . . • . . . . . . . • . • • . . . •• 81
Solid· Propellant Rockets. . . . . . . . • . . . . . . . . . . . . • • • • • • • • . . • . . . . . . . • . • . • .. 82
Liquid.Propellant Rockets. . . • • . • • • . • • . . . . . . . . . . • . • . • . • . • . . . . . . . • . . . . .. 92
Part V
HIGH TEMPERATURE MATERIALS FOR AIRCRAFT PROPULSION
Introduction .............•.......................•.....••..•••.......•.. 103
Materials for Gas· Turbine Power Plants .............................•.... 103
Materials for Ramjet Power Plants ....................................... 107
Materials for Rocket Power Plants ............................•.•.•••••.. 108
Materials for Atomic Energy Aircraft Power Plants •............•..••..••.. 111
Conclusions .................... .................................................................... . 112

iii
 PART I

GAS TURBINE PROPULSION

By

FRANK L. W ATTENDROF
 RESTRICTED

PART I

GAS TURBINE PROPULSION

DECEMBER 1945

INTRODUCTION
Gas-turbine propulsion is used as a general term covering the various means of
propulsion incorporating a gas turbine as its primary component. Included under
gas-turbine propulsion are the following systems:
(1) Turboprop, or gas turbine driving a propeller.
(2) Turbofan, or gas turbine driving a ducted fan.
(3) Turbojet, or gas turbine plus jet reaction.
In the pre,ent report brief comments will first be made on the reciprocating
engine, compound engine, free piston, and various gas-turbine systems. Then the
scientific background of some of the propulsion systems will be discussed from the
point of view of stressing future possibilities of development.
The primary purpose of this report is to show ways in which the gas turbine is
basically capable of considerable improvements toward increasing its potentialities
for high-speed aircraft application. In order to obtain improvements of this nature,
greater emphasis on basic and applied research is needed. This report stresses in
particular the items on which research and development should pay the greatest
dividends.

RECIPROCATING ENGINE
The chief advantage of the reciprocating engine at the present time is in low
specific fuel consumption, especially at partial load. This makes it particularly applica-
ble to long-range flight at present cruising speeds. Typical reciprocating engines now
in use have specific fuel consumptions which are in the neighborhood of 0.7 lb/bhp-hr
at full-rated rpm, decreasing to about 0.4S lb/bhp-hr at cruising conditions.
The principle characteristics of several current and proposed reciprocating
engines of higher horsepower of interest fflr comparison with gas turbines, are given
in Table I.
The largest reciprocating engine now under development is the SOOO-hp Lycom-
ing R-77S S. One of the chief aims of this engine is fuel economy for long-range bomb-
1
er application. The estimated minimum specific fuel consumption is approximately
0.37 lb/bhp-hr, the improvement being associated largely with increased compression
ratio. However, it is noted that there is no improvement in either specific weight or
power per unit frontal area of the R· 7755 over the R.4360. The ratio of dry weight to
take-off power is 1.211b/hp for the R-7755 as compared with 1.14 lb/hp for the R-
4360 and the ratio of take-off power to frontal area is 178 for the R- 7 7 5 5 as compared
with 198 for the R-4360.
Further increase of reciprocating-engine power involves further increase in
cylinder volume, which is already at a critical stage as regards number and arrange-
ment of cylinders, complexity of the drh'e, excessive frontal area, and difficulty of
cooling. There are development projects at present aimed at increasing the output
of the R-3350 to 3200 hp and the R·4360 to 4300 hp.
Some increase in fuel economy can reasonably be expected by utilizing higher
pressure ratios, improved fuels, improved metallurgy, etc., but engine manufacturers
consider that it would be difficult to attain a specific fuel consumption essentially lower
than 0.35 lb/bhp-hr.
A serious limitation of the reciprocating engine for high-speed flight is the cooling
problem, as pointed out in the discussion on airplane designs by Sears, Ashkenas and
Hasert (Part II of Aerodynamics and Aircraft Design, another report of the AAF
Scientific Advisory Group). The adiabatic rise in temperature due to impingement of
the air on the engine surfaces is proportional to the square of the air velocity, and would
reach a value of 90°F at 700 mph, and 180°F at 1000 mph. This cooling difficulty
coupled with the relatively large frontal area, would act as a serious drawback to the
application of the reciprocating engine to supersonic flight.
The motorjet or Campini system, whereby a reciprocating engine drives a ducted
fan, has several advantages, especially that afterburning can be used for temporary
increase in thrust, but the basic limitations are those associated with the reciprocating
engine.

COMPOUND ENGINES
The compound engine represents the beginning of a transition between the con-
ventional engine and the gas turbine. Compounding is an attempt to recover a con·
siderable portion of exhaust energy by operating some form of a gas turbine which
feeds back into the main power drive. Several projects now under consideration are:
1. Wright Aeronautical Corporation proposes a system in which the exhaust
blows forward through turbine buckets located on the tips of cooling fan blades in the
front nose cowl ring. The turbine drives the cooling fan, and the surplus power is
transmitted to the propeller shaft through a gear transmission. By this means, it is
estimated that the fuel consumption can be reduced from 10 to 20 percent.
2. Allison V·1 710-127. An exhaust turbine is mounted on the rear of the auxil.
iary stage supercharger, and is connected to the crankshaft system by means of an
extension shaft.
3. The General Electric Company at West Lynn have a development project for
an exhaust gas turbine feeding back into the crankshaft of a R·4360 engine.
2
 ,REE-PISTON ENGINE
In the compound engine described in the preceding paragraphs, the pistons of"
nciprocating engine are used partly to drive a crankshaft and partly to serve as a gas
generator to drive a gas turbine. The so-called free-piston engine represents the ex-
treme in the compounding principle, in that the pistons are used solely as gas-gener-
ators, and the products of combustion are used entirely to run a gas turbine.
The main principle of the free-piston engine is the utilization of floating pistons
which, instead of driving connecting rods and crankshafts as in conventional engines,
generate compressed and heated gases by recoiling to and fro between cushions of
charged air, with combustion taking place at the time of maximum compression of
each charge. The products of combustion are then available either to pass directly
through a discharge nozzle, furnishing jet propulsion, 01' they may pass through a
turbine, and generate power for driving a propeller 01' ducted fan. For the gas-turbine
application, one of the chief advantages claimed for the free-piston engine is low
specific fuel consumption, which is associated with the fact that combustion takes
place at a relatively high value of instantaneous pressure. Since partial expansions
take place in the cylinder after combustion, the temperature of the gases entering the
turbine is not excessively high. The free piston can also be used to compress air which
is mixed with hot gases, thereby resulting in a cooler mixture. These factors are favor-
able for turbine wheel construction and life. In addition, the partial load character-
istics are more favorable than the normal gas turbine.
The successful applications of the free-piston engine so far have been to stationary
power plants 01' marine engines, where weight and bulk were not important factors.
Professor Junkers had been experimenting in Germany with the free-piston principle
for several decades, and the Junkers Company has been making a free-piston marine
engine. Pescara in France has developed stationary power plants of the free-piston
type. Sulzer Brothers in Switzerland also build free-piston power plants.
The Japanese Navy, in 1940, became interested in applying the free-piston engine
to gas turbine - propeller combinations for aircraft propulsion, and in 1941 a model
free-piston gas turbine was placed in operation. It soon became apparent, however,
that too much development work would be needed before it could be useful in the war
effort, and it was later abandoned in favor of copying German turbojets.
In Germany, O. Lutz of the Luhfahrtforschungsanstalt Hermann Goring was
attempting to develop a free-piston engine for aircraft application. This engine was
claimed to be essentially lighter in weight and smaller in bulk than the Pescara and
Junkers types. This was accomplished chiefly by having" a series of pistons rotate and
oscillate within a torus-shaped housing. In a working prototype there are two banks of
three pistons each, which rotate about a central axis. Although both groups of pistons
rotate at the same mean speed, they have a superimposed oscillatory motion of such
nature that one piston group is always oppositely phased to the other. In this way there
are six working spaces, of which three are expanding while the other three are con-
tracting at the same rate. Valves are simple slots which require no further mechanism.
By the end of the war, the prototype had not been tested enough to indicate its feasi-
bility.
With regard to possible aircraft application of the free piston, it is believed that
a specific fuel consumption of 0.35 lb/bhp-hr may be attainable. Considerable re-
3
significant factor is the turbine inlet temperature T4. It is seen from the cycle diagram
of Fig. 2 that higher tenrperatures result in greater work output. However, as far as a
theorectical cycle without losses is concerned, the fuel consumption would increase
proportionately with the net output so that the specific fuel consumption would be
independent of turbine-inlet temperature. However, for the actual cycle with losses,
increased turbine-inlet temperature has generally a beneficial effect on fuel con-
sumption.
This is shown in Fig. 3, for temperatures of 1600 o F, which have already been
exceeded; 2000 o F, which several current units almost attain, and 2600 0 F for a future
goal. A flight Mach No. of 0.8 is used for this example. The assumed compressor and
turbine efficiencies of 85% have already been attained in axial flow units.
It is seen that the minimum specific fuel consumption shows a moderate decrease
with increasing temperature. The influence of turbine-inlet temperature on the power
output is much greater, as shown in Fig. 4, for the same range of temperatures. It is
seen that the power output would be increased 60% at 2000 0 F and over 180% at
2600 0 F for the same mass flow, providing the pressure ratio were also increased for
the higher temperatures. The magnitude of the power increase emphasizes the need
for intensive support of high-temperature materials development.

4. 'nRuence 01 Compressor and Turbine EIIiciency.

The influence of efficiency on specific fuel consumption as a function of flight speed
is shown in Fig. 5 for a constant pressure ratio of 6:1. The corresponding curves for
shaft horsepower are shown in Fig. 6. The values of 85% for component efficien.cies
are fairly representative of current axial flow units. Laboratory tests of experimental
units have indicated component efficiencies of 90% to be within reach. As an ultimate
goal, the values of 94% for compressor efficiency ,and 96% for turbine efficiency,
were used because these values assume that all secondary losses have been eliminated,
and only the unavoidable skin friction remains; in other words, they represent ceiling
values. As a matter of interest, the ideal curves for 100% efficiency are also included.
This gain in fuel economy and shaft horsepower emphasizes the need for compo-
nents development, involving new blade shapes, study of mutual interference between
blades, three-dimensional flow, compressibility effects, and the influence of centrifugal
force on boundary layer flow. All these subjects relate to the important field of high-
speed flow in rotating machinery, which should be given much more attention in the
future.

5. 'nfercoo/ing.
One means of improving the power output of a gas turbine is by intercooling
between compressor stages. By this method the power required by the compressor to
obtain a given pressure ratio is less, which leaves more of the turbine output for the
propeller. The nature of intercooling is such that its effectiveness increases with
pressure ratio.
The pressure ratios of most current units are too low to warrant incorporation of
intercooling, and its utility is greater for gas turbines with high pressure ratios. It is
likely to become of increasing importance since pressure ratios are becoming pro-
5
gressively higher. Interviews with manufacturers indicate that the majority did not
think that intercooling paid for the increased weight and complication, except for
units of high pressure ratio.

6. Regeneration.
One of the most effective means of improving cycle efficiency and decreasing
specific fuel consumption is by the use of regeneration. The reasoning behind this is
that the exhaust gas leaving the turbine still has an essential temperature head. This
heat may be partially recoveredby the incorporation of a heat exchanger in the turbine
exhaust. If this heat is used to preheat the air entering the combustion chamber, it is
obvious that less fuel will be eneded to raise the temperature of the combustion gases to
the allowable limiting value at the turbine inlet. By this means, it is estimated that a
saving of between 20 and 30 percent in the specific fuel consumption is possible.
Against this advantage must be weighed the disadvantage of increased weight, pressure
loss, and complications. Wright Aeronautical Corporation engineers, among others,
have made a theoretical study of regeneration, including weight estimates.
Sample estimates for a typical 5000.hp gas turbine are:
W ithollt Regeneration
Dry Weight 40001b
Specific Fuel Consumption at Sea Level .80 lb/hp/hr
Specific Fuel Consumption at 20,000 ft .58 Ib/hp/hr
With Regeneration
Dry Weight 5600lb
Specific Fuel Consumption at Sea Level .561b/hp/hr
Special Fuel Consumption at 20,000 ft .41 Ib/hp/hr
One of the chief factors hindering development of the regenerative system is that
this type of heat exchanger for aircraft application is in the early stages of development.
A German project for developing a ceramic heat exchanger was being conducted at
Gouingen, and should be reviewed for possible application. General research on the
development of light and efficient heat exchangers with low pressure loss for aircraft
application is recommended.

7. Relaeat.
A further increase in performance of the gas-turbine cycle can be obtained by
dividing the turbine into several separate stages and reheating the gas by combustion
between the stages. Both power and fuel economy are improved. Against this gain the
increased complication and weight must be weighed. Reheat for aircraft gas turbines
is relatively unexplored, because lightness and simplicity were necessarily prerequisite
to the first aircraft applications. All such methods of improvement should be explored
for the future.

8. Separate Propeller Tu"'ine.

The simple gas turbine had the fundamental characteristic of having its perform-
ance improve with rotational speed. This is opposite in nature to the reciprocating
engine, which fact makes the normal gas turbine relatively more disadvantageous
6
under cruising conditions than at maximum output. One method of improving gas
turbine performance at partial load is obtained by using two turbines, one driving the
compressor only without giving any net power, and the other supplying power solely
to the propeller. The improvement at partial load is brought about by the fact that the
turbine driving the compressor can always be operated at the most suitable speed for
the compressor, independently of the speed of the turbine driving the propeller. It is
obvious that the two-turbine system is more complicated than the single-turbine
system; however, it might be justified for large, long-range airplanes which operate a
major portion of the time at cruising speed. For such application, the advantage of the
two-turbine system must be compared with alternative plans such as utilizing a smaller
single gas turbine which has its maximum speed at cruising conditions, and obtaining
higher speeds for short periods by supplemental burning or other temporary methods
of increased performance, or in a multi-engine aircraft by having several units which
are closed off under cruising conditions and only used for high speeds.

9. Closed Cycle. ,
In the closed cycle as developed by). Ackeret and C. Keller for Escher Wyss in
Zurich, the gas passing through the compressor and turbine is recirculated in a closed
circuit, while combustion takes place externally and transmits heat to the gas through
a heat exchanger. This cycle offers several advantages from the theoretical point of
view; for instance, since the cycle is enclosed, it may be made gas tight and operated
at an] desired pressure level. By using air under pressure, it is obvious that a higher
output of power may be obtained with smaller dimensions of compressor and turbine.
Also, there is much better control of the temperature distribution entering th~ turbine,
which enables the turbine to operate at a somewhat higher level of mean temperature
than would otherwise be the case. A schematic diagram of a closed cycle is shown in
Fig. 7. The air is compressed in the axial 01' centrifugal compressor A and heat is added
at B through a heat exchanger associated with external combustion. The hot gas passes
through the turbine D and the residual heat is removed by means of the heat exchanger
F and transferred to the air leaving the compressor. The air leaving the heat exchanger
recirculates and enters the compressor without further change in pressure and temper-
ature. High efficiencies are obtainable with this cycle, as shown by curves of). Ackeret
in Fig. 8. It is seen from the figures that the level of efficiency occurs at relatively low
values of pressure ratio, which means that a compressor-turbine unit with fewer stages
could be used. The disadvantage of the system is the large amount of heat exchange
surface required, which would tend toward a unit much too bulky and heavy for present
consideration. However, the inherent advantage of good efficiency would indicate
research on this cycle to be desirable. It is recommended that estimates be made of the
over-all weight, size, and performance of this system. Another possibility which should
be investigated is the utilization of different gases for this cycle, such as freon, carbon
dioxide, argon, helium. etc.

10. Displaceme'" CompreuOf Units.

There are some proposed gas.turbine units, such as the Lysholm type, which utilize
a displacement type of rotary compressor in place of the more common axial or centri-
fugal. The displacement compressor had the advantage of high compression ratio,
7
but is essentially low in capacity. Therefore, in order to obtain the flow needed by
gas turbine units, a system is required which involves weight and complication.
11. Tur&olan. *
The advantage of the turboprop over the turbojet in specific fuel consumption for
subsonic flight speeds is due to the higher propulsive efficiency of the propeller. This
is, in turn, due to the lowered velocity of the slip stream compared with the exhaust
jet velocity of the turbojet. To improve the fuel economy of the turbojets, one can allow
practically all the available power in the combustion gas from the combustion chamber
to be converted into mechanical power by the gas turbine. The excess power over that
required for driving the compressor is then used to drive a separate fan which com-
presses air from the atmosphere in a duct. The compressed air then mixes with the
exhaust gas from the turbine and the mixture discharges through the tail pipe. This is
the turbofan as shown in Fig. 9. Due to the increased air mass used, the exit momentum
is larger than the turbojet with a corresponding reduction in the discharge velocity.
The propulsive efficiency is thus increase. and specific fuel consumption reduced.
For operations at transonic speeds, the turbofan has the added advantage over the
turboprop in that the air duct will slow down the air stream before arriving at the fan.
The relative velocity of the fan blade with respect to the air stream can be kept at a
favorable value for high efficiency. On the other hand, conventional propellers of high
efficiency at transonic speeds are difficult to design. British investigators have shown
that with the present design of propellers, the turbofan should give greater fuel eco-
nomy than the turboprop for flight velocities over 550 mph.
At supersonic flight velocities, the possibility of additional combustion in the tail
pipe of the turbofan should also be considered. For such operations, the turbofan is
not unlike the ramjet. In fact, if for a given air flow one can imagine a series of designs
with gradually decreasing amounts of air going through the compressor-gas turbine
system of diminishing size, then the ultimate power plant, with no air at all passing
through the gas turbine, is the ramjet. In this sense then, the turbofan with tail-pipe
burning is an intermediate between the turbojet and the ramjet. At supersonic speeds,
the thmst of such a system will be approximately that of a ramjet but the weight will be
slightly heavier than the ramjet. However, it will be much lighter than a turbojet of
equal thrust since only part of the total air flow goes through the compressor gas tur-
bine system, with the consequent reduction in the size of the gas-turbine. Furthermore,
the turbofan with tail-pipe burning can also develop thrust at low speeds. It is, thus, a
more versatile power plant than the ramjet. Therefore, further investigations of this
power plant are recommended.
12. Current Gas-Turbine Developments.
Brief data on several existing and proposed gas-turbine units are given in Table II.
TURBOJET
1. General Description.
A schematic diagram of a typical turbojet unit is shown in Fig. 10. It is similar in
principle to the gas-turbine cycle described in the previous chapter, except for the fact
* Contribution of Dr. H. S. Tsien to this section is acknowledged.

8
Ihat there IS no propeller drive, and all the net output of the unit goes into the jet. The
turbine extracts just enough heat to drive the compressor. fhe work developed by the
turbine, therefore, is equal to the work absorbed by the compressor. The remaining
energy of the hot gases leaving the turbine goes into thrust upon expansion in the
nozzle.
Calculations were made by M. Alperin of the specific fuel consumption and specific
thrust for a series of assumed values of turbine-inlet temperatures, compressor e~­
ciency, turbine efficiency, and pressure ratio. * With the aid of these studies, it is possi-
ble to draw general conclusions regarding the factors influencing the performance of
turbojets, and to estimate the order of magnitude of the gain to be expected for various
future improvements.

2. Fadors InRuencing Performance.

Since the thmst of a turbojet unit is obtained by expansion of hot gas through the
tail pipe, it would be expected that higher compression ratio would increase the effi-
ciency of the system. However, there are several factors which limit the validity of this
statement. First, there are lossef in the turbine and compressor, and these losses
increase with compression ratio. Therefore, a point is eventually reached at which the
increase of losses is greater than the theoretical gain. Also, it should be noted that the
over-all efficiency in flight is equal to the product of thermal efficiency of the gas turbine
unit alone times the propulsive efficiency. The propulsive efficiency, in turn, depends on
the ratio of jet velocity to flight velocity. The factors influencing the performance of a
turbojet are in general those influencing the gas turbine, as previously discussed.

3. InRuence of SpeeJ and Altitude.

Flight velocity influences the performance of a turbojet in various ways. The ram
pressure increases at the compressor inlet with increasing speed, the mass flow in-
creases, and the compressor and turbine reach a new equilibrium condition. The
differential between jet speed and flight speed also changes. It is difficult at present to
make general statements since for any specific unit the new equilibrium condition has
to be computed for each new flight speed. Curves for a specific unit, the 1-40, are shown
in Figs. 11, 12, and 13.
With increasing altitude the temperature of the air entering the compressor inlet
decreases. This is a favorable effect if the turbine-inlet temperature is maintained at a
constant limiting value, conditioned by the me,tallurgy of the turbine blades. The lower
compressor-inlet temperature, therefore, combined with the same tprbine-inlet tem-
perature, would mean an over-all increase in temperature ratio which is favorable for
cycle efficiency~ However, the Mach number relative to the compressor tip is increased
due to the lower value of sound velocity, which usually results in lower compressor
efficiency. Since the velocities in the unit remain approximately constant and independ-
ent of altitude, it would be expected that the mass flow would decrease approximately
in proportion to the density. The thrust, in turn, would decrease with the mass flow.
The influence of altitude on the performance of the 1-40 is also illustrated in Figs. 11,
12, and 13.
* The Research Branch of the Power Plant Laboratory, Wright Field, has made similar calcula-
dons, limited, however, to the region of subsonic Hight.

9
 4. Met~oJs of Obtaining Temporary Increase of Thrust.
Methods of obtaining temporary increase of thrust are especially important for
jet engines at both ends of the speed range. For take-off the over-all efficiency of the jet
unit is low due to the high residual knietic energy associated with the large differential
between jet velocity and airplane speed. Since the efficiency of the turbojet increases
with rpm, cruising conditions are closer to the maximum than is the case for a con-
ventional engine. Increased thrust is, therefore, desirable for temporary increase of
speed.
One method of obtaining a temporary increase of thrust is by after burning or
supplementary combustion in the tail pipe. The amount of thrust increases with in-
creasing pressure ratio. Therefore, afterburning is especially effective with high pres-
sure ratio gas turbines, and the thrust increases rapidly with flight speed due to the
ram pressure. In the report by Sears, Ashkenas, and Hasert, it was estimated that
although supersonic speeds were questionable for turbojets now in operation, after-
burning would furnish enough thrust to make supersonic speeds appear feasible. Due
to the importance of this question, it is recommended that systematic theoretical and
experimental studies be conaucted to explore the possibilities of afterburning, es-
pecially at supersonic speeds.
Another promising means of obtaining increased thrust is by liquid injection,
especially in combination with after burning. Injection of ammonia and fuels should
receive further attention, especially with regard to oxygen-bearing fuels for altitude
operation.

5. Estimates at Supersonic SpeeJ.

Experimental data on the performance of turbojet units above the speed of sound
are greatly needed at the present time for the predictions of future application of super-
sonic aircraft. Some theoretical estimates of turbojet performance at supersonic speeds
have been made by M. Alperin. The assumption was made that a shock wave forms in
front of the duct entrance of the turbojet, and that the flow through the turbojet is
subsonic. Pressure, density, and temperature changes across the shock wave were
calculated by the usual shock equations. With these new values of pressure, density,
and temperature at the inlet, the turbojet performance was calculated in the normal
way. Sample curves showing the influence of pressure ratio on specific fuel consump-
tion for several Mach numbers are given in Fig. 14. It is to be noted that the optimum
pressure ratio of the turbojet compressor decreases with increasing Mach number as
would be expected because of the increasing contribution of ram pressure. It is seen
from the curves that as Mach number increases a value will be reached where the op-
timum compressor pressure ratio is one, and the turbojet becomes a ramjet. The trans-
ition point between turbojet and ramjet depends on the assumptions as to turbine-inlet
temperature, compressor efficiency and turbine efficiency. The higher these values the
higher the Mach number at which a turbojet will be effective.
The influence of turbine-inlet temperature on specific fuel consumption and speci-
fic thrust is shown in Figs. 15 and 16 respectively, for a flight Mach number of 1.6.
It should not be concluded from the above, however, that turbojet units now in
operation are capable of propulsion at supersonic speeds. In the report by Sears,
10
Ashkenas, and Hasert, it is shown that using scaled-up test results of the thrust per unit
frontal area for the Westinghouse 9. SA turbojet unit, the drag of the frontal area would
become equal to the thrust at about 800 mph, so that the unit could not propel its own
frontal area above this pseed unless afterburning were employed. However, estimated
thrust values of the Lochkeed L-I000 appear sufficient for supersonic. flight, without
afterburning, and Lockheed estimates with afterburning show a 100% increase of
thrust. The curves shown in Fig. 16 emphasize the fact that a very considerable increase
of thrust for the same frontal area could be obtained if applied research can develop
methods of increasing turbine temperature limits. In addition, the utilization of after-
burning at supersonic speeds should be intensively investigated since, even with
present-day units, the increase of thrust should be sufficient to make supersonic flight
possible.
Internal duct flow for bodies traveling at high subsonic or supersonic speeds is a
relatively unexplored field. Tests by the NACA have shown promise of developing
effecient supersonic diffusers. Considerable applied research.in this field is needed. It
appears possible to develop diffusers whicp, instead of having a strong shock in front
of the duct entrance have a weaker shock inside the duct, thereby decreasing the loss.
Control of this shock by boundary layer suction should be further investigated. Cone
entry diffusers appear especially promising.

6. Comhuslion.
The combustion process opens a new field for analysis in that both thermodynam-
ics and aerodynamics must be taken into consideration. This forms, in fact, a new
science which has recently been referred to as aerothermodynamics. The fuel is sprayed
into the combustion cylinder through a series of spray nozzles, and combustion takes
place in the presence of a steady flow of air. Since the amount of fuel is small in propor-
tion to the air, the process involves an addition of heat content without essential change
of mass flow. The equations are, therefore, similar in nature to those for a shock wave.
If conditions upstream of the combustion process are denoted by a subscript 1, and
conditions downstream by 2, and H represents the heat added by the combustion
process per unit mass flow, then the three basic equations for mean changes through
the combustion chamber are:

For a given o:et of entering conditions the amount of heat added can be plotted as a
function of either pressure or Mach number at the downstream section. The pressure
decreases in a parabolic function as the amount of heat is increased. However, the
significant fact emerging from the analysis is that there is a definite maximum to the
amount of heat that can be added. The equations show that the maximum added heat
corresponds to a discharge Mach number M2 equal to 1. For this maximum condition
of added heat, a curve can be plotted of maximum heat versus inlet Mach number. This

11
is shown in Fig, 17. It is noted that the maximum heat which can be added increases
rapidly with decreased inlet Mach number. This would indicate that low entrance
velocities are desirable from the point of view of adding as much heat as possible.
However, it should be pointed out that lower velocities are associated with larger
combustion chambers which mean larger and heavier units, so that for this particular
item a compromise must be made. One of the chief problems in combustion is turbu-
lent mixing and diffusion to obtain an efficient and uniform combustion process. There
has been considerable aerodynamic research on the turbulent Bow of air alone, but
there is a need for research on the turbulent mixture of air and fuel droplets. This is
especially true at altitudes where the air deasity is very low.
One of the difficulties of present units is the nonuniformity of the combustion
process. The distribution of temperatures around an annular ring is irregular and high
points or hot spots occur. This means that the mean turbulent temperature has to be
kept at a lower value than would be the case if the temperature were uniform and
regions of higher temperature did not exist.
Combustion at altitude is a problem about which there is very little known.
Various factors become critical at altitude, for instance, the air becomes too rare for
proper diffusion and mixture; the spark becomes too weak for proper ignition; the
distance between the fuel nozzles and the point of ignition may be 'Wrong; the com-
pressor and turbine may be far enough off their design operating point to cause
pumping and instability; the relative Mach number at the compressor inlet increases
with the decreasing temperature, which may lead to increased losses and choking.
All these subjects should be the subject of intensive investigation in altitude com-
bustion laboratories and in high-speed, altitude wind tunnels.

7. UtilizGfion 01 Nuclear Energy.

It is expected that the application of nuclear energy will revolutionize propulsive
methods, but it is too early to make an enginee.dng evaluation of its possibilities.
Some speculation as to the application of nuclear energy to propulsion is given in a
report by H. S. Tsien.
In general, it would appear that the application of nuclear energy to gas-turbine
propulsion would remove the item of specific fuel consumption as a factor of consider-
ation, and range would automatically become unlimited. The chief question is one of
initial weight and the utilization of the energy at the proper rate. For supersonic turbo-
jet application, for instance, the use of nuclear energy by itself might not contribute
directly to enable the turbojet to By faster than sound. This would still be a problem of
providing greater thrust per unit frontal area, which is bound up with the basic turbo-
jet development problems, such as increasing turbine-temperature limits and improv-
ing compressor and turbine efficiencies. However, the item to be pointed out very
strongly is that if applied research can develop improvements in basic turbojet per-
formance to make the turbojet capable of supersonic Bight, then with the utilization of
nuclear energy the range at supersonic Bight becomes infinite. The possibility of future
applications of nuclear energy, therefore, appears to emphasize the importance of
intensive development of gas turbine components.
12
 8. Existing and Proposed Turbo;et Units.
Several of the more representative types of existing and proposed turbojet units
art: presented in Table III for American units, Table IV for British, Table V for Ger-
man, and Table VI for Japanese units.
RECOMMENDATIONS FOR FUTURE RESEARCH
1. Umitations 01 Propellers lor Hig#t-Speed Flight.
In the various studies of the relative performance of reciprocating engines plus
propeller drive, gas turbine plus propeller drive, and turbojet, the main factor which
hinders the comparison is lack of information on propeller efficiency at speeds above
500 mph. Estimates of high-speed propeller performance are usually based on two-
dimensional airfoil data. However, recent altitude flight tests indicate that the Jhree-
dimensional flow associated with an actual rotating propeller may not be as critical as
two-dimensional blade theory would predict. In addition, German wind-tunnel tests
on propellers with swept-back blades indicated improved efficiency at transonic rela-
tive speeds near sonic. Therefore, it is recommended that test data be obtained as soon
as possible in high-speed wind tunnels and in flight on the behavior of conventional
and swept-back propellers at transonic and supersonic speeds; also, that development
and testing of propeller sections specifically designed for supersonic speed~ be
expedited.
2. Interrelationship Among Size, Weight, Thrust, Pressure Ratio, Turbine-Inlet
Temperature, Speci", Fuel Consumption.
As discussed in previous sections of this report, it is evident that the performance
of a gas-turbine cycle is influenced greatly by the selection of pressure ratio, turbine-
inlet temperature, etc. However, theoretical advantages, as shown by such analysis,
must be weighed against the practical aspects such as size, weight, life, simplicity, and
maintenance. For instance, if the analysis for a certain application shows a gain in fuel
economy for an increased pressure ratio, it should be kept in mind that the higher
pressure would normally be obtained by increasing the number of stages. This would
add to the rotor weight which, in turn, would involve other weight increases in the
unit. Whereas this increase might not be justified for high-speed applications at short
range, it might well be profitable for long-range cruising. The danger of making gener-
al conclusions on the basis of specific cases should be avoided, and all aspects recon-
sidered for each new application.
3. Lighter Weight by Improved Construction and Materials.
The development of gas-turbine and turbojet units to date has shown that there is
still room ahead for improvements in the specific weight, or ratio of weight to thrust.
It is to be noted, for instance, that in the successive development of five different West-
inghouse turbojels there has been a consistent decrease of specific weight from unit
to unit as follows: .66 - .56 - .49 - .44 - .36. The British RB-41 has a specific weight
of .31. Manufacturers believe it possible to attain a value of .25 pound per pound of
thrust.
Further material research should be conducted on alloys such as beryllium,
aluminum, magnesium, etc. Parts such as casings, etc., might make use of ceramic
sandwich construction with resin binder.
13
 4. Compressor Research.
Axial compressors have been developed to a fairly high state of efficiency by using
many stages of relatively highly loaded airfoil-type blading. The use of lattice-type
blading with closely spaced and highly cambered blades, as developed at Wright
Field for supersonic wind-tunnel compressors, would greatly increase the stage pres-
sure rise, and, therefore, decrease the number of stages. This would be quite ad-
vantageous towards decreasing the weight and complication of a gas-turbine unit
providing high efficiency could be maintained. However, the chief factor limiting
the development of higher stage pressure ratio is the almost complete lack of experi-
mental data on the behavior of blade sections in lattice arrangement at high Mach
number. Since the stage pressure developed by a compressor is approximately pro-
portional to the square of the inlet Mach number, it is apparent that any gain in inlet
Mach number is very advantageous provided that no excessive loss of efficiency is
involved. High-speed blade research should be actively sponsored not only in two-
dimensional stationary lattice models, but also on three-dimensional rotating grids.
Boundary layer control applied to compressor blades has been considered by
the Power Plant Laboratory, Wright Field, but further investigation is needed to see
if the increased complication is justified.
Another possibility is the utilization of supersonic flow in the compressor. Encke
of Gottingen experimented with wedge-shaped- rotor blades at supersonic relative
velocities, but found the operating condition too far away from the design point to
warrant any conclusion to be drawn. Weise of the DVL has an experimental com-
pressor whereby the velocity leaving the rotor blades was supersonic. The stator vanes
behind the rotor were arranged to cause a normal shock which added to the pressure
rise. Estimates showed promise of considerable increase in stage-pressure rise in this
way, but preliminary experiments showed deviation from the design condition too
great to be conclusive.
Considerable increase in mass flow and associated thrust per unit frontal area
of a gas-turbine unit is believed possible by axial compressor research and develop-
ment. Wind-tunnel compressor design at Wright Field has indicated that smaller hubs
might be used than is usually used in present gas-turbine systems. This means in-
creased disk area and associated increase mass flow and thrust. At the same time it
is believed that the axial velocity components relative to the blades can be increased,
with still further increase of mass flow and thrust. To explore these possibilities re-
quires intense applied research in the field of compressor design, especially for three-
dimensional flow in rotating systems.
It is recommended that the subject of high-speed flow in rotating machinery be
given much more attention in the future. High-speed component test facilities are
needed.

5. High-Temperature Materials.
The development of high-temperature materials is especially important in view
of the considerable increase of thrust or power per unit frontal area which occurs
with increase of the turbine-inlet temperature. It should be emphasized that the fatigue
characteristics at high temperatures are of greater importance than simple rupture
14
strength. The development of ceramics should be given special attention because
they give promise of significant temperature increase. Research on the develop-
ment of ceramic alloys which can be used as the construction material itself should be
expedited.

6. Turbine Research.
Aerodynamic research on turbines is needed on problems of three-dimensional
flow, degree of reaction, influence of tip clearance and interaction with nozzle flow.
The present turbojet units tend to use impulse-type of turbine blading because it is
possible to utilize only one stage for the lower pressure ratios of such units as the
1-40 and TG-180. Reaction blading appears to be somewhat more efficient, but since
it has less pressure drop, more stages are required. There is an important question,
therefore, as to the interrelationship between efficiency and the number of stages.
Aerodynamic factors contributing to the forcing function for turbine-blade vibration
should be investigated. Utilization of hollow blades, especially in conjunction with
blade cooling or boundary layer suction, is an important research problem, in the
light of utilizing higher gas temperatures with associated thrust increase.
With respect to the cooling of turbine blades several German projects are of in-
terest. One project by Schmidt of the LFA was concerned with water cooling of the
blades. The steam resulting from the heating of the water was utilized in a steam tur-
bine which could either feed back into the driveshaft or be used to furnish auxiliary
power for the aircraft.
Another similar project was by Rietz of Gottingen who was experimenting with
sodium instead of water as the cooling medium.
Cooling of walls by seepage through porous material as mentioned in. the report
by Duwez should be investigated for application to turbine blades.
In general, research on blade cooling is recomended because of the possibilities
of increased thrust per unit frontal area. The cost in weight and complication should
be estimated throughout in order to have realistic evaluations.

7. Combustion Research.
This subject is considered especially important in view of the primitive state of
knowledge regarding even the fundamental processes of combustion. Investigation
should be made as to the physical principles at work and the factors of similarity
which make it possible to predict behavior in the di1{erent operating conditions. Es-
pecially important are questions of stability at altitude and behavior over wide operat-
ing ranges. For flight above the speed of sound, the effect of flow changes through the·
jet unit should be investigated as to their effect on combustion. The relative merits
of the annular type of combustion chamber versus the individual type should be studied
Interests of size and weight dictate small, compact combustion chambers. This, how-
ever, involves higher air speeds through the combustion chamber and sufficient
investigation should be made to enable reasonable design compromises to be reached
The mixing of fuel with air should be studied, especially under altitude conditions
where the air density is low. Also in dealing with combustion chambers, the ques-
tion of safety becomes important, and the construction should be considered from this
viewpoint.
15
 '0 Fuel Researc",
It is recommended that more study be given to the heat content per unit volume of
various fuels as well as the heat content per pound, since with the large quantities of'
fuel required by jet airplanes, the frontal area is a significant factor. Combustion effi-
ciency should also be studied, since a fuel with a high energy content may lose its
advantage if the combustion efficiency is low. Hydrogen is attractive from a heat value
standpoint and investigation is recommended as to the possibility for its utilization.
Development of fuels for high.altitude operation: It may be necessary to carry
fuel with a certain percentage of self-oxidizer to be used for altitudes above the limits
of normal combustion.
Of paramount importance for the future is the possible utilization of atomic
fuels. This is discussed in the report by H. S. Tsien. It is pointed out that although
the heat content of atomic fuels range between 3 x 10 8 and 3 x 1010 BTU/lb as com·
pared with 1.87 x 10 4 BTU/lb for gasoline, the significant factor for its utilization in
gas turbines, etc., is the rate of energy release, which is·about 5 x 10· BTU/hr/lb,
for a typical case. Although low, application to gas.turbine propulsion appears within
the range of feasible application.
Ther~ is a distinct need for engineering development work to apply the abstract
results of nuclear physics to engineering utilization of nuclear power for propulsion.

9. Cycle Researc,,"
It has been noted in an earlier part of this report that considerable gain in fuel
economy can be obtained by use of regeneration, reheating, use of a closed cycle with
gas under pressure, etc. However, the advantages are counterbalanced by increased
weight and complication. It would be highly important to know to what degree weight
and complication might be cut down by suitable research on heat exchangers and other
such items entering these cycles.
It is recommended that the component parts involved in cycle improvement be
subjected to systematic research regarding future improvements. Similar studies
should be made on the compound engine, the free-piston system, the turbofan, and
the use of new gases in csnjunction with a closed cycle.

10. Duct Re.earc,,"

At high speeds the aerodynamics of shapes at entrance and exits of ducts, as well
as the interior portions, become increasingly important, especially at transonic and
supersonic speeds. It has been found for instance, that an increase in diffuser effi.
ciency at supersonic speeds is obtainable by cone, entrances, and also by suitable boun-
dary layer control. The whole subject of high. speed flow in ducts should be the sub·
ject of combined theoretical and experimental research. This should also include the
effect of angles of attack or yaw on the flow, since unbalanced entrance condition can
ruin the aerodynamic performance of compressors Qr turbines, and if the nonuniformi.
ty is great enough it may lead to blade failures. Theoretical work on compressible
flow should also include the flow through both the rotating and stationary blades of
compressor and turbine.
16
 In this connection also, it is important to study means of varying the flow condi.
tions in the duct to suit different speed or altitude requirements. This is especially
true when supersonic speeds are obtained in the tail pipe since only one speed of air
passing through the duct can be obtained for anyone duct shape. Also, the use of
variable vanes for the compressor or turbine should be studied since one of the limita-
tions on compressor and turbine design at present is the necessity for having stable
operation over a wide range. In general, the interrelationship between turbine and
compressor characteristics should be investigated so as to eliminate danger of stalling
or pumping under certain flight conditions.

11. IlicreaseJ Thrust lor Talce-OR ancl High SpeeJ.

The providing of increased thrust for take-off is of great importance for jet units
since take-off is a critical condition. Liquid injection, afterburning, etc., should be
the subject of systematic research. Since the increased thrust due to afterburning is
considerably increased by ram, it becomes of significant magnitude at high speeds,
and should play an important part in obtaining supersonic speeds.
Other miscellaneous research problems are: the development of lighter and more
convenient means of starting jet units, including self-starting devices; lubrication re-
search, especially for larger units operating over wider range of thrust, temperature,
altitude, etc.; determination of optimum clearances; dirt and erosion problems; icing
problems; and the development of cheap expendable units for pilotless aircraft.

17
 TABLE I - RECIPROCATING ENGINES
Manufacturer Wright Aeronautical Pratt & Whitney Lycoming
Model R-3350-57 R-4~60-27 XR-7755-3
Type 18-Cyl 28-Cyl 36-Cyl
Single-Stage Single-Stage Dual Rotation
Single-Speed Variable-Speed (Two-Speed Reo Gear)
Supercharger Supercharger Fuel Injection
Single-Stage
Single-Speed
Supercharger
Sea Level Performance Take-Off 2200 hp Take-Off 3000 hp Take-Off 5000 hp
60% Cruise 1200 hp 60% Cruise 1500 hp Medium Cruise 2000 hp
Speed RPM 2800 rpm 2700 rpm 2600 rpm
2000 rpm 2144 rpm 1600 rpm
Air Flow 4.45Ih/sec 6.11 lb/sec
2.22Ib/sec 3.13 lb/sec
Specific Fuel .80 lb/hp-hr .73 lb/hp-hr .70 lb/hp-hr
Consumption .44 lb/hp-hr .46 lb/hp-hr .37Ib/hp-hr
Compression Ratio 6.85:1 7:1 8.5:1
Over.all Length 76.26 in. 96.75 in. lZl in.
Over-all Diameter 55.94 in. 52.5 in. 61 in.
Total Dry Weight 2758lb 3404lb 6050lb

TABLE II - GAS TURBINES (American Types)

Manufacturer G.E. Wright Northrup Frederick Westinghouse
Flader
Model T-31 (TG-100) T-35-1 T-37-1 XT-33-1 X25D2
Diameter (in.) 37 59 42 33
Length (in.) 115-3/4 140 192 173-1/2
Weight (lb) 1950 3800 4000 5000 2540
Power (hp) 2190 bhp 5000 bhp 4000 bhp 4000 est he 3300 shaft hp
+ 625 lb thrust + 1300 lb thrust at 35,000 ft at 35,000 t +980-lb thrust
Sea Level 0 mph Sea Level 0 mph at 400 mph at 500 mph Sea Level 0 mph
2780 bhp 5000 est hp at
+340 net 20,000 ft at
jet hp 500 mph
Sea Level 410 mph
RPM 13,000 7080 250 6000 12,400
Specific Fuel" 0.81 Sea Level 0.90 Sea Level
Consumption o mph o mph 0.45 at 0.50 at 0.615
(lb/bhp-hr) 0.61 Sea Level
410mph 0.58 at 20,000 ft 35,000 ft 35,000 ft
0.52 at 35,300 ft at 500 mph at 400 mph at 500 mph
410 mph
Mass Flow I 22 73 32 at 35,000 ft 40
(lb/sec) at 400 mph
-
Type
Compressor
14-Stage
Axial
2-Stage
Centrifugal
-Stage
xial
Axial I Axial

NOTE: These data furnished by Power Plant Laboratory, Wright Field, based on information available as of
2 November 1945.

18
 TABLE III - TURBOJET UNITS (American Types)
Manufacturer G. E. G.E. G.E. I Westing- Westing- Westing- Lockheed
house house house
---_. ..

Model J-35 J-33 i J-31 19XB-2A 9.5A 24C XJ-37-1

(TG-180) (1-40) i (1-16) (L-1000)
Diameter (in.) 38 50.5 41.5 25-5/8 9-1/2 32 x 38 25
Length (in.) 177 102-7/8 72 93-9/16 52-1/2 112-9/16 123.5
Weight (lb) 2400 1920 885 826 140 1075 1235
Thrust (lb) 4000 3750 1550 1600 275 3000 5100 Take-Off
(Mter-
burning)
RPM 7600 11,500 16,500 17,000 34,000 12,000 15,600
Specific Fuel 1.08 1.20 1.25 1.144 1.50 1.066 1.76 at
Consumption Sea Level
(lb/lb-thrust- . Take·Off
br) (After-
burning)
1.07 at 600
mph
Mass Flow (lb/sec) 73 73.3 32.8 29.0 6.5 52 34
Type Compressor . II-Stage 1 Double 1 Double 10-Stage 6-Stage 10-Stage 16-Stage
'Axial Entry . Entry Axial Axial Axial Inlet
Centrif- Centrif- 16-Stage Hi$h
ugal ugal Pressure AXlal

TABLE IV - TURBOJET UNITS (British Types)

Manufacturer de Havil- de Havil- de Havil- Armstrong Metropo!i-I Rolls Power Jets
land land land Siddeley tan Vicker Royce
Model Goblin III Goblin IV Ghost II ASX-1 (a) F2 RB-41 (b) W2/500
Diameter (in.) 49.5 49.5 53 42 33 49.5 41
Length (in.) 93.5 93.5 83 109 96.8
Weight (lb) 1530 i 1550 1950 1900 1500 1550 830
Thrust (lb) 33- I 38- 5000 2600 1800 4070 1650
3500 4000
RPM 10,500 11,500 10,000 8000 7500 12,000 16.500
Specific Fuel ' 1.15 1.08 i 1.08 0.96 11.04 1.03 1.14
Consumption
- (lb/lb thrust-hr) I
Mass Flow (lb/sec) i 82 50 47.5 80 34.7
No. Cumbustion 16 8 14 11 1 Annular 9 10 Reverse
Chamber Straight I Straight Straight Straight Straight Flow
'Through Through Through Through Through
Type Compressor Single • Single 14-Stage 9-Stage Double Double
Entry . Entry Axial Axial Entry Entry
CentriC- Centrif- Centrif- Centrif-
ugal ugal ugal ugal
Type Turbine l~Stage I-Stage 2-Stage 2-Stage I-Stage I-Stage
Pressure Ratio I 5.0 3.5 4.0 4.1
(a) ASX-1-Unit has air inlet between compressor and turbine. Air flows forward and turns 180 0 through com-
bustion chambers to turbine.
(b) RB-41-Run at 5000-lb thrust also, minimum Standard Fuel Consumption - 1.0 maximum - 1.05.
NOTE: These data furnished by Power Plant Laboratory, Wright Field, based on' information available as of
March 1945.
19
 TABLE V

TURBOJET UNITS TURBOPROP

.(German Types) (German Types)
-_ ....

Manufacturer BMW Junkers Heinkel Junkers BMW Daimler BMW

Hirth Benz
-
Model 109-003 109-004 109·011 109-012 109-018 021 028
Diameter (in.) 27 35 32 43 49 32 49.3
Length (in.) 124 152 99 194 165 138 228
-_ ....

Weight (lb) 1340 1640 1980 3520 5060 2560 7270

Thrust (lb) 1760 1980- 2860 6100 7500 Sea Level Sea Level
2200 o mph o mph
1950 hp + 4700 hp +
1100-lb 4850-lb
thrust thrust
Sea Level Sea Level
500 mph 500 mph
2750 est hp 12,600 est hp
RPM 9500 8700 10,200 6000
Specific Fuel 1.4 1.35 1.38 1.2 1.1-1.3 0.81 0.58-0.68
Consumption Ib/lb lb/lb lb/lb lb/lb lb/lb lb/hp-hr lb/hp-hr
thrust-hr thrust-hr thrust-hr thrust-hr thrust-hr
Type Compressor 7-stage 8-stage 1 Single l1-stage 12-stage 9-stage
Axial Axial Entry Axial Axial Axial
Centrif-
ugal
+ 4-stage
Axial
Type Turbine I-stage I-stage 2-stage 2-stage 3-stage

TABLE VI

TURBOJET UNITS MOTORJET

Oapanese Types) Oapanese Type)
Manufacturer Japanese Navy Ishakawa Jima Nakajima Mitsubishi Hitachi
Shibaura
Model NE20 NE 130 NE230 NE 330 TSU-11
Diameter (in.) 24.4 33.5 36 x 30 46.5 x 34.5 25.2
Length (in.) 106 152 135 157 86.5
Weight (lb) 990 1980 1920 2640 440
Thrust (lb) 1100 1980 1950 2860
RPM 11000 9000 8100 7600 9000
Specific Fuel 1.3 1.3 1.3 1.4
Consumption
(lb/lb-thrust-hr)
NOTE: This information obtained by interrogation in Japan.

20
-
~

TUR.BOP~OP
OR.
51 MPLE. GAS TUR.BlNE DRIVING A PROPELLER.

Figure r- TurlJoprop or Simple Gas Tu,w,.. Drivittg a Propellor

 GAS TURe'NE CYCl.E.
B C
r----~

·'-------~-D

SPE.CI f:IC VOLUME

,
I
I

\
\
\
\

,
I
I
I T-S DlAG2AM
I
I

Figure 2 - (S) Entropy

22
lNTAKE JET

COM&USTION
COMPRESSOR. TUR.8lN£

TUt2.BOFAN
Figure 9 - Turflolatt
 _ C>

O_OO.C»e»
0 0 0 0 0 0 _ .. 0
a • • a atGO 0 _ 0

Col
o

TURBOJET

Figure JO - Turboiet
 PART II

EXPERIMENTAL AND THEORETICAL PERFORMANCE

OF AEROPULSE ENGINES

By

HSUE-SHEN TSIEN
 PART II

EXPERIMENTAL AND THEORETICAL PERFORMANCE

OF AEROPULSE ENGINES
DECEMBER 1945

THE PRESENT STATUS OF AERO PULSE

The German aeropulse and the American copy of it for the flying bomb are
the first successful realizations of this type of power plant. The general dimensions
are given in Fig. 1. The air is sucked into the combustion chamber by the vacuum
created by exhaust of the previous cycle. The intake air passes the venturi where
gasoline is continuously injected. The explosion of the air-fuel mixture raises the
pressure in the combustion chamber to a high level and closes the spring valve at
the intake. The gas is thus forced to expand through the exhaust duct and is discharged
at high speed. This gives the propelling impulse. At the end of discharge, the inertia
of the gas creates a vacuum in the combustion chamber and the engine is ready to
start a new cycle again. The pressure in the combustion chamber is controlled by
the rate of fuel injection, and this, in turn, controls the discharge velocity of the gas,
and thus the propulsive thrust. To start the engine, a carefully adjusted amount of fuel
is sprayed into the cold combustion chamber, so as to create a mixture of correct
ratio around the spark plug. The spark plug ignites the mixture and the resulting
strong explosion starts the cycle. The flow in the combustion chamber and the dis-
charge duct is thus a pulsating one with very large amplitude, as shown by Figs. 2
and 3.

1. THE SPECIFIC FUEL CONSUMPTION.

The test results* on the German aeropulse are shown in Fig. 4, where the specific
consumption in lb/hr/lb of gross thrust is plotted against the flight Mach number of
the aircraft propelled by the power plant. The gross thrust is the total thrust of the
unit without deducting the drag of the duct. At a given Mach number there are a num-
ber of points, each of which represents a different fuel injection pressure or fuel flow
rate. As the fuel rate is increased, the pressure at the end of combustion is higher, and
this tends to increase the efficiency of conversion of the heat energy to kinetic energy.
* "Preliminary Testing of German Robot Bomb Engines," Memorandum Report No. TSEPL-
5-673·55, AAF (Nov 1944); "Twenty-Foot Wind-Tunnel Tests on the Jet Unit of a German
Robot Bomb," Memorandum_Report ENG.-51-6731-11, AAF (Aug 1944).

41
Therefore, it might be expected that the specific consumption will decrease with in-
crease in fuel rate. In other words, the thrust output of the unit increases faster than the
increase in fuel flow. This is actually borne out by the experiments, as the lower points
in Fig. 4, at a given Mach number, correspond to higher fuel rates. It is seen that as
flight Mach number or flight speed is increased, the specific consumption gradually
decreases. The trend is represented by the heavy curve in the figure. At stationary
conditions, the consumption is 5 lb/hr/lb-thrust. It decreases to 3.7 lb/hr/lb-thrust
at a flight Mach number of 0.6, or 450 mph.
To extrapolate the specific consumption curve to higher flight numbers, it is
necessary to have the guidance of theoretical calculation. Since the flow is nonsteady,
the complete anlysis should include the inertia effects, and the calculation will be very
complicated. To simplify the calculation the following assumptions will be made:
(a) The. pressure in the combustion chamber at the end of the charging process
or at· the beginning of combustion is one-half of the stagnation pressure of the air
stream, assuming isentropic compression.
(b) The combustion is carried out at constant volume and the combustion effi-
ciency is 95 percent.
(c) The discharge process is quasi steady. In other words, the flow at each in-
stant is assumed to be the same as that of a steady discharge through a Laval nozzle
of perfect design, with the chamber pressure prevailing at that instant.
(d) The specific heats of the air and combustion products are assumed to be con-
stants, but different values are used for air and for the combustion products.
(e) The mass flow of fuel is neglected with respect to the mass flow of air.
The basis for these assumptions will now be discussed.
The pressure in the combustion chamber at the end of the charging process de-
pends upon two factors: the diffusor efficiency of the duct, and the throttling effect
of the venturi with fuel jets. Since the duct is very short, most of the compression must
occur outside of the duct with diverging streamlines. Therefore, aside from the
possible effect of shock wave at supersonic flight velocities, the pressure in front of
the spring valve should be very close to the st~gnation pressure with isentropic com-
pression. On the other hand, the throttling effect of the spring valve and venturi is
very strong. Since the charging process is very fast, the flow velocity through the valve
and venturi must approach that of sound. The pressure at the valve and venturi must
be very close to one-half of the pressure in front of the spring valve. Due to the poor
aerodynamic shape of the valves and the venturi, necessitated by the requirements for
mixing of fuel and air, the pressure recovery must be inefficient. A rough approxi-
mation would be to assume no recovery. Thus the pressure in the combustion chamber
at the end of the charging process is one-half of the stagnation pressure.
Since the combustion is explosive and occurs in very short intervals of time, the
expansion of the product during combustion is negligible. Therefore, the assump-
tion of constant volume should be a satisfactory approximation. The actual com-
bustion process is quite complicated: During the expansion and dicsharge process,
the gasoline injected into the combustion chamber is vaporized by the walls and the
hot gas remaining in the chamber. The result is an over-rich mixture, which is pre-
45
sumably partially oxidized, i.e., the large hydrocarbon molecules are partially broker.
down to more active elements. When the discharge process is completed, the low
pressure then prevailing in the chamber will open the spring valve, and fresh air in
strong eddies created by the venturi will be admitted to the chamber. The mixing of
the fresh air with the partially oxidized and activated hydrocarbon vapor immediately
starts the rapid combustion with a rise in pressure. Of course, if detonation occurs,
much higher local pressures can be attained than by smooth combustion, as assumed
here. However, actual measurements do not indicate the occurrence of detonation.
Thus the assumption of constant volume combustion should be satisfactory.
The assumed discharge process really corresponds to that of a slow discharge
through a small nozzle fitted to a large chamber. In the slow discharge, the inertia
effect is small and can be neglected. Then the quasisteady process is an exact repre-
sentation of the physical phenomenon. The assumption adopted is thus to approxi-
mate the impulse of a rapid discharge process by the impulse of a slow discharge
process. This approximation should be satisfactory, as generally the impulse of a
jet propulsion device is little influenced by the rate of the process.
The last two assumptions of constant specific heats and constant mass flow are
generally adopted and are· believed to have little influence on the results.
The results are shown in Fig. 4. The curves correspond to different values of the
pressure ratio after combustion and before combustion. For a given pressure ratio,
the specific consumption first increases with the flight speed, but then reaches a maxi-
mum value and decreases with further increase in flight speed. The larger pressure
ratios give lower fuel consumption than expected. However, in order to explain the
experimental data, it seems that the maximum pressure ratio that can actually be ob-
tained in the German aeropulse is a function of the forward speed. At lower speed,
this ratio is smaller than four. This is in accordance with the pressure record shown in
Figs. 2 and 3. However, as the speed increases, the combustion can be pushed to higher
limits and the pressure ratio is increased, giving a decrease in the actual fuel consump-
tion. On the other hand, with gasoline as the fuel, the maximum com bustion tempera-
ture is roughly 5000°F. In other words, the pressure ratio cannot increase indefinitely,
but must be limited to a value lower than nine. Hence, the theoretical specific con-
sumption corresponding to the pressure ratio nine forms the lower limit of the con-
sumption of the German aeropulse. This is the basis for the heavy curve in Fig. 4,
representing the probable trend of consumption with increase in flight velocity. Ap-
proximately then, the specilic consumption of the German aeropulse is not likely to
be lower than 3 lb/hr/lb-thrust at supersonic speeds.

2. FREQUENCY OF PULSATION AND THRUST

The simple theory of aero pulse outlined in the preceding section, being based
upon a quasisteady discharge process, fails completely in predicting the frequency
which is directly associated with the non steady pulsating flow. In order to have an
estimate of the frequency, another very simple picture of operation will be adopted.
This is the assumption of very small pressure ratio, i.e., very small pressure amplitude.
Then the frequency can be calculated by the elementary considerations applicable to
small pressure amplitudes, i.e., the sound vibration in a pipe. Since experiments (Ref.

47
cited on page 41) show that the frequency of the aeropulse varies very little with
changes in the fuel rate or the pressure ratio; the frequency so calculated for very small
pressure amplitudes should be representative for that of large pressure amplitudes.
Consider the aeropulse as an organ pipe closed at the end where the spring valve
is located and open at the other end. Then the pulsation in the pipe can be considered
as a quarter wavelength oscillation with a maximum pressure amplitude at the closed
end and a zero pressure amplitude, but maximum velocity amplitude at the open end,
as shown by Fig. 5. If a* is the velocity of sound propagation and L the length of the
pipe, the frequency f is
a*
f = 4L cps. (1)

~
I
I
•
I
I

PRE.SSUR.E. 'DlSTR'IUT'ON AT 8EGlNNING

OF COMBUST\ON

Figure 5
For the application to the aeropulse, a* should correspond to the mean con-
ditions of the flow in the duct. It is estimated that the German aero pulse with no
forward velocity should have a frequency of f = 50.1 cps. This checks very well with
the measured value of 46 cps. Most of the difference is perhaps accounted for by the
combustion period, which, although short, is not negligible. At higher fuel rates,
the combustion period is lengthened. This tends to lower the frequency. Experiments
48
show that the frequency at high fuel rates is actually slightly lower, in spite of the
higher sound velocity at higher temperatures, due to the richer mixture.
The average thrust of the aeropulse depends on the rate of airflow and the mixture
ratio. The rate per second at which the air is taken into the duct is, in turn, dependent
upon the charge per cycle and the number of cycles per second. However, all three
factors are interrelated. For instance, if the number of cycles per second is very large.
the charge pressure in the chamber will be low due to the necessary rapid acceleration
to push the air into the chamber. In other words, the charge per cycle tends to decrease
as the frequency is increased. Furthermore, the combustion process also tends to
limit the mixture ratio that could be effectively used to lower values if the frequency
were increased as the combustion time grows shorter. Thus the average thrust of the
aeropulse, being an increasing function of the product of all three factors, has a maxi-
mum with respect to the rate of fuel injection and with respect to the frequency, which
depends on the length of the tail pipe. The prediction of the thrust of an aero pulse
is thus rather complicated. The situation is very similar to the case of the reciprocat-
ing gasoline engine. The calculation of horsepower output of the reciprocating engine
is generally based upon empirical data, as the effects of volumetric efficiency or breath-
ing capacity and combustion conditions are very difficult to calculate.
The experimental data available at present on the thrust output of the German
aero pulse is plotted in Fig. 6. The gross thrust is presented in two forms, C F and K F•
C F is the ratio of the maximum gross thrust to the product of the dynamic pressure
and the sectio~al area of the combustion chamber or the frontal area. KF is the ratio
of the maximum gross thrust to the product of the atmospheric pressure and the
sectional area of the combustion chamber. Thus
C = gross thrust
(2)
p 1/2 air density x (velocity)2 x frontal area
K = gross threat
(3)
P atmospheric pressure x frontal area
Both quantities are plotted against the flight Mach number and are related by the
formula,
. 2
Cp = Kp OMo2 (4)

where 0 is the ratio of specific heats for air and Mo is the flight Mach number. Due
to the fact that the aero pulse has a static thrust, the value of KF is not zero at zero
velocity. Therefore Cp become infinite when Mo = O.
As stated previously, the prediction of the thrust of aeropulse is very difficult.
Therefore, to estimate the thrust coefficients at higher flight velocities, two assumptions
have to be made:
(a) The frequency of the unit remains the same as at low speed.
(b) The maximum thrust operating conditions follow the heavy line in Fig. 4.
The first assumption agrees with the observational data available and leads to the
conclusion that the air mass flow through the aero pulse is proportional to the density
of the air in the combustion chamber at the end of the charging process. The second
assumption seems justified by the test data, as explained previously.
49
 With these assumptions, the thrust coefficient curves can be extrapolated as shown
in Fig. 6. It is seen that for the German aeropulse at supersonic speed, the value of the
thrust coefficient Cp is almost constant and remains at about 0.70.

POSSIBLE IMPROYEMENTS OF THE

EXISTING FORM OF AEROPULSE

To increase the thrust and to reduce the fuel consumption of the present-day aero-
pulse as described in the previous paragraphs, either the air flow or the combustion
pressure should be increased and the external drag of the duct reduced. To increase the
air flow, the effective flow area into the combustion chamber must be enlarged. The
resultant reduction in the throttling action of the grid valve will also raise the pressure
in the combustion chamber at the end of the charging process and hence give higher
combustion pressure and better fuel economy. For instance, by removing part of the
rib* in the grid of the valve, the static thrust can be increased from 660 to 880 lb.
The specific consumption can be similarly reduced from 3 lb/hr/lb of gross thrust to
2.8 lb/hr/lb gross thrust. To reduce the external drag, the unit can be mounted in-
ternally in the fuselage. But this is not possible with the present construction material,
due to the overheating caused by the lack of cooling air flow.
If air alone could be introduced into the duct behind the explosive air-fuel mixture,
then when the mixture is burned, it would act as a piston to push out the air column.
The total air mass per cycle .is thus increased, with resultant larger momentum and
better efficiency. The air and air-fuel mixture must be separated for two reasons: (1)
A very lean mixture will not burn properly. (2) Even if the mixture burns properly,
the explosion pressure would be too low for effective energy utilization. This form of
augmentation really already occurs in the present aeropulse to a certain degree, as
during the charging time some atmospheric air flows into the duct through the rear
opening. This part of the charge contains no gasoline (which is only injected into the
combustion chamber) and serves as the air for augmentation. For air to enter the duct
through the rear opening, the direction of the flow has to be reversed, and this be-
comes more difficult as the flight speed is increased. Therefore, this natural augmen-
tation effect decreases at higher velocities. To remedy this, it has been suggested that
two ducts be used, mounted one in front of the other, as shown in Fig. 7. Then during
the suction or charging period, the air flows into the se'cond grid B and fills the second
duct with augmentation charge. However, this idea has yet to be exhaustively tested.
The combustion pressure can also be raised by using a fuel which tends to deton-
ate. Of course, for such fuels the injection has to be timed properly instead of using
the present continuous uncontrolled flow. This may lead to the further necessity of
using a timed spark plug to ignite the mixture. Whether there would be a large enough
increase in performance by this method to justify the complications, only experiments
can determine. The weight of the present aeropulse is about 300 lb, for a thrust of
* This was done by Eisla and G. Dietrich of DFS at Ainring, Germany, during 1944.

51
r .....

- - - ~
- . -
R
ij

..
Air Velocity B

Figure 7

70 lb. With better materials and method of construction, this value could be re-
duced. This is another possible direction of development.

VALVELESS AEROPULSE_

An obvious improvement for the present aeropulse is the complete elimination of

the spring valve and the venturi. If the valve and venturi can be discarded, then the
strong throttling effect can be avoided. The result would be an increase in the chamber
pressure at the end of the charging process, to the stagnation pressure, or twice the
present value. This would give a higher density of ail' in the chamber, and hence larger
mass flow and higher thrust. Furthermore, the pressure at the end of combustion for a
given temperature rise will also be increased and thus more efficient expansion and
lower specific consumption will be attained. Of course, at low flight speeds a valveless
aeropulse cannot be very efficient. due to the partial expansion of the combustion prod-
ucts through the front opening and the loss of impulse or thrust of the unit. At super-
sonic flight velocities. such losses can be eliminated, as the explosion pressure travels
with a defini~e speed higher than that of sound. If high flow velocity can be main-
tained in the duct from front to real', the explosion pressure can never travel to the
front part of the duct. Therefore, the expansion of the hot gas will be directed only
toward the real'. In other words, the inertia of the high-speed ail' flow in the front part
of the duct acts as the valve, without the necessity of a mechanical valve. In fact, with
propel' design of the duct, this possibility exists even at high enough subsonic speeds. *

1. IMPROVEMENT IN SPECIFIC FUEL CONSUMPTION

By using the same simplifying assumptions as before, except that the pressure
in the combustion chamber at the end of charging process is now equal to the stag-
nation pressure, the specific fuel consumption can be easily calculated. The result is
shown .n Fig. 4. It is shown that here an increase in the combustion pressure ratio,
i.e., the ratio of pressure before combustion and after combustion, tends to increase
* "Note on Valveless Aero- and Hydro-Pulse Motors," AMP Memo 137.1M, AMG·NYU No.
120, Applied Mathematics Group, New York University, May 1945.

52
the specific fuel consumption. However, the difference for a rather large change from
pressure ratio nine to pressure ratio six is quite small. The parts of the curves for flight
Mach number less than unity are drawn in broken lines, to indicate that the simple
theory does not correspond to a real situation, due to the neglect of expansion and
escape of the combustion products from the front opening of the duct. Thus, in general,
the idealized valveless aeropulse will give a specific consumption equal to 2 lb/hr/lb
of gross thrust in the supersonic region. Actually, of course, minor losses are in-
evitable and a valveless aeropulse will have a specific consumption somewhat higher
than this, i.e., between 2 and 3 lb/hr/lb-thrust. The higher value is the specific con-
sumption of the valved type, for example the German aeropulse for the flying bomb.

2. INCREASE IN THRUST COEFFICIENT

For the valveless aeropulse the throttling effect of the spring valve and the venturi
is eliminated and the pressure of air in the combustion chamber at the beginning of
combustion is twice as high as for the valved type. Since the temperature of the air is
the same as in the previous case, the density of air in the chamber is also twice as large
as for the valved type. Thus the mass of air flowing through a duct of given size will
be doubled. If the heat added per unit mass of air is made the same by using the same
combustion pressure ratio, the total heat added will be twice as large. On the other
hand, the specific consumption is lowered by a factor of two-thirds, as stated in the
preceding paragraph. Therefore, the thurst produced for a given size of the duct will
be three times as large as the valved type. Actually, if losses occur, the difference will
not be as large. However, it seems that a Cp as high as 1.0 for a valveless aeropulse
in supersonic flight can be considered reasonable.

CONCLUDING REMARKS

In the foregoing sections, the possibilities of further improving the aeropulse

are discussed. It seems that with these developments a lightweight aeropulse could
be made, with a weight as low as 0.16 lb/lb gross thrust at sea level, especially at
high speeds. The specific consumption based upon the gross thrust could be lowered
to about 2 lb/hr/lb-thrust. However, to carry out these investigations, a supersonic
wind tunnel of large enough test section so that a complete aeropulse can be tested
with burning, is a necessity. The strongly pulsating flow around and inside an aero-
pulse indicates the close relation between the inside flow through the duct and the
outside flow around the duct. Therefore, only a complete model test with the model
submerged in an air stream can reproduce the correct flow conditions. Static tests,
or tests with air flow through the engine, would be misleading and unreliable.

53
 APPENDIX I

A Simple Theory for the Aeropulse

Let the subscript 0 denote quantities cottesponding to the free atmospheric condi-
tions, the subscript 1 denote the stagnation conditions when the spring valves are
dosed, the subscript 2 denote the conditions in the combustion ,chamber at the end
of the charging process, and the subscript 3 denote the conditions at the end of com-
bustion. Since the compression from free stream to the stagnation pressure is assumed
to be isentropic,

(5)

and the temperature ratio is

Tl 0-1 M2
- = 1 + - (6)
To 2 0

According to the assumption stated in the main text,

1
P2 = - PI (7)
2

The temperature Th being a representation of the total energy of the gas at rest,
must be the same as Tl, since no appreciable heat loss can occur.
T2 = Tl (8)
Since the combustion is assumed to be carried out at constant volume, the heat
added h per unit mass

h = C: (T3 - T2) = ~ C~ T. (1 - ~:). (9)

The primed quantities refer to the combustion products. Since the combustion is
carried out at constant volume

(10)

Therefore,

h = .!.. C'
0' p
T I (1 - P2PI). (11)

To calculate the discharge process, the quasi steady How is assumed. Thus if v
is the discharge velocity corresponding to p. then

v = r~ .P. [1 _(Po) 0'; 1] (12)

0'-1 P P

54
The impulse due to a discharge dm at this velocity is
dI = vdm (13)
Let m be the mass before the removal of dm, the ratio of density in the chamber after
the removal of dm and that before removal is then
m-dm
m
(14)

Similarly, the pressure ratio is

P + dp
(15)
P
Since the process in the combustion chamber is isentropic, we have

P +p dp (m -:ndm) 0 1
(16)

Thus
0' dm = _ dp (17)
m P
Now m = p V, where V is the yolume of the combustion chamber.
Therefore

... 1
dI
1 20' .f. [1
0'-1 P
_(Po)
P
-V]
0'-1
1 PV d P
0' P
(18)

To find the total impulse due to the discharge, we have to integrate dI for pressure
variations from the initial pressure pa to the final Po' Therefore

(19)

where '1 = p/pa.

But VPa is the total mass in the combustion chamber at the
2
beginning of discharge, and (0' ::)
1/ is the velocity of sound, Aa, correspond-
ing to the conditions in the combustion chamber at the end of combustion. Thus
if ve, denotes the "effective exit velocity," then

~ f, ~0/:1 JI~O'~1 [1 _(Po) 010~1

= /-1] d'1 (20)
Po/ Pa '1 0' pa '1 0'

For the case of 0 = 4/3, the integral is of the closed form

~ = 6..f6 ...1/1 _ (Po) ~ [1 _.!. (Po)~ .! (Po) f - ~ (Po)~] (21)

as 7 ps 5 pa 15 pa 15 ~
If the average mass flow is m, then the thrust is m (ve - vo ). Let H be the heat
value of the fuel per lb and '1b the combustion efficiency, then
s (22)

55
After some reduction, this equation can be written as
as ( P2)
s =' 3600 ;: 1 - Ps (23)
778H'tlb 6'(6'-1) !!' _ Mo (ao)
as as
In this formula the ratio of sound velocity is given by

(
as)2 == g ~'-1 (~) (Tl) (24)
ao Cp 0-1 P2 To
Fig. 4 is calculated on the following assumed values for the constants:
C~ = 0.276 BTU/lb;oF
Cp = 0.243 BTU/lb;oF (25)
6 = 1.405, 6' = 4/3
H == 18,700 BTU/lb
and
'tlb == 95%.

56
 APPENDIX II

Frequency oj Aeropulse

For the application to aeropulse. a* must correspond to the mean conditions

of the flow in the duct. Since the temperature in the duct at the end of expansion is
0'-1
(26)
Ta (~:) I I
the mean value u for the velocity of sound is
0'-1
1 o
u = 2 a ao (aa) [1 + (Po)
pa
-20'] (27)

For instance, if pa/p2 = 4, Mo = 0, then pa/pz = 2 (28)

with 0' 4/ a , h = 1.849 a o • The German aeropulse has an effective length, L =
10.3 ft, and for standard sea level conditions. ao = 1/20 ft/sec. Thus
1.849 x 1120
f = 4 = 50.1 cps. (29)
x 10.3

57
 APPENDIX III

Thrust oj an Aeropulse at High Flight Speeds

To extrapolate the thrust for higher flight velocities, it is assumed that the fre-
quency of the unit will remain approximately constant. Then the volume flow of air
will also be constant. However, the density of the air is proportional to
1
0-1
(30)
2) 0'-1
( 1 + -2 M0

Thus the air mass flow is proportional to the same factor. On the other hand, the heat
added per unit mass of air is proportional to the temperature difference T3-T 2, or pro-
portional to

(31)

Hence, the total heat added is proportional to

0'
(1 + 0~1 M~t'-l (:: 1) (32)

From the tests on the German aero pulse, it is found that the fuel rate for maximum
thrust at Mo = 0.515 is 2890 lb/hr. From Fig. 4, the corrsponding pressure ratio
pa/p2 is 5.9. Hence,
Pa _ 1
1 + 0.2025MJ) 3.469 ~ (33)
fuel rate at Mo = 2890 ( 1.0537 4.9

The thrust is then

pa _ 1
1 + 0.2025M2) 3 • 469 ~ ~ (34)
t h rust at M 0 = 2890 ( 1.0537 0 4 ·9 S

where s is the specific consumption in lb/hr/lb-thrust taken from the heavy curve
in Fig. 4. The value of pa / p2 can also be taken from the same figure.

58
 PART III

PERFORMANCE OF RAMJETS AND THEIR

DESIGN PROBLEMS

By

HSUE-SHEN TSIEN
 PART III

PERFORMANCE OF RAMJETS AND TH EIR

DESIGN PROBLEMS

DECEMBER, 1945

INTRODUCTION

The ramjet (Fig. 1) consists of a diffusor to decelerate and to compress the high
velocity air stream due to the rapid forward motion, a combustion chamber for add-
ing heat to the air mass, and a nozzle to discharge the combustion products at high
velocity. The pressure in the combustion chamber, being obtained by ram compression
only, is necessarily lower than that of the aeropulse moving at the same speed. In fact,
at static conditions, the pressure in the combustion chamber of a ramjet is equal to the
atmospheric pressure and no thrust can be produced. On the other hand, the continu-
ous operation of the ramjet leads to a more efficient utilization of the combustion
pressure. Furthermore, the inherent extreme simplicity of the power plant, as com-
pared even with the aeropuise, is very attractive from the engineering point of view.
Compared with the turbojet, the specific fuel consumption of the ramjet is higher,
especially at small flight velocities. The present available information seems to indi-
cate that for flight velocities corresponding to a Mach number greater than 2.5, or
for velocities over 1800 mph, the specific fuel consumption of the ramjet becomes com-
parable with that of the turbojet. Then the extreme simplicity and the lightweight of
the ramjet would indicate that it is the favored power plant for such very high flight
speeds, provided the air density is high enough to give a satisfactorily large thrust.
For these reasons, the interest in the ramjet at present is wide-spread. However, the
experimental research on this power plant is only at its beginning. The only available
complete data are those obtained for the hydrogen-burning Focke-Wulf ramjet (Ref.
1). The Bumblebee project of the Bureau of Ordnance, U. S. Navy, has demonstrated
the feasibility of a supersonic ramjet with a measured acceleration of the missile
equal to 1 g at flight Mach number 1.5. However, no detailed performance data are
available.
Historically, the concept of the ramjet as a power plant for locomotion is cer-
tainly not new. Due to its inherent simplicity, the ramjet was "invented" many times
during the past decades. However, a doser examination always reveals the high fuel
consumption and the inefficiency at low flight speeds in comparison with the con-

61
 ; -/ FLAME HOLDERS

1-4---- DIFFUSER ------i~- COM8USrJON CHAMBER -+j"'---i.N~Q~tZ.~Z:f'L!:E

SECTION THROUGH RAMJET

...figure J- SediOft Through Ram;et

ventional power plant such as the aircraft reciprocating engine. This is perhaps ex-
pected, as the ramjet is truly an aerodynamic engine depending upon the motion of.
the aircraft for efficient operation. The conventional reciprocating engine, however,
is an adaptation of the stationary power plant which can operate independent of the
forward motion. As the velocity of the aircraft increases, the inherent advantages of
the aerodynamic power plant, the ramjet, become more evident.
During the last years of World War II, German engineers and technicians devoted
a large fraction of their effort to ramjet research. Due to the lack of liquid fuel, they
tried to burn coal in the form of slabs loaded directly in the combustion chamber.
This is rather incon venient, especially for long duration of operation. The Japanese
started the investigation of ramjets as early as 1937. However, no fruitful result was
obtained.
It is the purpose of this report to examine the general performance of the ramjet
and to estimate the probable speed range for its efficient application. The discussion
of performance will be based essentially on the theoretical analyses of the staff ofthe
Jet-Propulsion Laboratory of the California Institute of Technology (Ref. 2) and by
the staff of the Bureau of Aeronautics, U. S. Navy (Ref. 3). The report will conclude
with a discussion of the problems of research and development, particularly the prob-
lem of combustion-chamber design.

BASIS OF THE THEORETICAL ANALYSIS

Since the ramjet is essentially a combination of diffusor, combustion chamber,
and nozzle, a correct theoretical analysis should be based upon the best available
information on the performance of each of these components. The essential data
required are the diffusor efficiency, the combustion efficiency, the pressure drop across
the combustion chamber, and the nozzle efficiency. The following is a critical evalua-
tion of this information, and thus forms a means of judging the reliability and the
accuracy of the results of the theoretical analysis.

1. Dilfusor.
The deviation from isentropic compression in the diffusor is caused by the loss
through skin friction on the wall of the diffusor, by the eddying dissipation due to
boundary layer separation, and by the irreversible process of shock wave in supersonic
flows. In subsonic flows, the shock loss does not occur and the losses are due to the
presence of the walls of the diffusor only. Therefore, as far as the compression of the
air stream through the ramjet is concerned, it is better to dispense with the diffusor
completely alld obtain the pressure rise by diverging streamlines ahead of the duct.
This is the so-called "outside compression" as indicated in Fig. 2. However, the diverg-
ing stream ahead of the duct requires very large fairings on the outside surfaces of
the duct to reduce the external drag. But even then the external drag is considerably
increased. Therefore, an optimum design requires the correct compromise between
the complete inside compression with a long diffusor and the complete outside com-
pression without the diffusor. Test results of NACA (Ref. 4) show that with proper
design the diffusor efficiency in subsonic flows, or the ratio of the actual pressure in-
crease in the diffusor to the theoretical pressure increase in isentropic compression,
can be made to be 85 % or higher. This is the basis of Ref. 3.

63
 Flow
Into
Ramjet

Figure 2

For supersonic flows, the matter is much more complicated due to the presence
of the shock waves. If the flow through the duct is small compared with the size of the
duct, then the flow velocity in the duct must be low and the shock wave is forced to
the front of the diffusor opening. For this case then, the pressure rise can be calculated
by assuming a normal shock from the supersonic flight velocities to subsonic veloci-
ty. After the shock, the flow is subsonic and the subsonic test data can be applied, i.e.,
85% diffusor efficiency for flow behind the shock. The over-all efficiency includ-
ing the shock is, however, much lower. At Mach number 3, the efficiency is only 30%.
This is the basis of Ref. 2.
However, this is really an oversimplification, as the shock wave does not always
occur ahead of the diffusor opening. If the flow through the duct is large, the shock
wave is generally oblique to the stream direction and occurs within the diffusor. Thus,
the situation is not unlike that of the diffusor after the test section of a supersonic
wind tunnel. However, for the latter case, there is a rather thick boundary layer at the
beginning of the diffusor while for the diffusor of the ramjet, there is no such boundary
layer at the entrance. It is well known that the effect of boundary layer is to cause
a premature occurrence of shock wave and thus is detrimental to the compression
process. If the shock in the entrance diffusor is weaker than that in the wind-tunnel
diffusor, then much better over-all diffusor efficiency than that used in Ref. 2 can be
achieved. The effect is shown very clearly by the recent tests of A. Kantrowitz and C.
duP. Donaldson (Ref. 5). These tests demonstrated that the efficient supersonic diffu-
sor should be designed with a contracting entrance section before the final expanding
section. Then the shock wave will occur in the neighborhood of the throat. This means
that the velocity of the air stream is considerably reduced by the contracting sec-
tion before the shock wave appears. Since the loss through a shock wave decreases
if the velocity in front of the shock is reduced, the efficiency of such a diffusor is con-
siderably higher than the diffusor with a normal shock ahead of the entrance as
assumed in Ref. 2. Thus the diffusor compression efficiency used in Ref. 2 is on the
conservative side.

64
 Another method of improving the diffusor efficiency is to design the entrance so
that a series of oblique shock waves form at the entrance section to reduce the flow
velocity before entering the expanding section of the diffusor. Since the loss across
an oblique shock is considerably less than that across a normal shock, the compression
efficiency will be raised. K. Oswatitsch (Ref. 6) has designed such a diffusor for Mach
number 2.9 witb a total pressure recovery of 60% (Fig. 3). This is higher than the

Divergent Section

Oblique Shock
111" now..

Figure 3 - o.wafifsclt l)i"ulOl'

compression efficiency used in Ref. 2. The difference or improvement by these new

diffusor designs is especially large at high Mach numbers. Thus the supersonic per-
formance of a ramjet as estimated in Ref. 2 is perhaps much lower than the maximum
obtainable.

2. The Combustion Chamber.

Due to the relatively low pressure in the ramjet combustion chamber, the thrust
produced will be small unless the temperature after the combustion is raised to a high
value. Furthermore, in the effort to reduce the external drag the sectional area of the
combustion chamber is also reduced with a resultant high-flow velocity through the
combustion chamber. These two factors of high velocity and large heat addition
distinguish the ramjet combustion chamber from the turbojet combustion chamber.
At present very little is known about the performance of such combustion cham-
bers as the tests are only beginning. However, one thing is certain: The static pressure
drop in the combustion chamber is a function of the degree of heating. For very large
heat addition, the pressure drop is many times that of small heat addition as in the
turbojet combustion chambel:i. The physical reason is the expansion of the gas due
to high temperature. Then the velocity of flow must be greatly increased according to

65
the increased volume. The increase of velocity, however, can be obtained only by a
large accelerating force corresponding to a large pressure drop.
The calculation in Ref. 2 takes this fact into account. The additional pressure
drop due to the eddying motion created by burning and baBIes is taken from prelimin-
ary test data of combustion research at the Massachusetts Institute of Technology.
In the first repon of Ref. 2, this additional pressure drop is assumed to be four times
the dynamic pressure at the entrance to the combustion chamber. In the light of later
experiments, this drop is reduced to one dynamic pressure in the second report of
Ref. 2. However, the best design of ramjet combustion chamber now tested indicates
that a half dynamic pressure drop is possible. Thus the calculation of Ref. 2 may be
considered as somewhat conservative. On the other hand, the calculation of Ref. 3
isbased upon a test on turbojet combustion chambers with low heat addition. There-
~re; for large heat addition, the pressure drop must be considerably larger than the
assumed value. For this reason, the value used in Ref. 3 is somewhat optimistic.
3. Noule.
Due to the non uniformity of the flow in the cross section of the nozzle and due to
skin friction, the expansion in the nozzle is not isentropic. Both calculations in Ref.
2 and in Ref. 3 use a nozzle efficiency of 95 %. * This value is quite reasonable and is
in agreement with the results of tests on rocket discharge nozzles.
4. Over--all Thrust Calculation.
In both Ref. 2 and Ref. 3, the strong interaction in subsonic flow between the jet
and the flow over the outside surface of the duct is not considered. Due to the frictional
forces or the forces of mon..entum transport between the high-speed jet and the sur-
rounding air, the velocity over the outside surface of the duct is considerably increased.
In other words, there is an effective ejector action of the jet. The net result on the per-
formance of the ,ramjet is not unlike that of the thrust augmentor for the rockets. The
theory of thrust augmentation is based upon the mixing of the jet and the surrounding
air stream. At low velocities, the increase in thrust of the rocket system due to the
augmentation can be as large as 60-80%. For the subsonic ramjet, similar augmenta-
tion exists with the resultant considerable increase in effective thrust. Physically this
thrust increase comes from the stronger suction force over the leading edge of the
diffusor due to increased flow velocity. The neglection of this effect in both theoretical
calculations of Refs. 2 and 3 makes the results conservative. This should be kept in
mind for comparing theoretical and experimental data.

CALCULATED PERFORMANCE
The calculated performance of the ramjet as obtained in Refs. 2 and 3 will now
be discussed. The results of the theoretical analyses are generally confirmed by the
available test data (see Appendix).

1. SpecHic Fuel Consumption.

Since the influence of the altitude on the specific fuel consumption is small, only
the result at sea-level conditions need be examined. Fig. 4 shows the comparison of
... Tbe definitions oj tbe nozzle efficiency Jor tbe two ca/cillations differ. However, tbe difference in tbe
final resNIl ;s 1'101 large.

66
 •
the two calculations. It is seen that due to the rather optimistic assumptions on the
pressure drop of Ref. 3, the specific fuel consumption calculated is lower than that
of the more conservative computations of Ref. 2. However, from the trend of the curves,
this difference only occurs in subsonic flight velocities. For supersonic flight veloci-
ties, both calculations will give a specific fuel consumption of three pounds per hour
per pound of gross thrust. Of course, strictly speaking, one must be careful in compar-
ing the results, as the assumptions regarding the combustion tempeutures are not the
same. Ref. 2 assumes a constant static temperature at the end of combustion chamber
equal to 3000°F while for Ref. 3 the ratio 1" of the total temperature at the end of com-
bustion and the atmospheric temperature is kept at a constant value for a given curve.
However, the difference due to variation in T is small and will not appreciably modi-
fy the general conclusion as stated previously.

2. Tltrusf Coefficient.
The results of calculation for sea-level conditions are again compared in Fig. 5
where the gross thrust coefficient is plotted against the flight Mach number. The thrust
coefficient is defined as the ratio of thrust to the product of dynamic pressure of the
free stream and the cross-sectional area of the combustion chamber. Ref. 2 assumes a
constant static temperature. Therefore, more energy can be added to the gas stream
as the flight velocity increases, since this additional energy is transformed directly
into velocity at the exit of the combustion chamber instead of into temperature rise.
The increase in the energy added is reflected in the rapid rise in thrust coefficient.
However, this rapid rise is slowed down and even changes to a decrease at higher
flight Mach numbers due to the choking of the combustion chamber. The choking
condition is the limiting condition for stable combustion, and limits the amount of
fuel that can be efficiently burned. Of course, this condition can be delayed by de-
creasing the inlet velocity to the combustion chamber. However, this would require
. an enlargement of the combustion chamber and the increase in the thrust coefficient
will be small.
For subsonic velocities, the result of Ref. 2 should be compared with that of Ref.
3 for 1" = 6.6. Then the maximum temperature will be approximately the same.
It is seen that the thrust coefficients of the two computations are nearly equal.
From Fig. 5, it is seen that the highest thrust coefficient is obtained in the range
of flight Mach numbers from 1 to 2.6 for a maximum temperature of 3000~. How-
ever, as stated in "Basis of the Theoretical Analysis," page 63, the assumed diffusor
efficiency for flight Mach numbers greater than 2.0 is probably too low. Therefore,
the possible performance of the ramjet at high Mach numbers will be much better
than indicated by Figs. 4 and 5. In fact for Mach numbers greater than three, one has
good reasons to believe that the specific fuel consumption will be close to two pounds
per hour per pound thrust and the thrust coefficient will be approximately O.B. It is
then clear that the ramjet is essentially a propulsive power plant for extremely high
speeds. In other words, it is an ideal power plant for guided missiles with speeds
over 2000 mph if the trajectory of the missile is not too high for supplying enough
atmospheric oxygen to support combustion.

68
WEIGHT OF A RAMJET
Since no actual ramjet has yet been built, it is difficult to estimate the weight
of such power plants. However, due to the extreme inherent simplicity of the unit.
the weight per pound of thrust should be lower than that of the aeropulse.

BOOSTING OF THE RAMJET AT LOW FORWARD SPEEDS

As stated previously, the static thrust of a ramjet is zero. In other words, the ram-
jet is not self-statting. Of course, the launching can be carried out by rockets until
the operating speed of the ramjet is reached. A more efficient alternative is, perhaps,
the ducted rocket. In this scheme. the rocket motor is installed inside the ramjet duct.
The rocket jet is used to induce an air How through the duct even at very low forward
speed. Additional fuel is injected through the ramjet injector and the resultant com-
bustion could give a large thrust. A preliminary calculation shows that at low forward
speeds the thrust of the ducted rocket could be several times the value for the rocket
alone. The corresponding specific consumption of both propellant and fuel is approxi-
mately six pound per hour per pound thrust. Therefore, the system is worth further
analytical and experimental studies.

RESEARCH AND DEVELOPMENT ON THE RAMJET

As stated previously, the performance study of ramjets is yet in the very prelimin-
ary stage. Complete experimentation should cover a wide range of Hight Mach numbers
and designs for the diffusor, the combustion chamber, and the nozzle. Of course, the
possibilities of boosting the thrust at low speeds should also be carefully: studied. Two
further specific problems of ramjet development can be stated.

1. AeroJynamies of Power Plonf 'nsfollation.

Aside from the question of optimum design for the diffusor and nozzle, the in-
stallation of the power plant in an aircraft leads to many problems. This is true es-
pecially for the case of compressorless thermal jets which require an extremely large
ratio of How of air and thus relatively large size of the power plant compared with
other components of the aircraft. Specifically, the following items should be studied:
(a) the interference between the engine duct and the lifting surfaces; (b) the inter-
action of the hot gas jet and the How around other components of the aircraft; (c) the
possibility of combining the engine duct with the body of the aircraft, i.e., "internal
installation;" and (d) the ducting of high-speed air stream from the entrance to the
combustion chamber in an internal installation.

Such investigation requires a supersonic wind tunnel allowing the burning of

fuels in the engine. This can be accomplished by an open-circuit tunnel or an air-
exchange system. The size of the wind tunnel should be large enough to allow the
use of large enough models for accurate reproduction of the prototype.

2. High.Energy Fuels.
If the heat value of fuels can be increased, the specific consumption will be corres-
pondingly smaller. Here the restrictions on fuels for compressorless thermal jets are
less stringent. For instance, if the combustion product contains solid particles, the
70
fuel will not be suitable for turbojet operation due to the possible erosive action of
such particles on the turbine blades. But such fuels can be used for compressorless
thermal jets. Furthermore, a small increase in the heat value has a very large beneficial
effect on the pay load of the aircraft as the percentage of the fuel weight for such air-
craft is necessarily very high. For instance, if the fuel weight is 70% of the total weight,
while the pay load is 10% of the total weight, a 1% increase in heat value will give a
7% increase in the pay load.
Due to the high gas-stream velocity and large heat addition, the problem of com-
bustion chamber design is very difficult. Experiments carried out by various research
laboratories, especially those by the National Bureau of Standards and the Chemical
Engineering Department of the Massachusetts Institute of Technology, have answered
many questions. However, one must realize that such tests are exploratory rather than
systematic. The fundamental problems of combustion under conditions of the ramjet
are essentially aerodynamic in character. Hence the correct solution can be arrived
at only by utilizing all the modern developments in fluid mechanics. This important
problem will be discussed in detail in the following section.

RAMJET COMBUSTION AS A PROBLEM IN FLUID MECHANICS

The combustion if given sufficient time will reach the temperature and composi-
tion determined. by chemical equilibrium. In a nonuniform mixture, there will be
locally overrich regions where the combustion temperature is particularly high and
the accompanying higher dissociation will reduce the efficiency of combustion. In
other words, a nonuniform mixture of air and fuel will reduce the heat liberation·
Therefore, for high combustion efficiency it is imperative first to obtain a uniform
fuel-air mixture.

1. Atomizofion of Fuel and the Mixing of Fuel and Air.

If liquid fuel is used, the fuel-air mixture is obtained, first, through atomization of
the fuel into droplets, then partial evaporation of the droplets and finally mixing of
the fuel vapor with the air stream and the diffusing of droplets to form a uniform but
not necessarily homegeneous fuel-air mixture. The problem involved is then essen-
tially a diffusion and mixing problem. It is well known that diffusion and mixing in
laminar flow is po very slow process, far too slow to be useful in a ramjet combustion
chamber. However, if the flow is made turbulent, then the fluid elements are under
rapid agitation and there 'is a much greater exchange between fluid elements. The
diffusion and mixing in turbulent flow is thus many thousand times more rapid than in
laminar flow. Therefore, turbulent flow conditions are necessary for the ramjet com-
bustion chamber.
Modern investigations on turbulent flow show that it is not sufficient to specify
just the degree of turbulence by the ratio of the fluctuating velocity to the mean velo-
city. One must specify the character of the turbulent flow. The character of turbulence
is generally stated in terms of scales of turbulence, [he correlation coefficients, the
decay factor, etc. All these characteristics have a close relationship to the diffusion and
mixing process. Therefore, the problem is not just to introduce turbulence into the
air stream but to introduce the right kind of turbulence. In fact, even more, One has to
introduce the right kind of turbulence with a minimum loss of pressure. Here the mod-
71
ern theory of turbulent fluid motion could be of invaluable help in solving this phase
of combustion problem.
To allow for more time for the mixing of air and fuel, the liquid fuel is generally
injected far ahead of the combustion zone. The recent designs of ramjets usually have
the fuel injectors located in the diffusor part of the duct. The best location for uniform
fuel-air mixture and necessary means of introducing the proper turbulence should be
a main subject of research. However, if diffusion and mixing is the main problem
here, there is no reason not to investigate the problem as such. In other words, the
combustion phase of the problem can be separated from the diffusion and mixing
phase of the problem. The mixture does not have to be ignited while trying to obtain
a uniform fuel-air mixture. In fact, the sampling of the gas stream would be much
easier in a cold test than in a hot test with combustion. One may even go another step
further: If a liquid approximating the liquid fuel in its essential characteristics such as
surface tension, vapor pressure, viscosity, heat of vaporization, density, etc., can be
found, this liquid instead of the fuel could be used in tests, and thus possibly avoid the
danger of explosion and other inconveniences. This approach to the diffusion and
mixing problem would make available the whole instrumentation and methods of
fluid dynamics which have been intensively developed during the past 15 years.

2. Ignition and Flame Front.

After having obtained a uniform fuel-air mixture, one then meets the combustion
problem proper, i. e., the ignition ofthe mixture, and the maintenance of a stable flame
with complete combustion. Generally it is recognized that to maintain a stable flame
in a gas stream of very high velocity, flame holders must be used. A flame holder is
nothing but a blunt body introduced into the high-speed gas stream to create a dead-
water region where the velocity is low enough for a small flame to remain there. This
small flame then acts as a source for continuously igniting the main high-speed gas
stream. If the stream velocity is q and the flame propagation velocity is VEt then the
angle of spread (j of the flame from the flame holder should be sin -I = (j (Fig. 6).
q

Flame Front

,.,gur. 0

72
Therefore, to shorten the combustion chamber, we have to introduce many ignition
or flame holders.
The question of flame propagation velocity is, of course, closely connected with
the chemical-kinetic process of combustion. However, fluid mechanical considera-
tions must also be introduced. First, the flow is not laminar but turbulent, i. e., in addi-
tion to molecular agitation of microscopic scale, there is also agitation of macro-
scopic scale with fluid elements as units. It is certain that the flame propagation velocity
is a function of the scale and degree of turbulence. This relation must be thoroughly
investigated. A beginning has been made by P. Chambre' and C. C. Lin (Ref. 7) in this
field of research, yet much work remains to be done.

3. Combustion in an Accelerating Gas Stream.

After the gas mixture passes through the flame front generally the chemical re-
action is not completed. It might be felt that the reaction after the flame front should
be a problem purely in chemical kinetics. However, this is not the case. One must
again introduce fluid mechanical consideration. First of all, the presence of turbulence
greatly alters the rate of diffusion of the reacting molecules. Secondly, due to the pres-
ence of walls of the combustion chamber, generally the products of combustion can·
not expand without at the same time increasing their velocity. This is the coupled
effect of heating and inertia forces, i. e., aerothermodynamic effects.
To illustrate the point, consider the heating of a perfect gas in a duct of constant
section. For simplicity, consider the velocity to be uniform across the section, so that
a one-dimensional calculation can be made. The specific heats of the gas are assumed
to be constant. Let p, p, v, T be the pressure, density, velocity, and temperature of the
gas respectively. Then the equation of continuity is
dp dv
-p + -v = 0 (1)

The momentum equation is

pvdv = -dp (2)
The energy equation is

dh "" d ( -l vO
2 P-
+ - ) (3)
2 0-1 P
where dh is differential heat added per unit mass and 'Y is the ratio of specific heats.
By simple elimination,
.k.!....
vdv = 1 dh (4)
M2- 1

where M is the local Mach number. Similarly, the variation of the local Mach number
is
OM2-1 db
d(M2) = (0-1) 2. _ 1 0;2 (5)
M2
Equations (4) and (5) show that for flow velocities smaller than the local velocity
of sound, both the velocity and the Mach number will increase with the addition of
heat. For flow velocities greater than the local velocity of sound, both the velocity

73
and the Mach number will deaease with the additional heat. Fo~ these quantities to
increase in supersonic flow, heat has to be subtracted from the flow i.e., make dh
negative. Therefore, it is impossible to obtain supersonic velocities at the exit of a
combustion chamber of constant section by continuous addition of heat if the Mach
number at the inlet is smaller than one. In fact, the maximum amount of hear that can
be added corresponds to that required to accelerate the gas stream to the local velocity
of sound. This limit can be called the critical rich-mixture limit or simply the choking
limit. This result is indeed somewhat surprising if one is limited to considerations of
chemical kinetics only. It is certainly a combination of inertia and heating effect.
Now it would be natural to ask: If the fuel-air ratio is gradually increased, keeping
the inlet Mach number constant, what will happen when the mixture ratio is larger
than the critical value stated above? One possibility will be the flash forward of the
flame to the exit'of the duct and combustion outside of the duct. If the flame velocity
is smaller than the exit velocity of the unburned gas mixture, the flame will be "blown
out-" A warning before reaching this limit is perhaps rough combustion with large
oscillations.
Thus it is clear that the three phases of the combustion process in ramjet, i.e., the
mixing and diffusing to. achieve a uniform fuel-air mixture, the ignition and flame
fronr propagation, and finally the combustion behind the flame front, are all closely
related to fluid mechanical considerations. The effect of turbulence of the gas flow
and the intimately coupled effects of inertia forces and thermal energy liberated by
chemical reaction must be emphasized. This point of view on the combustion problem,
as a problem in the combined field of fluid mechanics and chemical kinetics, is believed
to be essential for a fundamental understanding of the design principles of ramjet
combustion chambers.

REFERENCES
1. Pabst, Test Report of Focke-Wulf Co.
2. Tangren, R. F., "Estimated Ramjet Performance," Progress Report 3-1, Jet-
Propulsion Laboratory, California lnstitute of Technology, (November 1944).
Tangren, R. F., "Estimated Performance of Ramjets at Subson~c Speeds," Progress
Report 3-3, (August, 1945).
3. Bollay, W., and Redding, E. M., "Performance of Open-Duct Propulsion
Systems (Ramjets) at Subsonic Speeds:' Power Plant Memorandum No.5, Bureau
of Aeronautics, U. S. Navy (December, 1943).
4. Becker, J. V., and Baals, D. D., "High-Speed Tests of a Ducted Body With
Various Air-Oudet Openings," NACA Advisory Conference Report (May, 1942).
5. Kantrowitz, A., and Donaldson, C. duP., "Preliminary Investigation of Super-
sonic Dift'users/' NACA Advisory Conference Report No. L5D20 (May, 1942).
6. Oswatitsch, K., and· Bohm, H., "Luftkrafte und Stromungsvorginge bei ange.
triebenen Geschossen," Forschungen und Entwicklungen des Heereswaft'enamtes,
Bericht Nr. 1010/1, 1010/2 (1944).
7. Chambre, P., and Lin, C. C., "The Effect of Turbulence on Flame Propaga-
tion," Progress Report 3-S, Jet-Propulsion Laboratory, California Institute of Tech~
nology (November, 1945).
74
 APPENDIX

COMPARISON OF THEORY AND EXPERIMENT FOR A SUBSONIC RAMJET

The only complete test data on a ramjet are those obtained for the subsonic
ramjet designed by Pabst of the Focke· Wulf Co. (Ref. 1). The tests were made in the
A·9 wind tunnel of the Hermann Goring Research Institute at Brunswick, Germany.
The diameter of the ramjet is approximately seven inches. The .camjet was supported
by a streamlined strut. The drag of the strut was added to the measured thrust to obtain
the net thrust of the ramjet. The fuel used was gaseous hydrogen. Since the lower heat
value of gaseous hydrogen is SO,700 BTU/lb, while the lower heat value of gasoline
if 19,000 BTU/Ib, the measured specific fuel consumptions in hydrogen can be easily
converted to those in gasoline by the muttiplying factor SO, 700/19,000 = 2.67.
The reduced test data are given in Figs. 7 and 8. For the case with jet temperature
equal to 1472°F, one can compare the results with the calculated values in Ref. 2 for
static temperature at the exit of combustion chamber equal to 1 SOO°F and a nozzle
contraction ratio equal to 0.8. It is seen that the calculated thrust coefficient is lower
than the observed .value except at high speeds. The calculated specific fuel con sump·
tion is somehwat higher than the observed value except at high speeds. The calculated
thrust is thus too small. It is pointed out in the section "Basis of the Theoretical Analy-
sis," page 63, that the ejector action of the jet will tend to increase the net thrust of the
unit. This seems to be confirmed by the present data. The ejector action is stronger if
the jet velocity is higher or if the thrust of the unit is larger. This means higher velocity
over the outer surface of the duct can be expected at a given flight Mach number if the
temperature of the jet is higher. Therefore, at.higher jet temperatures, the shock wave
over the outer surface will appear earlier. This means that drag increase or drop in
thrust coefficient due to shock wave will appear at smaller free-stream Mach number
if the jet temperature is higher. This is shown by Fig. 7 to be true.
Thus the differences between theory and experiment can be easily explained. The
trend and magnitude of the specific consumptions and the thrust coefficients are, how-
ever, satisfactorily predicted by the theory. Hence the general validity of the assump-
tions made in the theoretical analyses is established.

75
 PART IV
FUTURE TRENDS IN THE DESIGN AND DEVELOPMENT
OF SOLID AND LIQ1JID FUEL ROCKETS

By

HSUE·SHEN TSIEN
 PART IV
FUTURE TRENDS IN THE DESIGN AND DEVELOPMENT
OF SOLID AND LIQUID FUEL ROCKETS*
DECEMBER 1945

TYPES OF ROCKETS AND THEIR

PRESENT APPLICATIONS

The two main types of rockets are the solid-propellant type and the liquid-pro-
pellant type. The solid-propellant rockets are now used or suggested for use in pro-
pelling artillery rockets, for the assisted take-off of aircraft, for the launching of fly-
ing bombs and missiles, and for the propulsion of large missiles. The liquid-propellant
rockets are used for assisted take-off and for the propulsion of very large missiles
and airplanes. While there is no essential difference in the operating characteristics
of these two types of rockets, and thus for any new application, the possibility of both
types should be investigated, there are certain facts which should be kept in mind. The
solid-propellant rocket contains all the propellant in the high-pressure combustion
chamber or the motor. Thus if the duration of the operation is long, the chamber
volume becomes very large and the weight of the chamber will be very large. There-
fore, for very long durations, i.e., durations in excess of 30 or 40 sec, the weight
of a solid-propellant rocket is heavier than that of a liquid-propellant rocket. How-
ever, this line of demarcation also depends upon the thrust of the rocket. The reason
for this variation is that for the liquid-propellant rocket the unit weight is a function
of the thrust. Larger thrust makes the unit weight smaller, especially in the case of
pump-fed rockets. The preceding value of 30 or 40 sec corresponds to a thrust of
approximately 4000 lb. In other words, for durations in excess of 30 or 40 sec and
for thrust in excess of 4000 lb, the liquid-propellant rockets are definitely superior
to the solid-propellant rockets. For durations or thrust below these limits, generally
one can say that for short durations at large thrust, such as artillery rockets, solid-
propellant rockets are more suitable; while for long durations at a small thrust, as in
airplane propulsion, the liquid-propellant rockets are superior.

* In nrder to cover the whole field of rocket development, liberal use has been made of material
discussed in greater detail in the two papers, L. P. Hammett, "Solid Prnpellants for Rockets and
other Jet-Propelled Devices;" A. J. Stosick, "Liquid Propellants for Rocket-Type Motors," in the
volume Aircraft Fuels and Propellants, a further report of the Scientific Advisory Group.

81
SOLID PROPELLANT ROCKETS
1. Pre••nt Statu. of the Solid Propellant Roelc.".
Solid propellants are often discussed from the point of view of thermal energy,
which is essentially connected with the effective exhaust velocity or the specific fuel
consumption. However, there are many other characteristics of solid propellants
which greatly influence their usefulness as rocket propellants. These characteristics
are as given in Table I: (a) density, (b) temperature sensitivity, (c) rate of burning,
and (d) exponents in the rate of burning law.
The density of the propellant is important in that it determines the volume of the
combustion chamber if the specific impulses of the propellant remain constant. In
other words, for a given weight a less dense propellant requires a larger combustion
chamber than a denser propellant. Since for solid-propellant rockets the main part
of the motor weight is the weight of the combustion chamber, a denser propellant
wi11lead to a lighter unit. For most solid propellants, the density is 100 lb leu ft.
The temperature sensitivity is really the sensitivity of the burning rate to the
temperature of the propellant before ignition, i.e., the ambient temperature. A high
sensitivity would mean a large variation of the chamber pressure for a given motor
design. To operate the rocket at high ambient temperature, the chamber has to be
designed to withstand the high pressure, and this leads to the heavy motor weight. At
low ambient temperature, the pressure is too low for stable combustion. Therefore,
if the temperature sensitivity is high, the rocket motor operating temperature range
will be limited. The double-base propellant ballistite is an example of high-sensiti-
vity propellant.
The rate of burning of the propellant determines the burning surface. It is seen
from Table I that this rate varies but little. The large difference in the operating dura-
tion from 0.' seconds for the artillery rockets to 30 seconds for assisted-take-off
rockets is achieved by the difference in design of the propellant charge. For short
burning time, the charge is designed in such a manner that it will burn with large
surface, and thus the volume or weight of charge burned per second is large. This is
the so-called unrestricted burning. The other type for long burning time is the so-
called restricted burning. In this case, the burning surface is generally restricted to
the cross-sectional area of the combustion chamber. It is thus immediately clear
that for a restricted-burning rocket, the motor could be loaded to a fuller extent, since
gas passage ducts are not required in the motor, and thus the motor will be more com-
pact and lighter in weight. From this point of view, then, a restricted-burning rocket
is preferred. This is clearly seen in Table II, where the performance of the actual
solid-propellant rockets is compared. The impulse to total weight ratios for the un-
restricted-burning motors are lower than those for the restricted-burning rockets.
The rate of burning of a solid propellant varies as a power n with the pressure of
the gas surrounding it, while the flow of the gas through the nozzle varies as the first
power of the pressure. This situation admits of a stable steady state only if n is less
than unity. The state is more stable and less sensitive to disturbing influences, such as
variations in the burning surface, the greater the value of .!. In this sense the difference
n
between the exponent 0.7' of most double-base propellants and the 0.4-0.' exhibited
82
 TABLE I

CHARACTERISTICS OF SEVERAL SOLID PROPELLANTS

Composite Propellants
Ballistite GALClT '3 GALCIT'8 GALClT61C fGALCIT6'
Slow Fast (Asphaltic) (Asphaltit:) (Asphaltic) (Asphaltic)
------
Specific Impulse,
Ib-sec/lb 210 160 170 170 174 186 177
~ -- --- ---

Exhaust Velocity,
ft/sec 6800 5150 5500 5500 5600 5900 5700

Density, Ib/ft3 101.5 101 112 104 108 110 110

--
Impulse per Unit
CD
Volume, percent 100 76 90 83 89 96 92
Co)

Temperature Sensitivi-
ty,percentincreasein
chamber pressure be-
tween 40° and 90 0 P 29 10 14 10 10 11 11
---

Temperature Range for

Safe Operation, of Limited by -40 to -40 to 30 to 120 15 to 120 -9 to 120 -5 to 120
(Laboratory Tests) Sensitivity 140 140
--
Rate of Burning,
in./sec 1.40 0.25 1.32 1.32 1.80 1.60 1.46

Exponent in Rate of
Burning 0.69 0.40 0.40 0.69 0.80 0.76 0.7
--
Chamber Temperature, 5000 to 3000 to 3000 to 3000 to 3000 to 3000 to 3000 to
OF 6000 3500 3500 3500 3500 3500 3500

Above data are all for 2000 psi chamber pressure, 14.7 psi external pressure.
by composite propellants, is a large and important one, which leads to materially
greater reproducibility and reliability with the latter material.

2. poss;b/e Improvements ;n Performance 0' Solid-Propellant Rockets.

The possibility of raising the performance of solid-propellant rockets by improv-
ing the energy of the propellant is not oustanding. However, by considering the mat-·
ter in its broad outline, it is evident that the aim of rocket design is to raise the impulse
produced per unit total weight, i. e., the weight of the propellant plus the weight of
the motor. Therefore, ~ne can achieve this by either decreasing the propellant weight
or by decreasing the motor weight. These possibilities are dearly shown for the par-
ticular application of artillery rockets in Figs. 1 and 2. The effect of variations of the
effective exhaust velocity and the motor weight on the maximum flight velocity of the
rocket, while keeping the pay load constant, is shown in Fig. 1. The effect of varia-
tions of the effective exhaust velocity and the motor weight on the pay load, while
keeping the maximum flight velocity constant, is shown in Fig. 2. It is seen that almost
the same percentage gain can be achieved by decreasing the motor weight as by in-
creasing the effective exhaust velocity. Hence the design factors are just as important.
in improving the performance of solid propellants as is the propellant itself.
In the following paragraphs, the different aspects of the lightweight design will
be discussed in detail.
a. LOW-COMBUSTION PRESSURE. Since the weight of a rocket motor
increases approximately in direct proportion to the gas pressure, while the effec-
tive exhaust velocity varies much less rapidly, a decrease in operating pressure is
favorable until it reaches the point where other forces than bursting pressure become
important factors in motor design, certainly as far down as 500 psi. In general, solid
propellants possess a lower pressure limit, which depends on size and geometry of
motor as well as on the propellant, below which they fail to burn, or burn irregularly
and incompletely. This is shown in Fig. 3, which gives the chamber pressure record
of the burning of a double-base propellant at low pressure. The normal burning pres-
sure of this propellant is approximately 1000 psi. The irregularity at low pressure
is evident. The lowest pressure for smooth burning is called the lower pressure
limit. Obviously a low value of this limit is a desirable property of a propellant, and
possibilities in this direction should be exploited.
The irregular burning at low combustion pressure can also be controlled by a
spring-loaded control valve, as shown in Fig. 4. It is the German practice to use one
of the valves with one or two conventional discharge nozzles, so that when the gas
pressure in the chamber is higher than the preset value, the valve is opened and the
pressure is released, and when the gas pressure in the chamber is lower than the pre-
set value, the valve is closed and the pressure in the chamber is built up. The smooth
combustion with the control valve is shown in Fig. 5 as compared with the rough and
intermittent burning of the same propellant without the valve shown in Fig. 3. Of
course, the addition of the control valve is a mechanical complication. However, the
possibility of lowering the combustion pressure and the consequent saving in weight,
especially for long-duration rockets, may well overbalance the inconvenience.
b. SMALL EXPONENT IN BURNING LAW. The effect of a small exponent
in the burning law on stability and reproducibility has already been pointed out in a

84
 TABLE II

PERFORMANCE OF SOLID PROPEllANT ROCKETS

Total Weight, Impulse, Impulse Weight Weight oj

Type Ib Ib/sec Ratio Propellant
.~--

3.5 in. AR·M5 33.8 1,820 53.6 8.5

5 in. AR·M7 33.8 1,800 53.2 8.5
CD
UI
Unrestricted 5 in. HVAR 88.4 5,320 61.0 24.0
Burning 11.75 in. AR·M3 690.0 32,900 47.6 148
7.2 in. T21 20.54 877 42.6 5.74

X8AS·I000 150 12,500 83.3 69

Restricted XlOAS·lOOO 200 16,000 80.0 90
Burning 12AS·1000 250 20,000 80.0 108
(Aerojet Co.) 30AS·I000 415 36,000 86.7 192
 Figure 1

EB'ect of changes in effective gas velocity. in propellant weight. and in motor weight. on veloci-
ty of 5.0 in. HVAR. Ordinates are rocket velocities.

Curve I. Effect of variation in effective gas velocity over range from 50 to 150% of value for
existing rocket (7130 ft/sec). All weights constant.

Curve II. Effect of variation in weight of propellant over range from 50 to 150% of value
for existing rocket (24.0 Ib). EB'ective gas velocity (7130 ft/sec). pay load (48.2 lb). and motor
weight (64.4 lb) constant.

Curve III. Effect of variation in weight of motor over rang~ from 50 to 150% of value for
existing rocket (64.4 lb). Effective gas velocity (7130 ft/sec). pay load 48.2 lb). and propellant
weight (24.0 lb) <;on stant.

86
 Figure 2

Effect of changes in effective gas velocity, in propellant weight. and in motor weight on pay
load of 5.0 in. HVAR. Ordinates are values of pay load.

Curve I. Effect of variation in effective gas velocity over range from 50 to 150% of value for
existing 5.0 in. HVAR (7130 ft/sec). Motor weight (64.4lb), propellant weight (24.0 lb), and
projectile velocity constant (1375 ft/sec).

Curve II. Effect of variation in weight of propellant over range from 50 to 150% of value
for existing rocket (24.0 lb). Effective gas velocity (7130 ft/sec), projectile velocity (1375 ft/sec),
and motor weight (64.4 lb) constant.

Curve III. Effect of variation in weight of motor over range from 50 to 150% of value for
existing rocket (64.4 lb). Effective gas velocity (7130 ft/sec), projectile velocity (1375 ft/sec),
and propellant weight (24.0 lb ) conStant.

88
Intermittent Burnin9

FiflVre 3 - ''''ermiHe,., Burning

Figure 4

Smooth Btlrnlnq

FiflVre 5 - Smootb Buming

90
previous paragraph. The advantage also allows the designer to use a lower factor of
safety in the chamber design. This naturally leads to a lighter motor.
The value of a low exponent is especially great in the usual type of artillery rocket,
in which a tubular or cruciform grain burning over all or nearly all of its surface is
contained in a long motor of small diameter. When an attempt is made to reach the
highest possible ratio of propellant weight to motor weight, the limiting factor be-
comes the space available for flow of gases toward the nozzle. If this is small, a pressure
differential is set up, the burning rate increases in the high-pressure region, and the
whole effect is 'magnified by an amount which is greater, the greater the exponent.
Consequently, a low exponent permits a higher propellant-motor weight ratio for a
specified time of burning. A rigid propellant is of course desirable for the same reason,
as is one resistant to distortion under applied forces, since any bulging of the grain
under the forces of setback and pressure drop reduces the port area.
c. LOW-TEMPERATURE SENSITIVITY. As stated before, the temperature
sensitivity is closely related to the permissible range of operating ambient tempera-
ture. A low temperature sensitivity is desired. Such a characteristic is also desirable
from the point of view of low motor weight, as the advantage of operating at a low
pressure vanishes if the pressure becomes four or five times greater at temperatures in
the upper part of the operating range, due to high temperature sensitivity.
It has been suggested by L. Pauling that a charge containing particles or strands
of small diameter, composed of a fast-burning powder of low temperature coefficient,
embedded in a matrix of slower burning powder of high temperature coefficient,
possesses, as a whole, the low coefficient of the particles. In this case the grain does
not burn by recession of a plane 'surface, but by the formation of a broken sutface
whose magnitude is determined by the behavior of the faster burning powder. This
procedure, already proven experimentally, makes it possible to eliminate the unde-
sirably high temperature coefficient of the double-base propellant by the addition of a
proportion of solvent-extended composite propellant.
d. RESTRICTED BURNING AND HIGH BURNING RATE. As mentioned
previously, the restricted-burning solid-propellant rockets are superior to the un-
restricted-burning rockets, due to better charge-loading efficiency. However, to
make a restricted-burning rocket for very short duration of burning, one has to
develop a propellant or a group of propellants with a much higher burning rate
than those available today. It seems that by addition of certain metallic powders,
such as in the Aberdeen propellant, this can be achieved. The actual research work on
this type of solid propellant is, however, just beginning, and the final satisfactory pro-
duct is still far away. As a tentative aim, one might set the limits of burning rate from
one to 100 inches per second. Of course, to develop these propellants, a closer under-
standing of the combustion phenomema is necessary. Such understanding will not
only greatly facilitate the research work on burning rates, but also will be very useful
in the reduction of the temperature sensitivity of the propellant. It is generally true
that the temperature sensitivity' is higher for unrestricted burning than for restricted
burning for the same propellant. Thus development of a propellant of high burning
rate, so that the burning can be restricted to the cross-sectional area of the motor, is
also desirable from the point of view of temperature sensitivity.
91
 e. OTHER POSSIBLE IMPROVEMENTS. To reduce the weight of the mo-
tor, one can also try to reduce the temperature to which the motor walls are heated
by the propellant gases, since the strength of metals decreases materially at tempera-
tures easily reached by the motor walls in conventional designs. Since the propellant
itself is a very good insulator, for a short-duration rocket, the charge might be made
to burn from a central perforation along the axis of the charge and gradually to burn
toward the wall of the motor. If this can be done, then aluminum, magnesium, or even
plastic walls might be used. A further possibility is the ceramic-coated wall, to insu-
late the hot gas from the metal itself.
Smoke elimination is sometimes desirable, as for ground-fired rockets, but is un-
important for others, such as airborne missiles. The same considerations apply to
flash, i. e., a luminous jet.

3. Concluding Remarlcs.
In this section, the status of the present solid-propellant rocket development is
briefly reviewed. The various possibilities of improving the performance, such as low
temperature sensitivity, low combustion pressure, and small exponent in burning law
are discussed. It is believed that the solution of these problems will greatly increase
the usefulness of the solid-propellant rockets, which possess the advantage of inherent
simplicity.

LIQUID PROPELLANT ROCKETS

1. Design Criteria lor Liquicl-Propellant Rocleets and CItoice of Propellants.

The two main applications for liquid-propellant rockets are the assisted take-off
of large aircraft, when a thrust larger than 6000 pounds is required for durations up
to 40 seconds, and the propulsion of large guided missiles, which usually need thrust
in excess of 20,000 pounds for durations up to 80 seconds. The design criteria for
these two main applications are not the same. For the assisted-take-off application, the
essential requirements are (a) reliability of operation, (b) simplicity of servicing, and
(c) long life for repeated application. On the other hand, the problem of fuel con-
sumption is not very acute, as the weight of the rocket propellant is only a very small
fraction of the take-off weight of the aircraft, and since it is consumed at the end of the
take-off process, it does npt constitute a load in the normal operation of the airplane.
For the missile application, the first requirement is high specific impulse for range.
Simplicity of design and ease of handling are secondary. On the other hand, if the
missile is intended to be used on a very large scale, the aspect of production economy
must also be considered.
Table III gives the calculated value or the expected value of the performance of
the different liquid propellants. The experimental values are generally lower by ten
percent than the calculated values, due to unavoidable losses. The nitromethane pro-
pellant and the hydrogen-peroxide propellant are outstanding in that the chamber
temperature is low. This fact, together with the advantage of having to deal with only
a single component, naturally leads to their adaptation for the assisted take-off of air-
craft. Nitromethane itself is somewhat shock sensitive. But this sensitivity can be con-
siderably reduced by the addition of a few percent of a desensitize~, such as methyl

92
 TABLE III
CALCULATED PERFORMANCE OF VARIOUS LlOUID PROPELLANTS
(at 300 psi chamber pressure)
Volu- Average
metric Chamber Exit Ratio of Molec-
Propellant Mixture Exhaust Specific Specific Tempera- Tempera- Specific ular
Ratio Velocity Impulse Impulse ture IIIre Heats of Weight of
It/sec Ib-sec/lb Ib- 0p 0p Products Products
sec/ft3
Liquid Oxygen, Gasoline 2.5 7,780 242 14,958 5,470 2,990 1.219 22.66
Liquid Oxygen 1.5 7,830 244 14,777 5,260 2,965 1.212 21.69
Ethyl Alcohol
Liquid Oxygen
Methyl Alcohol
1.0 7,587 236 13,702 4,760 ... . .. .. ....
Liquid Oxygen, Ammonia 1.4 8,000 248 13,161 4,950 . . .. ... . ....
Liquid Oxygen, Hydrazine 0.4 8,285 257 17,327 4,120 ... . ... . ....
Red Fuming Nitric Acid 3.0 7,091 221 18,763 5,065 2,746 1.220 25.40
Aniline
White Fuming Nitric Acid 3.0 7,035 219 18,184 4,900 2,635 1.221 25.01
Aniline
White Fuminf Nitric Acid
Furfuryl Alco 01
2.0 6,840 212 17,867 4,293 2,578 ... . ....
Mixed Acid
Monoethylaniline
2.5 6,780 210 17,569 4,600 2,800 ... . ....
Nitromethane .... 7,008 218 15,515 3,950 1,980 1.245 20.31
Diethyleneglycol, Dinitrate .... 6,865 213 18,484 4,078 2,061 1.236 21.79
Hydrogen Peroxide (100%) .... 4,710 146 13,184 1,794 714 1.249 22.70
Hydrogen Peroxide (87%) .... 4,065 126 11,039 .1,216 379 1.271 21.92
Hydrogen Peroxide (87%) 4.0 7,180 223 16,845 4,156 2,538 .... . ...
Methyl Alcohol
Hydrogen Peroxide (87%)
Nitromethane with
1.0 7,305 227 17,007 4,050 ... . .... ....
30% Methyl Alcohol
Hydrogen Peroxide (87%)
Nitromethane with
0.3 7,014 218 15,788 2,864 1,966 ... . ....
35% Nitroethane
Hydrogen Peroxide (87%)
Hydrazine Hvdrate
1.6 7,009 218 16,468 4,000 1,950 ... . ....
alcohol. Hydrogen-peroxide propellant has the lowest chamber temperature. In tact,
the chamber temperature is so low ( 1220 °F) that no cooling of the motor is required.
This is a great advantage. On the other hand, the inflammability of concentrated hy-
drogen peroxide with any organic matter, such as cellulose, causes difficulties in the
handling of the propellant, although these difficulties seem to have been overcome by
the Germans. A second choice of propellant for the aircraft application is a bipro-
pella nt, such as hydrogen peroxide plus methyl alcohol, or nitric acid plus aniline.
While these propellants have good performance, the two components complicate the
feed system. The higher combustion temperature also makes the cooling of the motor
more difficult. Although the performance of the propellant with red fuming nitric
acid is better than the performance of the propellant with white nitric acid plus a
small percentage of catalyzer, the nitric oxide fumes (N02) are objectionable for the
aircraft applications and the latter type is preferred.
For the long-range missile application, the most effective propellants are those
with liquid oxygen as the oxidizer. However, due to the difficulty of using liquid oxy-
gen as a coolant for the motor, which is subjected to very high heat flow due to high
combustion temperature, it is preferable to have fuel in the liquid state under the
conditions of temperature and pressure in the cooling ducts, and of such composi-
tion as to make adequate cooling possible. Thus the combinations should be liquid
oxygen plus ethyl alcohol, liquid oxygen plus methyl alcohol, or liquid oxygen plus
hydrazine. The last-mentioned combination is new and has not yet been tested.
However, it gives the highest performance and holds much promise.
The density of the propellant should also be. considered in conjunction with
the specific impulse Liquid hydrogen is rejected as a possible fuel on the ground
of its extremely low density, which necessitates large tank volume for storage. This
fact is particularly important for antiaircraft guided missiles, which must travel
at high speeds in relatively dense air. Hence the air resistance associated with the
large body required ">y the low-density propellant is a distinct disadvantage. This is
demonstrated in Table IV, where three winged missiles of the same explosive load,
and with three different propellants, are compared. The nitric acid and aniline pro-
pellant gives about the same flight duration as other higher preformance propellants.
This fact, together with the ease of manufacture and the spontaneous inflammability,
led the Germans to adopt this type of propellant for their antiaircraft rockets developed
during the last years of World War II. One disadvantage of the large-scale use of the
TABLE IV
! Liquid O2+ Liquid O2 + HNO a+
Propellant Liquid H2 CHaOH Aniline
Structure Weight and Pay Load, lb 3340 3130 3100
Propellant Weight, lb 3700 4770 5260
Gross Weight, lb 7040 7900 8360
Specific Gravity of Propellant 0.4 1 1.4
Specific Consumption, lb/hr/lb-thrust 12.6 16.2 18
Volume of Propellant Tanks, cu ft 150 77 61
Body Frontal Area, sq ft 19.4 12.9 10.8
Wing Area, sq ft 161 183 194
Flight Duration, min 19.5 20.5 19.5

94
nitric acid plus aniline propellant is, perhaps, the high cost of the aniline. However,
the German discovery of the possibility of adding other cheap inert fuels and then re-
taining the spontaneous inflammability by use of a catalyzer is most interesting and
should be exploited. The use of a mixture for the fuel also offers a means of con-
trolling other properties of the propellant, such as freezing point and viscosity, for
wide operating temperature range of the rocket.

2. Propellant Feed Systems.

a. COMPARISON OF VARIOUS FEED SYSTEMS. Since the propellant of a
liquid-propellant rocket is stored in tanks, a feed system is necessary to force the pro-
pellant from the tank to the combustion chamber, under high pressure. The various
systems developed to date (listed in order of decreasing weight for a given amount
of propellant and chamber pressure) are pressurized-gas system, gas-generator sys-
tem, and the turbine-pump system. In the pressurized-gas system, an inert gas such as
nitrogen or air is stored under very high pressure (up to 2000 psi). The gas is allowed
to expand through a pressure regulator and to act on the propellants in their tanks
to force them through the propellant injector to the combustion chamber. The pro-
pellant tanks are thus under high pressure and have to be built strong and heavy. The
weight of the inert gas and the high-pressure gas tank is proportional to the volume
of the propellant to be injected.
In the gas-generator system, the required higy-pressure gas is produced not from
a high-pressure gas tank, but from either the burning of a solid charge or the burning
of a part of the main prop'ellant. Since the high-pressure gas is produced continuously,
the apparatus is much smaller and weighs much less. If the gas is produced by com-
bustion of a part of the main propellant, the size and weight of the gas producer is
independent of the total amount of the propellant to be injected, but is directly pro-
portional to the rate at which the propellant is injected. Although this system still
requires pressurized propellant tanks, the weight of a large high-pressure gas tank
of the pressurized-gas system is eliminated. Consequently the weight of the whole
system is less.
The lightest of the propellant-feed systems is the turbine-pump system. Here the
propellant tanks can be built light, as they are not under pressure. The turbine is driven
by hot gas obtained from the combustion of a small amount of the liquid propellant.
The propellant is transferred from the tanks to the chamber by pumps driven by the
turbine. At present the only type of pump used is the centrifugal pump. These pumps
have the advantage of small dimensions and hence light weight. However, centrifugal
pumps cannot be operated at very high efficiency if the propellant flow is small and the
feed pressure high. Thus, other types of fluid pumps should also be explored, especially
with regard to the possible improvement in pumping fluids of high vapor pressure,
such as fuming nitric acid and liquid oxygen. The weight of the turbine-pump aggre-
gate is proportional to the rate of propellant flow, but independent of the total amount
to be pumped.
Thus the essential difference between the pressurized-gas system and the pump
system is that while the weight of the former is proportional to the total quantity of
the propellant, the weight of the latter is more closely reiated to the rate of propellant
95
flow. Therefore, for short-dpration large-thrust operation, the pressurized-gas sys-
tem is definitely superior. But for long-duration operation, the pump-feed system is
better. For a duration of 60 seconds and a thrust of 4000 pounds, the pump-feed sys-
tem is the lightest. However, the pump system suffers from the complexity of the
machinery and hence higher manufacturing cost. For an expendable weapon such as
a guided missile, where large-scale application is envisaged, the simplicity of the gas-
generating system may have many advantages when the man-hours required for its
production are considered. Thus the two feed systems that should be intensively'de-
veloped are the gas-generating system and the turbine-pump system.
b. PUMP DRIVES. For many applications of the liquid-.propellant rockets, the
power required to drive the pumps can also be derived from other sources. The
possible sources are: (1) mechanically connected with an auxiliary engine or with the
main power plant of the aircraft, if the rocket is only used for auxiliary propulsion;
(2) separate gas turbine driven by rocket propellant; (3) rotary rocket motors; and
( 4) windmill.
For assisted take-off of aircraft, or for superperformance of aircraft, the pumping
power can be taken most conveniently from the main power plant of the aircraft. How-
ever, two facts must be kept in mind. The first is that for the turbojet power plant there
is a delicate balance between the gas-turbine power output and the compressor power
absorbed. If a certain amount of gas-turbine power output is bled off for driving the
feed pumps, proper measures, such as variable discharge nozzle, must be taken to
insure the satisfactory operation of the turbojet system, i.e:, to maintain' the matching
of turbine and compressor with the pump in gear or not in gear. The second point i..
that the proper mechanical arrangement of the pump and the main engine requires
careful consideration during the design of the aircraft. Here coordination of effort
must be made. The rocket system should be considered as integral to the whole pro-
pulsive system, rather than as auxiliary equipment attached to the aircraft after the
aircraft has been designed.
Included in the same category of main engine drive are, perhaps. the separate
auxiliary engine drive and the gas-turbine drive using combustion gas from the com-
bustion chambers of the turbojet or the main gas-turbine power plant. The scheme with
the auxiliary engine is handicapped by lack of a supercharged engine of small output
for altitude operation. However, in general, this type of pump drive is complicated
and, must be considered as expedient only when the directly connected mechanical
drive discussed in the preceding paragraph is not available.
If the rocket is the only power plant on the aircraft, then the pumping power
must be furnished by independent means. At present, the most promising drive for
this purpose is a small gas turbine directly coupled to the centrifugal pump. The gas
turbine is driven by combustion products from either an auxiliary small combustion
pot or those bled off from the main rocket combustion chamber. This system is called
the turborocket. In either case, the combustion products from propellants are gener-
ally too hot for safe, long-duration operation of the small gas turbine and have to be
cooled by water injection. If hydrogen peroxide is used, then such injection of coolant
is not necessary. as the decomposition products of this propellant are at relatively
low temperature. The flow through the pumping turbine is roughly 2-3% ofthe total

96
flow. Thus, for small thrust of the order of a few thousand pounds, the axuiliary com-
bustion pot may have metering difficulties, especially when bicomponent propellants
are used. In general, these systems have the advantage of independent location of the
pumping plant and the main rocket motor. For aircraft or missile applications, this
freedom is sometimes very desirable. Furthermore, the system is completely inde-
pendent of altitude and can be operated even in vacuum.
A similar gas-turbine pump drive has been developed by R. H. Goddard. In
this system the gas-turbine wheel is driven directly by the rocket jet. This is achieved
by bringing the edge of the wheel into the exhaust jet while letting the greater part of
the wheel remain outside the jet. Thus the blades of the turbine are in contact with the
high-temperature exhaust only a fraction of the time. Hence proper cooling can be
accomplished in spite of the very high exhaust temperature. A preliminary estimate
shows that the equivalent propellant consumption due to the presence of a gas tur-
bine in the main jet is about the same as for the turborocket drive. However. here the
advantage of independent location of the pumping plant and the main rocket motor
is lost.
Another propellant-feed method is to combine the pumping plant and the rocket
motor into a single anit, such as the centrojet. Here the rocket motor is in rapid rota·
tiona At present, although the possibility of this system has been demonstrated, its
perfection as an efficient, dependable unit has to await more fundamental research on
the combustion and the cooling of a rotating combustion chamber. However, its
inherent compactness and simplicity demand further investigation in these directions.
It has also been suggested that a windmill might be used for the pump drive. For
high-speed Hight, it is more efficient to use a ducted windmill, so that the tip speed of
the blades can be kept low. At the design altitude, the equivalent propellant consump-
tion due to the presence of the windmill is approximately four percent of the total
propellant How; thus, it is comparable with the turborocket system. However, the
output of the windmill is dependent on the air-How condition. It seems difficult to
have a design that will operate efficiently at a wide range of altitudes. Ultimately, the
size of the windmill will be prohibitive at extreme altitudes. Therefore, this type of
drive must be considered as a special device with limited applicability.

3. Mo'or Construction and Design.

The two main problems in motor design are the optimum propellant injection
for efficient combustion and the cooling of the motor walls for long-duration opera-
tion.
The problem of propellant injection is closely connected to the geometrical
shape of the combustion chamber. The aim is to utilize fully the volume of the com-
bustion chamber, so that combustion is completed with the smallest possible dimen-
sion of the motor. This is important not only for reducing the weight of the motor,
but also for reducing the surface of the motor walls and thus decreasing the amount
of heat that has to be taken away by cooling. At present, little is known about the de-
tailed How pattern in the combustion chamber, and it is difficult to ascertain whether
the combustion chamber is fully utilized. For instance, if one adopts the concept that
the combustion is only a function of the length of time the propellant particles stay

91
in the chamber, then the length of the chamber should be kept at almost constant value
when the thrust rating of the motor is varied, as shown in Fig. 6. However, the path
of the gas in the combustion chamber does not necessarily follow the axial direction.
In other words, the length and the duration of flow of gas depend upon the flow pat-
tern in the chamber. The simple procedure as shown in Fig. 6 is too crude and does
not lead to the most efficient design.

-----Jl...... Increasilll tbrwst

Figure 6
For long-duration operation, i.e., duration in excess of ten seconds, the motor
wall must be cooled to maintain the structural strength. There are three possible
methods of cooling the motor: (a) part or all of the propellant passes through a
jacket which encloses the motor and the nozzle walls, such as the liquid rocket motor
developed by Aerojet Engineering Company in this country; (b) part of the propellant
is injected with a very small velocity along the walls, and a liquid film is created over
the surface, as a shield against the hot gas, such as in the V·2 motor; (c) part of the
propellant is made to seep through a wall of porous material and the wall is cooled
by vaporization of the liquid.
The first method is the so-called regenerative cooling. The difficulties which
might be encountered are the local stagnation of the cooling flow and thus hot spots
in the wall, and insufficient available quantity of coolant. An example of the latter
difficulty is the case of gasoline and liquid-oxygen propellant at optimum mixture
ratio, with gasoline as coolant. This difficulty, it seems, would rule out the gasoline
and liquid-oxygen combination as a practical propellant. Of course, one possible
remedy is to coat the wall with a layer of heat-resisting and heat-insulating ceramic
material, so that the rate of heat flow is reduced. The development of such a ceramic
coating is, however, only in its beginning. Much research has yet to be done.
The second and the third methods of cooling are similar in principle, in that both
depend upon the heat of vaporization to absorb the heat transferred to the wall.

98
Of course, part of the propellant evaporated will particiapte in the combustion, but
the combustion can hardly be as efficient as for the propellant injected through the
normal injector into the main combustion lone. Thus a certain loss and an increase
in propellant consumption is inevitable. On the other hand, the loss might be quite
small and unimportant. By comparing these two methods of evaporative cooling, tbe
more uniform distribution of coolant by the porous walls leads to a more efficient
utilization of the liquid. Initial experiments (Fig. 7) indicate that this method is very
effective. However, both methods (b) and (c) should be developed for more definite
comparison.

4. Concluding Remarks.
It is seen that while the future development in solid-propellant rockets leans
more heavily on the improvement in the propellant characteristics, the development of
liquid-propellant rockets depends more on research on the mechanical design.
The design of efficient large-scale rockets for heavy missiles requires especially in-
tensive enJdneering work on the combustion, cooling, and pump problems.

99
 PART V

HIGH TEMPERATURE MATERIALS

FOR AIRCRAFT PROPULSION
DECEMBER, 1945

INTRODUCTION
The basic requirements for materials to be used in aircraft propulsion devices
depend on the type of engine considered. To discuss these requirements it is conven-
ient to consider separately the three principal types of engine, i. e., turbojets, ramjets
and rockets. This subdivision into three classes is justified by the fact that the range of
temperatures to which the materials are subjected is different in each type of engine.
In a gas-turbine power plant, temperatures from 1200° to 1600 0 p are at present en-
countered. In a ramjet unit, a temperature range of from 2500° to 3 500 0 p is generally
considered possible. The case of the rocket motor is the most extreme, with tempera-
tures as high as 5 500 0 p in the combustion chamber.

MATERIALS FOR GAS·TURBINE POWER PLANTS

1. Higb-Temperafure-Res;sting Alloys for Gas-Turbine Power Plants.
The critical part of a gas turbine, as far as materials are concerned, is undoubted-
ly the blade. Very extensive work has been done during the last few years in order to
improve the high-temperature properties of alloys, and meet the severe conditions of
stress and temperature present in a modern gas turbine. Among the most important
variables to consider in the use of a given material are temperature of operation, maxi-
mum stress, allowable deformation during the expected life of the unit, resistance to
corrosion and erosion. Other factors, like weldability and machineability are also of
importance. A large number of alloys have been produced and tested for applications
in which the temperature of the blade itself is between 1200° and 1600 0 P. The la-
boratory tests include generally high-temperature strength, time required for rupture
under certain conditions of stress and temperature, rate or creep under constant stress
and temperature, endurance limit at high temperature, damping capacity and Charpy
or bod impact tests at high temperature. Other physical constants like thermal ex-
pansion and terma! conductivity have also been measured. As a rule, the most promis-
ing alloys on the basis of these laboratory tests were then tested in an actual unit. It is
l
not as yet clearly established which one of the above-mentioned properties is the most

103
important. Many failures have been encountered in practice that could not be clearly
connected to any particular weakness of the materials as shown in laboratory tests.
The correlation between actual performance and physical properties is still an import-
ant problem.
A general survey of the physical properties of the high-temperature-resisting alloys
available at present shows that both tensile strength and creep strength decrease very
markedly in the range of temperature between 1200° and 1600 0 F. The order of mag-
nitude of the maximum safe stress for operation at 1500° to 16000 F is 7000 psi. The
best alloys developed so far have, under those conditions, a creep rate of less than 1 %
in 10,000 hours, which is generally recognized as a safe upper limit. At higher tem-
peratures, very few results have been obtained, and only tensile strength values are
available for a few alloys. The order of magnitude of the decrease in strength is from
approximately 50,000 psi at 1600°F to about 10,000 psi at 2000°F. From these re-
sults it seems logical to conclude that the allowable stress at 2000°F corresponding
to an acceptable rate of creep would be much too low to meet the requirements of the
present gas-turbine design. This drop in strength with increasing temperature is the
obvious consequence of the small margin left between the required service tempera-
ture and the melting point of the alloys. More research work on high-temperature-
resisting alloys will undoubtedly improve their performance for temperatures below
1600 0 F. There seems to be little chance, however, that alloys can be developed for
higher temperatures.
The evolution of the gas turbine has been marked by continuous increase in inlet
gas temperature, which resulted in a corresponding increase in efficiency. If the field
of application ofall<?ys is really limited to a certain upper limit in temperature, consid-
eration should be given to other types of material or other methods of design.
The use of ceramic materials with high melting points, would probably raise the
temperature limit to 2000°F or higher. Another solution would be to provide an
appropriate cooling to the parts of the engine exposed to high temperature. This
approach to the problem will eventually require the development of appropriate
materials and should be considered from that point of view.

2. Ceramic Material. lor Gas-Turbine Power Plants.

In the broadest sense of the word; ceramic materials include all materials com-
posed of various combinations of oxides properly mixed together, processed by slip-
casting or by pressing, and fired at a sufficiently high temperature to insure a proper
bond between particles. This bond is generally obtained by fusion of a small portion
of the constituents. The structure of a ceramic material is essentially heterogeneous
and consists of crystals imbedded in a glassy matrix, the relative proportion of the two
phases varying within wide limits. The essential differences between the structure of
a ceramic material and that of:1 metal are, first, the existence of a vitreous phase in the
ceramic material, which is absent in a metal; second, the fact that the crystals in a
ceramic material are of the ionic type in which the stability of the lattice is due to elec-
trostatic forces, while the metal crystals are made of neutral atoms linked together by
a special kind of bond which is typical of the metallic state. The lack of plasticity in
ceramic materials is inherent to their structure, and is likely to remain one of the most

104
important drawbacks in their use. These fundamental concepts should be borne in
mind if successful use of ceramic material is to be made for mechanical devices. Ex-
tensive studies should be made in an effort to correlate what is known of the physics
of the solid state, and apply it to the study of ceramics. This would give the basic in-
formation so badly needed to plan a logical program of research.
To be of any interest to the gas-turbine designer, ceramic materials should have
sufficient strength at temperatures above 1600°F. Very little is known of the variation
of strength of a ceramic material with increasing temperature. A few tests have in-
dicated that the creep behavior is similar to that of metals. This similitude in the shape
of the flow-versus-time curve at a constant temperature does not necessarily mean that
the mechanism of flow is the same in the two cases. Creep in metals is mostly due to
plastic flow inside the crystals, while it is very probable that a viscous flow takes place
in the vitreous phase of a ceramic material. In addition to the strength requirements,
ceramic materials for gas turbines should also have a sufficient resistance to variations
in temperature. The thermal stresses induced in a material during a transient state of
heat transfer depend on the thermal conductivity, termal expansion and modulus of
elasticity of the material. These physical constants should be measured accurately and
correlated with the structure of the material. At present, very little information exists
on this subject. As a matter of fact nothing is known on the variation of modulus of
elasticity of ceramic materials with temp~rature~
The problem of thermal stresses is somewhat complicated by the fact that the
structure of the material is heterogeneous. The physical properties of the crystalline
and of the vitreous phases being different, local temperature stresses will exist
throughout the material, even in a state of thermal equilibrium. Microscopic
cracks will eventually develop in the material and will ultimately bring failure. This
phenomenon is likely to become a new type of fatigue due to temperature reversals
and will require a very detailed study.
An intermediate step in the trend toward using ceramic materials in a gas-turbine
power plant is the development of a ceramic coating on metals. Satisfactory coatings
have been obtained, and are expected to serve the dual purpose of decreasing the
corrosion rate and increasing the resistance to creep. The coating is only a few thou-
sandths of an inch thick and is not intended to serve as a thermal shield.
A further step in combining the properties of metals and ceramics is to produce
a composite turbine blade having a metallic center core and an outside ceramic shell.
In this case, the existence of a temperature gradient in the ceramic shell would allow
an appreciable increase in gas temperature without increasing the temperature of the
metalic part of the blade. The ceramic layer would act as a thermal shield, and would
not be subjected to an excessive stress. The technique of powder metallurgy has been
used for the production of such composite materials. By properly selecting the metal
and the ceramic powders, specimens have been obtained, thus showing the feasibility
of the method.
Along the same general line some investigators have proposed the addition
of ceramic powder (preferably a pure oxide) into a metal powder. The mixed pow-
ders are then compacted and sintered at high temperature. It is believed that the
creep properties of the mixed material thus obtained are superior to those of the

105
corresponding pure metal. No actual experimental data is available at present, andl
further research is necessary before any comment can be made on the chances of
success of this method.
The introduction of ceramic materials ina gas-turbine power plant will un-
doubtedly require some changes in design. These changes should take into account the
characteristic features of the ceramic materials. The shape of the blade should be
determined with special consideration for the methods of manufacturing. For these
materials, the difficulty of obtaining close tolerances is aggravated by the fact that
machining is at present very difficult, if not impossible. The method of joining the
blade to the motor will be a problem in itself. The maximum stress in service should
be reduced as much as possi ble, and stress concentrations should be minimized.
Substituting a ceramic blade for a metal blade in an existing unit would probably be
unsuccessful because it would subject the material to unduly severe conditions. In
order to assure the maximum chances of success, research on ceramic materials for
gas-turbine power plants should be integrated into the more Ileneral problem of
designing a high-temperature unit in which the advantages of the ceramic would be
exploited and its disadvantage minimized.

3. Tlte Coo'ing of Gas-Turbine Parts and ,lte Materia' Requiremen,s.

The cooling of the parts of a gas turbine exposed to high temperature constitutes
a possible answer to the problem of increasing the inlet gas temperature. If cooling
is included in the design, the material problem can be considered from a different
angle. Essential properties like creep at high temperature become secondary, and
other properties like thermal conductivity and thermal expansion become important.
Besides the conventional method of cooling a turbine blade by inside circulation
of a liquid, two less orthodox methods have recently been proposed, for which special
materials will be needed. In one of these methods the coolant is a gas flowing inside
the blade and being ejected along the trailing edge. The blade in this case is a hollow
thin-walled tube with an airfoil cross section. The second method of cooling recently
suggested consists of making the part to be cooled of a porous material, so that the
cooling fluid can be forced through it. In this case the cooling fluid may be either a
gas or a liquid. An analysis of the method and a few preliminary experiments have
demonstrated the efficiency of this method of cooling. When a liquid is used as a cool-
ing fluid, the temperature of the porous material can be maintained in the neighbor-
hood of the boiling point of the liquid without using an excessive amount of coolant.
If this technique is to be applied to turbine blades, the metalJurgist is faced with the
problem of developing porous alloys' having the required physical properties. Since
the temperature of the metal will be relatively low, the creep requirements will be much
easier to satisfy than in the case of an un cooled blade. The tensile strength of porous
metal, however, is relatively low even at room temperature, but if the cooling is as
energetic as it is expected, the blade may be kept at a sufficiently low temperature,
so that its strength is altogether greater than that of any alloy at temperatures above
1600°F. The endurance limit and the damping capacity are also to be considered.
As far as erosion and corrosion are concenred, it is probable that the film of liquid on
the surface of the porous metal will have a favorable influence.

106
 Porous metals are obtained by compafting and sintering metal powders. The
porosity is obtained by adding to the powder a substance which evaporates or burns
during the heat treatment of the compact. Such a technique is already used in the mak-
ing of porous bearings, but is limited to very few alloys. For gas-turbine applications,
the requirements will be much more severe than those encountered in the present
applications. This problem opens a very wide field of research. Powder metallurgy
is still a new science, and most of the basic phenomena remain without a correct quan-
titative explanation. Curiously enough, the fundamental problems involved here are
essentially the same as those encountered in ceramics, both sciences dealing essen-
tially with c~lemical reactions in the solid state.
While it is realized that the cooling problem is essentially a design problem, the
above considerations indicate it will give rise to extensive research in the field of
materials.

MATERIALS FOR RAMJET POWER PLANTS

The ramjet power plant is, without doubt, the simplest type of engine for jet-pro-
pelled aircraft. In contrast with the turbojet engine there are no moving parts in a
ramjet, and the problem of blade materials is therefore eliminated. There are other
material problems however, mostly due to the relatively higher temperature of opera-
tion. The gas temperature in the combustion chamber, and in the nozzle of the ramjet
engine, may be as high as 3500°F. Fortunately the pressures involved are not very high,
and as a consequence, the stresses to which the structure is subjected can be kept rela-
tively low.
At the present time, ordinary stainless steel is used in the construction of com-
bustion chambers for ramjets. This solution appears quite feasible if the temperature
of the metal sheets does not exceed some 12 00oF. In order to maintain the temperature
at this level, a certain amount of cooling of the walls of the chamber is required. This
is obtained by a circulation of cold air around the outside of the chamber. The most
important drawback in the use of steel is the difficulty in avoiding warping in service.
This fact is most troublesome in the case of units intended to operate over long periods
of time as compared with those required in aircraft propulsion. The main reason for
the uneven deformation of the steel walls is the existence of thermal stresses introduced
by a nonuniform temperature distribution. It is probable that parts of the structure
which are stressed at high temperatures yield plastically, and upon cooling a per-
manent localized deformation sets in and causes warping. Therefore, thermal stresses
should be reduced to a minimum. For a given design, these stresses depend on the
thermal conductivity, and on the thermal expansion of the material. A careful choice
of the material may therefore reduce their effect. Unfortunately, thermal conductivity
and thermal expansion are of the same order of magnitude for most of the high-
temperature-resisting alloys, and it appears that only slight improvements can be ex-
pected in the future.
The reduction of the metal wall temperature in a ramjet unit may be attained by
using a ceramic coating or a ceramic liner. Ceramic coatings have been developed
for protection of the steel exhaust pipes of aircraft engines. These coatings are rela-
tively thin (about ten-thousandths of an inch), and provide a reflective surface for the

107
infrared rays of the gas. A ceramic liner of low conductivity and sufficient thickness
would constitute a more efficient solution, but it would also be a more complicated
problem. Ceramic bodies with a sufficiently high melting point to stand the combus-
tion temperature can easily be found. The difficulty is to fabricate a large cylinder with
relatively thin walls, having a low density. Besides, such a liner would obviously be
mechanically weak, and would fail by cracking when exposed to variations in tempera-
ture. To avoid these difficulties, it might be necessary to have a liner made of separate
small pieces, each piece being strongly bonded to the metal structure. The question
of finding a suitable bond between metal and ceramic is therefore of primary impor-
tance.
The presence of a ceramic liner on the walls of a ramjet combustion chamber
might have a favorable effect on the combustion of the fuel-air mixture, because it
would maintain the walls at a relatively high temperature. This factor has not been
experimentally verified yet, and its importance is not dearly established.

MA TERIALS FOR ROCKET POWER PLANTS

1. General.
The importance of the material problem in the construction of long-duration
rocket-~otor power plants does not need to be stressed. The large amount of heat
generated by combustion at very high temperature in a small volume makes it appar-
,ent that no easy solution is to be expected. As in the case of a gas turbine, there exist
two different approaches to the material problem. The first is to find materials that
would stand the high temperatures developed in the unit, without introducing into
the design any special cooling device. The second is to try to decrease the temperature
of the materials as much as possible by means of cooling. At the present time the use
of the un cooled rocket motor is limited to applications in which the duration of
operation does not exceed from 30 to 60 sec. Its interest for aircraft propulsion de-
vices is therefore restricted to assisted-take-off units. The liquid-cooled type of en-
gine is required for aircraft power plants. Although both metals and ceramic materials
are very often used simultaneously in a rocket motor, it is convenient to discuss them
separately.

2. Metals fJIId Alloys lor Rocleet Power Plants.

In the design of a liquid-cooled combustion chamber, the choice of a metal is
primarily determined by its strength at high termperature, its thermal conductivity,
and its resistance to corrosion. The strength at high temperature does not have the
same importance it has in a gas turbine. In fact, the stress in a jacket-cooled cylindrical
chamber can be kept low, and creep at high temperatures is not often a cause of ma-
terial failure. The thermal conductivity is an essential property, and should be as high
as possible in order to maintain the inside wall temperature at a reasonable value.
The corrosion resistance is very often the decisive factor in the choice of a metal for
a liquid-propellant rocket motor. Most of the propellants used at present are highly
corrosive. Corrosion is most critical in the case of the combination nitric acid - anilne,
for which only stainless steel, pure aluminum, or chrome-plated alloys can be used.
Other propellants, like nitromethane and hydrogen peroxide are unstable when in
contact with some metals. In general, however, these factors limit the number of metals

108
which can be used with each propellant combination, but do not constitute a very
critical barrier in the development of rocket motors.
As a rule no serious difficulties are encountered at the present time in building
a liquid-cooled rocket combustion chamber, as far as materials are concerned, provid-
ing the metal temperature can be kept below a certain maximum value by a correct
cooling technique. This last requirement deserves some special consideration. Cool-
ing the combustion chamber consists essentially in removing a certain quantity of
heat from the .!petal walls by raising the temperature of the cooling fluid. In aircraft
applications, because of weight consideration, no auxiliary fluid can be carried for the
purpose of cooling. One of the propellants must therefore be used for cooling. The
maximum amount of heat which can be absorbed depends on the heat capacity of the
propellant and the maximum safe temperature to which it can be heated without
danger. In some cases, as in a nitric acid - aniline motor, either one of the propellants
can absorb enough heat without reaching a dangerous temperature. In other cases,
(nitromethane, for example) the temperature of the fluid after its passage through the
cooling coil is found to be dangerously high, as a consequence, the design of are-
generatively cooled motor having metal walls becomes a very difficult problem. The
logical solution is to line the combustion chamber with a material oflow heat conduc-
tivity and high melting point, in order to reduce the rate of heat transfer through the
walls. Ceramic materials are being considered for this particular application. The
characteristic properties of a satisfactory ceramic material for combustion-cham-
ber liners will be discussed in the next section.
The nozzle of a rocket motor is undoubtedly the critical part of the unit as far as
materials are concerned. The erosion of the throat by the hot gases flowing at high
velocity is still a rather obscure phenomenon. Studies of many nozzle failures due to
erosion have shown that in all cases the temperature of the nozzle, at least at the point
of failure, must have reached rather high values. It seems that erosion occurs only if
the surface in contact with the gas reaches some critical value. Nozzles made of very
soft material like aluminum have been used successfully when properly cooled. The
hardness of the material at room temperature is not a criterion. A sufficient hardness
at the temperature the surface of the nozzle is supposed to reach during operation seems
to be required. Here again the choice of material is related to the amount of cooling
provided.
For short-duration rocket units using solid propellant as a fuel, a solid copper
nozzle properly chrome-plated has been used successfully. Refractory metals have been
tested and molybdenum inserts in the throat of a copper nozzle are quite satisfactory.
The method of cooling by injecting a fluid through a porous material seems to
offer an interesting solution to the nozzle problem. In this case, the inside of the
nozzle is made of porous metal, and a certain quantity of one of the liquid propellants
is injected through it. It appears that the quantity of fluid required is small enough
to justify the feasibility of the technique. In the case of a solid-propellant unit, a small
quantity of water could be carried in the unit, and injected through the throat of the
nozzle. The difference in pressure between the chamber and the throat would be used
to force the water through the porous metal. The penalty in weight for carrying
the water might be counterbalanced by the elimination of the solid copper nozzle.

109
If successful, this new method of cooling will give rise to a systematic research on
the development of porous metals for nozzles of rocket motors.
In planning future research on metallic materials for rocket-motor power plants,
careful consideration should be given to the conditions under which these materials
are being used. Since cooling appears essential for long-duration aircraft applications,
the actual temperature of the metal can be maintained within reasonable limits. There
is, therefore, no need for alloys having exceptionally high strength at high tempera-
ture. Not too much emphasis should be placed on creep properties at high temperature.
On the other hand resistance to corrosion, low thermal expansion, and especially
high thermal conductivity, should be the criteria of quality.

3. Ceramic Material. lor Roc"e, Power Plan'"

The use of ceramic materials in the construction of rocket motors has so far been
very limited. Their use would certainly be justified since they combine a high melting
point with a low thermal conductivity. A ceramic liner in the combustion chamber
of lI. rocket motor would eliminate or at least simplify the problem of cooling. Por
rather short-duration applications, it is obvious that an un cooled motor is advantage-
ous from the production point of view. Por aircraft rocket power plants, an uncooled
motor has less chances of success, but in this case the ceramic liner may be used to
reduce the heat transfer to the cooling fluid. This factor is of special interest with
some propellants like nitromethane for which, at present, no regeneratively cooled
motor can be safely operated because of the danger of overheating the propellant
in the cooling coil.
The type of ceramic material required in a rocket motor may depend on a number
of factors. In general, however, it can .be expected that only very high-melting-point
ceramic bodies have a chance of success. A temperature of the order of 3600 0 P may
be taken as a lower limit for the melting point requirement. This consideration eli-
minates most of the commercially available ceramic materials. The most promising
ceramic bodies seem to be those made of pure oxides, or combinations of oxides, with
melting points above 3600 oF. Among these oxides the most interesting seem to be
those of aluminum, beryllium, zirconium, and thorium. The production of such ceramic
materials having very high melting points presents great difficulties because of the high
firing temperature required. It is not common practice in the ceramic industry to oper-
ate furnaces above 3000 o P. Test specimens will have to be pr~pared in laboratory
furnaces especially built to provide the necessary firing temperature.
To be used as a liner for a combustion chamber, a ceramic material does not have
to retain a very high strength at high temperature since it is completely supported
by the outside metal structure. The liner, however, should have a sufficient resistance
to rapid variations in temperature. This condition may be more difficult to satisfy
than that of having a high melting point. A careful study should be made of the condi-
tions-under which a ceramic material fails when subjected to changes in temperature.
When the critical factors of this problem are known, the material could be systematic-
ally improved.
Ceramic liners have already been used with some success. A zirconium oxide
ceramic, for example, can be used in a nitric acid - aniline motor for several minutes

110
without failure. Much improvement can be expected in the near future, and the de-
velopment of a satisfactory ceramic chamber liner for long-duration aircraft power
plants is not too far remote.
The problem of a ceramic nozzle is by far more difficult. The rather discouraging
results obtained so far emphasize the complexity of the question. In general, ceramic
nozzles fail by cracking, and show very definite signs of erosion. The very high rate
of heat transfer in the nozzle throat is without doubt the reason for the failure. The
thermal stresses introduced in the material by an excessive temperature gradient pro-
duce cracking. As a consequence of the low thermal conductivity of the ceramic, the
temperature of the surface in contact with the gas is very high, and probably above the
critical value where erosion cannot be avoided. The important physical properties
of the material in this problem are melting point, thermal conductivity, thermal ex-
pansion, and some less accurately defined characteristics connected with erosion at
high temperature. The relative importance of these variables should be carefully
studied in conne~ion with what is known about the heat transfer in the nozzle of a
rocket motor. This basic information will then serve as a guide in the search for
new materials for nozzle construction.

MATERIALS FOR ATOMIC ENERGY AIRCRAFT POWER PLANTS

Without knowing exactly what will be the design of the future atomic power
plant for aircraft, it is difficult if not impossible to speculate on the problems expected
in the field of materials. It is very probable, however, that temperatures at least as high
as those developed in the present rocket motor will be encountered. In the case in
which the atomic energy is used in connection with a thermodynamic working fluid
like hydrogen, the temperature in the chamber of the rocket motor is expected to be
of the order of 4000°F. It is not known yet how the heat generated by the distinegra-
tion will be transferred to the w~rking fluid. It can be expected, however, that a very
large release of energy within a small volume will produce extremely high tempera-
tures. It seems very improbable that any material will be found to withstand such tem-
peratures, and the solution may be achieved only by some methods of cooling. For
aircraft propulsion applications, for which the weight consideration is of great impor-
tance, the amount of coolant should be reduced to a minimum. The design of highly
efficient heat exchangers will undoubtedly call for special materials.
Another type of problem which appears to be of primary importance is con-
cerned with materials for protection against radiation. If atomic energy is used in an
aircraft power plant, an adequate shield will have to be provided for the pilot. Highly
absorbent materials are essentially heavy. In order to minimize the dead weight
carried for radiation protection, materials should be developed in which a com-
promise is made between the two contradictory requirements of achieving a high
absorption for radiation and a relatively law density. .
In the study of materials for atomic energy development, it will be essential to
consider the nuclear properties of the materials. The engineer will certainly have
difficulties in carrying on useful research without the help of the physicist. This rather
unorthodox approach to metallurgical problems opens an entirely new field of re-
search. There is a strong possibility that progress will be slow because of the lack
of scientific personnel having a sufficient background in nuclear physics combined
with an engineering knowledge of materials.
111
 RESTRICTED

CONCLUSIONS
The problem of high-temperature-resisting materials is one of the most essen-
tial in the future impro~ements in aircraft power plants. In the past, most of the
effort has been devoted to the development of materials for gas-turbine blades. Quite
satisfactory alloys are now available to withstand temperatures up to 1600o P. As a
consequence of the relation between the efficiency of a gas turbine and its operation
temperature, it is believed that the evolution of the gas turbine is toward higher and
higher temperatures. The physical properties of the present alloys at temperatures
above 16000 P are too low to expect any appreciable increase in oper~tion tempera-
ture. The two apparent solutions to the problem are the use of ceramic materials, or
the cooling of the blade. These two solutions give rise to new problems in the field
of materials.
Concerning special materials for ramjet and rocket motors, very little systematic
research has been done. In studying high-temperature-resisting materials, considera-
tion should be given to the special conditions encountered in these types of motors.
The approach to the problem is essentially different from that of the gas-turbine blade.
Ceramic materials and cooling should also play an important part in this program of
research.
ro make any research successful, a sufficient knowledge of the basic principles
involved in the problem is absohltely necessary. This is particularly true in the field
of materials. The physics of the solid state should be the foundation on which future
research should be based. While the metallurgist has already made efficient use of the
present knowledge of the structure of matter, its importance should be emphasized,
particularly in view of the increasing interest in relatively less explored fields like
those of powder metallurgy and ceramics. The introduction of the use of atomic
energy in engineering developments brings the final argument for the need of basic
research in the field of materials.

112