q 4..

- - 13

N,9
OQQQ9 ECTURESERIE AGRDIS 136
Rajtad arce Pouso

71--~T&\N
C.Pu I

7:7-73
 AGARD-LS- 136

NORTH ATLANTIC TREATY ORGANIZATION

"ADVISORYGROUP FOR AEROSPACE RESEARCH AND DEVELOPMENT

(ORGANISATION DU TRAITE DE L'ATLANTIQUE NORD)

AGARD Lecture Series No.136

RAMJET AND RAMROCKET PROPULSION

SYSTEMS FOR MISSILES

Accession For
SNTIS GRA&I
DTIC TAB 0
"Unannounced [p
110Just if Ictsion-..

Distribution/
"Availability Codes
:'i . ..... Avail nn /or.. .

The material in this publication was assembled to support a Lecture Series under the
sponsorship of the Propulsion and Energetics Panel and the Consultant and Exchange Programme

I AI- s A0 A -I IL I T I
 ..77-
- 2

h..

THE MISSION OF AGARD

The mission of AGARD is to bring together the leading personalities of the NATO nations in the fields of science and
technology relating to aerospace for the following purposes:

- Exchanging of scientific and technical information;

- Continuously stimulating advances in the aerospace scie-nces relevant to strengthening the common defence posture;

- Improving the co-operation among member nations in aerospace research and development;

--Providing scientific and technical advice and assistance to the North Atlantic Military Committee in the field of
i" aerospace research and development;

- Rendering scientific and technical assistance, as requested, to other NATO bodies and to member nations in
connection with research and development problems in the aerospace field;

- Providing assistance to member nations for the purpose of increasing their scientific and technical potential;

"- -- Recommending effective ways for the member nations to use their research and development capabilities for the
-* common benefit of the NATO community.

The highest authority within AGARD is the National Delegates Board consisting of officially appointed senior
representatives from each member nation. The mission of AGARD is carried out through the Panels which are composed ol
"expertsappointed by the National Delegates, the Consultant and Exchange Programme and the Aerospace Applications
Studies Programme. The results of AGARD work are reported to the member nations and the NATO Authorities through
* . the AGARD series of publications of which this is one.

U Participation in AGARD activities is by invitation only and is normally limited to citizens of the NATO nations.

The content of this publication has been reproduced

directly from material supplied by AGARD or the authors.

"Published October 1984

Copyright C AGARD 1984
All Rig~ts Reserved

ISBN 92-835-0360-0
 PREFACE

ThisXcture 6eries provided an introduction to modern ramjet technology an"9pplications to missiles were especially
tmphasized. The survey and characterization of various tyres of ramjets were followe~by the discussion of ramjet
components. Three of the lectures i n.ohasee 1otdadcealt with the main types of subsonic combustion ramjets. te ... .
, ' t 4
&as lctur
c devoted to supersonic combustion ramjets.,he experisnce gained from the research and development of
existing systems and components was covered in detail in all he lectures.

This AGARD Lecture Series was sponsored by the AG RD Propulsion and Energetics Panel and implemented by the
Consultant and Exchange Programme.
;j

I
 LIST OF SPEAKERS

Lecture Series Director: Mr B.Crispin

Messerschmitt-B61kow-Blohm GmbH
Unternehmensbereich Raumfahrt
Abt. RT 31
Postfach 80 11 69
D-8000 Munich 80
Germany

SPEAKERS

Mr J.G.Bendot Mr P.Cazin
The Marquardt Company O.N.E.R.A.
16555 Saticoy Street 29, Avcnue de la Division Leclerc
Van Nuys, California 91409 92320 Chftillon-sous-Bagneux
USA France

Professor F.S.Billig Mr E.L.Goldsmith

The Johns Hopkins University Royal Aircraft Establishment
Applied Physics Laboratory Aerodynamics Department
Johns Hopkins Road Clapham
Laurel, Maryland 20707 Bedfordshire MK41 6AE
USA UK
Mr T.D.Myers
United Technologies
Chemical Systems Division
P.O. Box 50015
San Jose, California 95150-0015
USA
 97 .1
r I"

7 CONTENTS

\\ 2'. 'Page

PREFACE iii

LIST OF SPEAKERS IV

-' Reference
INTRODUCTION AND OVERVIEW
by B.Cnispin 1

,*RAJETAIR INDUCTION ~iST DEIGN/OR TACTICAL MISSILE APPLICATION'

Sby J.GBenl-Gt, A.E.Hein~s Jr and ft.Piercy < ~"32
F-EGIE AIRFRAME INTEGRATIO /R RAMJET,4i<D
RAMROCKET IP WEEd M SILES/
$by E.L.Golds~mith '~3

041NTrEGRAL BOOST, HEAT PROTECTION, PORT COVERS Ar~<) TRANSITION,

Sby T.D.51yets s-<' I'l Z-a- 4

LES STATOREACTEURS A COMBUSTIBLE LIQUIDE

par P.Cazin 5

-SgPECIAL PROBLEMS,(OF RAMJET VH SOLID FUEL

SbyT.D&MyerS / - X/ - 6

'-SOLID PROPELLANT IIAMROCKETS

-3f. by It-L.Besser '~7

RAMJETS WITH SUPERSONIC COMBUSTION

by F.S.BiIlig 8

BIBLIOGRAPHYB
 INTRODUCTION AND OVERVIEW

by
- B. Crispin
MESSERSCHMITT-BULKOW-BLOHM GMBH
Unternehmensbereich Apparate
Postfach 801149
8000 MUnchen 80

SUMMARY

Ramjet propulsion has gained importance as sustainer of missiles which have to combine long range and high
speed. Its high specific impulse makes it favourable for this application. Moreover, military demands are
satisfied by its simple design. The lecture describes the different types of ramjet propulsion and charac-
"terizes them by explaining their main features.
After a survey of the development of the design principles, the fields of application are mentioned and
comparisons with other types of missile propulsion are made. The performance characteristics and the quali-
fication of ramjet propulsion for different missions are discussed.

The lecture concludes with remarks on the present state of development and on the points of emphasis of
further development.

1. INTRODUCTION

Modern concepts of guided missiles impose increased requirements for speed and range in order to improve the
stand-off conditions of the carrier and the penetration capability of the missile.

In many cases the requirements for engine economy are so high that they cannot be fulfilled by rocket engines.
Therefore, it becomes necessary to use airbreathing engines which offer a gain in specific impulse by about
the factor of five. Among these airbreath'ng engines, the ram-type engine is particularly well suited as a
missile sustainer because of its simple and rigid design. Compared with the turbojet engine, the ram-type
engine exhibits neerly no disadvantage in fuel consumption if the Mach number is high enough. This will be
demonstrated later.

The ram-type engine has two prominent characteristics, one being positive and the other being negative:
the positive feature is its extreme simplicity,
the negative characteristic is the lack of zero velocity start capability. Both characteristics have fascinati
engineers since the first work'of Lorin and various ramtype engines with different propellants have been
designed. In order to compensate for the lack of zero velocity start capability the combination with various
other engine types has been tried.

The multitude of systems shall not be discussed here, because this meeting will be confined to the aspects
of military applications. For practical application in military systems, mainly three propulsion systems
are of interest (Fig. 1):
the conventional liquid fuel ramjet (using Kerosene or a liquid high
density fuel)
- the solid fuel ramjet (with the solid fuel located in the ram-combustor)
the solid propellant ramrocket (using a solid propellant for energy
- I generation).
In the future, another type of ramengine will gain significance:
" the supersonic combustion ramjet.

Filure 2 illustrates the performance and adeq',,te range of application of turbojet, ram-type and rocket
engines. This figure shows the specific impulse of the particular engine type as a function of the flight
Mach number. The propellant may change from type to type, the common feature being the storability. It has
tolte underlined that only mean values of performance can be given by this kind of representation. Consider-
abl deviations will occur depending on the special design and the particular flight condition.
Pure hydrocarbons are assumed as fuel for turbojet and ramjet engines. Both engine types are almost equally
efficient above Mach number three. Below this Mach number, the performance levels of both engine types
diverge in the well-known manner. At zero speed, the specific impulse of the ramjet is of negative value
"while the turbojet reaches the maximum value. In the case of the ramrocket, a high energy propellant
containing about 50 percent boron is assumed. The performance curves of ramrocket and ramjet do not differ
substantially. The ramrocket, too, offers nearly no thrust at zero speed. Exceeding Mach number six approx-
imately, performance advantages will be gained by using the method of supersonic combustion.

This presentation demonstrates the performance superiority of the ramtype engine to the rocket engine. But
this superiority only exists in the middle of the Mach number range being considered and disappears at very
low and very high Mach numbers.

2. INTEGRATION
 1-2

generations. In the first generation, the ramjet engine was positioned within a nacelle outside the missile
dart. In the following generation, the engine already was integrated in the dart. The inlet diffuser took
an optimum position at the tip of the missile. But a considerable part of the dart volume was occupied by
the long air duct between inlet and combustion chamber. The third generation, which represents the integration
method of today, eliminates these disadvantages. The inlet system consisting of one or more inlet diffusers
is laterally installed at the dart, avoiding any loss of volume due to an air duct. In addition, the free
volume of the ram-combustor is used for installing the booster or a part of it. Thus an optimum of volume
economy has been achieved.
The sequence of generations is demonstrated once more by fig. 4 which deAls with systems being or having
been in service. The first generation is formed by the systems Bomarc and Bloodhound, the second one is
represented by the systems Talos, Sea Dart and Ganef.
By now, only one system of the third generation is in service:
Gainful (SA 6).
In the near future, a new missile system using a ramjet of the third generation will have to be added:
the French Air Surface missile ASMP
which will be the first ram-type engined system in service not belonging to the family of the surface-to-
air missiles (leaving out the drones).

3. PROPELLANTS
The most important criteria for the selection of the propellant are
the energy content
the density and
the storability.
Since the residence time of the combustion products within the ram-combustor is '.ery short (5 msec or less),
the requirement for good combustibility has to be added. For several applications, the aspect of smoke
generation may become important. In the case of hypersonic propulsion the heat-sink capacity of the
propellant has to be considered, too. The number of elements from which useful propellants may be composed
is rather small. This fact becomes obvious by a diagram correlating the heat contents and the atomic numbers
of the elements (a way of presentation used by several authors).
Fig. 5 shows the gravimetric as well as the volumetric heating values of the elements, since both are
relevant to missile propulsion. The statement of the graph is clear:
The elements with both high gravimetric and volumetric heating valuesare situated in the region of small
atomic numbers. It is evident, too, that only some light metals like aluminum, beryllium, - particularly
boron - and besides them carbon, can compete with the usual hydrocarbons. One remark should be added.
The noted heating values of the metals are valid only if their oxide produced during combustion is
condensed. Actually, this assumption may be not at all or only partly realistic. At hypersonic flight
conditions for example, the very high combustion temperatures may prevent the condensation of the metel
oxides.

4. PERFORMANCE
One of the preceding graphs showed the relationship of specific impulse and Mach number being a characteristic
of ramjet propulsion:
There is a steep increase of specific impulse in the low Mach number range, the maximum being situated at
a moderately high Mach number and a decrease in the high Mach number range. This characteristic shape of
the specific impulse curve derives from the ideal ramjet working cycle illustrated by fig. 6.
The ideal working cycle being identical with the "Brayton Circle" consists of
* isentropic compression of the ram air
* heat addition at constant pressure and
* isentropic expansion to the outside pressure.
Assuming idea., gas having constant specific heat, a simple formulation of the relationship between ideal
specific impulse and Mach number can be found. In the particular cases of the Mach number approaching
zero or infinite values, further simpllficationz are possible. Finally one can formulate:
* at low Mach numbers, the specific impulse is directly proportional
to the flight Mach number and
* at very high Mach numbers the specific impulse is inversely proportional
to the flight Mach number.
Though in reality, there are significant deviations from the conditions of the Brayton circle, the
discussed tendencies remain:
* at very low and at very 1igh Mach numbers, the specific impulse of
the ram-type engine approaches zero.
This fact is illustrated by fig. 7 for the case of a kerosene-fueled ramjet.

/ ... . ',- . ... .

 1-3

a certain change of the relationship between performance and equivalence ratio discussed above. In
addition, the range where equivalence ratio can be varied is confined by thrust re~quirements.
In the lower Mach number range, the increase of specific impulse with increaling Mach number exceeds
the increase of thrust demand. For this reason, a Mach number exists where fuel consumption of the ram~jet
per unit range becomes a minimum. By this way and taking into account additional Mach number dependent
effects to every mission, a flight Mach number can be attributed where the weight of the propulsion
system or - what is finally interesting - the eight of the missile system isminimized.
This fact is demonstrated by fig. 8 for missiles flying at very low altitude. The dependence of missile
weight is plotted as a function of the sustain Mich number for two different values of range without
taking into account the weight of the booster. There is a flat minimum near the Mach number of 2.5.
The consideration is completed by taking into account the change of booster weight. The results of calculation!
are shown in fig. 9. Again the low altitude flight mission is taken as an example.
Parameters of this diagram are two different ran~cs and two different
launch Mach numbers. Again, there is a flat minimunt no-o
near Mach number 2.
The position of the minimum is influenced to a certain amount by'several other parameters as for example
'y the flight altitude. But these influences are of second order. Generally, it can be stated that the weight
optimal Mach number lies between 2 and 2.5.

5. OPERATION
The off-design operational characteristics of the ram~jet or ramrocket are primarily governed by the interactio
of the inlet system and the sonic throat of the thrust nozzle. inthe case of the engine with constant geometr
the operation is influenced by the variable flight conditions (Mach number, altitude, angle of attack or
sideslip) and the rate of heat addition in the combustor.
The matching of inlet and thrust nozzle is illustrated by fig. 10. Some simplifications have been made in
order to concentrate on the dete~rmining factors: the changes of combustion chamber pressure loss and of mass
flow due to fuel addition have been neglected.
It is assumed that the relation between total pressure zt the end of the diffuser and air mass flow
can be described by the well-known rectangular graph. A two b:-.ck inlet with external compression and
- ~stable subcritical operation will have a characteristic similar tL. this.
The critical or supercritical air mass flow of the diffuser is linear proportional to the flight Mach
number during operation above the "shock-on-lip" Mach number. The dependence is of higher order below
"shock-on-lip" conditions. (The flight altitude isalways assumed to be constant.)
The total pressure at the entrance of the thrust nozzle isdefined by the values of air mass flow, cross-
sectional area of the sonic throat and combustion temperature. The dependence from mass flow is nearly
Wo linear if the combustion temperature is constant. In the case of constant heat addition (constant fuel
flow, the dependence from mass flcw can be approximately a square root function.
The synopsis shows a set of similar diffuser characteristics with the flight Mach number being the curve
parameter. The critical total pressure of the diffuser strongly increases with the mass flow. The total
pressure, necessary for swallowing the air mass flow has a less steep ascent with the increase of mass
flow.

At the point of intersection, the engine is operating at critical diffuser conditions. Going to higher
Mach numbers, the diffuser will become progressively supercritical. In the opposite direction, the
engine works with subcritical diffuser, the pressure level being defined by the maximum pressure recovery
of the inlet diffuser. (Inmost cases only part of the subcritical region can be used determined by the
l~imits of stable operation of the diffuser.)
After these statements introducing to the matching problem of inlet and nozzle, the thrust characteristics
of the ram-type engine shall be explained.
The equations describing thrust and thrust coefficient are formulated'in fig. 11 in a notation which
facilitates further transformation. The simplifications introduced before, for the reason of clearness
shdll be maintained. As an additional simplification, only operation above the "shock-on-lip" Mach number
will be considered.
At first, critical operation shall be discussed (fig. 12). The relationship between critical total
pressure of the diffuser and flight Mach number can be described by an exponential function of the Mach
number. The exponent isbetween 2.5 and 3.5. According to the simplifications made above, this total
-pressure is identical to the total pressure within the thrust nozzle. Since the static pressure at the
nozzle exit is discriminated from the total pressure only by a constant factor, the above mentioned
exponential function finally is appropriate to describe the dependence of nozzle exit pressure, too.

By choosing a medium value of three as exponent, the simple relationship given by fig. 12 can be
established. Mainly there is a linear proportionality at critical operation between the increase of
flight Mach number and thrust coefficient.
In the supercritical regime, the relationship between thrust coefficient and flight Mach number at constant
 1-4 ,

flow. With this simplification at first the statement is established that for constant heat addition
the nozzle exit pressure is proportional to the square root of the air mass flow. Because of the assumption
to operate above the "shocK-on-lip" Mach number, the air mass flow is directly proportional to the Mach
number. Finally, the expression of fig. 13 follows, relating the thrust nozzle exit pressure to the
square root of the Mach number. By using this result one gets the simple approximation of the supercritical
thrust coefficient at constant fuel addition given by fig. 13.

The results of exact calculation of the thrust coefficients of a rimrocket are presented by fig. 14 a.

The thrust coefficient at critic&l operation rises steeply with increasing Mach number. The curves of
the thrust coefficient at supercritical operation with constant fuel addition are mildly bent and decline
in the direction of higher Mach numbers. As long as the inlet diffuser is stable, the subcritical region
can be principally used. The calculated values of the thrust coefficient are indicated by the dotted
lines.

In fact as shown by fig. 14, the pre-entry-drag of the diffuser strongly increases because of the normal
shock spillage in the subcritical regime. As a result of this, only a marginal gain in the net value
of thrust coefficient is attained. Besides, fig. 14 b gives information about the minimum operational
Mach number of a ramjet. For that reason the drag coefficient of the missile (represented in the same
manner as the thrust coefficient) is presented, too. If subcritical operation has to be excluded because
of insufficient diffuser stability, the intersection point of the lines of critical thrust coefficient
4
and of drag coefficient designates the m iimum operational Mach number. But even if subcritical operation L
can be allowed, no remarkable change follows. This minimum operational Mach number is only a theoretical
figure and the transition Mach number has to be well above this valve if the flight mission requires
maneuvering or climbing in the early sustain phase.

An impression of the dependence of the acceleration capability on the margin between transition Mach number
and minimum operational Mach number is given by fig. 15.

This graph referring to a long range missile shows the acceleration (as multiples of the acceleration
due to gravity) as a function of the time after transition. For this calculation, the assumption has
been made that the transition Mach number exceeds the minimum operational Mach number by a value of 0.05.
It is obvious that a certain value (being nat too small) of this margin has to be provided because of the
dynamic characteristics of the missile.

6. COMPARISON OF RAMJET PROPULSION WITH OTHER TYPES OF MISSILE PROPULSION

6.1 Comparison with Rocket Propulsion

Ramjet propulsion offers the advantage of low fuel consumption compared with rocket propulsion, but the dis-
advantage of additional structure weight has to be accepted. Therefore, the range of the missile has to ex-
ceed a minimum value in order to justify ramjet application. The minimum range where the propulsion system
weight of a ramjet engine corresponds to a rocket engine, essentially depends on ambient pressure (flight
altitude) and Mach nurn.ier. Since the minimum range is approximately reciprocal to ambient pressure, the
simple manner of representation of fig. 16 becomes possible.

This figure shows the product of minimum range and ambient pressure as a function of the flight Mach
number. Basis of comparison is:

0 a solid propellant rocket having a specific impulse of about 240 sec

and
* a hydrocarbon-fueled ramjet or a ramrocket using a high energy propellant.

Since the dimension of ambient pressure is "bar", this graph directly informs about the minimum range
of ramjet application for low altitude missions.

Of course, the statement of fig. 16 is only of theoretical nature. In practice, the range has to
exceed these theoretical values cnsiderably in order to justify the change from rocket propulsion
to the more complicated ramjet propulsion.

An impression of the margin of propulsion weight which can be saved by using ram-type engines is given
by fig. 17.

Fig. 17 indicates the propulsion weight per unit of range as a function of range at sea level
conditions. A solid propellant rocket and ramrocket using a high energetic boron containing
propellant are compared. Similar results will be attained if the ramrocket is replaced by a
hydrocarbon-fueled ramjet. The calculated values apply to a missile of 0.1 m2 crosz-sectional area
flying at sea level conditions with Mach number 2.

The advantage of using ram-type propulsion for medium or long range missions is clearly indicated
by the graph. It must be added that only the weight of the sustainer propulsion module has been
considered. If boost progulsion is taken into consideration, the gap between the weight characteristics
of both propulsion types will increase further.

The influence of the Mach number on this comparison is illustrated by fig. 18, Pow referring to the
aspect of volume requirement of the propulsion system.

In this_9 raph, the length of the propellant section needed per unit of range, is shown as a function
 777 7-77 77' *

1-5

The statement is: The application of ram propulsion is the appropriate way to reduce missile length
at supersonic flighc missions.
6.2 Comparison with Turbojet Propulsion
In the subsonic and transsonic region, there is a clear superiority of turbojet propulsion with regard
to both fuel consumption and thrust coefficient. As far as fuel consumption is concerned, this superiority
lasts up to a Mach number of about 3.5. But in the case of missile application the required thrust
coefficients generally exceed by far- the values which can be provided by a pure turbojet. The extreme
thrust requirements only can be fulfilled by a turbojet if it is equipped with an afterburner. In this
case the afore-mentioned superiority with respect to fuel economy diminishes at a Mach number of about
2. On the other hand, a lot of penalties (weight, complexity, problems concerning the accommodation of
* - an integral boost motor and so on) arise for the application of the turbojet engine.
Situations may exist where advantages can be drawn from the self-acceleration characteristics of the
turbojet. But in general, an additional boost motor is needed providing a sufficient thrust level to
accelerate the missile as fast as possible to its sustain speed, thus minimizing the minimum operating
range.
Keeping in mind these considerations in the case of missile applciation nearly no argument remains for
using a turbojet instead of the much simplier ramjet if the flight Macn number exceeds a value of about
1.8 (even a lower value can be assumed).

7. FINAL REMARKS
At this time, development work or at least technology work in the field of rairjet or ramrocket propulsion
is going on in several countries of the NATO conmmunity. But still there exist some reservations against
the application of ramjet propulsion in the mind of the system engineers. This may be due to the following
reasons:
* The development period is long (a period of five years has to be anticipated
nowadays)
s Development and production costs are high. (Some components and the equipment
are expensive inmanufacturing. Compared with rocket propulsion
ramjet testing is very expensive.)
* Some additional restrictions have to be accepted by the designer of the
missile. (The restrictions concerning usable Mach number range, altitude
range, margin of angle of attack or sideslip, and minimum launch Mach
number are more severe as in the case of rocket propulsion.)
Some of the just mentioned facts are inevitably inherent in airbreathing propulsion and have to be
accepted in return of a higher efficiency. But the remaining problems could be overcome by a better
knowledge of the component technology leading from the variety observed now to more standardized solutions
of component design and construction.
With respect to this, an improved cooperation between the different developing companies and institutions
within the NATO cormmunity could be very helpful. Additional efforts have to be made in the following
fields in order to improve the prospects of application of ram-type propulsion:
- inlet stability
- sensibility concerning angle of attack and sideslip
- combustion stability
- durability of heat protection
- high density fuel maragement
- particle combustion efficiency
- throttling mechanis~m for solid propellants
- consumable port covers and booster nozzles
- supersonic combustion of storable propellants
This list does not claim to be complete, but some of the topics are appropriate to be subject of a cormmon
technology work of the NATO countries.

Nomenclature

A Cross-sectional area Mass flow

C0D Drag coefficient P Pressure
CF Thrust coefficient q Dynamic Pressure
F Thrust R Range
H Altitude S Entropy
I Enthalpy T Temperature
Isp Specific impulse t Time
M Mach number Equivalence ratio
1-6

Subcripts: Supercripts:

0 Free stream cond. 0 Total

D Design 4
f Fuel

REFERENCES

General references are listed in the bibliography prepared by the AGARD TIP (Supplement of L 136).

Pohl, W.D. Flugleistungsaspekte bei Stauantriebeni,

MBB-Bericht UR-699/84, 1984

Crispin, B. Betra.htungen zum Gesamtsystem eines Staustrahl- und eines Kombinationstriebwerks

MBB-Bericht UR-226-74, 1974

Voss, N. Verbrauchsoptimale Marschflugmachzahl

MBB-TN RC1-32/73, 1974

Weinreich, H.L. Overview of propulsion systems and related fluid-dynamic aspects

AGARD-VKI LS No. 98, Supplement, 1979

Besser, H.L. Investigation of bo;'on solid propellant combustion on ducted rockets

AGARD Symposium on Ramjets and Ramrockets for Military Applications,
26-29 Oct., 1981

Sarner, S.F. Propellant Chemistry

Reinhold Publishing Corpcration, N.Y., 1966

(liquid fuel ramjet) Specific Impulse of various propulsion types

diffuser efficiencies 71KE=0.9 Isubsonw combust.,on I
-7KE= 0.95 1supersorc comb~,s on I
xyo ramrocket. high energetic
propellant
-- turbo rocket M86 fuel

(solid fuel ramjet) 35000

30000 - corde

m t~u Iojetl:.

20000 '- - -- -.

(solid propellant ramrocket)

15000 .r...-(CRJ)

Fig. 1 - Types of Ram Propulsion ,

1000- omjet (SCRJ kerosone L_

0
-

2
-*---

1. 6
5z~1W;Izl
ymclj
8
1-

10 12 14 1
Mo ----
 Fig. 3
En'gine Inte~ration and
Booster Instalkptlon

2. GEDRATION

SOOSTEft

SOVAC ?5O K6 U4

Fg42 TALOS GMEF -

Ramjet History .2000 KG IS G,

GAIxFuL.-
SA6
550 KG
SEA DkART- -6

-S G KK

1. GENERATION 2.GNRTO .GNRTO

 1-8

4 GravitietriC Hleating Value

/fJ/g 7

100.000

50.000 L
C ~~~~Fg AlS5
iVC
S0.0 Elements()Uz

I sN

N-N

w~ ~ ~2 W L N-oriD

(2)N 5 ,x 0 ( . __ Fi-g.

I tIT"
Volumetric
Valu Heain
 1-9

00
2000C
0,13 Fig. 7

Specific Impulse of a
.. 0,5 Hyd rocarbon-Fueled
(M ] Ramjet
"" 15000 -

,oooc

' Q9
"KE
H ..I I km

' "5 00 0 .
0 3 2 3 5 6
MO -

S~800

Fig. 8 r_

"Weight of a Ramjet Powered

Missile as Function of
Sustain Mach Number
(Booster Weight Excluded)

"" ~~~ ~~600 .... ...

"20 2,5
M".
* 1-10

1200 - __ _
TRange 100 km I
Launch Macm Numbe

1100 Fig. 9
kg Launch Weight of a Ramjet Powered
_______ ______ IMissile as Function of Sustain Mach

70Okm;O0 Number, H = 0 km
9 0 0 - _____(Booster Weight Included)

2P0 2,5
M -----

inlet Diffuser- M const.

Fig. 10 Thrust Nozzle

Macigof Inlet and

cOnst. comb. temp.

*i \ubcritical /
critical

const. fuel flow /

_____________________________supercritical
rnh nmm 5

Operation Mode

subcrlitcal supercritical
0
Pp

critical total
~,~M~M
3 pressure of diffuser

total pressure of
nozzle
thrust .fuel
(const flow)

Mmi H const.
AlM
 1-Il

Thrust:
r - As1"M')" P, AoXM' -Po "A
P
C p

q MoI

Thrust Coefficient:

CF MI. X5
Po( x-- X;

Specialization:

o Const. Geometry - a M5 - const.

No ) MD - Ao * const. at supercritiCal cond.

," Simplification:

2 Supercritical Operation

Critical Operation Specialization: const. fuel flow

const. flight altitude

Assumed dependence of inlet total pressure from flight

"Nachnumter at const, altitude: P3. m5
m5p ' Mo ( No MO)
P2 2.5 to 3.5const.)

71 : o i o T - r" (simplification)
P55
'P5 const. (const. geom.) pi

P5 Mn0 P5. const. (const. geom.)

P5
Using n -,3 PO f"

l F ost I Mo const.2 M.CF

[ostconst.
=const.2
2

C, /

/ /
1-12

7 ~ A- critical operation

Fig. 14 a

1.6 1 3 . , , , , . ,

0.4

16 1,8 1.8 26 1.9,

28 . ,

MO

Fig. 15

Acceeraion ascto
Fucto ofcTime.r
Supercitigh
Inlet:, Margin of
3etHtrs
0kkm H

C, 0.0.5

0.0 p10 20 30c.

I -dIv~
 1-13

12
R2T Fig. 16
Po'R 10 % RAMJETS TS
RAMROCKE Regions of Application of Ram-
bar. km] 8 rockets, Ramiets, and Rocket
_ Propulsion Systems

ROCKET
2 PROPULSION -

SYSTEMS
0
12 H ---- 34

MI

15 I I Mo 2.0

R ISsolid prjpellant rocket H

Fig. 10
17
10 ?m 1Unit Propulsion Weight per
of Range

high energy ramrocket

-/I.

0 20 40 60 80 100 120

LENGTH! OF FUELSECTION_
PER UNIT RANGE

Awl SOLIDPROPELLANT
[] ROCKET

ROKE Fig. 18

o
/-
'"''"

--
"SOLID
"" " ""-,,.RAMROCKET

2"_.____,_-_
1MEDIUM
PqOPELLANT

LIQUIU FUEL
ENERGETICI

RAMJET EGINE
H0 DENSITY FUEL)
IHGH

"1'"_-e BORON RAMROCKET

0,0, 1.0 1,5 2.0 2,5 Me

FLIGHT MACMNUMBER
 r 777 7 -77 7. 7711-

2-1

RAMJET AIR INDUCTION SYSTEM DESIGN FOR TACTICAL MISS'LE APPLICATION

J. G. Bendot
A. E. Heins, Jr.
T. G. Piercy

THE MARQUARDT COMPANY

16555 Saticoy Street
Van Nuys, California, 91409, U. S. A.

INTRODUCTION/SUMMARi

Active development of the liquid ramjet engine was undertaken by the United States in the early 1950's.
Successful development of the BOMARC and Tales long range surface-to-air interceptor missiles was
accomplished in this time period with initial operational deployment taking place in 1957 and 1959,
respectively (Figure 1). Shortly thereafter, similar surface-to-air systems were developed and
operationally deployeo by the United Kingdom and the Soviet Union. The United Kingdom developed the
Bloodhound MKI and Bloodhound MK2 missile systems (Figure 2). Later it developed the Sea Cart shipboard
based surface-to-air missile (Figure 2). During the same time frame, the Soviet Union developed the
SA-4 GANEF mobile surface-to-air missile (Figure 3).

From the inlet designer's viewpoint, it is instructive to note that all of these missile systems
incorporated an axisymmetric air induction system into the nose of the vehicle or the nose of externally
mounted nacelle or pod mounted engines. By today's standards, the design of such inlets is quite
straight forward, well documented and development undertaken with little risk.
In the 1960/1970 time period the Integral Rocket-Ramjet (IRR) type of engine was successfully developed.
The operating principal of the IRR engine is described in Figure 4. Flight demonstrations were success-
fully accomplished in the following programs:

PROJECT CROW 1961

LOW ALTITUDE SUPERSONIC RAMJET MISSILE (LASRM) 1965/1966
ADVANCED LOW VOLUME RAMJET (ALVRJ) 1974/1976
SUPERSONIC TACTICAL MISSILE (STM) 1979

It is noted that each of the flight tested IRR systems (with the exception of the CROW test vehicle)
utilized a multiple aft mcunttd inlet system (see Figure 5). Furthermore, the Soviet Union developed
SA-6 GAINFUL surface-to-air missile system also utilizes a multiple aft mounted inlet system.
In recent years, we have seen a growing interest in the development of advanced tactical missile
systems. Specifically, long range air-to-surface, air-to-air and surface-to-air systems are of
interest. The generalized missile/propulsion system requirements for these missile systems are:

0 Long Range
* High Speed
* Demanding Launch Airiraft Compatibility Requirements
* High Packaging Density (Missiles are typically volune limited)
* Low Cost

These design requirements are best satisfied with use of ramjet propulsion. Both liquid and solid fuel
engines are of interest.
Four significant design/mission characteristics of these proposed new missile systems are of major
interest to the ramjet inlet designer:
1. The vehicle forebody will house the missile guidance system, autopilot dnd payload. A nose
inlet vehicle is no longer a viable design solution. (Missile systems specifically
designed for storege/firing from a shipboard eased box launcher may be an exception to this
statement).
2. The ramjet air induction system must be highly integrated with the vehicle to result in an
aerodynamically efficient, minimum weight design.

3. Integral rocket-ramjet engines are emphasized.

4. The vehicle flight profiles are typically highly transient in terms of flight Mach number,
altitude and angle of attack. Particularly in tha air-to-air applicaticn, vehicle maneuver
requirements dictate high angle of attack operation (see Figure 6).
The above described design/mission requirements were initially vict;ed with concern by experienced inlet
designers. They reasoned the following consequences of these requirements:
 2-2

1. Installed inlet performance would be degraded as compared to nose inlet systems.

2. Successful development of these inlet systems would be more difficult and therefore require
a longer, more expensive inlet development prcgram.

The United States has conducted inlet oesign/performance studies for a number of candidate advanced
ramjet powered missiles. By utilizing favorable interference design concepts (vehicle forebody and/or
wing flow field effects, precomp.-ession shrouds) and compromises in inlet design parameters (design Mach
number, contraction ratio, bleed flow) excellent installed inlet performance has resulted. In fact,
installed inlet performance is often significantly improved when compared to nose inlet systems,
especially at high angle of attack flight conditions. The inlet design concepts/trade-off paremeters
* considered in these programs will be discussed in this lecture.

DESIGN/PERFORMANCE TRADE-OFF PARAMETERS

The inlet designer must properly consider/evaluate (analytically and/or experimentally) a number of
design trade-off parameters to configure the air induction system which best meets specific vehicle/
system requirements. These inlet design parameters are discussed in the paragraphs which follow. To
facilitate these discussions, the air induction system (inlet) station notation and nomenclature used
throughout this paper is presented in Figure 7. In addition Figure 8 presents representative inlet
performance curves. The classical PT2/PT_ versus AI/AC iniet performance curve is often
referred to as a *candy cane" curve because of its characteristic shape. Figure 8 is representative of
an inlet which employs boundary layer bleed. With such an inlet the inlet airflow definitions
(A, /AC)Total and (Am /AC)Plenum are required.

Inlet Type

Typically, three inlet types are evaluated for application to advanced ramjet powered tactical missile
systems. These inlet types are as follows:
Axisymmetric
Two-dimensional (2-D)
Chin
Representative performance for these three inlet types is presented in Figure g. As shown in this
S
.... figure, axisymmetric inlet performance deteriorates rapidly with increasing angle of attack. In contrast
to the axisymmetric inlet, the two-dimensional and chin inlet designs exhibit increasing inlet perfor-
mance (total pressure recovery and engine airflow) with increasing angle of attack. These performance
characteristics are well matched to advanced tactical maissile system requirements.

Can the angle of attack performance of the axisymmetric inlet be improved? Yes by locating the inlet in
the favorable flow field generated by the vehicle wing. Using this design technique, axisynmmetric inlet
performance closely approximates 2-D inlet performance (Figure 10). However, it must be recognized that
this oesign approach has a serious limitation. The vehicle wing/inlet geometry established by wind
tunnel testing must be maintained. This significantly increases the problem of meeting vehicle stability
and control requiremeits. These requirements are significantly easier to satisfy if the wing and inlet
locations can be treated as independent variables.

If a 2-D inlet is located in the wing flow field, is its performance improved? The test data trends of
Figure 11 answer this question. Mid-wing performance is approximately equal to the no-wing data.
However, a bottom mounted wing improves inlet performance. Location of the inlet between cruciform
wings further improves 2-D inlet performance. While this increased performance Is highly desirable, the
fixed wing/inlet location problem discussed above presents the vehicle designer with serious stability
and control design problems.

Is there another way to improve 2-D inlet performance with
precomnpression shroud can significantly increase 2-D inlet
conditions (see Figure 12). The test data trends shown in
design which emphasized ramjet takeover thrust/drag margin
conditions.
increasing angle of! attack? Yes use of a
performance at angle of attack operating
Figure 12 are for a, precompression shroud
and high Mach numbe /high altitude cruise

Can the precompression shroud design concept be applied to the axisymmetric inlet? Experimental studies
conducted by NPCA (National Advisory Committee for Aeronautics) and NWC (U.S. Navy - Naval Weapon
Center) were disappointing.
For the reasons discussed above, modern high performance ramjet powered tactical missile designs
typically incorporate a chin (see Figure 13) or two-dimensional inlet system. Use of a highly
integrated two dimensional inlet is presented in Figure 14. Shown is an inboard profile drawing of a
candidate ramjet powered long range air-to-air missile design. Axisymmetric inlets are specified where
low development risk, low weight and low manufacturing cost are major vchicle design drivers.

Ramjet powered missile systems specifically designed for storage/firing from a shipboard based box
launcher or a submarine torpedo tube may be special cases. The Johns Hopkins University/Applied Physics
Laboratory conducted a design study for a SCRAIJET powered shipboard launched interceptor missile
(reported in AGARD publication AGARD-CPP-307). Rectangular and round launchers were studied. The study
considered the following inlet options: (1) an axisymmetric nose inlet; (2) a chin inlet; and
(3) a single bottcm mounted aft inlet. Missile storage constraints coupled with performance
 2-3

requirements resulted in selection of a nose inlet system. Inlet cowl area limitations -were a major
factor in this conclusion. In contrast, several studies have shown the chin inlet to be compatible with
ram~jet powered missiles designed for firing from a submarine torpedo tube. The inevitable conclusion
drawn from these studies is that the design of ramjet powered vehicles which are stored and/or fired
from severely constrained launchers must be carefully structured to insure compliance with system
performance requirements. Radar dish size and bank-to-turn versus skid-to-turn flight control must be
considered.
Numbter of Inlets/Inlet Location
Multiple (four) aft-mounted inlet systems have been successfully employed in three ramjet propulsion
flight research test vehicles and one operational surface-to-air missile (see Figure 5). This type of
air induction system functions well when Vehicle angle of attack/angle of yaw requirements are minimal.
These are typical operating requirements for supersonic/low altitude missile systems. However, as
discussed earlier in this paper and illustrated in Figure 6, today's advanced ramjet powered tactical
missile systems (particularly air-to-air missiles) emphasize high angle of attack operation during
guidance handover and terminal engagement maneuvers. Under these operating conditions, the performance
of a multiple (four) aft mounted inlet sys~i-em is poor as illustrated in Figure 15. Consequently,
advanced high performance tactIcal missile designs typically emrploy one or two inlets. Furthermore, if
a single inlet is used, it is usually bottom mounted. High inlet performance has been demonstr~ated with
forward (chin) and aft bottomr mounted single inlets. If two inlets are used, they are mounted
approximately 450 from the hor17ontal centerline. The arguments for these inlet locations is the test
data trends presented in Figures 16 and 17. Photographs of a high performance chin, aft bottom mounted
single and aft mounted dual inlet wind tunnel test model are presented in Figures 18, 19 and 20
respectfully.
Design Mach Number
The inlet design Mar'h number not only strongly influences design point performance but also off-design
performance. Figure 21 presents the classical maximum total pressure recovery versus design Mach number
plot. Although the total pressure losses of external compression inlets decrease as the number of
oblique shocks is increased, it has been found in practice that the use of discrete oblique shocks in
excess of two offers greater flow complication and less satisfactory results than Wisentropic" type
inlets.
Normally, ramj,2t inlets are fixed geometry and designed for a single Mach number. Therefore, these
inlets must accept the performance penalties associated with flying above and below design Mach number.
What is the effect of design Mach number on off-design performance? This effect, at zero degrees angle
of attack, is presented in Figure 22.
Low desicn Mach number inlets have higher capture area ratio (engine airflow) below the design Mach
number and lower pressure recovery above the design Mach number when compared to high design Mach number
,nlets. The experimental axisyimietric inlet data presented in Figure 23 confirms these trends. Low
design Mach number inlets emphasize ramjet 'takeover performance whi~le high Mach number inlets emphasize
high~ vehicle flight speed. However, it is important to recognize these performance trends can be
significantly altered at moderate to high angle of attack flight conditions. Typical performance trends
for isolated two-dimensional, mixed compression (internal and external) inlets at d moderate angle of 1
attack are presented in Figure 24. The trends are for inlets designed to be near the maximum amount of
external and internal compression at the design Mach number. In the middle Mach number range, the low
design Mach number inlet can have both higher pressure recovery and capture area ratio. This Is because
the higher design Mach number inlet, at angle of attack, operates unstarted and with or without detached
external shock waves over a large portion of the Mach number range. Therefore, selection of the inlet
design Mach number must carefully consider all pertinent inlet design requirements and is a significant
step in the definition of a new inlet design.
Contraction Ratio
Increasing the amount of external contraction increases the overall pressure recovery. This is
accomplished by increasing the Inlet initial cone or wedge angle and/or the amount of isentropic
compression. Ultimately a "compression limit" is reached which precludes comipression (externally) to a
pressure higher than can be achieved behind a strong shock. The reason for this is shown ir.Figure 25.
Static pressures must be balanced across the vortex sheet. This limits the degree of external
compression to the values shown in the figure. Generally speaking the consequences of designing an
inlet which exceeds the comprc-ssion limit is to spill air (decreased thrust and increased drag) at high
flight Mach numbers (Figure 25). Highly compressed inlets result in large flow angles at the cowl lip.
If the cowl lip isflow aligned or positioned at some fixed angle relative to the approaching flow, then
cowl drag increases with contraction ratio (cowl lip drag can be decreased by misaligning the cowl but
this usually requires inlet bleed with associated bleed drag). The optimum amount of external
contraction at a given Mach number can be found by computing net engine thrust (trading press~ure
recovery against cowl drag). The elements of this optimization analysis are presented in Figure 26.
Such a design optimization isoften accomplished for the ramjet takeover flight condition.
High performance aircraft typically employ variable geometry mixed compression inlets. Mixed compression
inlets employ external and internal contraction. At first examination the use of a large amount of
internal contraction would appear to benefit ramjet inlet performance. However, such is not the case
for two reasons:
1) Variable geom~try inlets are heavy and complex. Numerous ramjet design studies which
examined the use of variable geometry inlets showed little or no payoff, primarily due to
their added weight. However, hypersonic ramjet/SCRAMJET engines designed to operate over a-
 2-4

2) Ramjet engines are typically designed to operate over a large range of flight Mach numbers
and angle of attack. With a fixed geometry inlet and these demanding flight operating
conditions, only a small amount of internal compression can be utilized (I10%
contraction). Therefore, the maximum total pressure recovery performance presented in
Figure 21, which strictly speaking applics to external compression inlets, also generally
applies to fixed geometry ramjet mixed compression inlets.
Figure 27 illustrates the effect of inlet contraction on an installed bottom mounted, two-dimensional
inlet at angle of attack. It should be noted at high angles of attack, the low contraction ratio inlet
can be better in both pressure recovery and capture area ratio. The importance of establishing realistic
inlet design requirements early in the missile development program is apparent.

Bleed Airflow Rate

Typically, inlet bleed flow can improve the total pressure recovery of an inlet but at the expense of
drag. Therefore, in most cases, there is a bleed flow rate (and/or configuration) that produces maximum
net engine thrust. Such a bleed flow rate/configuration optimization is usually accomplished experimen-
tally for the critical ramjet tekeover flight condition. The performance trends shown in Figure 28 are
from experimental data for a bleed system with high bleed momentum recovery..

Use of an Aerodynamic Grid

r i During the flight test development of the BOMARC liquid fuel ramjet engine serious subsonic diffuser
flow separation problems were experienced, especially at large angles of attack. These flow separations
resulted in poor diffuser exit airflcw and, therefore, combustor entrance fuel-air ratio profiles. The
* result was poor combustor performance and often combustor blow out. A device known as the aerodynamic
grid solved this problem.

SThe aerodynamic grid is designed to choke as the rarajet engine is throttled to lean fuel air ratios.
This limits the downstream travel of the normal shock system into the subsonic diffuser - the cause of
F diffuser flow separation. The choked grid also redistributes the diffuser airflow much as a screen.
However, because of its carefully contoured shape, the effect on critical inlet recovery is small
(4PT2/PT2- .02).
Figure 29 describes several types of aerodynamic grids tested in development ramjet engines. The round
hole grid and the annular grid have found the most use.

Subsonic diffuser exit flow profile test data, with and without the use of an aerodynamic drag, are
presented in Figure 30. These test results speak f-r themselves.
The aerodynamic grid has been used in a number of operational and flight test ramjet engines. Examples
are BOMARC, Talos, Typhoon, Bloodhound, and ASALM. A grid will not be required if the design can
tolerate shock induced separation in the subsonic diffuser without adverse effects on the structure
(back burning) or on combustion efficiency.

Subsonic Diffuser Design

The design of a high performance subsonic diffuser is strongly influenced by flight vehicle geometric
constraints such as available length and combustor offset. "S"shaped diffusers are often employed in
ramjet powered flight vehicles and are particularly troublesome. Good performing diffusers (straight
and *S" shaped) have been developed using an effective conical flow expansion angle of 3 to 5 degrees.
Rapid inflections in diffuser wall contours should be avoided. If subsonic diffuser flow separation is
experienced the use of an aerodynamic grid will often greatly improve diffuser performance.

Inlet Drag
Inlet drag plays a strong role in the design and development of a high performance ramjet inlet/
diffuser. Mr. E. L. Goldsmith of the Royal Aircraft Establishment will address this importa.t subject
in his lecture.

Subcritical Stability

Figure 31 presents total pressure recovery/capture area ratio characteristics for a high performance and
a low pe-formance inlet design. Both inlets have the same supercritical A/AC. As is often true in
practice, the low performance inlet (PT2/PT ) has subcritical margin while the high performance
inlet does not. The constant slope lines ( 2/PT /A. /AC) represent lines of constant ramjet
combustor heat release or fuel-air ratio. Since it is undesirable to permit the inlet to buzz, a high
performance inlet is usually required Lo operate at some margin (5% to 10% in pressure recovery). This
requirement is due to accuracy or repeatability in test data or changes ir.aVlospheric conditions. This
"requires a pressure recovery limiter control. However, as can be seen in the figure, a lower
- performance inlet with sufficient subcritical stability can provide the same margin (or more), thus
eliminating the need for a pressure recovery limiter control. This subject will be further discussed in
the second portion of this lecture.
Vehicle Forebody Effects

For a given inlet geometry and location, the shape of the vehicle forebody can have a significant effect
on inlet performance. Whether this effect is large or small depends on inlet size, inlet type and
axial, radial and circumferential location. For completeness, it is pointed out that selection of the
vehicle forebody shape is usually dictated by airframe packaq!nq or radome slope error considerations.
 2-5

Typical vehicle forebody shape effects on inlet performance are presented in Figure 32.

An important aspect of vehicle forebody effects is the definition of the local flow field delivered to
the inlet. Rapid advances are being mad2 in the modeling and computation of these flow fields. This
important subject will also be discussed by Mr. Goldsmith.

Comparison of Wind Tunnel and Flight Test Inlet Data

The primary tool used to develop a high performance ramjet inlet is the high pressure (high Reynolds
number) blow down type of wind tunnel. A question often asked is "How well do wind tunnel and flight
test inlet data compare?" The test data typically compare quite favorably as demonstrated in-
Figure 33.

Additional Inlet Design Considerations

In addition to the major inlet design parameters discussed above, several other factors may significantly
influence the air induction system design process. These additional considerations are listed below:

* Vehicle/Engine Radar Cross Section

* Air Induction System Impact on Vehicle Aerodynamics
Inlet Pressure Margin
* Combustor Entrance Profiles
* Combustor Pressure Oscillations

* Inlet Fabrication Cost
* Inlet Weight
RECOMMENDED INLET DEVELOPMENT PLAN

The Marquardt Company has conducted inlet design!performance studies (including inlet model wind tunnel
testing) for several advanced ramjet powered missiles. In the conduct of these design/development
studies, a preferred or recommended inlet development plan h~s emerged.

The basic elements of this plan are presented below:

1. Establish inlet performance/design requirements.

2. Perform computer-aided inlet analysis and design studies.
3. Conduct inlet/vehicle system performance trade-off studies.
4. Conduct inlet wind tunnel test program:
e Installed Tests
e Isolated Tests
Flow Field Tcsts
5. Conduct vehicle force and moment wind tunnel test.
6. Estimate engine/vehicle system performance.

INLET SUBCRITICAL STABILITY/3UZZ

BACKGROUND

Virtually every inlet designer has experienced the phenomenon of inlet instabilit), commonly referred to
as buzz, in the development of a high performance supersonic inlet. D. D. Wyatt in his excellent paper
titled "A Review of Supersonic Air Intake Problems" (Reference 1) described inlet instability as follows:

"Generally the supercritical flow regime of supersonic intake systems

ischaracterized by a steady flow discharge rate. In some cases there
are nonstationary oscillations but these are generally of such a low
order of anplitude that serious structural or engine operation problems
are unlikely. It is probable that these low-order oscillations arise
from unsteady boundary-layer separation in the subsonic diffuser. In
many intake designs, however, the steady flow breaks down in the
subcritical regime. The terminal shock wave undergoes violent
excursions in position from well ahead of the cowl to far down in the
subsonic diffuser. This phenomenon is accompanied by severe
fluctuations in mass flow and pressure recovery."

We believe this description is too good to be improved upon. During recent development of a high
performance, dual aft mounted, two dimensioral inlet, we installed a small, flush mounted, high frequency
response, piezo electric static pressure transducer in the diffuser. A representative pressure trace
from this test series is presented in Figure 34. In this test subsonic diffuser back pressure was
increased (throttle plug movement) until buzz was experienced. Figure I clearly shows the noise output
of the inlet normal shock system during "steady state" operation and the large amplitude pressure
oscillations associated with classical "buzz". Thus Fiqure 34 clearly su~ports Wyatt's buzz description.
 2-5

inlet instability with relatively little success. In reviewing the inlet literature there are data for
a few inlet configurations which indicate some degree of stable subcritical operation, particularly at
the lower Mach numbers (M0 :2). By and large, through, the approach taken has been to design a high
performance inlet (which generally has little or no stable subcritical operation) and to avoid
subcritical operation by providing an inlet supercritical operating mode insured by a shock position
control. This control either limits fuel flow to the engine or opens up a bypass door downstream of the
i;,let minimum flow (throat) section. Such approaches are common in supersonic liquid-fueled ramjet and
turbojet applications, respectively.

Recent impetus to provide a margin of stable subcritical inlet operation has come from the following
sources: (1) In liquid-fueled systems, cost reductions could be achieved if the shock positioner
control could be eliminated. (2) The development of the solid fueled ducted rocket (ramrocket) for
tactical operation; has indicated a possible neeo for some margin of subcritical operation. There is
current interest in a throttleable solid fuel gas generator, however, the propellant flow rate will not
be controllable to the same extent nor for the same reasons as the liquid system and, in addition, the
flow rate is subject to large variations as a function of the gas generator soak temperature (frK
effects).
A third reason for reviewing the possibility of subcritical operation is shown in Figure 35. The dotted
line represents a typical pressure recovery-airflow characteristic of a high performance inlet. The
critical pressure recovery is point A; however, due to the lack of subcritical stability, an it 1let
margin is provided such that the inlet operates at point B. Now, if an inlet providing subcritical
stability could be provided with a critical pressure recovery equal to the der'.ted recovery of the high
performance inlet (as shown by the solid lines in Figure 35), comparable internal performance and thrust
would be generated. A brief look at the trade-offs between high performance and alternate inlet
approaches suggest that, in order to be attractive, the alternate inlet should provide a critical
pressure recovery within 10% of that of the conventional inlet. The amount of stable subcritical
stability required is not well defined but probably should be in the range of 10%-20% in terms of
capture airflow at the design Mach number. Due to the airflow matching characteristics of typical
ramjet engines, subcritical stability is required mostly at the lower Mach numbers of the flight
envelope. At the higher flight speeds the inlet generally operates with increasing amounts of
suparcritical operation unless a variable geometry exit nozzle is employed. Can inlets with a large
measure of subcritical stability be developed? The answer is not clear as discussed below.

SOURCES OF INSTABILITY

The sources of instability are generally well known and are discussed in the literature. Ferri and
Nucci in Reference 3, for example, explained a mechanism for buzz by the presence of a vortex sheet
entering the inlet. Daily (Reference 4) attributed instability to choking on one side of the vortex
sheet in the throat or minimum flow area of the inlet. Trommsdorff in Reference 5 noted that a strong
expelled shock, during subcritical operation, could cause boundary layer separation which could
alternately choke and unchoke the inlet throat. Ferri and Nucci also discussed the possibility of
boundary layer separation from the -nner cowl rather than the centerbody. Orlin and Dunsworth, in
Reference 6, present two inlet stability criteria. They argue that if either of these criteria are
exceeded, inlet buzz will occur. However, in order to make predictions as to when instability will
occur, a knowledge of the flow field downstream of the expelled shock and on each side of the vortex
sheet is required. Orlin and Dunsworth's stability criteria will be presented later in this paper
(lecture).

INLET LITERATURE SURVEY

The following paragraphs of this paper (lecture) present a brief and highly selective survey of material
that appears in the literature pertaining to the achievement of subcritical stability in supersonir
inlets, including comments given either to explain the cause of the instability or means of alleviating
it.

Boundary Layer Separation

Many researchers report that boundary layer separation on the centerbody (compression surface) reduces
the subcritical stability margin and/or initiates buzz. The NACA, in 1953, designed and tested side
inlets suitable for application to high performance aircraft. These inlets were designed for Mach 1.88
and 2.93 and were tested with various types of boundary layer removal systems. The inlets were half of
an axisymmetric design mounted on a flat plate to generate a boundary layer approaching the inlets to
simulate that generated on the fuselage. The cone half-angles were 25 and 3P degrees for the Mach 1.88
and 2.93 designs, respectively. Figure 36 presents a sketch of the cowl-lip scoop boundary layer
"removal system. It will be noted that the inlet centerbody was mounted flush to the flat plate. The
data in Reference 7 indicates significant improvement in critical pressure recovery, peak pressure
recovery, and subcritical stability when the height of the boundary layer scoop shown in Figure 36 was
increased above the boundary layer thickness generated on the approach flat plate. These test results
are reproduced in Figure 37. At the highest scoop heights tested the Mach 2.93 design achieved about
8.4 percent stable subcritical margin whereas the Mach 1.88 design achieved about 25 percent margin.
With the conclal compression surfac? mounted on a flat plate, the circun.Ferential pressure gradient
tends to direct the boundary layer off the cone toward the fiat plate and the cowl lip scoop thus
providing an effective boundary layer removal system. The test data of Reference 7 illustrates that
effective boundary layer control upstream of the inlet throat delays the onset of instabiliy related to
boundary layer separation.
 2-7 i.
Project Pigel Inlet Development Experience

In the early 1950's a series of axisynnetric inlet development tests were conducted in support of the
Rigel program. This 3amjet powered test vehicle was designed for cruise at Mach 2.0. In this
development proqram inlets were designed and tested with design Mach numbers of 2.0, 2.3, and 3.85. The
salient design characteristics and subcritical stability performance of these inlets is suimarized in
Figure 38.

Inlet subcritical stability test data often follow the trend shown in Figure 39; inlet subcritical
stability is minimum or nonexistent at the inlet design Mach number but increases significantly 'then the
inlet is operated at less than design Mach number. The Rigel test data presented in Figure 38 generally
follow this data trend. This subcritical stability data trend is one reason why inlets are often
designed for a higher, Mach number than the ramjet takeover Mach number.

Figure 40 is presented to show the effect of boundary layer bleed on the cone surface ahead of the
cowl. These data are fcr the Mach number 3.95 inlet when tested at Mach 1.92. It is clear that
boundary layer bleed significantly improved both critical pressure recovery and subcritical stability.

it is noted that all of the Rigel inlets were of the mixed compressibn type. Therefore, at the
Mach 1.66 test condition, these inlets operated unstarted (choked inlet). We believe this contributed
to their large subcritical stability margins but reduced total pressure recovery at this test condition.

University of Minnesota/R. Hermann's Supersonic Inlet Research

Rudolph Hermann, in his inlet textbook (Reference 11), reported the test results of an axisymmetric
mixed compression inlet designed for Mach 3.0. The inlet had a half-cone angle of 30 degrees with a
9.2 percent internal contraction and an internal cowl angle of 21.7 degrees (flow aligned). The
pressure recovery-capture area data for test Mach numbers of 2.0 and 3.0 are shown in Figure 41. When
tested at the design condition, Mo = MD= 3.0, the inlet had zero subcritical margin. However, when
tested at Mach 2.0, the inlet was unstarted (choked throat), and a 17 percent stable subcritical margin
resulted. This represented only the last data point tested and thus the margin may have been greater
than 17 percent. Hermann observed, that for an inlet operating unstarted, if the ecelled shock wave
becomes unstable it can only oscillate betwLen an upstream position and the supercritical expelled
position. TIis is because as the shock moves downstream a position is reached where the inlet throat
becomes choked. This prevents further motion of the shock such that it cannot enter the cowl therefore
the magnitude of the pressure variation is small.

It is noted that, for the inlet tested by Hermann, the triple point (origin of the vortex sheet) is
outside the cowl lip it Mo a 2.0 for supercritical and critical operation. As the inlet Is throttled,
the vortex sheet may cross the cowl lip but this will occur at large values of subcritical operation.

At Mo 8 MD= 3.0, the Inlet operates in a mixed compression mode and the triple point crosses the
cowl lip as the inlet is throttled. The vortex angle at the triple point is computed to be 4.5 degrees -

compared to the cowl lip internal angle of 21.7 degrees. This misalignment (16.2 degrees) is thought to
contribute to the lack of subcritical stability observed at Mach 3.0.

COMPRESSION RATIO LIMITS ANALYSIS

Orlin and Dunsworth in 1951 published a report on criterion for flow instability in supersonic diffuser
inlets (Reference 6). Stated briefly one of their conclusions was "The parameter determining the flow
stability through a supersonic diffuser is the rate of change of inlet static pressure (at the cowl
station) with mass flow. Stable subcritical flow occurs in the region of operation for which
dPl/dAo has negative values and the stability limit is reached when dPl/dAo becomes equal to
zero.
Two separate conditions lead to dP 1 /dAo = 0 and hence to instability. As the inlet mass flow is
reduced, the slope of the inlet total pressure curve with airflow may be negative, zero, or positive
depending upon Mach number and inlet geometry. Continuity relations between the freestream and cowl 7'-.7
station may be written to show that if the slope of the pressure recovery curve is negative or zero, the
slope of the static pressure will be negative. If the slope of the pressure recovery curve is positive
by a given dmount, the slope of the static pressure curve will go to zero. Figure 42 presents an
example of zero slope instability.
The second condition leading to instability is the reaching of a compression limit.* The following
sketch is taken out of Reference 6 to illustrate this point. It was assumed that the static pressure
across the cowl station CLB was
,'.7

-- L
8

A
 - 2-8

uniform. Through segment CL the total pressure PTO' is essentially that associated with a normal shock
at the local freestream Mach number Mo. The pressure recovery through segment LB in the sketch is
"somewh&t higher than PTO', having passed through an oblique shock and a strong shock at a Mach number
S"less than Mo. In this sketch if the static pressure at the cowl station should equal or exceed the
total pressure PTO' the flow across CL will stagnate and require flow reversal (hence instability).
SFigures
43 and 44 present examp1bs of .ompressinn limit instability. In these figures the term control
pressure is equal to PTO'/rT,.

r * Figures 43, 44 and 45 demonstrate that Orlin and Cunsworth instability criteria has merit. Powever, its
*". ,value s limited unless an analytical technique is available to calculate the variation of total pressure
recovery and cowl pressure ds a function of airflow.
Marquardt has used a simplified procedure using the equatirns of continuity combined with assumptions of
the expelled shock shape, to compute cowl pressure with variations in inlet airflow (Ao). This
technique is tedious and approximate at best. With the rapid advances being made in computational
aerodynamics, perhaps a good analytical model can be developed.

RECENT MARQUARDT TESTS

S- In 1974 Marquardt tested an axisynmtric inlet which operated in the unstarted mode at its design Mach
number of approximately 2.0. Ferri suggested this design approach in 1958 (see Figure 45 and
* Reference 1). The subcritical stability of this inlet was excellent; however, its total pressure
recovery was reduced when compared to high performance inlet designs.

Using the design philosophy of this inlet, its design Mach was increased to 3.0. The subcritical
stability of this inlet, at its design point, was zero. Study of the test data and shadowgraphs,
supported by analysis led to the following conclusions.

The achievement of stable subcritical operation becomes more difficult as the design Macd
number is increased. The reasons for this increascd difficulty are two-fold, namely, (1) the
increased tendency for shock-induced boundary layer separation on the compression surfaces, see
Figure 46, and (2) the inlet compression limits are more easily broached as the design Macl
S..number is increased.

2. There is evidence that the entrance of a vortex sheet into the inlet is not of itsel1
sufficient to cause inlet instability. There is evidence, however, that a highly negative angleo
vortex at the lip may cause boundary layer separation from the internal cowl lip. This situatior
"can occur only at Mach numbers equal to the inlet desi~n Mach number when the triple point is at
or near the lip. At Mach numbers above the design Mach number the triple point is below and away
from the inlet lip. At Mach numbers below the design Mach number the inlet must spill
considerable airflow before the vortex sheet crosses the inlet lip. As the expelled shock move!
forward, shddowgraph and schlieren photographs indicate that the vortex sheet angle rapidl3
departs from the two-dimensional value which must exist at the triple point. In the case of
axially symmetric flow the vortex sheet angle appears to approach that of an axially symmetric
flow field streamline. Thus, downstream of the triple point the vortex sheet angle is positive,
It is thought that the angle of the vortex sheet relative to the internal cowl angle is important,
with small displacements in these angles having less tendency for creating separation from the
cowl lip. This observation leads to the conclusion that the internal cowl should not be flot
"aligned at the design Mach number as is done in many conventional inlet designs. Rather, the
cowl should be aligned to produce a slight positive angle of attack with respect to the vorte)
sheet that results when the inlet goes subcritical a its design Mach numbev. In addition, ii
S","may be found that cowl bleed could enhance subcritical stability for those cases where internal
cowl boundary layer separation is a factor.
3. The Orlin and Dunsworth inlet stability criteria were applicable to the Mach 2.0 and 3.t
t designs. This provided valuable insight into the design modifications required to improve inlel
stability at Mach 3.0.
Modified Mach 3.0 inlet designs were tested with success. Inlet stability was achieved at the desigi
Mach number, without the use of boundary layer bleed. However, the use of bleed improved subcritica'
stability and total pressure recovery. As expected, however, inlet total pressure recovery was reducei
when compared to high performance inlet designs. Specific missile system performance trade-off studiel
are required to properly evaluate the merit of this type of inlet.

EARLY MARQUARDT EFFORTS AT BUZZ SUPPRESSION

As an outcome of an extensive review of available information in the field, Marquarot, in 1955, designei
a buzz suppression device which could be applied to .ny oxisymmetric supersonic inlet and which woulh
leave supercritical operation unaffected. The device consists of two parts (see Figure 47).
1. An annular slot, flush with the inlet center cone, located ahead of the cowl lip.

6"., 2. A chamber formed by the hollow inner body of the inlet cone, which conwunicates with thi
:'slot and is otherwise closed.
V:
"The mode of operation of the device is that the chamber acts as a pressure reservoir which absorbs 0!
- counteracts the pressure and shock wave fluctuations characteristic of the onset of diffuser instabilitv
The stabilizing effect on the expelled shock wave is illustrated in the schlieren photograph (Figure 48)
 w77

2-9

Aerodynamic models of two different inlets, a 50 degree cone and an isentropic spike, were tested over a
range of conditions with and without the suppressor and the results were eicojraging. Figure 49 shows
the improvement achieved through the use of the device in early tbsts. Of particular interest was the
relatively greater improvement at Mach numbers near design. Here, stability fell to very low values
without the suppressor, but with the suppressor attached, stability increased.
The buzz suppressor was also tested in 1956 on a ramjet engine at Mach 2.2/7 degrees angle of attack
* test conditions. Since percent spillage could not be measured accurately in the particular installation
employed, the indications are qualitative. However, the baseline engine, which could be made to buzz
and blow out at .072 fuel-air ratio, remained buzz-free at .074 fuel-air ratio when the suppressor was
attached.

"Abuzz suppressor patent (No. 3,046,733) was granted to The Marquardt Company in 1962. Figure 50 from
this patent, indicates that the surge chamber volume was divided up into four separate sections *sn that
each space operates independently of the other at angle of attack.*
, Although the buzz suppression research briefly described above was highly encouraging, ramjet engine
packaging requirements precluded its use. The ramjet air turbine driven fuel pump and fuel control were
-. packaged into the centerbody of the early ramjet engines such as BOMARC. The use of a buzz suppressor
has not been considered in recent years.
" CONCLUDING REKARKS

The arguments for a supersonic inlet which possesses a large amount of subcritical stability are
impressive. However, survey of the extensive experience in this research field leads to two important
i.'" conclusions:

1. Inlet designs which feature subcritical stability, almost without exception, operate at
critical total pressure recoveries significantly less than can be achieved with high
performance/no stability designs.
2. Currently there is no analytical technique that can be used to design high stability
inlets. The inlet designers' experience and the wind tunnel are the tools used to develop
such inlets. This is an expensive and sometimes dangerous path to follow. The development
of a new ramjet engine/flight vehicle is initiated assuming specific engine component
efficiencies and operating characteristics. If it is Initially assumed that a high
stability inlet can be developed and later test results say stability cannot be achieved,
program cancellation is a real consideration. Most engine/inlet designers, faced with this
possible situation, usually specify high performance inlet designs that do not require
subcritical operation to meet engine/flight vehicle performance requirements. If a high
stability inlet is truly required, its performance must be convincingly demonstrated before
initiation of the engine development program.

MULTIPLE AFT INLET PERFORMANCE

Multiple (four) aft-mounted inlet systems have been successfuilly used in several ramjet flight test
vehicles and one operational missile system (see Figure 5). We believe the reasons for selection of
this type of inlet system were as follows:

1. Vehicle packaging c~nstraints did not permit use of a nose Inlet system.

2. With a liquid fuel ramjet the increased shear contact area provided by four inlets dumping
into a common combustor results in high combustion efficiency with a short length combustor.

3. it is logical to assume that the performance of a ducted rocket ejector is maximized

through use of a large number of air inlets located uniformly around the periphery of a
centrally located ga generator (this subject is discussed later in this lecture).

4. The flight vehicles were designed for operation at low altitude, therefore, angle of attack
and/or yaw requiremenls were quite small.

As flight vehicle attitude and man uver operating requirements increase, the demand for inlet operation
at hilh angle of attack becomes a demanding requirement. What are the performance characteristics of a
multiple aft inlet system at large. angles of attack? Figure 15 presents the performance trends for an
aft mounted inlet system. The performance of such a system, at angle of attack, is poor. The test data
also indicate that a "+" inlet configuration is preferred to an "X" configuration.
Are these inlet performance trends supported by other test programs? Yes - Krohn and Triesch of DFVLR
(West Germany) reported similar results. Their test data for an "X" and a "+N aft-mounted inlet system
(half axisymetric inlets) is presented in Figures 51 and 52. Study of these data lead to the same
conclusions as discussed above.
it. flow.
Krohn and Triesch data also shows extensive regions of reversed inlet

.
"The reasons for the poor
With a vehicle operating
angle of attack performance of an aft-mounted inlet system is well understood.
at large angle of attack, the four inlets "see" significantly different flow
fields. These flow field characteristics are summarized as follows:
 2-10

I. The local Mach number and flow angle delivered to the bottom inlet are reduced when
compared to free stream Mach number and angle of attack. These charactcristics result in
high inlet performance.

2. The airflow is accelerated around the vehicle forebody such that the local flow angle and
Mach number delivered to the side mounted inlet is often larger than the flight vehicle
angle of attack. With most inlets this flow field results in degraded Inlet performance.
"3. The local fl'w delivered to the top inlet is the real problem. An inlet located on the lee
side of the vehicle forebody typically sees regions of flow separation, reduced total
pressure and vortices. The twin vortices are particularly troublesome. The flow field
characteristics on the lee side of an ogive forebody were experimentally mapped by
Triesch. The results of this test program are presented in Figure 53. The presence of the
forebody twin vortices are beautifully illustrated.

Aft-mounted inl-?t systems typically employ a single combustion chamber. Therefore, as engine back
pressure is increased, the inlet operating at the lowest total pressure recovery (usually the lee side
inlet) is driven critical and then subcritical/buzz. Under these conditions the inlet system spills air
with a major loss in engine thrust and in severe cases, experiences inlet flow reversal. Flight
operation under these conditions is not practical. What can be done to improve the situation?

Strakas have been used with some success in high performance aircraft. However, as discussed earlier in
this paper (lectore), we believe that high performance ramjet powered vehicles, which typically require
high angle of attack operation, should employ one or two inlets. If a single inlet is used, it is
usually bottom mounted. High inlet performance has been demonstrated with chin (forward) and aft bottom
mounted inlets. If two inlets are used, they are mounted approximately 45 degrees from the horizontal
centerline.

It has been implied throughout this discussion that a bank-to-turn flight vehicle guidance/control
"system willbe used in advanced ramjet powered missiles. With a bank-to-turn flight control system the
inlet(s) typically experience only small angles of yaw/sideslip. Obviously, this results In high inlet
performance. The use of a bank-to-turn flight vehicle guidance/control system has been demonstrated in
.- K-'"and several ramiet powered flight test/operational vehicles: (1) BOMARC A and '; (2) Bloodhound MK1 and MK2;
ASALM-PTV.

The use of a bank-to-turn versus skid-to-turn flight control system in advanced ramjet Powered missiles
is currently controversial. Rocket powered missiles, almost without exception, employ a skid-to-turn
control system. What are the factors that must be considered in the design of a bank-to-turn guidance
system?

1. What flight control algorithms need to be developed?

2. Can these algorithms be packaged into small flight vehicles?

3. Will the flight vehicle aerodynamic time constants permit use of a bank-to-turn system at
high altitude flight conditions?

4. What is the effect of radome slope error (aberration) on missile miss distance?
We believe the radome slope error argument is the key issue. High performance air-to-air missiles, with
a small warhead, designed to attack a maneuvering airborne tdrget are most vulnerable to this argument.
However, it is again pointed out that the ramjet powered B('iARC and Bloodhound surface-to-air interceptor
missiles employed bank-to-turn steering logic, were equipped with semi-active radar terminal guidance
and were operationally deployed *or many years. In fact Bloodhound MK2, after twenty years, remains in
operational status with the Royal Air Force, the Swiss government and the Singapore Air Defense
Command. BOMARC also continues in use as a high speed target vehicle for the U.S. Air Force.

The bank-to-turn versus skid-to-turn controversy cnntinues and is sufficiently important that a major
technical meeting, dealing only with this subject, will be held in September (1984. at The Johns Hopkins
University/Applied Physics Laboratory. The meeting is being sponsored by the Uuidance and Control
Information Analysis Center/ITT Research Institute.

Ramjet inlet performance is usually documented at 5 degrees, and in some cases 10 degrees angle of
yaw/sideslip. However, virtually no research in the United States has been directed to the development
of high performance ramjit inlets which tolerate operation at large yaw/sideslip angles.

'.% COMBUSTION DRIVEN PRESSURE OSCILLATIONS/INLET INTERACTIONS

Combustion driven pressure oscillations have been experienced in several small diameter liquid fuel and
solid ducted rocket engines (ramrocket). These pressure oscillations have generally been experienced in
two frequency regions. Low frequency oscillations typically occur in the 100 to 500 Hz range and high
frequency oscillations in the 1000 to 3000 Hz range. A Power Spectral Density (PSD) analysis of a
representative combustion induced pressure oscillation signal measured with a high frequency response
piezoelectric pressure transducer is shown in Figure 54.
The high frequency oscillations, commonly referred to as screech, are typically the first and/or the
second combustor tangential modes. This oscillation mode(s) generally increases combustion efficiency
 2-11

* Can the low frequency oscillation, generally the first or fundamental combustor longitudinal mode, .ccur
"at a frequency sufficiently low to "drive" the normal shock system of the ramjet inlet/diffuser? rhe
NACA, in 1950, demonstrated that a supersonic inlet could be driven critical and ilito buzz by both
.. combustion and mechanically produced pressure oscillations. In these tests it was observed that pressure
. oscillations reduced the diffuser exit static pressure, at critical operation, by one half the total
amplitude of the oscillation (see Figure 55). In addition, it was noted that critical iniet operation.
with pressure oscillations, occurred when the maximum instantaneous diffuser exit static pressure
equaled the steady state (no pressure oscillation) critical operation diffuser exit static pressure.
"* Ltquid fuel ramjet freejet engine tests conducted in support of the Naval Weapon Center GORJE Program,
in 1977, again convincingly demonstrated that large amplitude/low frequency pressure oscillations can
drive the normal shock system of a supersonic inlet into buzz. Figure 56 presents an inboa;-d profile of
the GORJE test vehicle and Figure 57 shcws the pressure oscillation/inlet buzz test data. Notice that
steady state instrumentation indicates a pressure margint of 10% yet the inlet was driven into buzz by
combustion induced pressure oscillations. The basic conclusion drawn from the NACA and GORJE engine
tests (and more recent test programs) is that the inlet/engine designer must provide for an increased
inlet pressure margin when combustion induced pressure oscillations are experienced during the develop-
ment cycle of a new engine. This increased pressure margin requirement is especially critical at .amjet
*.takeover conditions where ramjet inlet marqins are typically small. Representative ramjet inlet pressure
"margins during a mid to high altitude flight trajectory is presented in Figure 58. To compound this
"increased inlet pressure margin requirement, the correlation of pressure oscillation data shows rapidly
increasing amplitude as free stream total temperature (flight Mach number) is reduced (see Figure 59).
In other words, the amplitude of the low frequency combustor induced pressure oscillations are maximum
at ramjet takeover conditions.

"The combustioninduced pressure oscillation problem is being actively researched by a number of tkilled
investigators. The purpose of this paper is not be present a treatise on the subject but rather to
* relate the importance of this combustion phenomena to the inlet designer.

It is also of interest to the inlet designer that limited ramjet engine freejet test data strongly
/ , suggests that the boundary layer bleed system of a supersonic inlet sharply attenuates the combustor
. pressure oscillations that reach the inlet throat/normal shock system (see Figure 60). Obviously, this
acoustic attenuation process reduces inlet pressure margin requirements. This attenuation effect is not
well u-derstood and should be further studied.

DUCTED ROCKET EJECTOR PERFORMANCE WITH SINGLE, TWIN AND FOUR INLET SYSTEMS

Z, BACKGROUND
"The combined
cycle engine combines the best features of the rocket and ramjet engines into a lightweight,
efficient power plant. This type of propulsion system is particularly well suited to acceleration type
missions. Three prime candidate missions for this type of engine are (1) surface-to-air missiles,
*" (2) air-to-air missiles, and (3) ai.-to-surface missiles.

Today, there is strong interest in the ducted rocket type of combined cycle engine. This engine employs
d a solid fuel gas generator whose flow rate is fixed, or in more advanced designs, variable.
of pronlsion system i. also comnmonly referred to as a ramrocket or a rocket-ramjet engine.
This type

With prior ramjet and solid fuel rocket/gas generator development experience, the key technology areas
in developing a ducted rocket engine are the ejector subsystem and the ramburner (afterburner). The
*, ejector performance parameters of primary interest are the jet pumping ratio and the mixer length
required to achieve full mixing of the secondary and primary flows. Jet pumping ratio is defined as the
"- ratio of the flow total pressure at tie mixer exit/combustor entrance to the air total pressure at the
p diffuser exit/ejector entrance. With ejector systems designed for application to ducted rocket engines,
this pressure ratio can be quite high and can exceed one, thus, explaining the use of the term "jet
pumping ratio".

It is logical to assume that the performence o^ a ducted rocket ejector systsm (total pressure rtio and
required mixing length) is maximized through use of a large number of secondary air inlets located
uniformly around the periphery of a centrally located gas generator. In r'.ent years much interest has
centered on use of a symmetric four aft inlet system. This geometric arrangement offers high ejector
p performance; however, there are airborne missile applications which will not permit the use of this
design approach (see Figure 61). Shown in this figure are pylon and body mounted external carriage
launch aircraft/missile installations and a semi-submerged launch aircraft installation. These launch
.. aircraft/missile installations obviously lend the:mselves to twin and single inlet configurations.
"* Inlet performance also dictates consideration of single and twin inlet/air induction systems. The four
aft mounted inlet system has been successfully employed in several flight research and operational
missile systems. This system functions well when vehicle angle of attack and yaw requirements are
small. However, advanced ramjet powered tactical missile systems under study today (particularly
Air-To-Air Missiles) emphasize high angle of attack operation (see Figure 6). Under these operating
conditions, the performance of the multiple (four) inlet system Is poor (see Figure 62). In marked
contrast, single and two inlet systems, when located on the lower surface of the vehicle forebody,
feature significantly increased inlet total pressure recovery and capture area ratio with increasing
angle of attack. These performance characteristics are well matched to advanced tactical missile system
requirements. As a consequence of the launch aircraft/missile installation and inlet performance
arguments presented above the basic objective of the experimental program described in this paper was to
compare ejector/mixerperformance of twin and single inlet systems as compared to a s;ymetrical four aft

U
 2-12

EJECTOR TEST PROGRAM

Previous Marquardt testing with the liquid fueled Ejector Ramjet Engine (olso a combined cycle engine)
showed that subscale aerodynamic tests (no combustion) produced ejector design criteria/performance
which were largely corroborated by later combustion tests. This type of testing is low cost and lends
Itself to a highly flexible test program. Therefore, in this exploratory multiple inlet ducted rocket
ejector test program, a model aerodynamic test program was conducted.

The ejector test set-up used in this program is shown schematically in Figure 64 .and Figure 65 is a
photograph of the test set-up. Two ejector test items were fabricated. One test item simulated a four
aft' mounted inlet configuration while the second test item simulated either a twin or single inlet
configuration. Both ejector test items were designed to accept removable ejector heads. This feature
permitted variations in ejector nozzle geometry including evaluation of single and multiple ejector
nozzles. Only sonic ejector nozzles were evaluated.

For this aerodynamic test program the ability of several inert gases to simulate the solid fuel gas
generator efflux were ccmpa,-ed. These gases were nitrogen, carbon dioxide, helium, ambient temperature
air and heated air No single gas met all simulation criteria. Heated air was selected as the best
compromise. A SUES burner was used to heat the primary air; the burner fuel was gaseous hydrogen.
Ambient temperature air was used as the secondary test fluid.

A simulated ramburner or m!xer/combustor was located downstream of the ejector test item. The length of
this simulated combustor could be changed by insertion or removal of cylindrical spool sections. A
large number of static pressure taos were located in the mixer/simulated combustor test hardware. In
addition, two total pressure rakes were located downstream of the ejector test item. Initial tests
employed a fixed ramburner exit nozzle; however later testing employed a vwriable area exit nozzle
(moving plug) to more properly simulate engine back-pressure effects. Representative ejector total
pressure ratio data measured with use of the variable area exit nozzle are presented in Figure 66.

It has been substantiated by other test programs that, when jet mixing takes place in a constant area
"duct between
fluids initidlly at very different pressure and velocity (or momentum ratio), the degree to
which mixing approaches completion is indicated by the axial wall static pressure gradient. Where the
gradient passes from positive to negative, mixing is essentially complete and the losses due ;o wall
friction again dominate. Measured diffuser/mixer static pressure distributions for the four inlet
configuration tested are shown in Figure 67. It is clear from the static pressure data that mixing is
essentially complete two to three diameters downstream of the ejector. Additional test data showed that
the required mixing lengths for the twin and single inlet configurations were slightly longer (see
Figure 68).

Fifteen ejector nozzle/inlet configurations were evaluated in this test program. Testing emphasized the
twin and single inlet configurations because data showed these configurations to be sensitive to ejector
nozzle geometry. This was particularly true for the single inlet configurations. Representative twin
and single inlet test data are presented in Figure 69. Ejector/m!xer performance sensitivity to changes
in ejector nozzle geometry is clearly evident. The ejector/mixer total pressure ratio performance
measured in this test program is summarized in Figure 70. Data are shown for the four, twin and single
inlet configurations with secondary to primary flow ratio as the primary test variable. Test data for
the best performing ejector heads were used to construct this figure.

CONCLUSIONS
The major conclusion drawn from study of the experimental data and supporting analysis developed in this
program is that a high performance, ducted rocket ejector/mixer subsystem can be developed using any
number of well developed/well integrated inlets. However, it is noted that ejector mixer performance
with twin, and in particular, single inlet configurations, is more sensitive to ejector nozzle
geometry. Although not the subject of this paper, it is pointed out that the basic findings developed
in this exploratory development program were largely corroborated in later ducted rocket combustion
tests.

REFERENCES

1. Fabri, J. (Editor), "Air Intake Problems in Supersonic Propulsion", Published by Pergamon Press,
London-New York-Paris - Los Angeles, 1958.
2. Oswatitsch, K. L. (Translation), National Advisory Committee For Aeronautics, "Pressure Recovery
for Missiles With Reaction Propulsion at High Supersonic Speeds (The Efficiency of Shock
Diffusers)", June 1947, NACA TM 1140.
3. Ferrn, A and Nucci, L., National Advisory Conmmittee for the Aeronautics, "The Origin of
Aerodynamic Instability of Supersonic Inlets at Subcritical Conditions", 1951, NACA RM L50K30.
4. Daily, C. L., "Supersonic Diffuser Instability", Journal of Aeronautical Sciences, Volume 22, No.
11, 1955, pp. 733-749.

5. Trommsdorff, W., Durchgefuhrt in der Abteilung "Triebwerksaerodynamik" des Institutes Fur

Angewandte Gasdynainik der Deutschen Versuchsanstalt fur Luft- und Raumfahrt e.V, "Untersuchung an
Triebwerkseinlaufgen", August 1964.

... ......
 2-13

7. Piercy, T. G. and Johnson, H. W., National Advisory Committee for Aeronautics, "A Comparison of
Several Systems of Boundary Layer Removal Ahead of a Typical Conical External - Compression Side
Inlet at Mech Numbers of 1.88 and 2.93", 1953. NACA RM E53FI6. -
8. Orlin, W. J. and Bajek, J. J., The Marqiardt Company, "Results of Series I Tests of a One-Quarter
Scale Diffusey Model of the Marquardt Model C28-2.0 Cl Ramjet for the X55M-N6 Project Rigel
Configuration 7 Missile at OAL During the Period 11 to 15 December, 1950", March 1951. Marquardt
Report 5093.

9. Humenick, W., The Marquardt Company, "Results of Tests of a One-Quarter Scale Diffuser for the
Marquardt Model C28-2.3A Ramjet at CAL, Daingerfield During the Period 19 to 23 March, 1951",
June 1951. Marquardt Report 5150.

10. Humenick, W. and Doyle, A. F., The Marquardt Company, "Results of Tests of the Marquardt Model
AM-17 Aerodynamic Model for the. X55M-N6 Project Rigel Configuration 4 Ramjet at OAL During the
Period 25 June to 6 July, 1951", October 1951. Marquardt Report 5159

11. Hermann, R., "Supersonic Inlet Diffusers and Introduction to Internal Aerodynamics", Published
by Minneapolis-Honeywell Regulator Company, Minneapolis, Minnesota, USA, 1956.
12. Connors, J. F. and Meyer, R. C. National Advisory Committee for Aeronautics, "Design Criteria for
Axisymmetric and Two Dimensional Supersonic Inlets and Exits", January 1962. NACA TN 3589.
13. Kuehn, D. M., National Aeronautics and Space Administration, "Experimental Investigation of the
Pressure Rise Required for the Incipient Separation of Turbulent Boundary Layers in
Two-Dimensional Flow", February 1959, NASA Memo 1-21-59A.

9.

I%
 2-14

S: .... - _ .... _ _ ,_ _ -- _uD

BOMARC TALOS
SURFACE-TO-AIR INTERCEPTOR MISSILE SHIP LAUNCHED INTERCEPTOR MISSILE
OPERATIONAL IN 1957 OPERATIONAL IN 1959

FIGURE 1. OPERATIONAL U.S. RAMJET POWERED MISSILE SYSTEMS

'Af

SEADART
ShIP LAUNCHED INTE9CEPTOR MISSILE
OPERATIONAL IN 1975
BLOODHOUND MK1 E MK2
SURFACE-TO-AIR INTERCEPTOR MISSILE
OPERATIONAL 1959/1964
 2-15

~
'00

SA-4.'GANEF
MOBILE SURFACE-TO-AIR INTERCEPTOR MISSILE
OPERATIONAL IN 1964 (JANE'S WEAPON SYSTEM)
FIGUR~E 3. OPERATIONAL USSR RAMJET POWERED MISSILE SYSTEMS

INTEGRAL
FUEL CONTROL ROCKET BOOSTER\ ROKE NOZZLE

\DOMEPORTSPECIFIC IMPULSE
DOMEPORT-240 SECONDS

BOOST PHASE - The Integral Rocket-Ramjet is accelerated to ramjet takeover velocity by rocket propellan~t contained in the
ramjet chamber.

FUELT-COVER NOZZLE

TRANSITION PHASE - Ejection of the rocket nozzle and port cover allows air to enter the combustor where the ramjet fuel is
introduced and burned.

SPECIFIC IMPULSE
-1200 SECONDS
an i.- ,niA c. i, er lt A
M..L.nI ir n... . :.:...~.L. ~ L. .. cM rk a t *rc
2-16

-LASRM FLIGHT TEST

VEHICLE (U.S. 1965)
ALVRJFIH;ES3EIL

FIGURE 6. MULTIPLE
AFT &mOUIVT D
l L TRA J T P O UL i
N S T M

TERMINAL

GUIDANCE

HAND OVER
CRUISE 0-520-
HIGH Mc,
a f-
54N
g

A L TI T t D EENGAGEMENT

a-oo RAMJET-
TAKEOVER
a-D

RA N
 2-17

Ao4

. ... ....

PT- TOTAL.PRESSURE.
......
...
...
....
M~ MAC NUMBER .

CATURE
A-INLT REA O
........

.... . . .. .

BOUNDARY) SUPTCRITICEL

SUCRITICAL

BUZZ

SUECRITICAL

Lu

COMBUSTR TOTAINLET ABREFLO

AIRFLOW (A00/Ac)Total
-J (Ao/A,)Plenum

CAPTURE AREA RATIO (Amo/Ac) -w-

FIGURES8. REPRESENTATIVE INLET PERFORMANCE CHARACTERISTICS -

 2-18

""2-D 2-D 8&

"CHI-N-CHIN

PT21PTo A Ao cAXI

Mo
_
CONSTANT
I _<_r

00 '0-..0 _o"-"ao

IIGURE 9. AXISYMMETRIC, TWO-DIMENSIONAL AND CHIN INLET PERFORMANCE

(ISOLATED)

I
IAXI

2-D AND AXI

PT21P ) Aow/Ac

Mo= CONSTANT

-~a cD-aa
FIGURE 10. AXISYMMETRIC INLET PERFORMANCE - WING FLOW FIELD

/"
 2-19
r

BOTTOM WING
CRUCIFORM

WINGS

NO WING -MID WINGS

aa, .4

FIGURE 11. EFFECT OF WING ON INSTALLEI) CAPTURE AREA RATIO

ISOLATED TWO-DIMENSIONAL INLET

Ao -. -. QAc

WITH
SHROUD

_/N0 SHROUD

PT2 WITH A 7
SHROUD Ac /

/WITH ~ *SHROUD

WITH
%
a
00 ~ SHROUD
- NO SHROUD
--- a 100 NO SHROUD

Mo -Mo

FIGURE 12. EFFECT OF PRECOMPRESSION SHROUD

 FIGURE 13. RAMJET POWERED FLIGHT TEST VEHICLE

GUIDANCE/FLIGHT CONTROL PAYLOAD RAMIJET FUEL

SEEKER AIR INDUCTION ROCKET BOOSTER, RAMJET

SYSTEM COMBUSTOR

FIGURE 14. ADVAN'CED AIR-TO-AIR MISSLE DESIGN

 2-21

FOUR SYMMETRICAL INLETS

PT2/PTM Aa)Ac a

'NOTE:
ONE OR MOPE iNLETS
ARE CRITICAL, OTHERS
SUPERCRITICAL I 0

00 ROLL ANGLE 450 ROLL ANGL.:

00 450 00 450
VEHICLE ROLL ANGLE

FIGURE 15. MULTIPLE AFT MOUNTED INLET SYSTEM PERFORMANCE

Single Inlet Angle of Attack - 100

00.

1800
0* BOTTOM
445
45* 45*

1800*
135 5

900
1350
_________ ________ 180 _ _ _ _ _ __ _ _ _-

FIGURE 16. EFFECT ON INLET CIRCUMFERENTIAL POSITION

 2-22

Mo4.0

Ma) 2.

44

14

~~~FIGURE
BOTTO
17. MO NTETID TWODMNSINAL INESTMOELFRAC
 /
N,./

' 2-23

1
4'- . -- M#

. s

4,"4
 2-24

1.0

ISENTROPIC
0.8 N PK

OCO-

0.4
U.'
0
a-

DEINccI UBRM

FIUR
1.MAIUMTOA PESUE EOVRYVRSSDESIGN
MACH NUMBERM

Isolated Critical Performance

/O ,,.LVMD
Angle of Attack = 00
HIGH MD

PT2/PT 0 oA

/L

li Mo0 Mo0

FIGURE 22. EFFECT OF DESIGN MACH N'UMBER/ANGLE OF ATTACK =00

 2-25

S-- 00
COMPRESSION WAVES
\6 .
M0 SHOCK
1.0
0.9
~
--------

0.7
0.7 UEBLOW,
DESIGN4- '
0. "INTIAL WEDGE
OR CONE

0.5 ___, _-__ ON DESIGN

0.4 ON
PT 2 /PTO
CRITICAL 0.30

INLET ___________

0.2 - A59" ABOVE DESIGN BELOW DESIGN r

Z AM55C

0.1 A

2.0 2.4 2.8 3.2 3.6 4.0 4.4 4.8

INLET MACH NUMBER, MO
ABOVE DESIGN

FIGURE 23. OFF-DESIGN INLET PERFORMANCE

Isolated Critical Performance

Two Dimensional Inlet Mixe-d Compression Angle of Attack -100

"LOW
MD OW

"HIGH MD
PT2/ 0To AolAc /
HIGH MD
j//

Mo Mo

FIGURE 24. EFFECT OF DESIGN MACH NUMBER/ANGLE OF ATTACK = 100

 2-26

S% Shock
-- s\ /m
TURNING
ANGLE. 9 40 -ammmiftwu~vVre
IDEGREES) I I Shoa
20-se*
1,0 _MACH NUMBER. M" wave

1.0 Interscction zone

* CAPTURE 0.1 (cowl removed)

* AREA RATIO,
AIAC 0.6

0. AH TB RI M1I
.66F dt EE S EM OU V11 %
AFTER 2. -
COMPRESSION. 'L

1.o0I ,
2.0 2.5 3.0 3.5 4.0

LOCAL APPROACH MACH NUMBER, Mo

FIGURE 25. INLET EXTER1NAL COMPRESSION LIMIT

EXTERNAL COMPRESSION
Moo- CONSTANT, ac. 00

OPTIMUM

CDCI I DCEXTERNAL
COMPRESSION
LIMIT

CONTRACTION RATIO- AclA1

FIGURE 26. EFFECT OF INLET CONTRACTION RATIO/ANGLE OF ATTACK =00

 2-27

Installed Critical Performance

ATwo Dimensional In~let Mc = Constant

AAc

Ac/
f HIG

/ ~HIGH Ath

PT2IPT Go GH / AaDI Ac

/HIGH Ac/
LOWO
L

-00 0*

FIGURE 27. EFFECT OF INLET CONTRACTION RATIO/VARYING ANGLE OF ATTACK

Mixed Compression
Angle of Attack 00 Moo = Constant

FNET
OPIU

PT2/PTOD

INLET BLEED FLOW-

FIGURE 28. EFFECT OF INLET BLEED FLOW RATE

 2-28

SQUARE-HOLE GRID ROUND-HOLE GRID

RING GRID/FLAMEHiOLDER

__PADIAL GRID ANNULAR GRlID

GRID CONFIGURATIONS

FIGURE 29. AERODYNAMIC GRID CONFIGURATIONS

MACH HIJMBER _____ ____

K ~~~~0.8
0 _____. _ ___ ___

- 0.6A 6T1I8
~TAV0
0.1

0.04 - '9 2.5CI.

PTMBMAR
"B INLETc
0.02

SOLID SYMBOLS - GRID IHSTALLED

0.011 _ 1__
____

0.20 0.25 0.30 0.35 0.40 0.45 0.50

DIFFUSER MXT. MACH NUMBER

FIGURE 30. AIRFLOW PROFILE CONTROL WITH AN AERODYNAMIC GRID

 2-29

HIGH PERFORMANCE INLET BUZZ

NO SUBCRITICAL STAiLITY

LOW PERFORMANCE INLET IN_

WITH SUBCRITICAL STABILITY

INCREASING "-"
F/A BUZ..
PT2 :

" MARGIN",."
'.4

/ / .-..
AA/IAc
FIGURE 31. INLET SUBCRITICAL STABILITY PERFORMANCE COMPARISON

TWO-DIMENSIONAL INLET.-,
-- - F= 3:1 VON KARMAN OGIVE -'
HF F-2:1 TANGENT OGIVE A

-. 4

.. ,=3

oleo

L INLET LOCATED 6 DIAMETERS FROM NOSE -'

a~ cD

C^=nn
^V0 ~ili r00nU~r "Un
 7.77777777.

2-30

ANULE OF
FLIGHT ATTACK (DEG.)
FLIGHT 02 0 -6
& 2 04 00
003 0 5 5'
0 6 0 6 a010

0)15 15
0 7
FROMPi RISE
oFROM
T0 T ~ 1

4ALL DATA ADJUSTED TO4

U ZEROANGLEOF ATTACK 010
ZE AGLE F cc WIND TUN/NEL
DATA

ANGLEOF4
ATTACKIDEG.)0

-6 -

MACH NUMBER __ ---___- _ __--__

MACH NUMBER

FIGURE 33. COMPARISON OF- FLIGHT TEST AND WIND TUNNEL INLET TEST DATA

f I BUZZ

0 THRTTLE PLUG POSITION 1

tA
1, i F

- KULITE TRANSDUCER SIGNAL.

FIGURE 34. SUPERSONIC INLET DYNAMIC CHARACTERISTICS

 2-31

I HIGH PERFORcMANCE INLET

NO SUBCRITICAL STABILITY

LOW PERFORMANCE INLET

WITH SLIBCRITICAL STABILITY
Bz

MRI

F BUZZ

0
____ ___ ______ ___ __

FIGURE 35. INLET PERFORMANCE COMPARISON

INTERNAL
SPUTT7ER
PLATE WOODEN ITRA

FIGURE 36. CONICAL EXTERNAL COMPRESSION SIDE INLET

2-32

.6DESIGN MACH No.= 1.88 .6DESIGN MACH No.= 2.93

CL a.

a.0. a. BOUNDRY-LAYER

SCOOP HEIGHT 48 P/ARAE

8w PARAMETER cc 031
0 0 1.3610
wi .82 010L
cc 1125
Ej

04
S.74 - - CL

.2

.5 .6 .7 .8 .9 1.0.6 7 .8 . 10
IiI'LET MASS-FLOW RATIO, mD)/mliD INLET MASS-FLOW RATIO, mD/ml,

FIGURE 37. INLET CHARACTERISTICS WITH INCREASING BOUNDRY LAYER REMOVAL

AXISYMMETRIC NOSE INLET DESIGNS

INLET DESIGN CONE BOUNDARY SUBCRITICAL STABILITY*

MLODEL NO. MACH NO. ANGLE (DEG.) LAYER BLEED ma = 1.66 M 0 = 1.92

1 2.0 250 NO 18% 6%

2 2.3 230 NO 44% 11%

NOT
3 3.85 200 NO TESTED 7

YES 16% 13.5%

*ANGLE OF ATTACK = 0

FIGURE 38. PROJECT RIGEL INLET DEVELOPMENT EXPERIENCE

I ANGLE OF ATTACK - 0'

tc

z -INLET DESIGN
],---MACH NUMBER

C.,

FLIGHT MACH NUMBER

 2-33

M- 1.92/cro 00
1.0 ~~IMILET
DESIGN MACH NO. =3.85 ____

'V WITH BOUNDARY LAYER BLEED

NO 3OUNDARY LAYER BLEED

UA0.8
0

w0
w 07

I..s

0.24 0.26 0.28 0.30 0.32 0.34-

FREESTREAM CAPTURE AREA RATIO (A 0 /A3 )

FIGURE 40. PROJECT RIGEL INLET TEST DATA

1.0
NO BUZZ
0 SUBCRITICAL

0.8

M 0 =2.0 0

0
CC 0.6 ____

wuBZ

0.40'

0 0.

LL .30r, CONTRACTION RATIO

=~n =1.094 .-

eBJp lInt) -21.71

SEE REFERENCE I77

0.0oII _____

0.5 0.6 0.7 0.8 0.9 1.0

INLET CAPTURE AREA RATIO (Aoo/Ac)

ii
08u1$nCA4 nR IMi rf-lfau~r%
A R~a~ LIPUJ
A -4 n IftI3 C-T -rCOT L~raIII TO
 2-34 r .I

SINGLE CONE INLET 1250)

DESIGN MACH NO. - 2.3
W TEST MACH NO. = 1.66

0.9 'CONTROL PRESSURE

0.8"

00

ILI cc 0.7

wt

0.96 -0

I I ~~~~~0.90
- . _ _..

0*

0.25 0.01 ! .3% 0.3

CAPTURE RATIO (Aoo/A3) ;'

FIGURIE 42. ZERO SLOPE INSTABILITY - BELOW THE CONIROL PRESSURE

.=..

SINGLE CONE INLET (500)

DESIGN MACH NO. - 2.0
TEST MACH NO. = 1.92
"NORMAL SHOCK
CONTROL PRESSURE
w . .. .
0.76 __ _ .
o. . . . .
Ius
cc:rl 0.74-
-o.
0o 0.72-

0.70 .

- L
Lu
-,

S80.94-

0.92-0

cc 0.28 0.28 0.30 0.32

z CAPTURE RATIO (Aoo/A 3 )

-. ........ . a s a e e e e *. an. * *. . n,~ ;,

-V . ei lI I i liIA I I ~ l A l p AI IIIi RIi l A A IA ilI1

 2-35

DOUBLE CONE INLET 1400-700)

DESIGN MACH NO. = 3.3L 0
TEST MACH NO. =2.05
0 NO BLEED
BUZZ
POINTS 0 BLEED

S;CONTROL 0.9-I
8075PRESSURE 89
8
0O..p

uj0.70- 0

I~Ll

S0.9-
o go

~0.8*u
Lu
0.7

0.5CAPTURE RATIO (Aoo/A 3) .7CAPTURE RATIO (A 0 /A3) 03

FIGURE 4. NORMAL SHOCK CONTROL PRESSURE INSTABILITY - DOUBLE CONE INLET%

FIGURE 46. CHOKED INLET CONCEPT PROPOSED BY A. FERRI (REF. 1)

 //

.2-36

', 10.0-

9.0- Mc
"-2.75

8.0- MO
"MC
Jc
7.0- -2.5

6.0- CONE HALF-ANGLE

STATIC PRESSURE -2.25
"RATIOACROSS
NORMAL SHOCK
"" MO. P/Pcr
5.0

-2.0

4.0- FLAT PLATE

7 INCIDENT SHO~ K
1.75 R6 0 =5 x 10

"d3.0 - SEPk.RATION CRITERIA

REF. 13

2.0 _"RESSURE RATIO AT

POINT OF SEPERATION

X 2.0 2.4 2.8 3.k 3.6 4.0

MO

FIGURE 46. CONE STATIC PRESSURE RATIO AND SEPERATION CRITERIA

4:
"
it.
 2-37

INIETWITHOUT
SLOT AND SURGE

TOHURGMCHMBE

0.090OLU 0.MPTYUS ER
ANDT
SEALEDPACT
TAS SURGE CHAMBER

FIGURE 47. BUZZ SUPRESSOR INLET TEST MODEL

Mo =2.35 a =0 0
ISENTROPIC INLET

LAST STAaU POINT WITH(IJT %AGIEDWAOM

'01

~PG
W,,FCHiANI 3R CLQT

LA.STSTABLEPOINT WITH 9JRGE CRAt.'I3Ef

* //

2-38

kA
,0 30 ISENTROPIC SPIKE INLET - DESIGN MACN NO. 2.0
'.
:. I I
ANGLE OF ATTACK

r,- 0,

lu 2.2 2.3 2.4 "" 2.5

MACH NUMBER 2., 2.7
0 WITH SUPPRESSOR

0 WITHOUT SUPPRESSOR

W U CONE INL-T - DESIGN MACH NO. 2.57

W

00 ANGLEOF ATTACK

2.2 2.3 2.4 2.6 2.6 2.7

MACH NUMBER

"FIGURE 49. INLET MODEL TEST RESULTS WITH BUZZ SUPRESSOR

p.4

A~
' ACOUSTIC BUZZ SUPPRESSOR

% " FILED MAY 29, 19W

"- INMVINTOI
* EDWARD.. FRENCH
* ST,4"2 J. MoNtOM
F7 147 _
 2-39

0 0
Mo 2.50 2
o 4
A6

C3 REVERED FLOW41

UJ _ __ _ - 12
C/) 14

-2

MASS-FLOW RATIO -

FIGURE 51. DFVLR MULTIPLE (4) AF7 MOUNTED INLET TEST DATA -"X" CONFIGURATION

0101
.8 - 2. - -:k 0 -2

A6

*Li REVERM FLOW

uj4 10

Z +12
cl-

-2

J fl~~ASS-FLOW
RATIO .-

FIGURE 52. DFVLR MULTIPLE (4) AFT MOUNTED INLET TEST DATA - + 'CONFIGURATION
2-40

x/d)-

M 0 2.0
MR
REYNOLDS No. 3.10'
V= PT-')TO

+x /d 12
 2-41

1st LONGITUDINAL 1st & 2nd

MODE TANGENTIAL_
I-- -MODES

POW ER-ER --

SPECTRAL--
DENSITY
(UNIT 2 /Hz) - - - ---

0.010- - -- - --

0F

10 100 1000 5000

FREQUENCY (Hz)

FIGURE 54. REPRESENTATIVE RAMJET PRESSURE OSCILLATION POWER SPECTRAL

DENSITY TEST DATA

PRESSURE
CONFIGURATION MACH ALTITUDE
NUMBER (FEET)

0 ANNULAR REGENERATIVE BURNER 1.87 45,000

o LIQUID FUEL-INJECTION BURNER 1.87 45,000
* ROTATING-DISK OSCILLATOR 1.87 45,000
0 ROTATING-DISK OSCILLATOR 1.87 55,000
wu '4 3.6 INCH RAMJET 1.92 47,000

0Lu

cn
0 0
U. 0i 0. . .3 04 05 0. . .
RATIO OFTT=MA0.U3(EKT PA)FPUSTO
TO FRESRA OTLPESR
2-42

FOURAXISYhMIETRIC INL.ETS
HYDRAULIC NOBOUNDARYLAYERBLEED
COMMANDBGROUP

AUTOPILOT SECTIO.N \PRESSURE V1ALIVE GENERATOR\ MOTOAR

IGIE
TAOTE UALAT

PqUN-U

BURN
OUT FUULL
INJECTOR
RING

FIGURE 56. GORJE PROPULSION TEST VEHICLE

85 'INLET WIND TUNNEL

TEST DATAI
CRITIICAL BZ
8C-PRESSURE____
0-
- __10% 24P "* a tIII aPII i
F75MARGIN MIN MAX

=) 7

LL 0 P FROM 0-GRAPH TYPICAL 0-GRAPH TRACE

STEAY STATE DATAI

65 =ER-TOPAK AMPLITUDE
0.8 0.9 1.0 1.1 1.2
EQUIVALENCE RATIO I

FIGURE 57. EFFECT OF PRESSURE OSCILLATIONS ON INLET MARGIN

 2-43

HIGH
U ALTITUDE
uj CRUISE

0 IACCELERATION
U)
w RAMJET TAKE OVER

FLIGHT TIME

FIGURE W. REPRESENTATIVE RAMJET INLET PRESbURE MARGIN HISTORY

RAMJET
w TAKE OVER

gJo

UA

.J CRISL

0.

FRE-TRA TOTATTEPERTUR
FIGUR
5CRUISTRhiUE RSSR SILAINTSTEAACRRLTO
 77..

2-44

0.65

0A

0.60 AP1.7 -

MAX AMPIU~hD EGN

PRESSURE s

U)

IJ
; AP.7 E O T -EA
M LT D

a ILETWID TNNE DATAT

TSTED
FUE4AI RATIO-

WIT
INLE LAYERIBUEE
BOUNDARYEA

IA
*AR TUNELUETEDAT

CFRATION
AINR
FUEALAR

FIGURE 61. AIR LAUNCHEDSDUCTED RSILTOCKE POWNERE MARISSLE DPIMAENSIONA

 2-45

BOTTOMMMOUNTED
2_D INLETS

AXXISYMI'I.TRIC
~T2 AXISYMMETRIC

~~TuTE MOUTEETSET

FIGUR62. NOSTLED INLET PROMNECMAIO

FOURTE
INLETS

EOEUO NOZLFTA

~
M ~ ~ N
MINLETDINET

00 00

FIGURE 62. INSALET/EJECTO TESTRMNC

CONFGU ATIOSO

STA.
2-46

BY PASS DUCT

-- 44

VENTURI 4

VARIABLE AREA

VENTURI EJECTOR TEST ITEM

STILLING
CHAMBER ORIFICE STILLING CHAMBER

FIGURE 64. EJECTOR/MIXER TEST SETUP

a--

FIGURE 65. TEST SETUP - REAR VIEW

 2-47

1.0 CONFIGURATION

0.9
Total Pressure
Ratio
42

0.7

0.6
0 0.1 0.2 0.3 0.4
Primary to Secondary Flow Ratio
WpIW 5

rIGURE 68. TYPICAL EJECTOR/MIXER PERFORMANCE TRENDS WITH FLOW RATIO AND
MIXER EXIT MACH NUMBER

.4

1.0 __ ---0,__ 4
0__

STATIC
PRESSURE
RATIO 0 0
0.9 x0
p -00

wptis
0.-8 .254
.182
o .136
o3 .116

0.7 _ _ _ __ _ * 0

DIAMETERS FROM EJECTOR HEAD -L/D

FIGURE 67. FOUR INLET DIFFUSER/ MIXER STATIC PRESSURE DISTRIBUTIONS

2-48

p2 . 0 pD P2.5 2. 7 3 3. ..

1.01''

STATIC
PRESSURE
RATIO 0.0 ___ a hi
PxIPT 2

0.8 ,S~ FOU~ IoxNLETS .:

A TWO INLETS I,'.
(3ONE IN,.ET-,F'
"2 0 2 4 6 8 10 12 14
DIAMETERS FROM EJECTOR HEAD - LID

FIGURE 68. FOUR, TWO AND ONE INLET STATIC PRESSURE DISTRIBUTIONS ".

MIXER EXIT MACH NUMBER (M 4 120.25

EJECTOR EJECTORJMIXER TOTALPRESSURE RECOVERY -, *
* CONFIOURATION
TWO INLET PT41T2
P/T : .. 4
CONFIGZURATIONS Wp/Ws - 0141 Wp/W 0 3, J
-a

0.943 1.075

(~
)0.941 toes9

0,942 lIVS ;-A

SINGLE INLET
CONFIGURATIONS

0.913 1.035

0,917 1.044

0,923 1052_,_

- .5-5

FIGURE 69. REPRESENTATIVE EJECTOR/MIXER PERFORMANCE

;7' ..
 2-49

L0

1.10 MIXER EXIT MACH NUMBER (M 4 )=-0.25

1.00

TOTAL -- N NE
PRESSURE
RECOVERY 09 _____

0 .0 .00.30 0.40
PRIMARY TO SECONDARY FLOW RATIO-Wp/WS

FIGURE 70. EJECTOR/MIXER PERFORMANCE~ COMPARISON

a.
 3-1

SOME ASPECTS OF ENGINE AND AIRFRAME INTEGRATION FOR

RAMJET AND RAMROCKET POWERED MISSILES

by

E.L.Goldsmith
Royal Aircraft Establishment, Aerodynamics Department
Bedford MK41 6AE UK

SUMMARY
Ideally the subject of integration of a ramjet or ramrocket engine with an airframe
covers all aspects of the differences between the performance of a complete configuration
of a missile with air breathing propulsion 1nmtR11ed and the test bed performance of an
( engine added to the external aerodynamic characteristics of a body, wings and controls but
without nacelles.
The particular aspects that are covered herein concern the nature of the flow
around a long circular body and the internal performance of single and multiple intakes in
this flow. To understand this performance, the performance of isolated intakes at inci-
dence and yaw conditions has been reviewed, together with ways that have been suggested
for improving performance by use of variable geometry. The second half of the paper is
concerned with external or airframe aspects of Integration ie the measurement of forces
and moments on missiles with and without nacelles and the estimation of drag components
associated with nacelles.
NOMENCLATURE
A area Re Reynolds number

a speed of sound r,R radius

b wing span x,y,z spatial coordinates

CD drag coefficient a angle of incidence

(based on Ac or Am)
I ratio of specific heats
CN normal force coefficient
(based on Ab) 8 surface elope

d diameter 6 boundary layer thickness

D drag 4 roll angle/circumferential
position measured from
L length windward bottom generator

M Mach number 8 nose angle

P total pressure tl,PR pressure recovery P

p static pressure
P density
SR gas constant
0 SUFFICES

* b body OP operating condition

c cone/capture plane crit critical flow condition

d duct t total or stagnation

f at the end of the subsonic w,p wave

diffuser or in the combus-
tion chamber, W diverter
"* skin friction
L local
CP centre of pressure
S survey station
S1 lip position
free stream
S-
 3-2

1 INTRODUCTION
"The topics of airframe engine interaction and the closely allied, but not synono-
*\ mous, one of airframe engine integration, have been subjects of international and of
national conference papers since the advent of Zet propulsion t # P . In early designs of
jet propelled aircraft this interaction was often of a loosely coupled nature. With the
air intake in the nose of the aeroplane (as on the Oloster Whittle jet, the F86 and the
P100 etc aircraft) the interaction was more a 1 oblem of design than of aerodynamic per-
"formance. The designer's problem was the location of fuel, armament, radar (and not leact
the pilot) when the fuselage was largely occupied by duct and jet engine. When two
engines were needed the problem was more difficult, as witnessed by the convolutions of
the ducts of the British Aerospace Lightning aircraft, and engines were often placed in
separate nacelles on the wirgs as on the Messerschmitt 262, Gloster Meteor and more
recently SR71 aircraft. Here again, the interaction was more on overall aircraft design,
wing mounted nacelles affecting rolling moments of inertia, engine out yawing moments etc.

Many factors have been responsible for a closer integration of engine and airframe
* for military aircraft, so that engine or engines remain in the body but the duct does not
* "completely obstruct the main carrying capacity of the fuselage. All current military
"* aircraft other than the large long range aircraft such as B52 and B1 adhere to this
formula.
The history of missiles with air breathing propulsion has followed a similar evolu-
* tion. Early liquid fuel ramjet designs of the 1950s were a mixture of pod mounted
"* (Bomarc and Eloodhound Figs 1 and 2) and nose intake designs (Talcs Fig 3) and even in
the 1960s, Sea Dart (Fig 4) continued this approach. With the need for more volume fc!'
stowage of fuel and warnead in the body and the particular needs of the guidance and
homing radar to occupy an unobstructed position in the nose, more recent missile designs
have almost exclusively followed the aircraft trend. Intakes have been placed on the body
D ~back
sides,along
top or
the bottom or in all
body length, as four positions
in the and 5).
ALVRJ (Fig usually
From located half
the point of way or ofeven
view further
intake and
engine efficiency, undoubtedly the best place for en intake is under the body and close to
the nose as in ASALM (Fig 6). The missile must then be controlled as an aircraft by
banking to turn (twist and steer) so that whatever manoeuvre is pursued the intake always
"remains on the underside of the body at positive incidence. Arguments between the merits
* and elements of twist and steer versus Cartesian control for air-breathing weapons will
probably ccntinue indefinitely.

With engines in the body and intakes on the side and/or top and bottom of the body
interaction between body and intake aerodynamics is closely coupled. At small angles of
incidence and in the absence of an adequate bleed or diverter, intake internal performance
is largely determined by the severity of the interaction of the intake shock system with
the body boundary layer. At larger angles of incidence the body flow field has a domi-
nating influence on intake internal performance and on the quality of the flow delivered
to the engine. Drag, lift and pitching moment increments due to the addition of nacelles
can be comparable to those due to wings and body. Other more subtle influences are felt
for instance in the design of the S-shaped ducts that are required to tirn the air into
* the body. If this portion of the duct can be kept short then less useful volume in the
- body is consumed.
The object of successful engine-airframe integration is to make these closely
coupled configurations at least workable and possibly even competitive with nose mounted
.- intake or single intake bank-to-turn configurations. The object of this paper is to
asummarise and correlate, where possible, some of the published (and unpublished) material
that is relevant to this aim.

2 BODY FLOW FIELDS

2.1 FLOW AROUND SLENDER CIRCULAR SECTION BODIES AT INCIDENCES FROM 0' TO 20"
For ogive nosed axi-symmetric bodies, flow changes in typical intake locations
compared to freestream conditions are fairly minor in the incidence range 0"-4" or 5' in
relation to the very considerable changes that take place between 6' and 20'.
Nevertheless, if an intake is placed in the nose flowfield effects are not insignificant
and indeed with the right choice of nose shape can be favourable to intake performance.
Fig 7 shows the variation in Mach number on the body surface with change in body shape at
M 2.0 and a 0. The tangent ogive shape has advantages over only the first half of the
nose length. Fig 8 Ehows how the effect of the nose flow changes with distance away from
the body surface and position downstream from the nose. Radial changes near to the
tangent point of the nose profile are significant but are not so further downstream. At
small angles of incidenc differences in local Mach number and flow angle start to appear
"between top (leeward) side and bottom (windward!) of the body which again vary with radial
"and longitudinal position. The magnitude of uhese changes are illustrated in Fig 9. As
incidence increases the boundary layer thickens considerably on the lee side of the body
(Pig 10) and at incidences typically in the range 6"-10" (depending primarily on body nose
angle), boundary layer ac(immulation on the lee side forms into two discrete vortices
(Fig 11). When this happens, the flow around the upper half of the body becomes complex
and only windward and side positions can be expected to have reasonably uniform local
 3-3

A complete flow field is shown in Fig 14. Intake centrelines are usually in the
region r-r /r of 0.3-0.5. On top and bottom generators local values of incidence are
half to tw8 thirds the freestream value. On the body side local incidence is 50-60%
greater than freestream incidence. In the upper quadrant, local incidence is of the same
order but the flow is now very non-uniform with local cross flows that may be supersonic
and accompanied by embedded shocks. Even away from the core of the vortex where the
intake might be expected to avoid ingesting very low total pressure air, the combined
effect of high sidewash and upwash will be deleterious to any type or orientation of
intake.

Experimental evidence on longitudinal location of vortex separation (Fig 15),

together with vertical and lateral location of tha vortex cores (Figs 16 and 17), is
highly variable. This is probably due to the influences of Reynolds number, position Rnd
size of boundary layer trip and Mach number as well as the main geometric influence of
nose fineness ratio. Most experimental observations however agree that vortex cores are
located circumferentially between 20" and 40" from the top generator position (Fig 18) and
this vitally affects intake performance in this region. Although the exact position of
the twin symmetrical vortex cores is difficult to predict there Is even more variability
when at a higher angle of incidence the vortex cores become asymmetric and can vary in
position with time. Fortunately, as shown in Fig 19, the combination of nose fineness
ratio and incidence that leads to the production of asymmetric vortices is generally out-
Sside the region of interest to missiles with air-breathing engines (Ln/db 2 to 4 and a up
to 20").

2.2 THE INFLUENCE OF BODY MOUNTED STRAKES

When a body-mounted intake is rolled from the top position it must inevitably pass
through or close to the vortex cores that are shed by the body at incidence (Fig 20).
Howevor, If sharp edged strakes are placed along the length of the body parallel to the
centreline then these strakes will fix cross flow separation at a specific circumferential
* position. This position then moves round as roll angle is varied instead of remaining
fixed in space as in the case with the unstraked body (Fig 20). Measurements have been
made from vapour screen photographs taken with a camera mounted on the sting and
downstream of a straked body. Data for a 10" at M* 1.6 is shown in Fig 21. Vortex
positions are shown relative to the viewpoint of an observer on the top strake looking
upstream as the body is rolled from the top position to the side position (# 180' to 90").
S
, The particular orientation of the strake shown is for + 165" but the envelope of the
/* positions applies for all values of roll angle. It should be emphasised that these are
riot the absolute positions in space of the vortices but are the positions relative to the
i] body as roll angle is changed. An indication of the size of the vortex cores as Judged
from the photographs is given by the width of the envelopes enclosing the vortex centre
positions. It is clear that a location midway between the strakes is not a good position
for an intake at this incidence of 10" as it will be affected by the vortices shed by the
strakes on either side of it at the appropriate positive or negative roll angles. However
there is a small region directly over the top of the strakes which is free (or nearly so)
of vorticity at all positive and negative roll angles. This would be the best location
for an intake (Fig 20). A similar situation would be obtained by the use of a square-
section body, where the best position for an intake would be at the corners of the body
(Fig 20).
Variation of pitot pressure for a radial position corresponding to the centre of a
typical intake at L/db 7.5 is shown in Fig 22 as roll angle is varied at incidences cf 10'
and 14" both with and without strakes. Without strakes a large trough of low total
pressure at about + 150' appears as the pitot tube passes close to the vortex centre.
With strakes this trough almost disappears at the lower incidence and is much reduced in
extent at 14" incidence.

3 INTAKE AND ENGINE MATCHING AND DEFINITIONS OF INTAKE PRESSURE RECOVERY

Before considering intake performance, some digression is needed on the bases of

comparison of pressure recovery. The intake flow states of supercritical, critical and
"subcritical are too well known to require explanation. Pressure recovery at the critical
flow condition n has been a convenient performance parameter for use in aircraft
where variable gjSStry intakes and only small excursions from cruise incidence at super-
"- sonic speeds has also made this condition close to the matched intake and engine flow
Sstate. This is often not co however for highly manoeuvring missiles with fixed geometry
intakes.

Matching of intake and engine flows for liquid fuelled ramjets is a relatively
". simple procedure.
Weight flow through a streamtube of cross sectional area A is:

W - pVA paMA pMA

p

*R
 77~~~ 7- AF V71 -.

3-4

'1-
A* M 1 + 1 ]. )

[Y+1]- tT
1-T '2'
W -AP (1) ..

At supersonic speeds when the nozzle at the end of the combustion chamber is
choked:
Yn+1
Wnn -n.A
nas (2)z
t
n
At entry to the combustion chamber:

yf+l
T
W =. +1Yt A /ANN (3)
f~~~ - -2 ;,A)

If the added fuel weight and the difference in y and R between the incoming air
and the exit gas stream after combustion are ignored then: ,
AnPn AfPf IAA . ."
W - Wn and th erefo re - . - ..

tn tf

and if nozzle exit area An is constant

A . constant - (4)

tf

Similarly the incoming free stream tube flow W. , Wf so that

Ttf
A () .' .~. ~ ,* ___-_.'A

and hence / 1 because T T

A ,t.
thus n constant n (5)
c f -

as Pn/Pf 1.0
Engine operating curves for different values of heat addition Tt /Tt on the intake
characteristic of A /Aversus Pf/Pare thusstraight lines thrn,.:b thenorijin with
differing slopes. cva ti

In particular from equations (4) and (5) it follows that:

 3-5

TABLE I

Variable Varible
t t Mf coperating Intake

Ttn Ttf M point

If more fuel increases constant decreases decreases )A /A decreases at

is added )cons~ant P /P or
)P /P incarases at
If altitude is constant decreases decreases decreases )cgnstant A /A .Intake
increased from )moves fromesu~er-
sea level )critical to suberiti-
towards the )cal or from critical
tropopause )to subcritical
)conditions

If flight constant increases increases increases )A /A increases at

velocity )c~ns~art P /P or Pf/P_
increases at )decreases Et
constant )constant A /A
altitude )Intake moves irom
)subcritical towards
If less fuel decreases constant increases increases )supercritical
is added )conditions

From this it can be seen that (a) it may be possible to keep the intake at approxi-
mately the sat-e condition as the ramjet engine accelerates and climbs from a ground launch
(b) for a given temperature rise there is a tendency for acceleration to be limited as
intake recovery falls in the supercritical regime and thrust decreases.

The ramjet engine is opposite to the turbojet in its reaction to ambient and to
engine conditions eg reduction of turbojet rpm on closing the throttle is associated with
reduction in fuel flow and results in the intake moving towards the subcritical regime
whereas closing the throttle and reducing fuel flow in the ramjet results in moving the
intake towards the supercritical flow condition.

If the ramrocket and solid fuel ramjet are viewed as devices that increase the
total temperature in the combustion chamber the same matching precepts apply. For the
ramrocket the matching process could be complicated by the fuel rich under-expanded gas
from a central nozzle restricting the passage of incoming air from the intake duct. In
that case, part of the matching process could be similar to that encountered between
secondary and primary itreams of a supersonic ejector nozzle.

Fig 23 shows match points at the intersection of intake characteristics and the
straight line constant M (constant temperature ratio) curves of an engine characteristic
as angle of incidence chgnges at constant freestream Mach number. The parameters ncrit'
the constant engine demand nlimiting a and nra
operating are
aemd made clear
la byytemthn
the matching

symbols used. noperating or nOp is a more sophisticated version of n limiting a in that it

follows a constant engine demand line up to a limiting incidence a such as a 2 and then by
decrease in fuel flow the operating point is made to move across engine demand lines to
stay within the stable flow limit of an intake. nutilisable as the suffix suggests is
French in origin. It is the pressure recovery at zero incidence imposed by the need for
keeping constant engine demand lines at incidence within intake stable flow limits. When
its value is lower than the zero incidence value the ldecrement is the penalty in pressure
recovery incurred at zero incidence for operating the intake in a stable condition at
incidence.

In this paper ncrit and hop are used. The former because it is widely available in
the literature and the latter as it is a more realistic matching condition but
nevertheless exploits the advantages of a ramjet than can be throttled.

4 EFFECT OF FLOW ANGLE VARIATION AND OF BODY FLOWFIELD ON INTAKE PERFORMANCE

4.1 EFFECT OF FLOW ANGLE VARIATION ON PERFORMANCE OF ISOLATED SUPERSONIC INTAKES

It is important to know the effect of flow angle variation on isolated intake

performance before studying effects of the body flowfield.

For supersonic intakes with wedge compression surfaces that are normal to the plane
of incidence, probably the most important influence on pressure recovery is the variation
of shock loss with change in attitude of the compression surfaces. Fig 24 shows the
 3-6

indicates that extra-to-shock losses are controlled so that pressure recovery change with
incidence closely parallels the variation of shock recovery. If the wedge compression
surface Is parallel to the plane of incidence variation of shock recovery is not easily
calculable. Plow over the yawed wedge compression surfaces is largely influenced by the
presence or absence of swept endwalls that enclose the compression wedges. Flow expanding
round the windward endwall may increase local Mach number in this reion and the wedge
shock will approach or enter, the cowl lip. On the leeward endwall the opposite will
occur. Reduced pressure recovery will result from coalescence of wedge shocks within the q
cowl lip and possibly from separation and vortex formation along the inside of the swept
windward endwall. Removal of both and in particular the windward swept endwall reduces
the fall off of pressure recovery with increase in angle of incidence (Fig 25). However,
this increases drag and reduces pressure recovery at zero incidence due to supersonic
spillage at the ends of tha compression surfaces.

The flow pattern for an axisymmetric centrebody intake is generally more complex.
The initial cone shock wave angle changes by much less than the change of uentrebody
centreline angle so that the windward lip moves towards the cone shock and the leeward
lip away from it. Thus the variation of pressure recovery is largely affected by the
shock wave angle a relative to the lip position angle 01 at zero incidence. If 0 a 01
then intersection 'f cone and normal shock occurs within the windward cowl lip as vncil
dence is increased, with a consequent loss in pressure recovery. The effect is clearly
shown for single and two cone centrebody intakes in Fig 25(a). As freestream Mach number
is reduced from the shock-on-lip Mach number (MsoL) there is a reduction in both

terminal normal shock Mach number and a movement of the conical shocks away from the cowl
lip which lessens the fall in recovery with incidence (Fig 27(a)). Decrease in 0.
(increase in MSOL) to remove the cone shock from the cowl lip however results in a longer
compression surface. This gives an increasing thickness of boundary layer which tends to
accummulate on the leeside of the centrebody and results in lower recovery and increased
engine face flow distortion (Fig 28). The corresponding changes to maximum flow ratio
are shown in Figs 26(b) and 27(b). .
4.2 REDUCING THE SENSITIVITY OF ISOLATED INTAKES TO CHANGE IN FLOW ANGLE

It is not difficult to devise ways of decreasing the rate of fall of pressure

recovery as incidence increases in one direction. Intakes for highly manoeuvrable
aircraft or missiles can be placed under the body or wing, so that local flow 2 ngle at
intake entry plane is practically invariant with change of angle of incidence . At
supersonic speeds, increase in angle of incidence also decreases entry Mach number,
usually with little decrease in local total pressure. The same result can be achieved by
the addition of a plane surface above and in front of an intake. This is the principle
used In te staggered or raked side intake proposed for highly manoeuvrable strike-fighter
aircraft * A rather larger pre-compression shroud has been shown to increase pressure
recovery and capture flow for a rectang mr intake at supersonic speeds in Ref 25
(Fig 29). The uzi of swept or unswept f .. wr plates for half axisymmetric intakes and a .
visor with axisymmetric intakes have bec- suggested (Ref 26, 27). At positive incidences
above about 6' the half axisymmetric Intaxe is generally superior (Fig 30) to the
axlsymmetric intake with or without a visor.

A slightly increased recovery is obtained at high incidence by pivoting the centre-

body of an axisymmetric intake so that it is aligned with t1'e incident flow. This is
further enhanced by raking the entry plane so that the lower lip is clear of the conical
shock (Fig 31). If the intake has to be equally efficient at both positive and negative
incidences then the further complexity of pivoting both the centrebody and cowl is needed.
Fig 31 shows this combination to he very efficient up to the highest angle of incidence
tested (14'). Another variable geometry arrangement is the asymmetric compression surface
housed within a circular section nacelle . The only tests of this arrangement showed
inferior performance at a o" to the equivalent axisymmetric intake but better performance
at incidence (Fig 32). Addition of boundary layer control on the compression surface and
a more refined design method could probably enhance the performance of this intake which
in concept should roll so that the compression surface is always in a favourable position
with respect to the incident flow angle.

4.3 EFFECT OF BODY FLOWFIELD ON SINGLE INTAKE PRESSURE RECOVERY

Single intake pressure recovery is a function of angles of incidence and roll, of
the longitudinal position of the intake on the body and of the boundary layer bleed or
diverter height. At incidences up to about 6" before twin body vortices are formed
(particularly if the intake is not far downstream of the body nose tangent point) the
accummulation of boundary layer air on the leeside ensures that minimum recovery occurs at
the # 180" position (Figs 33 and 34). At a 10' the inpestion of a body vortex into the
intake results in a minimum recovery at the + 130"-150 position in accordance with body
flowfield data (Fig 34).

The underbody intake (# 0') has an innate superiority at all incidences over all
other position3 (Figs 33 and 34), including the nose intake (Fig 26). However, despite
the decrease in boundary layer thickneos with increase in incidence on the windward side
of the body (Fig 10), the effect of body cross flow leads to rapid .eterioration of
recovery as roll angle increases from 0' to 45' at a 10' (Figs 33 a-d 314) S
 3-7

Wedge intakes follow the same trends as halr axisymmetric intakes (FiS 35). The
beneficial effect of throat bleed and the deleterious effect of placing the intake on a
flat face on the body should be noted (Fig 36). An inverted two wedge intake follows the
same pattern with respect to variation of recovery with roll angle as the other intake
types 'Fig 37). The graph summarising recovery variation with roll angle at a 10"
(Fig 38) is notable for showing the similarities rather than disparities in performance as
intake type and freestream Mach number are varied which confirms the unive:sality of the -
influence of ingestion of a body vortex. The graph summarising side intake performance
(Fig 39) in contrast shows the wide variation in recovery obtained with different intake
types in cross flow. This assumes great importance in the performance of multiple intakes
as will be shown in the following section.

4.4 MULTIPLE INTAKE PERFORMANCE

The performance of multiple intakes is inferior to that of single intakes for the
same incidence and roll angles for two reasons. The first is associated with the extra
loss due to mixing (and in some cases due to 'dumping' ie a sudden enlargement loss) of
the flow from the separate ducts at the entry to the combustion chamber. The second, with
the matching that is imposed on individual duct flows by compatability conditions at the
mixing plane which results in each intake operating at a different point on its individual
characteristic of recovery versus flow.
Some indication of the relation Vtween losses due to mixing and the angle at which
"*:the duct flows merge is given in Fig 40 . Overall loss for multiple intakes at 0"/180" *1
roll angle as incidence Is varied appears to collapse reasonably in to bands dependent on
intake type by plotting the ratio of kinetic energy efficiency to that at zero incidence
(at critical point conditions), Fig 41. Kinetic energy efficiency is defined as the ratio
of the kinetic energy of the diffused stream after isentropic expansion back to freestr.am
to that in the freestream. Its relationship to pressure recovery Is shown in Fig 42.
Variation of critical flow ratio with incidence is as important as the variation of
pressure recovery and this is shown for # 0* or 180" in Fig 43.

The influence of body vortices at angles of incidence abcve 6"-8" Q-d roll angles
of 20*-35" from the top position is as large as it is for single intakes. Fig 44 shows
values for recovery variation with roll angle at a 8" for four axisymmetric intakes at two
different diverter heights and with strakes placed between the intakes. At this incidence
the affect of strakes is to anchor the body vortices in such a position that places them
below the intake. At a 12" however (Fig 44) the stronger vortices shed by the strakes now
impinge on the intakes and performance suffers accordingly. As with single intakes, the
influence of increased diverter height is much smaller at this incidence than at lower
angles.
Figs 45 and 46 show contours of constant ni over a range of both incidence and
roll angle for rectangular intakes mounted conven onally and inverted. Midmatching
between intakes leading to unstable operation does not allow measurements to be made at
some combinations of roll and incidence angles, particularly for the intakes mounted
conventionally. Tle matching of individual intake flows has a large influence on the
performance of multi-duct arrangements. Fig 47 shows that for two intakes at the top and
bottom positions pressure recovery is almost invariant with change of incidence up to 16%.
If four intakes are combined however the mean recovery of the top and bottom intakes falls
with increased incidence at almost the same rate as the mean of the two side intakes.

The matching condition of equal static pressure where the four intake flows
coincide results in the high performance at the top and bottom positions being reduced to
the performance of the side intakes by forcing the top and bottom intakes to operate in a
supercritical condition (Fig 48). Similarly when the body is rotated so that the two
leeside intakes are ingesting the body vortices then these intakes are forcing the
supercritical operation of the two windward intakes. The total extent of supercritical
operation imposed upon one intake when operating in the presence of the other three as
roll angle is varied is shown in Fig 49.
If the vortices can be prevented from being ingested, multiple intake performance
will be raised at the # 140"-160" position. However the poor performance of the side
intakes at the * - 90' position will still result in a minimal performance at this
condition.

5 MEASUREMENT OF FORCES AND MOMENTS ON MISSILES WITH NACELLES

5.1 TECHNIQUES OF MEASUREMENT AND EXPERIMENTAL PROCRAMMES

Considerable experience has been obtained in the measurement of forces and moments
on aircraft models that have nacelles with free flow ducts. Attempts to increase accurary
of measurement of external forces by mounting nacelles only on a balance have not been
successful for a variety of practical reasons. Apart from these reasons measurement of
forces on part of a complete configuration is not to be recommended because interference
forces can be carried on parts of a model that are non-metric as well as on parts that are
metric. It is better to measure forces and moments on a complete configuration and build
up incremental values by the addition of components such as nacelles, wings, controls etc
step by step. Models are almost invariably held by a rear sting with the ducted flow
 can be attached to either the metric or the non-metric pzrt ,f th :n.)-!!and cannot be
varied whilst the tunnel is operating. To provi,-e 6. rf-lert exit a.Pa :o ensure
supercritical operation of the intakes, the boatta. I 'e, . - n:. .les '-nd body may
have to be modified or truncated. Incremer'- I orceu %ptrtICLLa. - aecclet!d with
correct geometric representation of the hoattail and bare reg-,s ana t1-13ffect of'
underexpansion of tho jet issuing from the -eal c'l *' r .,:zles ,as -pposed to the free
flow annular nozzle of the model) have to be measured a sepa-- . Iel with augmented
jet total pressure.
Similar programmes of measurement of forLc.s i.- ,i..^ents -n " ileb piuS succezsive
additions of wings, controlc and nacelles have been "' .1 . A, ONEAA and RAE. The
models are all mounted on five or six component str;IT, La.ar-es that are shielded
from the duct flow, and measurements of base rress" % na internal turces are made to
correct balance measurements and obtain external forc;.b anM moments. The extensive tests
reported in Refs 36-38 were conducted with single and twi.. latake configurations with and
without a wing and tail controls. The intakes were both -ect.ngular and axisyimetric and
were tested with one value only of exit Pr~a f-1- 50). fE:cts in tunnel S: P~dane by ONERA
and in the 8ft x 8ft Tunnel at RAE ( - -d)htve been "one n models with ebout twice the
body diameter of the NASA model. The ONERA model ( 51) has a fixed choked exit with
exit contraction attached to the metric part of the model. Momentum at the choked eAit is
estimated from static or total pressure measurements upstream of the exit. Tn the RAE
model (Fig 52) the exit momentum station is upstream of .ha non-metric variable exit area.
Skin friction forces are estimated on the small portion rf the ou'pr wall of the annular
duct that is downstream of the rake end added to the , change evaluated from
freestream to the rake station. The total pressure rakes have large clearances from both
outer and inner duct walls so that metric parts of the model are not contacted when the
model deflects under normal force loads at high incidence. A calibration using a static
flow rig such as the Aircraft Research Associations Mach Simulation Tank (MST) is
necessary for the accurate evaluation of internal momentum and mass flow from the rake
pitot and static pressure measurements. The current range of model geometries and test
conditions is shown in Fig 53.

5.2 NORMAL PORCE AND CENTRE OF PRESSURE CHANGES DUE TO ADDITION OF NACELLES

Although nacelles are separate entities that are added to the bAsic body, results
of Refs 36-38 (unlike ONERA and RAE measurements) do not include body alone data. Normal
force and centre of pressure estimates for a cylindrical body with a tangent ogive nose by
the semi-empirical British Aerospace ABACUS method have been shown to be accurate.
Estimates using ABACUS and incremental forces measured due to the addition of tails on the
single intake configuration have enabled incremental normal !orces to be derived for both
single and twin intake configurations (Fig 54,55). Rectangular nacelles act as more effi-
cient wings than circular nacelles over the whole incidence range and, as would be
expected, the addition of nacelles and tails shifts the centre of preasure (measured from
the body apex) rearwards from the body al)ne position. Normal force increments measured
at ONERA on single and twin intakes at incidences from 0* to 7" agret very well with NASA ,..".
measurements. The addition of a single underbody intake results in a 3mall deterioration
to the lifting efficiency of the body. Increase in freestream Mach number from 2.5 to
3.95 leads to reduced normal force slope from O to 10" incidence but thereafter the slope
is little changed (Fig 56).
RAE and ONERA measurements have mainly been made on four-intake arrangements.
Variations with incidence of incremental normal force and of centre of pressure position
are shown in Fig 57 and exhibit the same trends as those observed for the two intake
configurations. Variation of normal force increment with roll angle for four rectangular
intakes is quite small (Fig 58).
Fig 59 summarises the changes in nacelle normal force increment in comparison with
wing normal force increments with net span which is relevant to launcher dimensional
constraints and with net planform area. On a planform basis obviously the nacelles are
inefficient with respect to wings but on a span basis this is not so evident.

6 DRAG OF MISSILES WITH AIR BREATHING PROPULSION

6.1 COMPONENTS OF DRAG AND THEIR MAGNITUDE

Total missile drag is made up of wave and skin friction drag of the four main
components, body, wings, controls and engine installation together with interference
between these components and the drag of miscellaneous items such as aerials, leaks,
surface roughness and excrescences such as cable ducts etc. Engine installation drag
components are cowl, pre-entry (or additive as it is known in the USA) intake spillage,
body boundary layer bleed or diverter, coopression surface bleed and afterbody.

These are illustrated diagrammatically in Fig 60 and all except afterbody drag will
be considered in more detail. The concept of pre-entry or additive drag is well known and
follows from the conventional definition of engine thrust as being the change of pressure
and momentum forces for the flow in the ingested streamtube from freestream to nozzle
exit. If engine and intake matching ensures that the intake is always operating in a
critical or supircritical condition then, depending on shock-on-lip Mach numL-r MSOL,
the intake is operating at datum flow (Fig 60(a)) or at maximum (or full) flow with
 3-9

confined to cowl wave drag D owl and skin friction drag. If at maximum flow, compres-
sian surface shocks are ahead ofothe cowl lip and there is a wave drag due to pre-encry or
supersonic fore-spillage, Dpreo, to be added to cowl wave drag (Fi6 60(b)) so that:

Dexto Dpreo + Dcowl

If the intake is opera Ing in the subcritical regime with the cowl lip shock
detached (Fig 60(c)) then:
"D" + D DpD + D +D
ext Dpre cowl pre cowl spill
or
DspllI Dpre - Dpre + Dcowl - Dcowl

The change in cowl wave drag ADcowl ( Dcowl - Dcowl ) is usually a small negative
0
quantity that slightly offsets the large increase in pre-entry drag (Dpre Dpre ) that
occurs when all or part of the fore-spillage is subsonic. Designers of missiles with
ramjet engines generally try to avoid operating intakes subnritically. However, if tho
intake has excessive (ie non-self starting) intertial ccntraction then the critical and
supercritical flow regime is identical to a subcritical flow regime as regards external
urag.
An illustration of the wide range of values for drag components for two designs
(nose mounted intake and side mounted intakes) ie shown in Table II.

TABLE II

Component Nose mounted intake Body side mounted intakes

Wave % Skin friction % Wave% Skin friction %

Cowl 35 2)
Diverter 12 ) 1
Afterbody 15 2 ) .:

Wings and controls 15 7 21

Body I4 ) 52

Miscellaneous 14

A more detailed br.,akdown of intake drag for body mounted nacelles has been given
in Ref 39 and is shown in Table III.
TABLE III

Intake type Rectangular % Half axisymmetric % Axisymmetric %

Component I
Cowl and endwalls 6.0 6.5 6.0
Pre-entry 1.5 2.0 1.7
Divertee or bleed h/d
0.033 2.5
0.053 2.0
0.066 6.5
0.1 8
Compression surface 5
bleed

Some of the large differences in componeist drag are due to different standards of
 3-10

6.2 ESTIMATION OF DRAG

6.2.1 AXISYMMETRIC AND HALF AXISYMMETRIC COWLS

Historically some of the first predictions of wave drag for circular pitot intakes
were obtained using linear or first order theory. Using the linearised equation:

- 82 32t + . o .
2 ar-ax r ar

where * is a perturbation potential defined by

SU - and - . .' ,
u ax p ar
and u - U+up, v VP

Up and v P U, up and vp c- a

Lighthill 0 O, 1 derived expressions for pressure coefficients op 2slender bodies with

both continuous and discontinuous variations in body slope. Fraenkel extended this to
open nosed bodies of revolution and obtainel expressions for drag of both conical and
parabolic cowl shapes. Willis and Randall applied slender body theory to a family of
curved profiles defined by: .

R . Rm - (R, - Rc)(1 - x)nwhere n ) 1

which includes the conical (n - 1) and the parabolic cowl (n - 2) (Fig 61) to obtain a
general expression for drag (if n is an integer and greater than unity) as:
2R A A
CD= -m-)n 2 n + 'n nm
n + + n + log, e ?-.- :

where an, Bn and yn are functions of n and are plotted in Fig 62.

2 ( e2
The conical profile (n - 1) has two discontinuities in slope (at the beginning and

C %~m 2
J'
1 cAcA e
end of the profile) and the expression for drag is:

( )+ + l og e 2 _ mloge
mm :''""
," - ' '.','."

A simpler expression for the wave drag of conical cowls was deduced by Ward using
quasi-cylinder rather than slender body appro-:Imations as:
C D /4 2 U1 L3 -"""
CD ac 1
8-.U (.-:)

where U (x) is shown plotted in Fig 63 and 6c is the cone semi-angle in radians and R is a
mean radius. Warren and Gunn 4 5 showed that the simplest mean radius - (R + HR/2 does
not give plausible results when RC becomes much smaller than Rn and the opeR noAe body

tends towards a solid cone. A weighted mean radius p2 _ (R2+R2)/2 gives a better result
2 2
for area ratios R /R from 1.0 to 0.2 so that the expression for drag coefficient based on " .
maximum area becokes."

2 1n 2c

Fraenkel 4 6 used quasi-cylinder approximations to derive the following expression .. :$

for the drag of parabolic profiles: '"%-

CD ,,,
2 +)2
26
.(L\
) i
...

where T(x) is also plotted in Fig 63. '-.,. .

LbiS the length of the full body and L is the length of the open nosed body which
' ,, , ' l i4
is adart of the full length
s body.
 3-11

slender body and quasi-cylinder approximations for conical and parabolic profiles are
given In Ref 146. Fracnkel gives a similar presentation combining slender body (for
OR /L - 0.05 to 0.3) and quasi-cylinder (for 0Rm/L - 0.6 to 1.0) values modified by the
Introduction of exact values for conical and parabolic pointed bodies at Ac/Am . 0.
Values from Ref 147are compared to the slender body resultsf Willis and Randall
and results from characteristics calculations for conical profiles 4fin Fig 614.
Both first order and Van Dyke's second order theooy become increasingly inaccurate
as Initial angle of the cowl approaches the freestream Mach angle. Cowls for centrebody
intakes are usually Inclined Initially near to the local flow direction from the conical
compression surfaces and hence their initial slopes can be close to or in excess of the
freestream Mach angle and other methods for the evaluation of wave drag have to be used.
The most accurate of these is the method of characteristics followed by (in order of
decreasing accuracy and increasing simplicity) the methods of second order shock-
expansion, generalised shock-expansion, tangent-wedge, and impact theory. In all these
methods it in convenient to divide the cowl into a number of straight line segments and to
evaluate the pressure coefficient on each segment before summing to obtain the wave drag.
Fig 65 illustrates the essentials of each method. Appendix B of Ref 149rontains all the
Information necessary to calculate the geometry of points on the input ray 1-6, at the
field points 3-ill (defined as the Intersection points of first and second family Mach
lines that are equally inclined to the streamline direction) and at body and shock wave
points 7 and 12.
Presley and Mossman in Ref 149give results of applying all these methods to the
same family of cowl profiles studied by Willis and Randall for one length and one area
ratio Ac/Am
Experimental values from a series of pressure plotted elliptic profile cowls that
can be approximated by the Willi3 and Randall family are given by Samanich in Ref 50.
These values of wave drag are shown plotted in Fi b6 in the linear theory form of
CD(L/Rm) 2 versus OR /L which effectively collapses all the results from different initial
slope cowls and Mach numbers for a given area ratio. Values from shock-expansion
calculations for the same profiles are not quite so well collapsed. The same values for
wave drag together with results from the Presley and Mossmam characteristics calculations
are plotted in Fig 67 using the hypersonic similarity parameters 0.7 CD M-2and 1M.So
The parameters also effectively collapoe both measured and characteristics values into2
curves that appear to be only functions of area ratio. However variation with area ratio
is different for the characteristics and the measured values as illustrated in Fig 68
probably because fineness ratio does not occur explicitly in the hypersonic similarity
parameters. Although fineness ratio does occur in the supersonic similarity parameters,
Fig 69 shown that characteristic values are not properly collapned for variation of both
Mach number and fineneeG ratio.
For the fineness ratios 0.14-2.3 and area ratios 0.5-0.8 of the experimental
measurements the shock-expansion method is probably sufficiently accurate for most
purposes until the Mach number at which the cowl shock wave becomes detached is
approached. Under these conditions if the cowl initial angle is not too high, linear
theory appears to give a reasonable re~sult (Fig 70).
6.2.2 PRE-ENTRY AND SPILLAGE DRAG
Pre-entry drag can be evaluated either by computing the change in pressure forces
and momentum for the flow in the ingested streantube from freestream conditions to the
intake entry plane or by inti-grating the preselare multiplied by projected area along the
stagnation streamline from freestream to Intake cowl lip.
Evaluation of pre-entry drag due to supersonic forespillage by integrating the
piressure along the stagnation streamline Is particularly simple for a two wedge or cone
compression surface:
DPre a (p1 -.
p)(A, -..
A) + (p2 -p.(e~

Some typical values for cone, wedge and double wedge compression surfaces with
compression surface shock waves focussed on the cowl lip at M% 2.14 are shown in Fig 71(a).
When the intake is operating with subsonic forespillage and the cowl lip 'shock is
detached evaluation of pre-entry drag (Fig 71(b)) for a two wedge or cone compression
surface becoiae:

Dpre a (pl-..)(Al-..A) + f (p -p.)dA

NS
and is more difficult because the variation of pressure along the stagnation streamline
from the detached normal shock to the cowl lip is not known. However as was stated in
@,..4.. ~lvn-"A%^i nra-ft*.,v Ara,e (illA ton ftihmofnin forP.Fhtn1 cP. in acromnanied by
 3-12 6
stagnation point
spill
- f
NS
(p-p.)dA + ADo
cowl
- (PNs-)(Ao-A.)
(P""""'

le the underestimate of pre-entry drag obtained by not accounting for the rise in static
pressure from the value behind the normal shock to the stagnation point value compensates

, Do, 10 A_
for the decrease in cowl drag (ADcowl is negative). This can be written alternatively as:

CD iiOIc
A
spill -c -c.
"MI

This method of 3st1mation obviously does not take account of differences in ADcowl
due to differences in the shape or sharp lipped cowls. It is found, generally, that above
M 1.6-1.7 the ADcowl term is small and for spillage up to 20-30% the effect of variation -

of cowl shape is very small.

The pitot intake assumption obviously applies directly for compression surface
intakes when the compressionr surface shocks are on or inside the cowl lip and spillage
drag evaluated on this basis at M 2.4 is shown in Fig 71(c). If the compression surface
shocks are outside the cowl lip a~d their intersection point with the detached normal
shock also remains outside the entering strearrtube as the intake spills, then on the same
basis (Fig 71(d))

Dspill - (pl-p.)( 1 -A-) + (P 2 -P.)(A-A ) + ( )(Ac -A 2 )

The position of the normal shock has to be found to evaluate A1 and this is done by
assuming a linear movement of the cowl lip shock from attached at full flow to the shop
position that would be obtained for a blunt body equivalent to the intake at zero flow.
The shock wave geometry can be treated in differing ways for a multiple shock ccmpression - -"
surface and analytical expressions for two-wedge intakes are given in Ref 53. - -:

6.2.3 DIVERTER AND BLEED DRAG

The boundary layer on the body can be prevented from entering the intake by raising
the intake from the surface of the body and then diverting or ducting away th'boundary
layer in the space between the intake and the body surface. In both cases the boundary
layer diversion or extraction will be associated with a drag.

BLEED DRAG
A ram scoop bleed may be preferable to a diverter if the main flow duct has to turn
into the body immediately downstream of the entry plane so that the angle of the diverter
wedge has to be large or if air is needed for some purpose such as cooling electronic
components. The bleed duct will perform in the same manner as the main engine air duct.
A correct choice of exit area will enable the intake to run full ie in a critical or
supercritical condition and at supersonic speeds the exit area to the bleed duct will be
choked unless the bleed pressure recovery is very low. A divergent nozzle can be added
downstream of the choked throat to expand the bleed flow to freestream or possibly base
static pressure and discharge the bleed air at supersonic velocity if sufficient bleed
total. pressure is available.
Bleed internal drag can be written as:
2
DBL ABL . 2+ x
re - / .-
CDqAA (Pex +'. ex 'qr AB
DBLn qA-C - pa / A.,%"

If the bleed air is choked at exit

ABL p P~ AB-
CD n- - 0.5283 ex - - ) 0.7396 A ,-q
Bint A . )~ ~'i
where the choked exit area for the bleed flow is:

ABL:x ABL-/A .
AC - e cB ": '" '''

A fA..
0 L x . ." *,o

BLe

If the bledd air is discharged at supersonic speed where ex

---
 //

3-13

(ABL - BL x L P ABL
cD - 2 x e eq ."-
BLint ex ".','-','$,

where MBL and hence -BL

ex and (A e follow from:
ex ex

PBLex P,/P
ex
77- P /P,
BLex BLex
and ABL 0
ex /\c
CTe ABL c

Values fo~r CDBL calculated from these two equations are shown in Fig 72 54 as

functions of M., bleed intake pressure recovery PBL /P. and bl(eer4 flow ABL /Ac- For a
normal ehock ram scoop bleed, bleed pressurz rccoveS values are shown typi'ally in
Fig 73'
The same equations can be used to calculate bleed dr.g for a compression surface
throat bleed (Fig 74). For a flush throat bleed, total pressure is found to be roughly
equal to the static pressure behind the final normal shock of the main intake and this
pressure! ,ecovery PNs/P. Is shown calculated for a l0O-i0" wedge compression surface in
Fig 73,.

DIVERTE . DRAG
The wave drag of wedge shaped boundary layer diverters (Fig 75) has been presented
in a correlation shown in Fig 76 (from Ref 55) for three different diverter height0o"8
0.75, 1.0 and 1.25. The same basic data (derived from NASA reports of the 1950s) o,',*
has been smooth and plotted in carpet form for h/8 1.0 in Fig 77(a). Change of diverter
drag with change in h/6 is shown in Fig 77(b) and illustrates some of the scatter of the
data.
Compressible flow flat plate boundar, layer data gives a good estimate of diverter
skin friction drag if the skin friction coefficicnt is based on total wetted area inside
the divertcr passage and the Reynolds number is based on the length of the diverter
passage. Alternatively Ref 55 gives diverter skin friction drag as:
CDf A .. -
CD fw.
fw fw p
where Cfw and the multiplication factor CD /Cfw are obtained from Fig 78. The Reynolds
number in this case is based on frestream gonditions and length of boundary layer upstream
of the diverter entry. Ap is the frontal area of the diverter and Aw the wetted area of """
the diverter passage.
7 CONCLUDING RE.MARKS
Although the internal flow efficiency of nose intakes and underbody intakes (when
combined with a bank-to-turn control system) will always be supericr to body mounted
multi-intakes, the overall efficiency taking into account external flow forces, volume
utilisation and control reaction time may not be so clearly demonstrable. The
considerable interaction effects of the body flow field on multi-intake performance and
nacelle contributions to external e'rces have been illustrated. Effects of body cross
flow and vortex ingestion have been highlighted. At the present time it is not clear what
type of intake (rectangular, half axisymmetric or axisyffnetric) or orientation is best
suited to these unfavourable flow fields at matched engine and intake airflows.
ACKNOWLEDGEMENTS

The author is indebted to his colleague Dr I. McGregor of the RAE and to

Mr C. Jell of BrItish Aerospace Dynamics Group (Filton) for their help in the preparation
of this paper.
REFERENCES
* - i~....... I "4..4 .. -. 4 ~.O.a. 4Al.i ~ Rrn~n1 A~r~v~ei~ 1Iti
AttV .T"hIV'fla
. [
 3-14

3 M.R. Nichols. Aerodynamics of airframe-engine integration of sapersonic aircraft.

AGARD C? Na 9, Pt I.
4 Integration of propulsion systems in airframec. AGARD CP No 27.

5 W.F. Davis, R. Scherrer. Aerodynamic principles :or the design of jet-engine

Induction systems. NACA RM A55F16.

6 L.E. Hasel, J.L. Lankford and A.W. Robins. Investigation of a half conical scoop
inlet mounted at five alternate circumferential locations around a circular
fuselage. Pressure recovery results at a Mach number of 2.01. NACA L53D30

7 L.E. Hasel. The performance of conical supersonic scoop inlets on circular

fuselages. NACA L53I14a.

8 J.L. Lankford. Preliminary results of flow surveys about an inclined body of

revolution. NAVORD Report 6708.
9 W.C. Ragsdale. Flow field measurements around an ogive cylinder at angles of
attack up to 15 degrees for Mach numbers 3.5 and 4.0. NOL TR 72-198.

10 L.H. Jorgensen, E.W. Perkins. Investigation of some wake vortex characteristics of

an inclined ogive-cylinder body at Mach number 1.98. NACA RM A55E31.

11 B.E. Tinling, C.Q. Allen. kn investigation of the normal force and vortex wake
characteristics of an ogive cylinder body at supersonic speeds. NASA TN D-1297.

12 M. Cooper, P. John and L.E. Hasel. A pressure distribution investigation of a

fineness ratio 12.2 parabolic body of revolution (RM10) at M - 1.59 and angles of
attack up to 36. NACA RM L52G14a.
13 J.F. Mello. Investigation of normal force distributions and wake vortex charac-
teristics of bodies of revolution at supersonic speeds. Journal of the Royal
Aeronautical Society, Vol 26, No 3, March 1959.

14 T.E. Chamberlain. Aerodynamic forces, separation lines and vortex trajectory on

slender bodies with several nose shapes. Massachusetts Institute of Technology
Thesis, June 1966.

15 M.R. Mendenhall, J.N. Nielsen. Effect of symmetrical vortex shedding on the

longitudinal aerodynamic characteristics of wing-body-tail configurations.
NASA CR-2473.

16 C.S. Brown, E.L. Goldsmith. Measurement of the internal performance of a

rectangular air intake with variable geometry, Part II. The affect of incidence.
RAE Technical Report 71119.

17 C.S. Brown, E.L. Goldsmith. Measurement of the internal performance of a

rectangular air intake mounted on a fuselage at Mach numberL from 1.6 to 2.0.
RAE Technical Report 72136.
18 M.A. Beheim. A preliminary investigation at Mach numbe., )1 of a diffuser
employing a pivoted cone to improve operation at angle of attack. NACA RM E53130.

19 M.A. Beheim, T.G. Piercy. Preliminary investigation of a shield to improve angle

of attack performance of a nacelle type inlet. NACA RM E57G25a.

20 J.F. Connors, R.C. Meyer. Performance characteristics of axisymmetric two-cone and

isentropic nose inlets at Mach number 1.9. NACA RM E55F29.

21 J.P. Connors, G.A. Wise and 3.C. Lovell. Investigation of translating double-cone
axisymmetric inlets with cowl projected areas 40 and 20 per cent of maximum at Mach ,

numbers from 3.0 to 2.0. NACA RM E57C06.

22 R. Hawkins. The effects of incidence upon the performance of 24*-31" double cone
intakes. Bristol Siddeley Engines Report AP 5081.

23 J.E. Hawkins. YP16 inlet design and performance. AIAA Papec No 74-1062.

24 J.A. Ross, I. McGregor and A.J. Priest. Some RAE research nn shielded and
unshielded fusalage mounted air intakes at subsonic and supersonic speeds.
Aerodynamics of power plant installation. AGARD CP 301.

25 JG. Bendot, A.E. Heins. Ramjet air induction system design for tactical missile
application. 5th International Sympoaium on Airbreathing Fngines, Feb 1981.

26 M.'. Beheim. A preliminary investigation at Mach numbei, 1.91 of an inlet

configuration designed for Insensitivity to posittve angle of attack operation.
 28 N.E. SamAnich. A.W. Cubbison. A pivoting cowl snd spike technique for efficient
angle or attaok operation of supersonic inlets. NASA RM E'"Jlla.

29 J.P. Connors, R.R. Woollett. Preliminary investirrtion or an ailsy-metric swept

none inlet of circular projection at Mach number 3.85. NASA RM E5426.

30 A.S. Valerino, C.B. Pennington and D.J. Vargo. Effect of cirumferential location
on angle of attack performance or twin half contal scoop type iilets mounted
symmetrically on the RMl0 body of revolution. NACA RM E53fm9.

31 E.J. Kremzier, R.C. Campbell. Angle of attack supersonic performance of a

* configuration consisting of a ram.p-t pe scoop inlet located either on top or botton
of a body of revolution. NACA RM E54C09.

32 E.J. Kremzier, R.C. Campbell. Effecc of fuselage fences .n the angle of attack
supersonic performance of a top inlet-fuselage configuration. NACA R1 E54JO.

33 G.A. Mitchell, B.O.Chiccine. Investigation of high angle of attack performance of

a 14' ramp type Inlet in various circumferential body locations. Mach number range
1.5 to 2.0. NACA RM E57C12a.

34 D.J. Vargo, P.4 Parks and O.H. Davis. Investigation of a hign performance top
inlet to Mach numter of 2.0 and at angles of attack to 20% NASA RM E57A21.

35 R. Hurd. Ramjet ei.gine intake performance data. Rolls Royce ON 18687.

36 C. Hayes. Aerodynamic characteristics of a series of single-inlet air breathing

missile configurations. NASk TM 84557.
37 C. Hayes. Aerodynamic characteristics of a series of twin-inlet air-breathine
missile configurations I. Axisymmetric inlets at supersonic speeds.
NASA TM 84558.

58 C. Hayes. Aerodynamic characteristics of a series of twin-inlet alr-breathlng

missile configurations II. Two-dimensional inlets at supersonic speed3.
NASA TM 84559.
39 0. Laruelle. Prises d'air four missile probatoire de statofusee. L'Aeronautique
et l'Aitronautique No 98 1983-1.

40 M.J. Lighthill. Supersonic flow past bodies of revolution. R&M 2003, Jan 1945.
41 M.J. Lighthill. Supersonic flow past slender bodies of revolution the slope iof
whose meridian section is discontinuous. Quarterly Journal of Mechanics and
Applied Maths, Vol I, Part I, March 1948.
"el42 L.E. Fraenkel. The theoretical wave drag of some bodies of revolution.
R&M 2842, May 1951.
43 J.H. Willis, D.0. Randall. The theoretical wave drag of open nose axisymmetrical
forebodies with varying fineness ratio, area ratio and nose angle.
RAE TN Aero 2360.

44 G.N. Ward. Supersonic flow past slender pointed bodies. Quarterly Journal of
0 Mechanics and Applied Maths, Vol 2, Part I, March 19119.
45 C.H.E. Warren, R.E.W. Gunn. Estimation of external drag of an axially symmetric
conical nose entry for jet engine at supersonic speeds. RAE TN Aero 1934.

46 -. E. Fraenkel. The theoretical wave drag of some bodies of revolution.

RAE Report Aero 2420.
447 L.E. Praenkel. Curves for estimating the wave drag of some bodies of revolution
"based on exact and approximate theories. RAE TN Aero 21811.

48 E.F. Valentine. External drag estimation for slender conical ducted bodies at high
Mach numbers and zpro angle of attack. NASA Technical Note D648, March 1961.

49 L.L Preslty, E.A. Mossman. A study of several theoretical methods for computing
the zero-lift wave drag of a family of open nosed bodies of revolution in the Mach
number ranre of 2.0 to 4.0. NACA Technical Note 4368, September 1958.

50 N.E. Samanich. Pressivre dray of axisymmetric cowls having large initial lip angles
at Mach numbers from 1.90 to 4.90. NASA Memo 1-10-59E.
51 L.E. Praenkel. The external drag of some pitot-type intakes at supersonic speeds,
Part I. RAE Report Aero 2380.
Ap WE. Moeckel. Approximate method for predicting form and location or detached
 3-16

53 I. Mcaregor. Some theoretical Oarameters relevant to the performance of

"rectangular air intakes with double-ramp compression surfaces at supersonic speeds.
"RAE TR 71232.
54 E.L. Coldsmith. Boundary layer bleed drag at supersonic speeds. R&M 3529.

aP 55 K.F. Little et al. Internal aerodynamics manual.

Corporation NR-68-H-434, Vol I-III.
North Amr-ican Rockwell

"56 T.O. Piercy H.W. Johnson. An experimental investigation at Mach numbers 1.88,
3.16 and 3 .A3 of pressure drag of wedge diverters simulating boundary layer removal
% systems for side intakes. NACA EM E53L14.

57 H.W. Johnson, T.G. Plercy. Effect of wedge type boundary layer diverters on
performance of half conical side intakes at a Mach number of 2.96. NACA RM E54E20.

58 P.C. Simon, K.L. Kowalski. Charts of boundary layer mass flow and momentum for
inlet performance analysis. Mach number range 0.2 to 5.0. NACA TN 3583.

59 R.C. Campbell, E.J. Kremzier. Performance of wedge type boundary layer diverters
"for side inlets at supersonic speeds. NACA RM E54C23.

60 L.E. Stitt, B.H. Anderson. Friction and pressure drag of boundary layer diverter
systems at Mach number 3.0. NASA TM X-147.

O .M Stationery office 1984.

fi:

I,.?

-4
|.

I:2;
 3-17

IE I

rnJ
LnN

-e Ln

Ln tA
C-,)

j L5
3-I N

Tangent Radial Local Mach number

ML ogive position f. ML
2.0 Ka~rman

1.9 , Newtonian 1.7 ... j

.. _- .______ .__iCone _____-_____.
1.8 moo 2.0 1 2 3 4 5 6 7 8X/db
LN
1. 3.0 Radial A., Local Mcch number
1b 2.6 position ,, A ML

1.0 0.2 0.4 0.6 0.8 1.0 25.':.i

rb 0. . 2.4
0 1 2 3, 4 5 6 7 'd '-
0.2 0.4 0.6 0.8 1 0 8X/db
FIG 8 VARIATION OF LOCAL MACH NUMBER
-LN ,, WITH RADIAL & LONGITUDINAL POSITION
FIG 7 VARIATION OF SURFACE MACH AT M 1.7 & 2.5
NO WITH CHANGE OF NOSE SHAPE (REF 5)
Leeward Side LS 364
2.1 -Radial Windward 640 b
30 ositiv2l. 1_________0___

30LN/db 0-P
1.9 35
2.1 3 4 6 S
7
34'68 /db 2,0-
1.0 135
Radial --- Leeward 0 .

Spos"tion Windward 0Incidence CC

1.8
S ' *I * 7 " FIG 10 DISTRIBUTION OF BOUNDARY LAYER
THICKNESS AROUND CIRCUMFERENCE OF BODY
1 2 3 4 5 6 7 8 X/db AT M2.o1 (REF 6)
FIG 9 CHANGE OF LOCAL MACHNO WITH
LONGITUDINAL POSTION AT 003 & 60 & N2.0

665 -or-
=60 / M.oo\ 2.00
C ==~ 12 '
oc04 =t=
-o M 00
= 6=2.0
L/db L/db
3 6.5 -6.5
2 4.0 -- I 4.0
p p 6:

Leeward Side Windward Leeward Side Windward

FI(; 11 FORMATION OF VORTICES ON LEE
SIDE OF BODY (R.EF 7) 4.0 Sd
CC20Sieid
SdL ML
16 L./ 3.0 Windward
12
SLN/d
MW 235"""
'b 2.0
L 1 400 .
ndwar d 1,39 1.0
0 4 A 12 1& 20 , I
 /

3-19

Ls 3,64

-J
3".5

0
I..--Jo
ML- M.190

2. 133 -O Gn Ref.
_Side 1.Incidence
0 2 4 6 8 10 12 18- \0 0w
Ogvcy~inder 20 1
IncidenceL 20 14

FIG 13 VARIATION OF LOCAL MACH NO 16 1, 20 14

WITH AN4GLEOF INCIDENCE (REF 6) 1
18 /d P6. 9 R.*dNo
. I5
s 35 1 I.S. -""'

I GS ) 03 10
- Cjv ,,.- 135...
135tW 10 0

8I-
8 "" Ref. 12

0X 0
-2-
.C '"'
2 4 6 d 10 12 14
o Vortex separation location
FIG 15 LONGITUDINAL LOCATION OF VORTEX
K" SEPARATION POSITION

4 0
. 2 Vertical 0 ,'
G
'01 zs position
0.5 lb 1,8IV
., I- / 1&Oagive cyine ":
tt ..... I / . l0 Cone cylinder -

, -
xWsin -c

0 1 2 3 4 5 6 7 8 9
FIG 16 VERTICAL LOCATION OF VORTEX CORES
FIG 14 PREDICTED & MRASURED FLOWFIELDS AT (REF 15)
M.3.5 (REF 9)

X/b :5.25
1.0 .0 X/db 52 7/b
.5 Lateral position."'-
of vortex 0
Z/db Cb La-a postio
0.6 -o12 I 0 15
A M"
o 1.7"" [
0.4 a 0.6 -
~18 e2.0 0.5
"21 S 2.5
0.2 ,2.- 0 X2 nX
024 0.4 sin-
''
00 0.3 3 I S 6 7
Y/dbO. 3 0.4 0.6 0 0.2 0.4 0.6 Y/db 01 23 4 56 78 9
 7_4

-77~~~. -77 77 777 7

3-220

Ln
10
10"
8
6 Asymmetric vortices

2 1
vortices 2 30tx pn f
ifre**. " "-
16 0Vortex
envelops Vortex
12 20 24 28 32 36 4.0 44 from top etral-'"l 'Zonef.,, otxevlp rm.-.-

Incidence 0C 0 vortex envelope 111 l1

top strake I-v. 0)

FIG 19 CONDITIONS FOR FORMATION from port strnke

6 Vortex envelopefrom
OF ASYKMETRIC VORTICES (REF 15)

Centre of vortex
Le4~bod 7.6 0.
d toa
Roil

FIG 21 VORTEX POSITIONS VIEWED UPSTREAM

FROM THE TOP RAKE POSITION AS MhE BODY
lisOtD,>.>. IS ROLLED

tO1.01 1'
P/P - With strakes /p it

S.0.6 0.6
"No strakes
1Is 150* 120*0 90.4
Pilot tube

Based on vapour screen Oiegrammafic

mealsurements -
FIG 20 VORTEX POSITION FOR CIRCULAR 9 './"-1
BODY WITH A WITHOUT STRAKE & FOR SQUARE tgot
loo
SECTION BODY
FIG 22 VARIATION OF PITOT PRESSURE WITH
ROLL ANLE WITH & WITHOUT STRAKE

Pressure rcovery Pressure recovery

Critical Recovery penalty incurred at
Stability pnt zero incidence as a function
limit of required incidence capability

;./G7 point
, 4 / "''-:-:"":.... ,-... .4..,F
O'44

Mass flow ratio 010 Gc or X3 0-4 " 6

 3-21

I PRt)A--. -

0.'4

I~r
d(P

9. - F/ 7

FIG 24 VARIATION OF CALCULATED SHOCK & MEASURED

RECOVERY WITH INCIDENCE FOR RECTANGULAR INTAKE (REF 16) 2v I~1 I,
.... FIG 25 VARIATION WITH INCIDENCE OF RECOVERYt-

CCII,(RELATIVE TO RECOVERY AT ZERO INCIDENCE FOR

IPR* NRECTANGULAR INTAKE WITH DIFFERENT k.NJWALLS (REF 17)
pa.
I PR i~oN.Will . sh..f

N (A.IACIde

is042 0 2.0 0

a2.59 20 20
isI 260
wlsdsc
I
lip *I_
___ d3 _ _

(.0 1.00
20 2o0 ISO
201 20. (0 2
I 4 110

9 2 1 a IS 12 (04 FIG 26(b) VARIATION WITH INCIDENCE OF MAXIMUM

FIG 26(a) VARIATION WITH INCIDENCE OF RECOVERY FLOW RELATIVE TO ZERO INCIDE.NCE VALUE FOR
RELATIVE TO RECOVERY AT ZERO INCIDENCE FOR AXISYNNETRIC INTAKES
AXISYNMETRIC INTAKES
1.0

9.00 s .0 2
'S 20 35 3

O (.17 OR2012:!J
2.44 zo 30 21
2 is.0
2 .0 J
t(07 A0 31 20

2 4 24 it 200 .

FIG 27(b) VARIATION WITH INCIDENCE 07 MAXIMUM

FLOW RELATIVE TO ZERO INCIDENCE VALUE FOR TWO
_______________________________ CONE CENTaEBODY INTAKES
2 4 0 to A. to is 14 '6

FIG 27(s) VARIATION WITH INCIDENCE OF RECOVERY JTM

RELATIVE TO ZERO INCIDENCE VALUE FOR TWO CONE ~"
CENTREBODY INTAKES

P.R. At

263 nO 1.0

2 Wlth shroud * 0
-No shroudC.
I..p I's 2.5 2.2 2.4 2.0 C's
3-22

A.I,,441w1,k c -iC4i - 04

Asal 11l

FIG 30 COMPARISON
- OF PERFORMANCE AT
-' 2 4
'SINCIDENCE OF HALF

0 iiy.W s20 AXISYMMETRIC INTAKES

- -- WITH & WITHOUT A
- 041 ~COWL
~ SHROUD
04 ov4th 4o 21 200
It", plot.

0 A.I.',t,-t 2s1 VI -

a6 r 14 41

-4. l ,-d.44jf.o*

4144 rA. z-z '*-,-

0.7%

Al~n~d 94FIG ~ 31 EFFECT OF

- A1.2n44 two a CENTREBODY PIVOTING
0-0.0-,4A
& COWL RAKING&
hg4.4~~ ~ It
.. 00031S4S~g4
&W PIVOTING ON AXI-
too4d ~~0020 SYMMETRIC INTAKE4
0.1 4 4 PERFORMANCE AT
44 .~s.4 0 INCIDENCE
0.2

12 W.44

0.90

Capture 4
nlow
100

3 6 9 C* .
 3-23

0Q 1350 dbY
I I0 ,!it 18 0

180 /6.
2.0i 0. 70 7.5 0.108 16 L

- ~ ~ - 1.0 06

*. Ri nge 0. 1@80.~

0 '40 so 120 160)Roll angle

FIG; 34 VARIATION OF RECOVERY WITH ROLL ANGLE FOR SINGLE HALF AXISYXIIETRIC
INTAKE ON RM1O BODY AT 60 a1 100 &
?Q,.99 (RE.F 30)
0.8 Yd S, 6
Ticrit 7.5 0.12 16 190 or 14* ramp

Nbol tam 19* FIG 35 VARIATION OF RECOVERY WITH INCIDENCE

otop 140 FOR SINGLE RECTANGULAR INTAKE AT TOP & BOTTOM
stop
, 19POSITIONS ON R.M10 BODY AT %2.0 (REF 31) t
0 4 8 CXO12

0.9

0ci.8 -/ - b C/A b Sw '/ 6

0.7 6.2 0.051 40 1.0

0.6 nbottom i7/.

0.5 *top with 5%.

0.4
0 a. a 12 16 20

0.1 -. 4.8 40 1.35

Icrit. postn, l.d1ra bleed %o
0.7 obottcm -8;/-.
ot09 I*/ perforations
06*top 0/ 61 62
0. to 10.4 11.1
0.5 2
0 4 a 12 1 20FIG 36 VARIATION OF RECOVERY WITH
INCIDENCE FOR SINGLE RECTANGULAR
hroo yawINTAKE
100j19, (WITH COMPRESSION SURFACE
BLEED) ON A BODY AT M 2.0 (REFS

0.7 'I
5 0.218

0.6 FIG 37 VARIATIO'N OF RECOVERY WITH

3-24 -.

Throat 3ody
bleed L/db M0 0 YzConical centrebody 30Ref.
S.. Inverted I r 2
2 ramp yes 5.0 2.3 1.0. "-""'%
-. Half conical
- No 7.5 2.0 1.0
0.9 -. Half conical No 7.5 1., 14 wedge intake 33
-- - Large lf conical No 1.8 0.9 (no swept endwalls)
crit. oconventional/ramp Y2 2.0 9 100 - 8, wedge
0.8 Conventional/ramp No 8.0 2.0 (PR_ %
0.7 (with swept endwall)

0.5 -7 '-4 06ik 0

FIG 38 SL4ARY OF RECOVERY VARIATION WITH16 20

0
ROLL ANGLE FOR SINGLE INTAKE AT a-1 0 FIG 39 SUMMARY OF RECOVERY VARIATION
WITH INCIDENCE FOR SINGLE SIDE (4900)
INTAKE

1 Tjn1ca0
Co nic centrtbody 100 -
PR.
P.86
23. 5 I 1WE100
Pf/, Io
110d 095 -0- - - n1 imntake 0 95 ". . ",
0 -" Inverted two- - - 070
"1.00 dimensional 062
2ndynan 0.71) " - ime09 0n54s-na-'O
06 .0 0 2 0.16.d1
0.1 .0 0 63F F 4 A T OF C
O.N oIF42 R -ET....
I 10
\ 'xhead \ / i // -KI

0.96 X . -mm 0ti 0-75

AC80 T0
0.876 2 x dynami
0 2 6 .10 12 14Incidence .A1.0 20 2.2 2.4 2.6 2.8 3.0 M1
0head
018 A-dnai M CPET FIG 41 VARIATION OF COHESSION
1xisynimetric FAX
0.&
0 01 0,2 03 0, 0.5 06 Mf EFFICIENCY WITH INCIDENCE FOR PRESSURE RECOVERY
FIG 4u TOTAL PRESSURE LOSS DUE FOUR INTAKES IN I180N/O/ POSITION 0.06 .
TO MIXING OF FOUR FLOWS (REF 35)-.--
0.0 0.7 -
Max
1 / ,~~xisymer
meri
&1"
& ,"' """v..' "

''Oll -dmensiooall erer..'.-.....

Large ..
0.76
-'--.'.'; 0."-

90 . 0
6 . 170
1
0 - 15 -+, 410
Idegre-s)
16
13 1@xsmer
0B7e " -4

FIG 43 VAR IATION OF MAXIMUM CAPTURE / .. ' .' .. ] -

AC0.7806
Z
FLOW WITH INCIDENCZ FOR FOU'l INTAKES Small diverter"" """"
SoooRolheight hadb 0
he5h - h/d 0..141'4

W @ r

INd 01800/0 210 70S0I50 14TION..

S--

Igo 170 160 5S0 140 130 " '." "

Roll angle",,...'. "
L/db AC/Ab 0 .21ISO 170 160 150 140 130 ". . '
7.5 0.42 Roll angle '" "
 3-25

15 // 0/ -FIG 45 VARIATION OF f'lo,WITH INCIDENCE & ROLL

15 90 AGLEFOR OURCONVENTTONALLY MOUNTED RECTANGULAR
0CFo ntability ~ INTAKES ON A BODY AT M.2.3 (L 5. 0 h 0. 7 AC 0. 87) .

(deg) 0eio db. A

100 10Corventionally mourted Intake duct flow Main duct flow
intkesmeasuring stn. measuring stn.

0 ?O.775 0.775 0.775, 6.775, 07,

180 170 160 150 140 130 106m

0 (degrees) Inverted intake mounting
508mm (medium bodyL
15 /0
508 .53

l 0d 620 .604 SS 2
5 257L6mm (tonbd

flp 0 9.670
0 ?.791 0,791 * .791 , .7911 0.79~ T ,ide slot
180 170 160 150 11-0 130 30838,75 1.33

FIG 46(a) INVERTED INTAKE 0(ers)61.25 bled-.155a

L h. . SL. Initial cowl
5.0 (a) M 2.3 undersurface
15 1.7 MSOL?.3 Shock on lip angle =13" ()M 26Soko i
design design
(de S. ,; FIG 46 VARIATION OF n. 0p WITH INCIDENCE & ROLL
-
.590T~o .e
So7 ANGLE FOR FOJR INVERTED RECTANGULAR INTAKES ON

.61062 .6,-632 A BODY AT M 2.3 AC/Ab 0.87)

5 6 70 15 .54*2 S.k40
.670 .677 S~94 .C
.75(dg 6 546

0 - 770 *.770 0.770 0 .770,* 01770 1de05t9

180 -170 160 150 140 130 .620 UL --- -- 12.84

566 .65 .652
FIG 46(b) INVERTED INTAKE a 0 (degees
1.25 M.6O2. .66
0 61068,0 6980.698,0.8

180 170 160 150 140 130

0 (degrees)
0.8 L 2h
+FIG 46(c) INVERTED INTAKE 5SOL25M .

06fw ssr a t ss d e

04 067

06068 0.7 08 09 10 06 . . .

0.7 n cveyi combusrtionchamberafteor

J 0.4
IN
F RAKA CPE AT(! 2.0 Relap
t iveflw A &/A Relativ flowimo
Aarm
 3-2 ". . . ... .

3-26

Single intake Twin Axisymmitric Intakes

Axisymmetric Rectangular Twin Rectangular Intakes

0C Rt~db
200 5 Os m.
106 76.2 14b .
295 .20.
3.50

Axisymmetric Intake Rectangular Intak- -

FIG 50 NASA MODEL FOR MEASUREMENT OF FORCES & MOMENTS ON BODY WITH NACELLES

Variations

Numnber
of -
nacelles

R Type of cyD
M.* O jjjL C
-bIar intake
6
0.6- -6to 60 1 14 I- 2 .10
but3.4 .12 12& -3 x10 ,. L.ngth of nacelle r - . .. i. % . .

mainly *5,6.17. 8.5 L "db -

2.0- .

Without boattail With boattail

FIG 51 ONERA MODEL FOR MEASUREMENT OF FORCES & MOMENTS ON BODY WITH NACELLES

Sting & balance shield

Halt cerotalrentrebody

=-"ri2 5 to'gent egq. / "\ ..

tached stin.-.g..
.to
Los 34r12 we 1i. pitot tubes end I static tube par arm)
note Optional bas, flow

baace controls lone Sloqng cone pressure

straii gauged) throttle pilot tubes

Wing Span o A.R tpor Win A.B,C.D9

Ab r mth in. i9 with naelles
or interdgitatod
A 0 2505 1017 0326 Parabolic Reotudl intake$
puN 001
0 0721 1546 01151 0 401
C 0673 11.03 0021 0418 paabli
o 03, 1208 7910,33 N!,.,,,o: :;:
o556 1074 0 0."65
,730 L..,,.::3..:..:..
- 0642 0296 3524 05
Par; Is a
 3-27

_, Body alone
Body wings A,B,C,D and E
Body * wings * controls

Large (AC/Abody 0.5) half 0

axisymmetric intakes in two
-___-- -_. longitudinal positions; in
line wings and controls

Large (AC/Abody O.5)

S. 0
rectangular intakes;
interdigitated wings
and controls ----

SIntake /x Reynolds No. Reynolds No.'"

motypy tyeAbody x bi (on body dia
/d on intake
nof1524mm) height/radius

1.8 Half conical 0.225 6.25 1.278 1.25 x 106 0.221 x 106
centrebody
2.0 0.375 1.653 0.271 . p
2.3 0.5 1.733 0.313
0.5
Rectangular 0.5 1,885 0.392
double wedge
compression
surface
FIG 53 RAE MODEL GEOMETRIES & TEST CONDITIONS

Notof Incidence CC
5 Ma Rf. itaes 8 12 16 20
0180 2.5 38 2 ,lp
rV130
A 90 d4 a With tfs
slO0 2.0 0NERA
. 0 2.0
2 ~ -2
"2 a Without tagts
0 2.5 36 1. .

AMN I FIG 54 VARIATION WITH

nf INCIDENCE OF -6 90"=
INCREMENTAL NORMAL -
FORCE & CENTRE OF ,-.- --
PSSURE POSITIOF DUE -8
TO THE ADDITION OF
SINGLE & TWIN -10 ",-,. J
RECTANlGULAR NACELLES...,.:..
5 10 15 20 TO A BODY
Incidence C

Incidence .
12
V8 16 01

0 ". Rif. intakes I 7b

4 130

c 3
1 oC
nacettes
36 1
~-
FIG 55 VARIATION WITH
INCIDENCE OF
INCREMNTA NORMAL
-4

-8
[.r:90
+-...

180...:..
20
0TO PRESSURE POSITION
TAR AFE ITTON OF DUE -810
 77

3-28

Nnacelles JI 39 3C
3 -nacelles
/ ~2 -

2 /00

FIG 57 VARIATION WITH _____________

20 INCIDENCE OF NORMAL 0 4 a 12 16 20
FORCE INCREMENT & OF l~ncidnc
0 5 10 j;* 95 CENTRE 01' PRESSURE Xc - - 0
Wscide'nce EIC POSITION FOR 4 INTAKE -
FIG 56 EFFECT OF MACHNUMBER CHANG CONFIGURATIONS db
ON NACELLE NORMAL FORCE INCREMENT

Mo vb -4F
a4 1885
1.38. 035 2.3*I
AC 4 1.885 0.5 2.3 16

3
-8

2AN A

Io .o s -- j -- til it.]ngRAE

1010 1 50IS 140 130 Wig 2.

Roll an 11/
10 170 160 150 14? 0130 .04 06
xcp ananan

-2 - Rectangut;,r intakes.01.aSIn

-4 - --- - r}Bd .02

-6 0

0 ~A
0 1 Net span 2 0 5 10
FIG 58 VARIATION WITH ROLL ANGLE Ntpa omae
OF NORMAL FORCE INCREMENT & CENTRE dbNtpAn omae
Or PRESSURE POSITION FOR 4 A
T
RiCANULR ITAESF C 59 VA!'IATION OF NORMAL FORCE INCREMENT
WITH NET SPAN & NET PLANFORM AREA

rression siorface Afterbody & base

- Bieed drag drag
-Diver
__________ er drag-

Intake at fut" flow Intake withf_ .subsonic.

ie with supersonic- forespiltageat Mo <MSOL
forspltge (b) Increased..j~Reue c
Pre entry or pre n drag
~~'---- addtiv 4r~<~wldra Intake with subsonic
C~pr% or1;111ccwoIntake at datum flow forespittage at Moo ?>M-
Wedge/Cone . -

0, Ocowl
s *C shoko
 3-29

to, ra o f C A, I_

,,8.6 2

i'" FIG 62 VARIATION

SOF FUNCTIONS al &

C-1 po-l eqa. It-(Tm .R

FIG 61 COWL PROFILES

FI/G 64*COMPAR
A \ " ow

S~~~~~~&
/S"t,. OF CHAkRACTERITC
LINEAR THEORY ,",. \ nA.

1.0 100 20 COWL

%az / 10. 'f

FIG 63 VARIATION OF FUNCTIONS ". --

1 & T(X)
,(x) WITH,-
tat forwly
Macy, line
Streamline Initial inputipln
SShack
point I_ Lhock-e__ni ____r_____
IsI ogIetnl: pl is static pressure behind
/Oblque shock fore M oo and deflection 5,
I. 2nd 1.1d 0ly. sogment: Mi defines PrAlndtl-Meyer angle V,
,.ch line 1'0 2.I 'otfield V3 -V,-
(60-61) defines M2 and p2

2 2nid order shock e xpansion

",Body point .
After Porandtl-Moyor expansoin, pressure on segmenrt
falls exporntntially nowards cone surface pressure
M .. " (cone Angle .,.M M.) - Ref 12
- ----- Tangent wodg*: pn Is stalIc pressure behind oblique
shOck far IM- MMand deflection aS

FIG 65 METHODS FOR CALCULATING Impact theory: , n

COWL PRESSURE DISTRIBUTIONS
.o.9

0. Measurernonti eR05eSneh.Ck-#.sO~i Man

A. JAC/Arm
nor- 5 1" 2.0 .lOS- 33.96 0.351 irCharcteriotlcs
L0I An o+ 3.1 1,1g -33.g v alues
C...
Co91 a O ,-- O023 -- 19.7 C. ft 07 4.0 1.48-23.66 Ref ,9
a) 0.? 5 27.5
2 .1.0.022-
: -2752alO.?C 0
Characteristics

oat r.?
7* r. 6,
I a 07.0 0163 el0:
s0c 197 0.0 ao
27.S 1.453
I 34 1.210 hI,: anufennonts

"Rm
from

0.3 v, 04
,.Ao.45
2.0
o 2.3,.
.
0.0.

-,0.2 1.0

0.2 OAS

0..
6 7 0t111
2 7 Oo 2
0ftj Iky 2 h'~
3-30

C. 1,0

I&-- 'L .

FIG 68 LIMI[AIONS O -L A -

HYPERSONIC SIMILARITY PARAMETERS 1

ft0.5* .,. P 1 .0 Ls 2.0 2.6 &

FG 7.0 COMPARISON OF LINEAR

THEORY, SHOCK EXPANSION &
MEASUREMENT FOR ELLIPTIC

tj~COWLSIO REF
~I 50
1.00.3 0.8

~ Pol 01 A)
0.60. 0.004 63.
2. _o AAco__ .

0.4 0

0. 0.4 0.6 0.8A 1.0c Co.2a

Spronic exit
(a)() 2eoct
010 geI -A

0.2~~0.
0o.6tea 0. 01Mea o.f.0

at inak th.oa Mac No1

WEDGE COPRSSO SURAC 1. FIG 73 BLEE 2.E3SR
RECOVER
 3-31

30 h/6 S/6
I,1.o0 0Io 4 h/5; .2J5 S/--0
20 - '1 2 ---
It Wedge in Wedge in
c10 0supersnic spersonic
Wedge diverter Co P I flow
stra sin I 3 -

0.6
2.0 3.0-

1.0 ' , 1.0

0. 0.6 - - - -

Wedge diverter 0.4 - 0.4o-, - [-

swept splitter 0.04 0.06 0.1 0.2 0.4 ,1.0 2.0 3.0 0.06 0.1 0.2 0.4 1.0 2.0 3.0 4.0
M -1 sin 6w/2

FIG 76(a), (b) & (c) WEDGE 30 h A 0.75 S/6 = 6

DIVERTER WAVE DRAG (REF 55) 20 / .

Cow( lip diverter 10 Wedge in

FIG 75 HALF CONICAL

CENTREBODY DIVErTTERS Sil2 .

2.o

0.6

0.4%
0.04 0.06 0.1 0.2 0.4 1.0 2.0 3.0
6S h/ 1.0; 0 , S/6 , 1.0; 600 - S $ 180
moo . <Re 1.6 to 16.4 x 106 per npetre

0.4
o ,4 0.3
hi
q__
1.5 2.0 2.5WEDGE
MFI77a&(b D)IVERTER

0.2 " p 38
0.1 - 0.10
___. 0 (a)
04
i.0 0e .

1.5 0 10
"(b) 0.75 w.h/6
0.3 Ref.
5.
06
57
"0.2
nL
"0.1 ['0.300
[" r=4~ A o-os-f
Skin friction
coeff" From Ret, 5

0. -- 20.0 - - 0.001 M;ch No 0 1

M =3.16 0. 9 - -- .00
0.11 1
0.00002 3
0.3 "t Ref 4 5r
' Cf.E a 56 100001
0.2 M 2.96 .Ef W 0 Re"nod No. Reo
1.0 0 60 Renod No-
0.1 -02 0
0 ~ .-1..2 .. - OFIG L" 0.03 0. 0
78 WEDGE DIVERTER
SKIN FRICTION DRAG
"0 __._-'_, Reynolds No.- area ratio AW -0.2
S. ... ..... ,;^parameter A IReIL
 4-I

INTEGRAL BOOST, HEAT PROTECTION, PORT COVERS AND TRANSITION

by
T. D. Myers
United Technologies/Chemical Systems Division
P.O. Box 50015
San Jose, California 95150-0015

SUMMARY

The integral rocket ramjet (IRR) employs a dual-purpose combustor that serves initially as a rocket
ccmbustion chamber for the integral booster and subsequently as a ramjet combustor. The subjects of this
lecture are the requirements and basic design approaches for the integral booster motor, heat protection
for the integral booster motor, heat protection for the combustor case and nozzle, port cover and
transition systems. These requirements and design approaches for the integral booster motor will include
the special considerations that must be made for the various types of integral rocket ramjet propulsion
ist2ms that include the liquid fuel ramjet. solid fuel ramjet and ducted rocket engines. .

The integral boost motor includes the solid propellant grain, igniter, ejectable nozzle and release
mechanism and interface bond system. The integral boost motor must accelerate the missile from the Mach
number and altitude conditions at launch to a suitable velocity for ramjet takeover, typically Mach 1.5 -

to 2.5. The integral boost motor weight is very important since it is typically 20 percent of the missile
weight for air launched and 25 percent for ground launch missiles. Furthermore, the combustor volume
required for the integral boost motor generally exceeds the size required for high performance ramjet
operation. Thus minimum booster length yields the highest ramjet fuel load and thereby range.

Since gravity drag losses are a significant factor in establishing the integral boost motor perfor-
mance, it is also desirable to use the highest thrust compatible with the vehicle structural design and
radome temperature constraints to minimize burn '.ime. These three primary integral boost motor design
requirements lead to tne need for high burn rate propellants with good physical properties to permit high
volumetric loading efficiencies. Typical integral boost motor design approaches are discussed that pro-
vide these capabilities. The use of nozzleless motors as a design solution to simplify the integral
booster is also discussed.
o

Heat prote~tI,1 for the integral rocket ramjet combustor case must orovide for long duration with a
high temperature, oxidizing environment. The material selected for heat protection of the IRR combustor
must also provide a good structural and bonding interface for the integral boost motor. The heat pro-
tection material must withstand both the high temperature and pressure boost motor environment as well as
the long duration, low pressure ramjet operating conditions. Several heat protection materials have been
investigated for applicatior. to IRR combustors that include silica phenolic, quartz phenolic and a sili-
cone elastomer material, DC 93-104, that was specially formulated for IRR applications. The requirments
for the heat protection and general lessons learned from development testing with the various materials
is discussed.

The IRR propulsion system requires port covers to seal the interfaces between the combustor and air
inlet. Depending on the IRR configuration the port cover may be a large unit covering a single air Inlet
or several small units covering multiple air inlets. The port covers must be designed to structurally
support the high integral boost pressure between 1000-2000 psia and still be capable of being easily
ejected by the ram air inlet pressure in a reproducible and reliable manner without damaging the heat
protection materials lining the combustor case and nozzle. Depending on the IRR configuration, the port
cover may, have a direct structural and bond interface with the integral boost grain. Typical design
solutions to the port cover requirements are presented, including monolithic and segmented designs. A
novel solution that has been conceived for solving the port cover design problem, employing a hinged
concept will also be discussed.

Design and analyses of the transition sequence from rocket to ramjet r des of operation represent one
of the most interesting and challenging aspects of the IRR propulsion system. The transition sequence
typically begins with ejecti'n of the port covers from the IRR combustor and includes boost nozzle ejec-
tion, ramjet fuel valve opening, initiation of ram air turbine operation and ramjet igniter actuation.
All of these events must be carefully sequenced to ensure successful transition inma very short time,
typically 300-500 msec. Moreover one of the major considerations in achieving a successful transition
Is an accurate accounting of all energy sources, including residual boost grain sliver, heat protection
material ablation products, ramjet igniter gases and the ramjet fuel transient flow rate. The heat
protection material ablation gases frequently include methane, hydrogen and other combustible hydrocarbons.
The heat release from these gases can contribute 30-40 percent of the energy needed for normal ramjet
operation at takeover. Since inlet design margins are typically minimum at takeover,failure to consider
these important energy sources can and has led directly to IRR flight failures.

INTEGRAL BOOST MOTOR

The integral rocket ramjet (IRR) propulsion systems represent a sophisticaLed technology objective
for advancing tactical and strategic missile
ronracnt thp tvnlral ennlne confiourations. applications. The three basic ramjet types shown
Intearation of the solid rocket booster motor within the
figure 1
 4-2

and volume advantages when compared to the podded or tandem ramjet configurations. When compared to
conventional boost/sustain solid rocket motors with the same weight and envelope, the IRR can provide
a 200% to 300% increase in powered range.

Naturally, the advantages for the IRR configuration do not come free. In this instance, the price
is primarily related to the unique complexities associated with the interfaces between the solid booster
motor and the ramjet engine.

The booster motor for the integral rocket ramjet must:

1. Accelerate the missile from ground or air launch speed to a Mach No. satisfactory
for ramjet takeover.

2. Provide high volumetric loading to minimize combustor volume.

3. Provide rEquired AV gain within 4 to 6 sec to minimize gravity drag losses.

4. Provide minimum motor residuals to allow rapid transition to ramjet operation.

5. Provide ejectable boost nozzle to permit rvmjet engine to operate at low

ramair pressures.

6. Cause minimum damage to the ramlet engine TPS.

A typical integral booster motor is seen in figure 2. Operation of the IRR begins with ignition of
the integral booster motor after the missile has been ejected from the aircraft. The booster motor must
"accelerate the missile from the Mach number-altitude condition at launch to a velocity suitable for ram-
jet takeover, typically a Mach number from 1.5 to 2.5. The booster must operate over a wide range of
launch conditions, ranging from ground launch to air launch at Mach number of 0.5 at low altitude up
to 0.7 to 0.8 at midaltitudes. The booster motor propellant weight is typically 20% of the missile
weight for air launched -pplications. For ground launch application the booster fraction is approximately
25%.

The booster motor burns for approximately 4 to 6 sec at chzmber pressures between 1.000 to 2,000 psia.
During the tailoff transient of the booster motor, a chamber pressure transducer signals a control
mechanism which releases the booster nozzle under the force of the residual booster pressure. A port
cover plug, installed at the interface between the combustor case and the turn and dump section of the
inlet diffuser, is subsequently ejected by the force of the ram air pressure when it exceeds the decaying
booster chamber pressure. The port cover is ejected aft through the combustor case and ramjet nozzle.

After the port cover has been released, ram air is free to flow through the combustor case and
nozzle. Since the vehicle drag forces are high at the transition from rocket to ramjet, it is desirable
to minimize the time from rocket shutdown to rarajet steady state operation. Typical transition times
are from 500 to 1,000 msec. A grain design with minimum slivers and minimum residual or combustible
materials is critical to avoid either an excessive loss in system performance capability or an undue
risk of inlet unstart which could cause a failurd in ramjet takeover. The quantity and rate of heat
release from the booster motor residual materials has a direct impact on the design of the ramjet fuel
control system and indirectly upon the inlet if pressure control probes which must be installed in the
inlet are rcqIred. Design of a booster motor having no combustible residual materials, for instance,
could significantly simplify the design and cost of the ramjet fuel control system.

High Volumetric Loading

Maximum IRR range is obtained with a solid propellant boost grain of minimum volume since the solid
propellant volume required for an optimum ramjet takeover velocity is larger than the chamber required
for satisfactory ramjet combustion efficiency. Thus, a high density solid propellant in a grain config-
uration of high volumetric efficiency is an 'important design objective for the booster motor. A 1%
increase in propellant grain length and combustor length could cause a 5 to 10% loss in IRR vehicle
range due to loss of ramjet fuel storage.

Short Burn Time

Since the booster motor volume requirement is a key controller in establishing the maximum IRR r nge,
it is also necessary to minimize gravity drag losses and thereby reduce the required booster motor size.
The approach is to minimize the booster motor burn time within the structural limits of the vehicle.
Typically this approach results in integral booster motor burn times on the order of 4 to 6 sec to main-
tain vehicle acceleration limits of 15 to 30 g. The maximum thrust limit required for the booster due to
missile structural limitations, in turn, influences the size of the booster motor.

The ramburner case must be fabricated from high temperature alloys such as L-605 or 1-718 to with-
stand both the high external aerodynamic and internal ramjet combustion heat loads. The ramjet operates
at a maximum combustor pressure of approximately 150 psia at wall temperatures up to 2,000OF. Generally,
the case wall thickness for this design condition permits operation of the booster at Dressures up to
the MEOP of 1,500 to 2,000 psia. Since these high temperature alloys are heavier than typical ;olid
rocket case materials, the booster motor must operate as efficiently as possible within the specified
maximutm MEOP limit to avoid undesirable penalties in case weight or booster performance. Therefore, a
neutral chamber pressure trace and a propellant with minimum value of TK will yield a booster motor with
highest performance.
Propeiiant Selection
 7 71 77
7.o 7I-'7

4-3

generally within the range shown in Table 1. The ARCO R-45M HTPB polymer provides outstanding processa-
bility and pot life with the 87% to 89% solids propellants currently being developed for booster applica-
tions, provided that curing agents and pot life extenders are selected to minimize the catalytic effect
of burn rate modifiers on the isoc)anate-hydroxyl cure reactions.

As Table 1 shows, the required burn rates of 0.8 to 1.3 in/sec at 1,000 psi are in the medium-high
range and require use of burn rate catalysts to achieve these rates. Catalyst selection is one of the
key formulation tradeoffs at the present timB. Iron oxide, Ft203, is the most widely used catalyst in
solid propellants, but it has the disadvantage of yielding pressure exponents above 0.5 and temperature
coefficients, ITK, of 0.20%/F or greater. Since high pressure and temperature sensitivities result in
performance penalties, there is current interest in the utilization of anhydrous ferric fluoride as a
replacement for iron oxide. This additive provides n and nK values at the lower end of the range shown
in Table 1. However, studies have indicated that HTPB propellants using FeF 3 may have poorer aging
characteristics than counterparts formulated with Fe203. More work is required to evaluate these
tradeoffs and investigate approaches to improved aging stability.

The high web fractions characteristic of IRR booster grain designs coupled with the broad -650 to
165 0 F environmental temperature range require propellants with excellent mechanical properties.
Typical requirements include greater than 20% multiaxial strain capability at -650 F iasured in strain
evaluation cylinders. Table 2 shows mechanical properties of an 88.5% solids IRR booster propellant
candidate tested at Chemical Systems Division which meets the established require.'ents for this type
of propellant.

Grain Design
The primary objective in selecting a grain design for an IRR propulsion system is to achieve
maximum volumetric loading within the structural limits of the propellant. There are, in addition,
several considerations which are unique to an integral booster motor:

- High volumetric loading

- Neutral pressure trace
- No boots in forward closure which must act as ramjet recirculation zone
- Short burn time
- Minimum sliver and residuals

Figure 3 will be used to illustrate the design considerations to be made to the sample grain design.
The keyhole grain design provides the following key features.

High volumetric loading of 92%

H.

2. Keyhole slot provides stress relief without boot in forward

combustor closure.

3. Nearly neutral pressure/thrust profile (1.055)

4. Theoretically sliverless F
5. Minimum boost residuals as large exposed area of liner is consumed.

Propellant Grain/Liner
The propellant grain/liner interface with ramburner is a critical interface which has presented a
unique new set of technology requirements. The propellant grain/liner interface with the liquid fueled
and ducted rocket combustor are very similar as seen in figure 4. The solid fueled IRR presents a
completely different interface since the booster grain/liner must interface directly with the solid ram-
jet fuel. In many solid fuel ramjet (SFRJ) IRR designs a dual boost grain is required for maximum
volumetric efficiency. Each of the two grains may require different propellant burn rates as well.

In the liquid fuel ramjet (LFRJ) IRR and ducted rocket the interface is with the TPS which lines
the case wall. Three primary TPS materials are currently being used in IRR combustors:

1. A silicone elastomer DC 93-104

2, Silica phenolic

3. Quartz phenolic

The structural and chemical interfaces are vastly different between the silicone elastomer and
silica or quartz-phenolic materials.

The propellant/grain liner structural and chemical
Tables 3 and 4. The structural interfaces between the
presents no problems relative to support of the grain.
satisfactory bond strengths with the silicone elastomer
have been evaluated for the DC 93-104 matprial.
interface considerations are summarized in
booster grain/liner interface with DC 93-104
The major structural problem is to achieve
material. Many interface material bond systems

Booster Igniter

I I I I I I I I
---- - .. - .. -.
.... .7 -7.. -. -'.-: ."'.-.

4-4

closure as well as the need to install a port cover at this interface. In addition, the booster igniter
must provide fast, uniform ignition of the internal grain surface to avoid nonuniform burning and
undesired slivers at booster grain burnout. An igniter material such as MAG-TEF is highly desirable
to avoid overpressure during the ignition transient which could result in an undesirable weight penalty
for the ramjet case or e:cessive pressure rise rates with the cold L605 case material.

Nozzleless Motors

The primary advantage of the so-called nozzleless solid rocket motor lies in the elimination of the
ejectable booster nozzle and its attendant complexity and cost. Nozzleless solid rocket motors employ
simple circular port grain desiqns as seen in figure 5 and are characterized by highly regressive
pressure-time curves but nearly neutral thrust-time histories. Figure 6 shows a typical set of pressure
and thrust curves for a nozzleless motor design. Although nozzleless motors have much lower specific
impulse values than conventional motors using the same propellant, this loss can be largely compensated
by packaging additional propellant in the volume usually occupied by the nozzle. In addition, a divergent
aft conical surface serves to accelerate the exiting combustion products, yielding a small improvement
in delivered performance.
Nozzleless motors normally require significantly higher propellant burn rates than conventional
"motors designed for the same mission. In addition, propellant pressure exponents should be less than
0.50 in order to obtain designs which are competitive with nozzled booster configurations. These
additional requirements and :onstraints add further difficulty to the problems inherent in formulating
satisfactory bcoster propellants.

Table ' summarizes the outlook for various categories of nozzleless booster propellants. Feasi-
bility studies of nozzleless ramjet boosters using high burn rate aluminized propellants are currently
being conducted under Air Force contract. Since these propellants use liquid ferrocene Lurn rate
catalysts, some problems in long term aging may be expected. However, there is no reason that high
density-impulse zirconium propellants should not be as satisfactory for nozzleless designs as the alumi-
nized propellants, and the same perfornance advantages noted earlier for conventional motors can be
expected. Up to thib time, no e,'fort has been expended in adapting low signature propellants to nozzle-
less IRR booster concepts. Because of their inherent high burn rates, such propellants will be parti-
cularly susceptible to acoustic combustion instability. Very probably, some compromise will be necessary
in requirements for smokelessness in order to develop nozzleless boosters with these types of propellants.

HEAT PROTECTION

Thermal protection of the integral rocket ramjet engines requires new materials and construction
to withstand the difficult thermal, structural and chemical environment. A comparison of the thermal
protection system (TPS) design consideration for rockets and integral rocket ramjets is shown in
Tablet' The unique aspect of the thermal prote-tion design is that all of the virgin and charred
materiul properties required either by a solid rocket or liquid rocket must be combined for the integral
rocket ramjet engine. In particular, the requirements for a high strain capability, chemical/structural
booster compatibility, high charred material mechanical strength and oxidation/erosion resistance
make thermal protection a signiftcant IRR design challenge.

The requirements for the thermal protection systems for rockets and ramjets are compared in
Table 7 . Because of the lower chamber pressure levels of the ramjet the heat fluxes and combustion
temperature are significantly lower than the rocket. The ramjet, on the other hand, operates for
sigi,ificantly longer durations with external aerodynamic heating in an ox dizing/high viscous drag
environment. Steady state combustor operation with a fully charred insulatlo, material is a unique
requirement of the ramjet.
Early podded ramjet combustors employed air cooling to thermally protect the combustor and nozzle.
The advent of the integral rocket ramjet with the dump combustors and high L/D combustors made it
generally impractical to apply air cooling. Ablation cooling is the most attractive solution to thermal
protection of ramjet engines. A large number of candidate ablative materials were evalucted as
potential ramjet insulation materials.

Thermal fnalysis

The thermal analysis model recommended for use in analyzing the integral rocket ramjet thermal
protection systems is the two zone model shown in figure 7.

Combustor State

The combustor environment is defined by a chemical equilibrium calculation based on measured fuel-
* .-- . -............- 7 -D `'' `'``-. -..

4-5

to-air ratio (f/a), fuel composition and enfhalny, and air temperature. Calculations are performed
using the ACE chemical equilibrium program. Correcticns for combustion Inefficeincy, when known
from test data, are applied as enthalpy losses. Gas emissivity is based on the calculated equilibrium
C02 and H20 concentrations.

Boundary-Layer Model

The convective transport of neat and mass to the surface is modeled using a boundary-layer computer
code developed by R. J. Flaherty.p This model predicts convective heat transfer coefficients that are
comparable to values calculated by other boundary-layer programs. 3' 4 The boundary-layer program incor-
porated the effect of axial combustor length into the analysis and provided a continuous model down
through the ramjet nozzle. Nozzle heit transfer correrations do not have this length effect explicitly
included. The boundary-layer model also considers the effects of pressure Mach number, wall contour,
wall temperature and gas properties.

Charring Lining Heat Transfer Model

The actual thermal response of the charring wall and Inconel case is modeled by the CMA computer
code? This program is a one-dimensional, transient, charring heat transfer model. It includes con-
vective heat and mass transport with corrections for blowing effects, radiation to and from the surface,
and heat conduction into the solid wall material. Pyrolysis of the wall material is described by three
independent Arrhenius-type equations; pyrolysis product gases were permitted to react at the surface of
the char. The program can consider multiple materials with intermaterial gaps transmitting heat by
radiation and gas :onduction. The material properties uses the same properties as those reported for
MX2600 silica phenolic. 6 Separate sets of temperature-dependent thermal conductivity and specific
heat were input for virgin and char material, respectively.

The reaction of the pyrolysis gas with the ramjet combustion products is assumed to proceed
to equilibrium at the char surface, with the resultant enthalpy change included in the solution for the
surface temperature. The net enthalpy change then appears as a reactive "heat" flux in addition to
the convective, radiative and solid conduction components of the surface heat transfer. Since the
pyrolysis gas products are fuel-rich, the reaction is strongly exothermic whenever unreacted oxygen is
in the combustion products, arid is usually the case in ramjets. There has been some question as to
whether the pyrolysis gas-ramjet reaction actually occurs at the surface as this model suggests.
Experimental results show that it does occur in some sitations.

Two-Zone Modeling
It has long been accepted that a rec6.culation zone exists downstream of the step expansion from
the inlet. This recirculation zone is characterized by higher fuel-to-air (f/a) ratios and higher flawr
temperatures than the overall combustor. To apply the charring heat transfer model to this zone, both
the local f/a ratio and the local temperature must be predicted.

An extensive study of recirculation zone f/a enrichment showed a correlation between overall
equivalence ratio 0 and recirculation zone equivalence ratio ORZ for center dump combustors. While the
correlation is not precise (and especially for charring linings) it gives a first approximation to the
recirculation zone state. The correlation can be used to determine a new local f/a ratio, and the
equilibrium chemistry calculations can be used to determine a new local gas temperature for the recircu-
lation zone.

As a first approximation to convective heat transfer, a short reverst boundary layer is assumed
to develop within the recirculation zone similar to that in the aft region.

The starting point for the aft boundary-layer becomes a semiempirical parameter in the two-zone
model. However, the aft predictions are in general not sensitive to the exact location of the boundary-
layer start. The remaining unspecified parameter is the recirculation zone gas emissivity, which becomes
the main point of empiricism in the present approach. Gas emissivity is determined from the charred
engine firing and then applied to the values derived are consistent with those derived from other fuel-
rich ramjets.

Two-Zone Correlation with Test Results

Full Charred Lining. The conventional boundary-layer model still matches the aft region thermocouple
response. The aft zone can be considered to prevail in the region downstream of the 0.45 combustor
station (see figure 8).
The forward rec!rculation zone prediction is significantly higher because of increased fuel
concentration. The l)cal equivalence ratio is approximately 1.5 for all phases of the duty cycle.
Equilibrium flame temperatures are singificantly higher. The best fi'; t: avrage forward case tempera-
tures over the entire test duration of 175 sec correspondsto a local efrtctive emissivity of 0.2, a not
unreasonable value considering the highly rich fuel mixture. This recirculation zone model extends
across approximately the tIrst quarter of combustor length.

VirIn Linn. When the same two-zone model is applied to the actively charring lining, a reasonable
matchIs stll obtained at both zones. Th3 aft region boundary layer again shifts to the right. The
midlength peak in the thermocouple band suggests a slightly further aft starting point than for the
fully charred lining.

The forward zone is similarly modeled to the fully charred lining; however, a temporarily higher r
f/a ratio is used during active pyrolysis due to the addition of fuel-rich pyrolysis gases. The oveiall .

I l I .... I
4-6

The two-zone model is able co fit both locations In both tests because the charring heat transfer
model considered reactive as well as convective and radiative heat fluxes. In the actively charring
lining, the forward and aft heat fluxes are similar in total magnitude, but quite different in
nature. The aft region is mainly reactive because of the exothermic pyrolysis gas reactions and is
only secondarily convective. In the forward region, the higher temperatures and emissivity raises the
convection and radiation but the reactive term is reduced a comparable amount because of the local lack
of excess oxygen. The net result is a similar total heat flux, both fo'ward and aft.

When the lining is fully charred, the reactive flux disappears. This reduces the aft heating
considerably, but has little effect in the forward zone when the flux was convective and radiative.
Under these conditlons, forward-to-aft variation become characteristic. If the initial run continues
long enough, this effect will be apparent. The hot forward zone is also seen in metal wall combustors
where again no reactive flux is present. Thus, the flat axial temperature profiles, which are seen with
all of the pyrolyzing lining materials, result from a hot, fuel-rich forward zone and a comparable,
cooler, but more reactive, aft zone.Table 8 summarizes these differences in heat flux and resultant
temperatures for the two zones of both tests.

More precise modeling of most of the present approximations should not change the validity of this
approach. Any recirculation zone f/a correlation that would give a fuel-rich forward zone would be
consistent with this model. Any aft boundary-layer starting point in the midregion of the engine would
also be consistent. Changes in recirculation zone convective heat transfer coefficient would result
in a revised empirical emissivity for both forward zones. Material property changes would also apply
to both zones of both tests.

In conducting preliminary design studies of a ramjet propulsion system it is important to consider

the relative effect of different configurations on thermal protection requirements. An an example,
in figure 9, two candidate combustors having the same nozzle area ratio (A ) but with significantly
different L/D ratios can satisfy thi basic mission requirement. The high EI0 combustor,-however,
has 50% higher heating and 100% higher combustor gas velocities. This significantly more severe
thermal environment could result in a longer and more costly flight engine development effort.

Several general conclusions can be drtiwn from recent thermal protection system development efforts
that have involved bench test, subscale test and fullscale durability testing of several thermal
protection systems. These general conclusions are summarized beiow.

Bench/lab tests useful as qualitative indicators

Can identify unsatisfactory TPS material/construction
Cannot be directly scaled to full size

Analysis required to identify requirements

Fundamental TPS laboratory/subscale test data acceptable.

Testing required to demonstrate specific requirements, such as

Weight loss vs temperature and time

Strain capability of virgin/charred TPS material

Correlation of information from various tests

Prohibitive cost/schedules for F/S IRR TPS durability tests systematic approach

Improved thermal/structural desired TPS modeling techniques are needed

PORT COVERS

Among the key elements in the IRR concept is the air inlet port cover. The port cover design
requirements are summarized below. Depending upon the configuration of the ramjet comtustor, the port
cover may be a large unit covering a single axial inlet or several smaller units covering multiple air
inlets. The cover, or covers, must withstand booster chamber pressures of 1,000 to 2,000 and still
be capable of being ejected reproducibly, reliably and without damage to the ramjet combustor, nozzle
or thermal protection system. To minimize the need for active control devices, it is desirable for
the port covers to be self ejected due to ram air pressure during booster tailoff. Finally, the port
covers must be suitably insulated from heat generated by burning of the booster propellant.

The Integral rocket ramjet port cover design problem is illustrated in figure 10. In the top
inset the overall transition sequence from rocket to ramjet operation is depicted. In the example
shown the booster propellant is cast over the multiple side mounted inlet dump openings. In this
configuration the boost propellant retains the port cover during captive flight (if inlet criverc are
not used). In alternate center dump configurations, for example, the boost propellant may not be
directly cast against the port cover and special retaining devices must be used to prevent the port
cover from being prematurely ejected by the ram air loads during captive carry.

The apparent simplicity of the basic port cover operation belies the true complexity of the design
engineer's challenge. The port cover interfaces with many of the other major IRR subsystems, creating
several specific interface design problems. These port cover design problems are deflned in figure 9.
In addition to the basic port cover structural, seal and design problems,some additional design
requirements are frequently imposed. These are associated with providing signals tc the flight vehicle
 4-.7

Design concepts currqntly in use include: the monolithic cover, the segmented cover, and the
frangible cover. (The simplest type of cover (a monolithic metal cover) is found on the Supersonic
Tactical Missile STM). The STM utilizes four smaller covers (figure 11A) over the combustor size
wall air inlets. In the design, ejection of the cover(s) necessitates the expulsion of fairly large
metal pieces, creatin~g the possibility of damage to the combustor or nozzle. Although numerous tests
proved these designs do not damage the ramjet combistor, other designs are belog~irivestigated to decrease the
damage p6tential by reducing the size of the ejected material.
The Modern Ramjet Engine (MRE) used a segmented inlet port cover (figure 11B) to reduce the size
of the ejected particles. This cover was made up of several structural beams placed across the air
inlet and held together by a rubber-like retention boot. At ejection, the boot released the "beams"
and allowed them to individually exit the combustor.
Both monolithic and segmented port covers have been successfully employed. However, several
newer concepts are of consider~ble Interest because they offer the possibility of further reducing the
size of ejected pieces or eliminating them entirely. one approach is to mechanize the port cover so
that it can be opened without being ejected. For example, the cover could be hinged (figure 11C),
louvered or sliding. These approaches would be particularly good for small dump openings. Moreover,
the port structural elements could potentially be designed to act as flameholders.
Another means of reducing the size of the ejector is the frangible cover. This cover is made
from high-strength chemically treated or heat-treated frangible glass. When the glass is broken (by
a hard metallic~ pin that penetrates the hard outer surface or by a small detonating device), it
breaks Into rock salt shaped granules about the size of the origin~l glass thickness. This method
seems best suited for combustors with center dump.
TRANSITION
The transition~ sequence in the integral rocket ramjet propulsion system created n~ew
design require-
ments, unique to this system.
During rocket operation, the combustor must function as a conventional rocket combustor, closed at
the forward end and with a suitable "rocket" nozzle at the aft end. During transition, the combustor
must reconfigure itself to a ramjet combustor. open at the forward end to allow air to enter the combus-
tor, and with a large throat nozzle, suitable for ramjet operation at the aft. Two schemes are employed
to permit this dual operation of the combustor chamber - the "blow-in" port cover and the ejectable
rocket nozzle.
In addition to their steady-state functioning, the port cover and ejectable nozzle design must
consider the ability of these compcnents to quickly and reliably allow transition from rocket to ramjet
operation. After tailoff of the booster, drag forces acting on the airframe with engine off quickly
slow the forward speed of the ramjet vehicle, typically losing on the order of 0.1 Mach No. per sec.
Longer transition times require increased booster loading which in turn causes undesirable ramjet fuel
losses and consequently significant range loss. Transition to ramjet takeover must, therefore, *be
done very quickly. Although transition sequences may vary, the basic transition operations following
booster tailoff remain the same. When the booster rocket chamber pressure decays to a value where
positive vehicle acceleration approaches zero,which is approximately equal to ramjet takeover combustor
pressure, the rocket nozzle release mechanism is activated, and the residual booster pressures force
the nozzle rearward, ejecting itfreely through the ramjet nozzle. During this same time period, the
air inlet pc;rt covers are forced inward when the inlet ram air pressure exceeds the residual chamber
pressure. The covers are expelled through the ramjet nozzle by the force from incoming ram air flow.
Ramjet fuel control and ignition systems then sequence the ramjet fuel and igniter. Ramjet fuel ignition
is further aided by the remaining burning residuals of both the rocket grain and the thermal protection
system combustibles. All this must take place in a few hundred milliseconds or exce~ssive forward
velocity will be lost to drag forces.
The propulsion transition sequence for the Supersonic Tactical Missile illustrates a typical
transition design (figure 12).
A typical propulsion transition sequence is shown in figure 13. The design transition sequence
is initiated during booster tailoff when inlet ram air pressure exceeds booster combustor pressure
(.&P - 5 psi nominal) forcing the air Inlet port covers Inward. Movement of any one of the four port
covers actuates a port cover sensing switch which energizes the nozzle snapring retention system relay.
Upon relay closure, voltage is applied across the propagation cord initiator bridgewires, firing
the initiator squib which in turn ignites the propagation cord. Closure of the port cover sensing
switch also initiates a nominal 300-ms time delay relay that is subsequently used to open the ramjet
fuel valve and fire the ramjet igniter. Upon firing the propagation cord, the booster nozzle snap-
ring is released and residual chamber pressure, in conjunction with incoming air pressure, ejects
the booster nozzle. Residual propellant slivers and liner continue to burn until the slivers are
essentially consumed at the end of the 300-ms delay. Closure of the time delay relay initiates fuel
valve opening and ramjet igniter firing current which is followed by fuel flow, ramjet ignition,
and steady-state ramjet operation. The ramjet igniter fires for approximately 1 sec, releasing 1,500K
Btu/sec of energy.
While the transition sequence must be accomplished quickly, it must also be timed very carefully
with the other sources of energy release during the transition period. Early ramjet light-off can cause
excessive chamber pressure, resulting in inlet "unstart". Late light-off allows excessive vehicle slow ,
down, also resulting in unstarted inlets.
4-8

The prlrary integral rocket-to-ramjet transition interface considerations are shown inTable 1.
Critical interfaces exist with the inlet, fuel control, combustor/nozzle and booster motors. Maintaining
a supercritical design margin or, if necessary, stable subcritical inlet operation during transition is
a primary requirement. Several early IRR flight tests resulted In failures due to failure to anticipate
the energy release during transition. Careful analysis and resolution of this problem during the
early 1970's cleared a major technical obstacle for IRR systems. The interface with the fuel control
frequently involves providing pressure control signals to limit maximum combustor pressure, in thereby
maintaining the desired inlet margins. Another solution may be to pre-program the fuel flow rate during
transition so that inlet margins cannot be exceeded.

During transition, combustor operation must account for energy sources from the ramjet igniter, ramjet
fuel and combustible pyrolysis gases from the thermal protection system.

Residual booster motor energy sources that must be accounted for during transition include Che
propellants sliver, residual propellant liner and, if present, stress relief boots.

A typical account of the total ramjet energy sources during transition is seen in Table 10. The energy
contributions from each potential source are identified as a function of transition time. The heat release
from each one of these energy sources is shown also as a function of time, leading to the predicted combus-
tion pressure (P 4 ) rise. During the first 50 msec of the transition sequence the residual energy sources
provide 7100 BTU/sec compared to 16,200 BTU/sec expected from the ramjet fuel flow. Failure to accommodate
this high energy release could easily cause an inlet unstart.

EJECTABLE BOOSTER NOZZLE

Several designs for the ejectable nozzle have developed and at least three nozzle release concepts have
been successfully tested: the use of multiple explosive bolts, the use of a Marman type clamping device,
and the use of a snapring retention system,

The use of multiple explosive bolts on the Low Altitude Supersonic Ramjet Missile (LASRM) represented
a conventional approach to nozzle retention as does the Marman type clamp (figure 14) used on the Air
Launched Low Volume Ramjet (ALVRJ). Both systems attain bccster nozzle alignment by direct mechanical
interfaces provided by the fasteners and the use of explosive bolts (common in aerospace use) to release the
booster nozzle. The systems are reliable and due to the use of ordnance common to aerospace use are easily
adapted for use in ramjet vehicles. They also share a common disadvantage - relatively high weight and
complexity.

A later concept incorporated on the Modern Ramjet Engine (MRE) and the Supersonic Tactical Missile (STM)
is the snapring retention system (figure 15). The advantage of this system is significant weight savings
over the other systems and it was demonstrated to be extremely fast, repeatable, and reliable by several
test programs. The system was successfully flight tested on the STM in 1979.
The snapring retention system utilizes a high strength snapring to restrain rearward movement of the
booster nozzle. Booster nozzle alignment is attained by close tolerance alignment surfaces machined inti
the nozzle assembly. Replacing the explosive bolts used in the other systems is a length of mild detonating
cord positioned in a groove beneath the snapring. Ignition of this cord releases large quantities of
combustion gas which compress the ring and force it from the retention groove. As in other systems, the
booster nozzle is thrust rearward by residual chamber pressure. Tests have shown this system to release
in less than 10 msec without imparting rotational or tumbling motion to the booster nozzle.

While the snapring and other concepts offer promise for the IRR of today, advanced concepts are in
development. Several concepts that have been advanced to eliminate the ejectable booster nozzle include
a nozzle fabricated to erode at a rate predetermined to give near maximum performance for boost and ramjet
combustor pressures. Such nozzles may be designed to "step" from one throat size to another as the chamber
pressure decreases or they may be designed to erode in a continuous manner approximating the chamber
pressure decay from boost to ramjet operation.

CONCLUSIONS/RECOMMENDATIONS
After expending a significant development effort over the past decade, a solid engineering data base
exists for the design and analysis of the integral boost motor, thermal protection system, port cover and
transition devices.

High performance and reliable operation of these components can be assured through knowledgeable
application of this comprehensive data base.
Further development efforts in this area should conicentrate on: (1) extending the data base on the
shelf life of the integral boost motor and (2) further simplification of the IRR transition sequence,
leading to a totally passive (no vehicle controls required) transition system design to reduce cost.
 4-9

REFERENCES
1. Kendall, R. M., "A Gene oroach to the Thermochemical Solution of Mixed Equilibrium-Noriequllibrium
Homogeneous or Heteroge /stems," NASA CR-1064, Aerotherm Acurex Corp. Report (1967).

2. Private communication with R. J. Flaherty, United Technologies Research Center, East Hartford, CT
(1971). .
3. Flaherty, R. J., "Arbitrary Pressure Gradient Integra' Technique for PredIcting Boundary Layer and
Thermal Parameters," J. Aircraft, Vol, 11, 243 (1974).

4. Elliot, D. B. et al., "Calculation of Turbul;ipt Boundary Layer Growth and Heat Transfer in Axisyrnnetric
Nozzles," JPL Tech. Report No. 32-387. p, .-.

5. Moyer, C. B. and Rindal, R. A., "Finite Difference Solucion for the In-Depth Response of Charring i
Materials Considering Surface Chemical and Energy Balances," NASA CR-1061, Aerotherm Ac.irex Corp.
Report (1967).

6. Schaefer, J. W. et al., "Studies of Ablative Material Performance for Solid Rocket Nozzle Applications,"
NASA CR 72429, Aerotherm Acurex Corp. Report (1968).

TABLE 1 TYPICAL PROPELLANT FORMULATIONS FOR CURRENT IRR B00STER DESIGN

T11285
Formulation Range Property Nominal Vaiu'e -

87 to 89% solids HTPB o sec 262

3
17 to 18% Al Density, lb/in. 0.064
1 to 3% fine-ground AP Burning rate at 1,000 psi, 0.8 to 0.3 -.-

ln./sec... -

Fe20 3 or FeF3 catalyst Pressure exponent 0.40 to 0.55

ZK, %/F 0.16 to 0.21

Hazards Class 1.3

Structurpl Characteristics

* Mechanical pioperties suitable for -65 to 165 F environment

* Tradeoffs of ballistic properties versus aging stability being studied

o4,.,
F
"4k "1 '..

4-10
TABLE 2 MECHANICAL PROPERTIES OF TYPICAL HTPB TACTICAL PROPELLANTS
1,1286
Propellant No. I Propellant No. 2

J Temperature, F Temperature. F

Property 165 70 -65 Property 165 70 4,

Tmc, psi 110 148 581 Tric, psi 102 147 822
Emc, % 35 42 22 Emc, % 27 29 12

ER, % 37 44 36 ER, % 28 31 15
Eo, psi 5M2 628 9220 Eo, psi 455 671 11,800

TABLE 3 PROPEI.LANT GRAIN/LINER STRUCTURAL INTERFACE rONSIDERATIONS

"T11296
, Provide suitable structural support for grid to avoid excessive strain

- Booster pressure loads must not cause TPS liner failure, particularly
at low temperatures
S.Propellant boots in forward combustor recirculation zone are unacceptable
- Port cover in side-dump combustor must not cause grAir/liner movement

- Booster motor MEOP must exceed structural margins of combustor case,

which is sized by ranmiet loads at hiqh temperature and pressure

* Booster motor must not cause structural damage to SFRJ fuel qrains

TABLE 4 PROPELLANT GRAIN/LINER CHEMICAL INTERFACE CONSIDERATIONS

".r*qrface material system must provide strong chemical bond with TPS:
"BIT > I!0 psi, peel > . plh

. Prooellant constituents must have long-term chemicJl cor.'attbility with TPS

e TFS constituents must not degrade interface material bond strength

 4-11

TABLE 5 PROPELLANT CATEGORIES FOR ADVANCED IRR BOOSTERS

NOZZLELESS MOTOR DESIGNS
__._. .... .... . . .. .. _T11288

Propellant Type Characteristics Status Problem Areas

High-rate HTPB 86 to 88% solids Feasibility Ballistic proper-
"with Al HTPB with liquid d~monstration ties and repro-
ferrocene cata- in process ducibility; pro-
lysts to achieve pellant aging
burn rate >1.5
in./sec, pressure
exponents <0.5
High-rate HTPB 91 to 93% solids Propellant Propellant aging;
with Zr HTPB with -40% Zr development in Zr purity and cost
and liquid ferrocene process
or FeF3 catalysts
Reduced smoke 86 to 88% solids No work in burn Combustion insta-
class 1.3 HTPB without Al; ra,.e regime of bility; aging
catalyzed to high interest
burn rate
Minimum smoke High energy binder No work in Achieving burn
class 1.1 with HMX/RDX and process; may rate and pressure
rate enhancing addi- require new in- exponent; combus-
tives gredient tech- tion stability;
nology low temperature
_ _ _ _ _ _ _ _ _ _properties

"TABLE 6 THERMAL PROTECTION SYSTEM DESIGN CONSIDERATIONS

T11289
Integral
Solid Liquid Rocket
Property Rockets Rockets Ramjets

Virgin material
High strain capability X X
Liqht weight X X X
Low thermal conductivity X X X
Oxidation resistant X X
Erosion resistant X X
Chemically/structurally compatible X X
with booster liner materials
High gas temperature capability X X X
Charred material
High mechanical strength X X
Light weight X X
Low thermal conductivity X X
*..' Oxidation resistant X X
'..' Erosion resistant X

TABLE 7 REOUIREMENTS FOR THERMAL PROTECTION SYSTEMS

T11290

External
Rocket Duration, Pressure, Heat Flux, Temperature, Aerodynamic
Type sec psi Btu/ft -sec F Environment Heating
Solid 60 to 120 500 to 1,000 500 to 1,500 6,300 Oxidizing None
particle
versions
Liquid 100 to 600 100 to 500 700 to 2,000 5,000 tm 5,500 Oxidizing None
eM44+,4t inn tn 1.-lnn in tn isn 10 tn 30 2.500 to 4.000 Oxidizino/ 500 to 1,200 F
"4-12

TABLE 8 PREDICTED AND MEASURED THERMAL RESPONSE

-_ _T11293

"f"Forward Aft
Recirculation Boundary-Layer
Property Zone Zone

Typical boundary conditions

* f/a 0.10 0.039
* Tc, R 4,300 3,700
* HTC, boundary-layer relative length 0.075 0.05 to 0.60
. HTC, blowing effect reduction, % 5 (maximum) 6 (maximum)
* Effective emissivity 0.15 0.036

Charring lining
* Typical heat flux (predicted), 13.5 13.0
Btu/ft -sec
* Convective, % so 30
* Radiative, % 50 10
0 Reactive, % 0 60
* Inconel case temperature
* Predicted/measured 1.043 1.006

"Fully charredlining
e Typical heat flux (predicted), 13.6 8.0
Btu/ftk-sec
e Convective, % 60 90
. Radiative, % 40 10
* Reactive, % 0 0
e Inconel case temperature
* Predicted/measured 0.974 0.992

"TABLE 9 INTEGRAL ROCKET RAMJET TRANSITION INTERFACE CONSIDERATIONS

"T11294
"Parameter Consideration

Inlet * Pressure margin

* Supercritical design margin
"* Subcritical stability limit
Fuel control * Fuel flow control accuracy
* Pre-programmed fuel flow rate and time
Combustor/nozzle * Ramjet igniter
* Fuel flow rate
* Thermal protection system

Booster motor e Propellant sliver

* Residual propellant liner
Boot
 -7 P. !*--0' - 717"

4-13

In u ~-I CI

4E C co 40
%a
CL 00 '800; C;N4O

U,

- r4

U.,

Cu008 8 00

ac Et -4

to 44 u

'oCA 04DCCD *

-J4

LU 41 u
0 j E 00

4A CC

0
IA0
Cu) CQ C, 0

LL.0

"I Cn Un to
-4 "W

LU Ln

% 4A
5-
4A I
U 4 ) -
u nJC 4 O
I.v - -CC

00
an~

5-**L-
41 C

goC
4-14

Podded combustor
A. Initial Podded Ramjet Configuration

Booster

Ramjet combustor
B. Tandem Roc ket Ramjet with Submerged Nozzle

Integral booster

1-

integral rocket ramjet combustor

C. Internal Rocket Ramjet

Figure 1. Evolution lofIntegral Rocket Ramjet Configuration

33057

04in IgntorFrangible forward support

Port cover (four pi aces) W ire m esh (M A O.TEG)ri g A tI n ors f a da m

Ectbebooster nozzle assembly,
CaseInslatin. nap rigrelease mechanism
WiremeshOC 3104silion*Sol*andarm device

boser Pransducer
Fueltmanidoldsbos

traPortcer brs

Top View

Figure 2. Typical Integral Boost Motor

33054
 4-15

Liquid Inlet Boost propellant - Ramjet nozzle

combustor
Liquid fuel ramjet
Cas generator
propellant olppean

S I". JI \V

Comb-ustor
Ducted rocket
.,Ramiot fuel Boost propellant

Combrusto?
Solid fuel ramjet

Figure 3. Types of Integral Rocket RainJets V- 08171-RI

ci1,500- - -

V; 110

3 Surface area history

NoI

= 0 12 3 4 5 6 7
Bum distance, In.

0.5 In. Unit of

Property Measurement dd

4.23-in. dimtrWeb 'hickness, In. 4.31

'te Propellant weight, In. 538.25
Grain length diameter, In. 2.53
1681i. imee Port/throat ratio (geometric) 1.84
16.8-Indiaete - ..- ' bla ratio (outer to Inner radius
ratio) 4.0
0.75I Volumetric loading 0.92
41.13In.-Neutrality (maximum
to average area ratio) 1.055
42. In
 4-16

Figure 5. idozzleless Motor

33056

* 0
0 a

1.600 Two-zone prediction

I- B/ Therlnoco.J; taoand

S0.75 - Boundary~layer
tfln location

0.5.
0.50~~on ;artaing sot0;~~ position

Initial case loni eratur llj Iiilca eprtr

0.25 % 0.25
00.15 0.30 0.45 0.60 0.75 0.90 1.05 0 0.15 0.30 0.45 0.60 0.75 0.90 1.05
Relative axial location, UL reference Relative axial location, UJL rcferenc-s
A. Virgin Lining, B. Fully Charred Lining,
Two-Zone Correlation -Two-Zone Correlation

Figure 6. TWo Zone Thermal ?.bdel Eat-i Correlation

33057

Recirculation zone state Combustor state, same as one-zone model

(ln,;reased fla,
equilibrium Tc, ;.nd composition
empirical emissivity)
Bouindary-layer model
(new starting location)

-c Charr Ing lining heat transfer model,

same as one-zone model

A. Forward Recirculation Zone B.Aft Combustor and Nozzle

 4-17

1,600 ..... 55F:-

1,200 - -'-
-55 F- ...
0.1

800
07

000L
0 1 2 3 4 5 6 7 8
Time, see
A. Forward Head Chamber Pressure vs Time

20,000 -

a 12,000

" 8,000 14 /
2I 70 F - ..

4,000 ii"
0 L"
0 1 2 3 4 5 6 7 8
Time, sec
B. Thrust vs Time

Figure 8. 'ozzleless Motor Ballistics-33059 33059,

LID = 1.5

LID 5.5

Same A5

Parameter High LID Low LID

Cruise chamber pressure, psia 10.5 13.5

Combustor Mach No. 0.43 0.21
Cruise heat transfer coefficient (H!CP) 0.0195 0.0135
Integrated surface heat flux at end 7640 5025
2
of Identical trajectory, Btulft
Combustion chamber gas velocity, ftlsec 1210 600

Conclusions:
(1) High LID comb',stor has 50% higher heating
(2) High LID combustor sees 100% higher gas velocity
4-1 8

Dual-urposecombuston charnorb Nnaxlesas

b grtainek ElectIntlePosn
cover
Asttet fuel AirAr i

Air

Ranmet
fuel

Soost Operation Transition Ramjet Operation

P.e 2100 pale 300 mssc Maximum
Operating Sequence of Integral Rocket Ramjet
Port Cover Design Problems Spacial Nonejectable Port Covcr Problems
"*Must sealInterfacebetweencombustor andair Intel over - 65to 165F
enoros tmental
mns (2100pols)

"*MustbeopenednapId I< 300mosseiwiht minlimum~ pressurediffernetial * Mustnotailect basiccombustor

flanwltolding and mixiengprocesses

" First priority mustbegivento "Ielactuation vs. comnmandedport coweopening * mustnotcausesignhtlc-nt I> 3%)pressureloea
for lmoicty,
elibilty* ow ostwW Mastnot alect anypartlole grealer than 0.03In. Indiamneter
tor taunoblches
"oMust Prood suitable boitdingistntcturet Interlase torIntegralsolid bootter aermaltsafety

*Mandatoryfor *tde dump Additional Problems

- opiona forcentr dup.* Msy berequiredto providecontrol signalsfor rerojetenginestert
"*Mustbelighttweightendnot damageramjetinsulsiato Opentuel lootatiort'at"
Firssigntsto turbine sterlcarrirdge
"oMustbeableto nithtaotndrenm
air external I< 100 peliduring catv
toads Pslongta to reennjtIgniter
Fir
tfplis *Fire signalto duededrocketgasgeneretor

F1t betr port cover,IneelatinIsrequiredduring boost

not consr
genedoese

Figure 10. IRR Port Cover Design Proble V259

Contour sdr plate

..Bg seal

12594

116- B. Segmented Port Cover

77 Hinge support,
fine required
il~ng* od-.,

Opening pai
ofd
dore Liner
AlicflqnlInettistion.
$ rew If 'O in.

Hard oxide O-ring

1'1261-2 coaling all Poricoe
around port Linesr shaped charge

A. Monolithic Port Cover C. Hinged Port Cover

Figure 11. Port Cover Types

33061
 -2,

4-19

Dual purpose combustion chamber Booster grain

Ramiet fuel Nzl lm

Ine otcvrBooster nozzle Nozzie

Ramjt pot cverclamp
nzzleInlt
Air1

Booster
Dual purpose combustion chamber Ramjet fuel Ramiet nozzle nozzle
Rame Aiue B. Transition

Ramjet nozzlj-
C. Ramjet Operation
Figure 12'. Integral Rocket Ramjet Comtbining Booster and Ramjet in One Combustor Chamber
33062

1200 --

Booster grain
web 9.0.
P.C. sensor setting____ ___

sgasnozzle
Tdfg~er mechanismfie
-..
2 800rele3se
1
device squib
IhgeI Ramjet igniter on Ramjet Igniter off
lB oort nozzle fullyelcdNoia me
IPort cover
bulolyi
.2=
E
If
i
Jovelucted
ej1e
Total fuel flow rate. -
2.64 Wbiesc
- Wair +Wilnr--
t~~~WT
f-
l Wi 60.7lblsacZ Chamber pressure, pals
II I.~ ~~~~~~
Steady stae cold ir"
a II- Pdu ocmuto
--- t, 220 I flow combustor It,50. 0

-
to 200 300 I
400 5001 600 700 800 900 1000
Timts, maec. 0

Figure 13. Typical Transition Sequences for High Altitude Launch A.

V-28398

.. S'
4-20

SExplosive bolt assembly

Segment (fcur each) -Band assemb~y

Figure 14. Marman Type Booster Nozzle Retention System

33063

Booster nozzle
Metal Insert

.Deonator (alternate location)

Detonator
(proposed location)
:Snapring

,, Nozzle shell

Pyrocore ejection cord

Ramjet nozzle

Figure 15. Snaprlng Type Booster Nozzle Retention System

S. ...... 33064
 5-1

LES STATORF.ACTEURS A COMBUSTIBLE LIQUIDE

par Ph. CAZIN

Coordinateur Missiles

Office National d'Etudes et de Recherches Afirospatiales (ONERA) '

29 Avenue de la Division Leclerc

92320 Chatillon
FRANCE

RESUME

Les stator~acteurs A combustible liquide, utilisfis pour propulser lee missiles supersoniques,
permettent de rapondre A certaines spdcifications op~rstionnelles, telles que la recherche de
portfies 6levfies, jointes 41une grende flexibilitf des tr~ijectoires. Ces moteurs trouvent sinsi lour
euploi doen un crfireat' partic'ilier, qui asedistingue de c,!lui des statofusfies ou des turborlecteurs
supersoniques :leur domains eat r.elui des missions qui recqaigrent de fortes variations d'litd
ou do vitesse.
lea performances Snerg~tiques 6levge3 de ces machines sont obtenues au moyen de techniques ou de
technologies sp~cifiques que lVon se propose d'examiner, A savoir

"* Les combustibl~es liquides : k~rosane, combustibles liquides tras denses, combuRtibles boueux,

"* Systimes d'alimentation :per pressurisation, par turbopompe. Avantages at inconvinients de cheque
systame. Conception des r~servoirs.

* Pguatonetinjection :lois de raguletion requises suivant lea missions. Diffirents types

"dnElations
*Chambre do combustion :conception, dessin gfindrel, zones de recirculetion, localisation de
l'injection, protection thermique, contreintes dues A ls pr~sence d'un acc~l~reteur integrG.

"* La combustion :performances, rendements et instebilit~s. On 6tudie plus particuliArement les

ceuses des instabilitais de combustion, leure cons~quences rt lea rem~des envisageables pour y ..

rem~dier.
"~Les bance dVesseli des statorgacteurs A combustible liquid., evec leurs cerectfiriatiques
spicifiques.

En conclusion, un epervu sur leas ossibilitls de ces moteurs suivant lee missions, our lea
principalec 6tapes d'un dfiveloppement at sur l'avenir de ces techniques. est proposa. --

SUMMARY

Liquid-fueled rami ets provide a means of satisfying certain performance requirements in supersonic
saSsmile propulsion, such 5s combining long range with a large flight path flexibility. These power
plants thus have a particular application quite distinct from that of ramrockets or supersonic
turbojets t missions requiring major altitude or -;elocity variations.
The high energy efficiency of these engines is obtained by specific techniques end technologies
which we propose to investigate:

*Liqsuid fuels : kerosene, very dense liquid fuels, fuels in suspension, etc..

FuP'eling system :Advantages end disadvantages of pressure vs. turbopump systems, tank design.

*Regulation and injection :Regulation laws required for diffe.*ent missions, various injector
types.
*Combustion chamber :Design, general chaps, recirculation zones, injector location, thermal
insulation, stresses due to integrated booster.
*Combustion : Performance, efficiency and instability. In particular, we will go into the causes,
consequences an possible remedies for combustion instabilities.

*Test stands :Specific characteristics of test stands for liquid-fueled ramjets.

5-2

NOMENACLATURE

A, Section

Indices i0 infini amont

1 section de captation de la prime d'air
2 diffuseur prime d'air
2' chambre de combustion, avant front de flamms
3 chambre de combustion, apras front de flame
4 col de la tuyare
5 sortie tuyare
i conditions d'arrlt

D. diamatra missile (chambre de combustion)

1I INTRODUCTION

Le mloteur fusae A propergol solide a Stf et rostera encore longtemps le principal mo:?en de
propulsion des missiles tactiques :bien adaptg Aux courtee portges, 11 seatcapable de fortes acctlgrations
et ourtout i1 eat relativeuznt bon marchS.

Maim ii a aussi certains inconv~nients :par exemple, pour augmenter sensiblement is portgs d'un
missile A moteur fusae, ii faut, soit lui faire suivre une trajectoire balistique A haute altitude, avec lea
probl~nes de guidage et de dAtection que cola suppose, soit, pour des trajectoiree A basse altitude,
augmenter fortament as masse at eon encombrement.

C'est pourquoi, pour lea missiles lanc~s A une certaine distance de 'our objectif, ii faut se
tourner vers d'autres systames de propulsion pius effica..ccc, elest a dire plus 6conomiques et capables de
produire uns poussfie continoe cleat is domains de is propulsion agrobieg avec 1. tuvbor~acteur, Ie
statofosge I combustible solide et 1k statorgacteur A combustible liquids. Apra* one rapids comparaison
entre iea possibilitgs de ces trois familles, qui ont chacune un domains particulste auquel elles sont bien
adapties, on as propose d'examiner lea techni.ques et lee Lechnologies requises pou la prop~ulsion des
missiles par etatoriactsur A combustible liquids.

2 - LES DIFFERENTS CONCEPTS

D~couvert sn 1911 par le Fran~ais R. LORIN, is principe de is propulsion par 5tator~acteur set
extrfusment simple :ii consists A tirvri parri de la recompression rfslisable en vol dane une prime dlair,
pour obtenir, sprAin combustion, une force despous~ge resultant de 1'6vacuation du debit d'i capt6, A one
vitesse suplrisure A calls du vol. Mais ii a un inconvlnient tn'Atant pas autonome, is statorgacteur ne
pout fonctionner qu 'apr~s une miss en vitesse par un propulseur auxilisirs ou "Acc~llrateur".

2.1. -Rappel hietorigue

Depuie 1s fin de ls Deuxilme Guerre Mondiale, jusqu'A 1965 environ, il y eut dane plusieurs pays un
effort de recherchs important qoi aboutit A des-ombreux essais en vol de missiles expfirimentaux et A
qusiques d~veloppsments opdrationnels. principalement pour des missions surface-air ou aol-aol.

Do cstte Apoque tras fertile, on peut citer tout d'abord lee missiles op~rationnelsa mdricains
UOMARC et TALOS, britanniques BLOODHOUN.) et SEA DART et sovi~tique SM4. En France, maigr6 T~lus des200 essais
so vol qui ont permis d'acqufirir de solides conneiseances techniques et technologiques, ii.nly out aucun
dfiveloppemsnt op~rstionnel :c'eec de cette 6poque que datent lee missiles exp~rimentaux CT 41 et VEGA des
Nord Aviation, R431 de MA'VRA-NORD-- AVIATION, STATALTEX de l'ONERA, etc...
Tous ces missiles avaient des configurations voisines, A savoir
(Fig. 1)

prime P'air axisym~trique situfie A l'avant

acc~lirateurs largables, montle en tandem ou lat~ralement

- chambre do combustion A accroche-flammes mAtailiques (gouttiares, c8nes perc~E ds trous ... )

-combustible liquids kfiros~ne.

Puts sntre lea anndes 1965 et 1970, ms produlsit one d~croissance des recherche* our is
statorlacteur classiqus, moddrdment supersonique, so profit du statordacteur hypersonique A combustion
supereonique (klrosane 00 hydrogane). Lee Applications alors envisagfies Stsient :i'avion Mach 6/7, Is
lancsuT atmosphlrique rficupdrable, Is missile tactique aol-air ou air-air I vitesse supirisure A Mach 6.
 5-3

#00

Fig. 1 :Configuration des arn~es 1960/1970 ritssilt tArinvrxtal

fran4;ais SrTAhLME (ONERA)
2.2. - Tendances actuelles

Vero lea aunnies 1970, on a assistt, surtout a'ixEtats-Uni aet an France, A l'apparition de nouveltes
configurations a~rodynamiques et A ia miss au point de technolcigies deatinacs A rendre le statorlacteur plus
attractif et plpia compatible avec lea exigences opfirationnelles. De plus, c'est en 1973, lots de la guerre
du Kippour, que le missile SA.6 sovi~tique propula4 par statofusi~e A acc~laratpur int6,qr6, a mcnt~r6 our le
terrain, l'efficacitg des formules nouvelles. Enfin lea progrAs effecntu~s daz.9 le domaine des 6quipei*.ants de
guidage, de pilotage at des calculateurs embarruas peruetten.' maintenant d'eovisager de fagon plus r~aliste
une extension sensible du domaine de vol des missiles (port~e, vitesse, d~nivel~e, diversification des
trajectoiies, manoeuvrabilitfi, ..

C'sa ainsi que, sous 1'impulsion de ces nouvellca contraintea op~rationnellet, mont apparus depuis J
lea annfes 70, lea concepts suivants
*Configurations a~rodynamigues A prises d~i lat~rales (at non plus frontales) &fin de icendre le missile
riodulaire (structures simplifi~es et maintenance facilit4e) at d~gager l'ogive (montage p'.,sible de tout
type d'eutodirecteur filectzxmagn~tique).
*Statorfscteur avec acc~llrateur int~grA cette technologie permet de concevoir un missile beaucoupiplus
compact et surtout monofitage et ainsi d'gliminer lea probllmes de s~curitd lies A la retomb~e d'un
acc~lfirateur d'aproint.

"* Chambrede combustion "tourbillonnaire" , c'est A dire dans lequelle la flamme n'est plus stabilisfe par
des accroche-flammes mttalliques mais par des diepositifs purement earodynamiques, laissant ainsi la
libert6 de remplir la chambre avec le propergul d'accllfiration.

"* Statofus~e
A combustible solide zdens ce concept, on a recherchfi A simplifier encore Ie systame
propulsif, de telle sorts que le stockage et 1A maintenance sa'pparentent A ceux d'une munition ou d'un 6
missile classiq-ie A propulseur fuste. Il en existe deux variantes principales

- la plus cimple eat le statofus~e A ginfirateur iLt~.grA ("Solid-fueled ramjet"), dana laquelle Ie
combustible solide du statorgacteur eat plac6 directement dana la chambre de combustion.

- le atatofusfe A g~nfirateur s~par6 ("iu~:_ted rocket") qui eat plus ou momns simple salon 1& nature du
combustible et l'existencc ou non d'un syst~me de modulation.. -k
Cas nouveblles technologies ont rendu cc systfta propul~sif trL~s attractif, du point de vue militaire,
en supprimant o'ienia~ttfnuant sea principaux inconvdnients. C'nat pourqnioi de nombreux pays se sont, 1cnL$s
doen cette vvie dep_is lee annges 1970 ;on peut ains2. citet, en se limitant aux missiles essay~s en vol en
Occident

- sux Etatb linis, des prototyp~es ou missiles exp~rimentaux :l'Advanced Strategic Air-Launched Missile
(ASALH), l'Advanced Low Volume Ramjet (ALVRT).

- en France, l'activitfi set trls importante puisqu'ella concerne, non seulement des missiles exphrimentaux,
tels qua le statofus~e A gln~rateur sApari ("Ducted rocket") essay6 an vol en 1976/77 et le statofus~e
"ofrustique" A g~n~rateur intfigrfi ("Solid fueled ramjet") essayG en vol en 1.983/84, minis surtout l'ASMP
(Air-Sol Moyanne Port~e) qui sera bient8t le premier missile op~rationnel de cc type done le sonde
occidental (Fig. 2) ; il utilise un stator~acteur A combustible liquide.

- enfin la France at l'Allemagne prfparent le missile ANS (Anti-Navires Supersonique) destin6 A succ~der."7
A la famille EXOCET. Deux systlmes propuleifs cont en comp~tition un statorfiacteur I combuatible
liouide (Agrospatiale/ONERA) et un stato~us~e A combtistible solide dop6 au bore (2438), dont un modlle A
-5-

P~ig. 2 t Configuratim r~cente a ndssile

~, frangais Air-Sol (AP6rospat1iale)

11 faut rioter que I'intgrft marquf des Europtens, e. particulier de 1& France, pour cernformules
nouvelles, s'expl'que, notawoaent pour lee missions air-surface, per lee foibles dimensions des avions
Europ~ex... i loruque V'on compare lea ponsibilitis d'emport d'un BI ou d'un F15, I celles d'un Mirage IV ou
d'un Suver-Eterdi~rd, il eat bien fivident que le, exigences d'am~lioration de la compacit6 du missile mont
.41 plus 6leyfee en France qu'aux Etats-Unis I

2.3. - Possibilitfs de'sstaa~orgacteurs A combustible liquide, par rapport A celles des statofugges e~tdes
turborfacteurs su;,ersonicues

Chaque type de moteur a des avantages et des inconv~nients selon Ilemploi op~rationnel auquel i1 si
desting. Ainsi, pnu- mieux cerner lee possibiliths du statorgacteur A combustible liquide 11 taut Ie
* comparer aux moteurs qui sont lea plus proches, 1. statofusfe I gnAnrateur s~parg at le turbor~acteur
* supersonique. En effet, le statofus~e I g~nfirateur int~gri ("Solid fueled ramjet"), plus ap~cifiqu* des
petits missiles ou des courts. portfies, entre beaucoup momns en concurrence avec Ie statorgacteur a
combustible liquids.

2.3.1.Statorgacteur et statofus~e A Og~nrateur sdtsrg

Les domaines d'emploi, & savoir lee trajectoires de moyenne portfe, sont essez proches 2c'est

Le tableau ci-dessous montre lea possibilitgs de cernmoteurs

a AVANTAGES 2 INCONVENIENTS

iMissile A stato- 2*Taux de modulation :0Complexitt (stockage2

t r~acteur I combustible 2 4154 , donc fortes :et alimentation du
Iliquide :variations d'altitude :combustible)
:ou do vitease
a 2*CoOt plus filevA
0* Rgulation prfcise :
NaSt sou is (manoeuvre.s
Ira ides

0* Performanceo massiques
a~ Glevess

t Miss'le A stato- 2* Simplicitf (surtout :0Performances moins

t fustse A g6n~rateur :si absence de modulation) fleyfes (surto-,t A 2
2 sparA :cause du taux de modu-
1 2 lation limiti)
a 0tstockage aisA

2 2* ~compacitf Sventuellep-2a
a :t ment plus flev~c
 5-5

Pour illustrer ia proximit6 des domaines d'einploi, is figure 3 compare l'encombrement et Is masse de
*quatre missiles supersaniques air-surface (port~e A basse altitude :100 kmn)
* deux sont propulsfis par stator~acteur A combustible liquide (k~ros ane de densitfi 0,8 et combustible
* * ltquide de densit6 1,0).

* deux sont propuJais par des statofusaes A combustible solids discret (densitfi 1,2) et A combustible solids
* dop6 au bore (densitg 1,7).

~'' Ma5ss*a
45r16 d~part (kg)

Fig. 3 QOrrparaison statoreacteur/statofusde. Ids-

sion :Air-Surface-btuch 2 -Z =0 D) 0,35 mn

I.? 900

* Bien qus lee 6carts de masse et de longueur soient relatIvement inod~rfis, on constate l'intgrlt des
* combustibles denses pour ia compacit6 du missile.

* ~Mais, une comparalson plus prdcise entre ces deux formusles doit tenir compte d'autres critAres
complamentaires tels que

* - diversification des trajectoires,

- discr~tion optique du Jet,

- co(It du d~veloppeinent et coOt du missile en sanie,

- possibilit~a industrielles,

-etc...

Eile d~pend donc fortement de Is mission envissgge.

2.3.2. Stator~acteur St turborgacteur suxversonitme

Lee domaines d'enploi ne sont pas aussi prochee. Mais puisque, pour lea missiles tactiques
subsonigues de type air-iser ou mer-men, on trouve une cohabitation entre des moteurs trAs diffdrenta, A
savoir ia fusae A propergol solids (EXC"'ET, K0RM0RAN, GABRIEL, ... ) et Is turbordacteur (HARPOON, OTOMAT,
SEA EAGLE, ... ), on peut penser qu'elle puisse se poursuivre en supersonique, en particulier posir des
raisons industrielles.

CVest pourquoi, le tableau suivant montre lee principales possibilitas de ces deux moteurs

AVANflAGES : INCONVENIENTS

Missile A statorfacteur *plus rapids : plus lourd et plus :

A combustible liquide plus manoeuvrant encombrant pour lee :
*dfnivel6es plus :port~es sup~rieunes A:
Alev~es 150 km (vol A basses
I altitude)

2Missile A turbo- :* faibie consomniation ~*complexitfi du

t r4acteur supersonique : syst~me propulsif
:' ventuelleisentI
fonofoteur pour ~ *coft 6levt
missions depuis avion
~*domains mains Atendu
I Zen :
z ~- vitesse msximale
* - altitude
- d~nivelle
2 3 - manoeuvrabilitg
 La figure 4 qul presente, suivant le moteur adoptA, la pouseSe et la trainfie du missile en fonction
de Is vitesse de vol, ficlaire 1e domaine d'emploi

- le stator~acteur est apte aux grandes vitesses (plus de Macb, 2) et aux trejectoires diversifi~es.

- le turbor~acteur supersonique basse altitude testers limit6 A des vitesses sensiblement inf~rieures A Mach
2 (Mach 1,6 ?). 11 prendra son essor d'autant mieux que lea port~es augmenteront et qu'il sera capable de
voler en sub-, trans- et supersonique.

F~dAN)

2 Fig. 4 Pouss~e et 7tain~ede irdmssiles tutt r~acteur et

-~A stator~acteur (Z =0)

1000o

2.4. Domains d'emplol du st,,tortacteur A combustible liquids

En r6eumg, pour la propulsion des missiles, le statordacteur set particuli~rement bien adaptg aux
aissions A moen pr~ (jusqu'A 150 kilom~tres A basse altitude et plusieurs centaines de kilom~tres A
ha!ute alt itude), A vitesse hautement supersonique ,avec de fortes variations d'altitude et de vitesse
Done ce document, on 9'intgresse exciusivement au stator~acteur A combustion subsonique qui seat
*capable de fonctionner depuis Mach 1,5 (*) jusqu'4 Mach 6, mais qui eat surtout utilis6 op~rati onnellement
dane Is domains de Mach 2 A Mach 4 en effet, corns le montre la figure 5, ii sat possaible d'obtenir dans
catte zone des periormances quasi-optimales, en simplifiant le moteur par utilisation d'une g~om~trie fixe
(r~fgrence 1). Dane cet example oills Mach de vol d'adaptation eat 6gal A43,2, la perte d'impulsion
dpficifique du moteur A g~om~trie fixe n'est que 10% A Mach 4, mais elle d~passe 20% A Mach 5.

(a)
/ eombr,$*

Fig. 5 t Performunces optirmles d' tm statordacteur au10.

keros&me (richesse :1, ASA 3) A6

Mach k

% Pour utiliser un tel moteur, 11 faut maltrieer des techniques ou technologies particuliares,
* sxaminges ci-aprls, A savair

* - la wise au point d'une famille de combustibles liquides

- lea rsearvoirs, at lee systftes d'alimentation, de r~gulation et d'injection

- lee chambrcs de combustion

3 - LES COMBUSTIBLES LIQUIDES

Las missions qui font appel A un combuatible liquids sont csractdris~es

soir par une longue port~e, le combustible liquids permettant une meilleure impulsion sp~cifiqiie que
cell. des combustibles solides (c.'r cernderniers contiennent g~n~ralement des produits oxydants).
- uit par un domains de vol 6tendu (en particulier en altitude), ie combustible liquids 6tant plus apte que
* lsa combuatibles solides A de fortes variations de d~bit
Le combustible liquids le plus utilis seatbien entendu ls k~ros~ne -de type aviation- en raison de
son faible coilt et de son approvisionnement ais6.

---- ----
-- ACJ...---- ------- l 4 ------
 Maio pour Ilutilisation our missile, le problilme du coOt eat beoucoup momns aigu quo pour lea
applications aux avions ou hfilicoptlres, puisque I. moteur no sort qu'une fois. A titre Indicatif, 1. prix
d'un kilograom de propergol classique eat frdqueommnt plus do 100 fois plus filev6 que celui d'un litre do
kiroslno I
V'est pourquoi, does plusiours pays, des recherches Cot 6td monfes dana Ie but do ddvelopper des
combustibles liquidos plus of i'caces quo lo kdrosdne, c'ost & dire ayant los caractdristiques suivantom

a) Si possible, pouvoir calorifique masmique #gal ou suphirieur I celui du kdrosbno

b) Esmontiellemont, masse volumique nettemont mupfirieure

* Certains do ces combustibles pouvent Stre utilisds aussi bien sur turbordacteurs queomur
statordacteurs. Maim 1'omplol our missile rlpond 9 des critdres de choix trim splcifiquos.

3.1. Critlres do choix

Solon I& mission, Con choisira un combustible qui riponde on prioritfi A tel ou tel critaro. Parmi
ceux-ci, on pout citer
a Performances SnergItiquss: recherche do combustibles I haut pouvoir calorifique massique (ou A foc.te
impulsion apdcifique),c'est A dire un pouvoir calorifique inffdrieur, au isoins figal 110 000 KCCal/kg
(soit 18 000 BTU/lb)
o asse volumique et impulsion volumique .Pour augmenter Is coriapacitl d'un missile, ou 11volume donn6 pour
amdliorer ma portle, on recherche une augmentation do l'impulsion volumique :le proc~dd le plus of ficace
eat gfintralemint un accroissement de Is masse volumique. Des combustibles liquides de masse volumique
comprise entre 1 000 et 1 100 kg/m3 ont ainsi d~jA Atfimis au point en vue d'une utilisation Sur missile
tactique.
Ponctilnnement a haute altitude :cortains combustibles boueux, tols coux qui contionnent des particulos
do bore, br~lent plus difficilement A basse pression. 11. peuvont donc ne pas Stre raetnus pour des
missions A haute altitude.
*Dilatation : I& masse voluminue varie avoc is tempfrature. Ainsi mu? un missile dovant fonctionnor a ios
tempiratures extrdmes, un kdrosAne classique me dilatera en volume d'onviron 7 A 6% pour un 6cort do 100
degrlm centigrades :le rfisorvoir devra donc Itre conqu pour permettro cette libre dila tation.
* *Viscositf :la recherche de combustibles do plus en plus denmes s 'accompagne gInfiralement d'uno
augmentation do la viscositt, en particulier aux basses tempgratures. C'est cs phdnomane qui limite
Ilutilisation do combustibles tram denses sun missiles tactiques, en particulier pour lea versions air-sol
ou air-surface non stockfies en soute, en raison des conslquences pratiques sur lea systames do cheose ou
loe pompos.
A titre indicatif, des viscositfis supfirieures 1 150 centistokes posent des problames ou nficossitent
une technologie adaptfie.

Maim cortains combustibles boueux peuvent Itre utilisgs malgr6 uno viscositt statique beaucoup
plus forte, car dotds de propriftfs thixotropiquem au repos. ces boues possLdent uno viscositi filevfie
qui chute considdrablement dim qu'elles mont soumises A une force do cisaillement, c'est A dire au moindre
mouvemont.

*Point d16clair :pour facilitor l'ellumngo ot la combustion, en particulier lorsque 1. combustible est
portdi A trim basso templraturo (missile en charge externe d'un avion ef fectuant un vol prolongA A haute
altitude), 1. choix as portera plutdt verm un combustible ayant un point d'dclair A basso tempdrature.

Maim parfois ce choix no pout Atre fait pour des raisons do sficuritS, par exemple pour limiter lee
risque& d'incendio A bard des navires (Wrefrnco 2). Une telle contrainte pout obliger A choi8ir un point
d'Aclair supdrieur A 60*C.
*Point do conggla14ion :il est bien 6vident quo celui-ci doit Itre infdrieur aux templraturem lee plus
L ~basses rencontrdos au cours du stockago ou de l'emport.
oDistrltion t certaines missions opdrationnelles nficessitent l'utilisation do combustibles assurant une
bonne discrdtion du let propulsif, soit dans 10 domaine 6llctromagnfitique, soit surtout dens 1. dowiine
optique (visible et infra-rouge). Dee tells cl~ntrainte pout conduire A prdfirer des combustibles no
contenant quo des hydrocarbures (uniquoment carbone et hydroglne), plutat quo certains mdtaux
(aluminium, magnfisium) o~jmgtallofdes (bore).
*StockabilitA lea combustibles pour missiles tactiques doivent 8tre aptos I un stockage de longue durge
A,1intlieur du rfiservoir, Or, il no s'agit jamais d'un produit chimique simple ipar oxomple, le
kdrosane, pourtant tras courant, oat un ddnivA du pdtrole dont I& composition chimique oat tram variable
suivant lee diff~rents traitements effectugoaen raffinenie et Ie provenance du brut. C'est pourquoi
certaines normew ont fitfitablies pour loa carburants A stockage de longue durfie.

loss principaux problames rencontris mont lea suivants

-oxydation en prdsence d'air dissout, ce qui peut avoir une action corrosive Cu favorisor l~apparition
TI - C.1p...4, .S..l, ,nAsie, rP,,nl 1eA~AAdu riservoir at l'adionction
5-8

- corrosion de certains mat~riaux piastiques ou de joints d'6itanch6itfi, ce qud.conditionne in

conception du r6servoir. Par ailleurs, ii existe des additifs qui limitent les risques de corrosion.

- pr~sence de micro-organismes (bactfiries, champignone, moisiesures, levuree) qui peuveaft se dfivslopper

au contact du combustible. Ils peuvent alors former des boues et des amas de filaments qui risquent
d'obstruer diff~rents organes. La prdsence d'eau dans le combustible en favorise in proliftration.

* Dcantation : certains combustibles de haute densit6 sont pr'8par~s A partir d'une suspension de particules
solides lourdes pulv6rie~es dans un support organique liquide ; is stabilit6i de in suspension 6tant &lors
assurfie par in gf11.fication do support de fa~on A bloquer le mouvement des particules.. 11 faut donc
* 'assurer qu'aprAs un stockage prolongfi ii n'y sit pas de risque de d~cantation des parricules solides
susceptibles de modifier lee caractaristiques du combustible au niveau des injecteurs.
*Aultres propri6t6s :cette liste n'est pas exhaustive :suivant les applications, d'autres proprigt~a
sphcifiques peuvent Atre souhsitfies. A noter en particulier
- I& prfsence d'additifs anti-giace pour fiviter in formation de cristaux pouvant obstruct les filtres
ou lee canalisations

- l'adjonction de diasipateure d'Slectricitfi statique pour rendre is combustible plus conducteur

- lea contraintes de co~t du combustible qui n'influent pas uniquement sur is co(It du missile de
n~rie mais aussi sur is coOt du dfiveioppement en raison du grand nombre d'essais effectufis (voir
paragraphs 6)
- i'absence de toxicit6, non seulement du combustible, sais des produits de an combustion avec l'air.
Cette contrainte seatimposde aussi par Is d~veioppement du missile plut8t cue pnr son utiliesation
op~rationnelle.

3.2. Principaux combustibles liquids.

De nombreuses recherches sont effectuges dans pluaieurs pays, en particulier aux Etat3-Unis
(r~f~rence 2), &fin de mettre eu pcint des combustibles liquides, .oit plus f4nergfitiques que is kfirosane,
soit surtout beaucoup plus dense..

Ainsi trois principales families gout apparues

a) carburants A base d'hydrocarbures provenant du

o::ptrole ou de in houille k6rosanes TRO, JP4, JP5,
dicaline, t~traiine,
b) carburants A base de produits chimiques de synthase : RA3, 1.35,JP9, JP10... Pour de tels produits, in
proportion d'atomes de carbone par rapport aux atomes d'hydrogane est plus 6lev~e, de fagon A augmenter
l'impuision volumique.

c) combustibles g6lifigs injectable. ou "boueux" il s'agit vOeliquides, tel. les pr~c6dents, dane lesquels
on ajoute une forte proportion de particules solids. tras fines en suspension ; par exemple : carbons' ou
hydrocarbure aromatique, bore, alumini.,m, magnAsium (le bfiryllium 6tant 41iming car toxique et
coOteux).

Covrme Ilindique le tableau 1, is bore seatintgressant par son fort pouvoir calorifique volumique,
alors que i'Aluminium et is Magnfsi8U= permettent plut8t d'sugmentsr is poussfie sp6cifiqus (rdffirence 1 et.
3).

Point Point Pouvoir calorifique Rapport

El6ment :de fusion :6bullit. Densitf KCal/kg KCal/Dm3: KCal/kg :de m~lange:
..C0Cair :(avec air):

*Aluminium : 660 2465 t 2,70 7 350 :19 845 1940 :0,264

Beryllium 1280 2970 : 1,85 :15 900: 29 410 :2100 :0,132
* Soe : 2300 250 : 2,34 13 800 :32 300 :1450 0,0
*Carbone : 3650 :4200 : 2,25 7 830 17 620 630 0,080
*Magn6-tium 648 1104 : 1,74 : 5 910 :10 270 :2080 0,352

TAMLEAU I Additifs poxx conbustibles liquides boueux

Doupoint de vue Anerg~tique, ii y a int~r~t I ajouter is plus possible de particules solides dane
Is liquids. Main pratiquement, cette tendince set freinfie pair

islsouci de conserver un produit fluids et injectable

- in capacitfi des particuies nolides A brdler dens is chambre de combustion (en particulier, pour
cells. de bore, ii faut qu'elles rencontrent suffisamment de zones chaudes issues de is combustion du
support liquids)
 5-9

A titre d'exemple, le tableau II donne lea caractfiristiques de queiques combustibles liquides. Sane
Itre exhaustif, car ceirtaines donnies ne sont pas accessibles dans la littdrature, il illustro la
recherche de comustibles A fort pouvoir calorifique volumique. On pout conatater en particulior quo
- l'auguentation du pouvoir calorif~que volumique oat mouvent lifie A une proportion Alevio d'atouos de
carbone

- pour lea applications a&roporties, on privillgie un bas point de coagglation

- pour lea applications our naviro, ce @ont plut8t lea probli~nes do sicuriti incendie qui guiJent 1. choix
(point d'iclair ilevg)

- cortains combustibles "boueux" permettent d'obtenir des pouvoirs calorifiques trim Alevis. Deux exemples
do coonproduita, objets de recherches en France gout indiquis. Dotls de fortes propriftls thixotropiques,
uls out une faible variation du curportement I l'iujectiou, doen un grand intervalle do tompirature.
3 iydrocarbures Corps do Synth~se Combustibles boueux I

: nO :JP5 sUtralin.: 3.34 :JP9 :CSD 1ST i AK 70-30 BX 5545 3

S : (on JPI)3 i*Shexans3 s a t

Poma.lation 3C 110R C
3C3 I
1
X 'C
C8 ~ 3U
C
20
C 2 H1
3
C1
'CO06'1C
1i4C 118:Anthracafto
1
i Bocs+karosan* s

.?ouvoir calorifique
3Loiaalour volunique 1 8180 3 324 1 9247 3 9323 :9456 310400 t 10300 3 12000
t (KCaI/dsj) 1 3 ,1 1 t 3 S

3 onsiti A 15%C 0,79 t 0,83 0,95 30,94 :0,94 t 1,04 3 1,06 3 1,25
:Viscostit cinmastique a9 17 3 11 :60 3 24 : 105 :Produita thixotropiqte. 3
a - 40*C (cat) I I I 3

.Point do congilation -53- <-46' -40- <-o0 <-54* 3 -60' <-40- <-.40

:Point d'tclair ('C) S 3 5 0 36' 24 a34 - a - a

3
I
3Applications 3Air Force: Navy a S Navy 'Air Force:' lionopirationnals failt llobJet3
a at 3- aviosas :Avions et: Non :-Taslo AI.CM 3*ssay~a on France sur : do racherches
SObservations :.alaaissatoparation:- SLCN turbor~actour et s on Franc.e
a I
3 3 ack tTomahawk i statortoctour3
3:carburant:carburant, C3RJ I0.1 I
tcboasique~clsosiqua: t I 1 ,03

TA~ELAU II : Qnfb~istibles liqudes &fort pouvoir calorifique volundque

4 -ALIMENfTATION, REGULATION V.TINJECTION DU COMIUSTIBLE

L'utilisatiqn do combustible liquide our missile conduit A des technologies particuliires, an raison
do coutraintes opgrationnolles spfcifiques, tellos quo

-Stockage do trsloued e( 10 ane), avec cycles thormiquos, sane modification do 1'fitat du

combustible ot !o l'6tauc-la2ititd du -rfservoir.

t.! ,!.ul4on surte doen toutes lea conditions de facteur do charge (jusqu'41 10 ou 20 g en
!ucombustible...
trnvr.) en 6vitant le ballottement du liquida ; en of tot, ii ne faut pas socouor lee 6quipomont3
eubarquga, ni surtout d6samorcer lea pompes ou lee vannos . avec 1. risque d'extinction du statorgactour
qui *'eu .uivrait. Cleat pourquoi ii ost prifirable de recourir A un dispornitif "actif" dlexpulsion du
combustible (pistoa ou vessie pressunisfie). W
-Systame do r~gul~ation capable dlss~urer une grande plage do variation do d~bit A titre indicatif, un
misaile devaut 6voluer dana uno gaiwne 6teudue d'altitude, avec des trajectoiros divorsififies, aura
pratiquoment bosomn d'une rigulation do dibit dana un rapport do 1 A 15 ou 1 A 20 pour couvrir Ilensemble
du domaine do vol.

Par ailloura, d'autros contraintee plus ginfirales gout A psroudre en compte

- recherche d'uue granie compacitfi , ce qui conduit A optimiser le systime do chasse du combustible et la
conception du riservoir,

- rtcberche d'uue riduction do la vuluirabilitfi du missile A des impacts do belles ou d'Aclats, ce qui pout
couduire A un fractiounement du rfisorvoir,

- aet...
5-10

4.1. SystAmes d'alimentation

11 existe de nombreuses techniques p~ur expulser le comsbustible du rfiservoir et alimenter le moteur.

Si lVon se limite A celles d~veloppfies ces derniAres ann~es, on peut dire qu'eiles se groupent en deux
families

*lea syst~mes de cheese A partir d'ur gag sous preesion

*lea syatAmes A turbo-pompe.

Ces deux techniques ont 6t d~vaolpp~es avec succas :schgmatiquement, la premiAre eat
particulibrement bien adapt~e aux problAmes de stockage de lonjue durge, tandis que Is sezonde eat
prAfhrabis pour les vol. de longue portfe.

IA* missiles lea plus anciens, Cu certains prototypes ont utilisS un riservoir d'air ou d'azote
comprim6 sous piusieurs centaines de bar. Ce gaz, aprAs passage dens un d~tendeur 6tait envoy6 dans un
reservoir A piston ou A vessie.

Maio cette techrnologie set mal adaptfe aux conditions opfirationneiles des longs vols A bases
altitude 0s que is port~e alaccroft, la masse et l'encombremsnt deviennent prohibitifs. CVest pourquoi
d
on utilise maintenant un gaz fourni par un g~n~rateur pyratechnique, soit A propergol solide, soit A ergois
liquides. Dans lee deux cas, lee gaz produits doivent 6tre suffisammerc froids pour ne pas d~grader lea
mat~riaux du piston ou de is vessie.

* La, figure 6 illustre uns telis technique : Les gaz "froidri' dfilivr~o par le g6ndreteur poussent un
piston qui chaoses is combustible vers les injecteurs. Cette cinf j.ration est bien adaptfie au stockage de
longus durge, car is combustibie peut Atre totalement isoiA et sans contact avec des mat~riaux organiques
ou plastiques, une membrane m~taiiique claqueble scellant complatement le r~servoir. 1l faut, bien s~r,
privoir A l'arriare du r~ssrvoir uns enceinte d'expansion pour assurer is iibre dilatation du combustible.
Cette forruie aupprime totaisment lea risques de ballottement ;main *Ise ss ncompatible avec
l'utilisation de r~servoirs multiples Cu de forms irrAguliAre.

Pis ton, Er"ens dl.xpansio

wsev-f
cwub -$bb-ItFi. Alirrentaticsn par cfndrateur de gaz et

d itanchists dorigulation

L'utilisation d'un g~nfirateur de gaz prAsente des aventages opfirationnels, mais ella conduit auasi A
certaines contraintes pour is rdguiation du d~bit. En effet, ce dernier sat fonction de l'6cart sotre ia
pression dane sler~servoir et cells qui r~gne dens la chambre do combustion. 11 exists donc

- uns pression minimale dene slerfservoir, suffisante pour une injection corrects loreque is pression
dens la chambre eat &lev~e (par exemple vol A grands vitesse A bases altitude)

- uns pression maximale &fin de limiter is. contraintes, ou lea dlformations, dane is structure du
rlservoir.
C'sst pourq'xoi, dens is cao de missions de longue durfie, comportant de grandes variations d'altl'.ude
Cu do vitesse, c'..st A dire de grandes variations ds d6bit de combustible, un tel syst~ms suppose
Sun fractionnement en plusisurs g~n~rateurs ds gaz fonctionnant successivemeot (limitation de is pression
insximals)

*l'utilisation d'injecteurs A section variable (voir paragraphs 4.3.1.)

4-1.2. yattes u ft=rbo.

Le transfert du combustible du rfiservoir vera lee injecteure peut 6tre assurl par une turbo-pompe.
Celle-ci aspire is combustible et fournit la pression n~ceesaire A une bonne injection. La turbine qui
sotratne I& pompe sat aliment~e par un prllavement dPair dens is manche (de l'ordre de 1% du dfibit d'air
qui traverse Is moteur), eon Achappement slfetun laipression atmosph~rique. On arrive ainsi A un
syst~us beaucoup plus lliger que dans is cas oilis pompe serait entratnle per uo moteur 6lsctriqus aiimentfi
par des batteries.

Dans cette formuis, le combustibls seatglnfralement atockS A l'intgrieur d'une vsssis soupie
celie-ci eat pressuriseAs laipreasion de l'air entrant dens Is chambre (fig. 7), cette pressurisation
permettant do "gaver" is turbo-pompe et sinai d'Aviter lee risques de cavitation.

Uoe tells configuration eat bien adapt~e aux missions de longus durls, Cu A l'utilisation de
 5-11

I&1mise au point de la turbo-pompe qui dolt, .kpris des annIem de stockage, faire preuve d'une grande
fiabilitS done des conditions opirstionnelles d'emploi trams variables, par exeaple en tempfrature.
I&lconceptirin du riservoir et le choix des matriraux permettant un stockege de longue dunfe. En
particulier, 1e zatiriau de Ia vessie dolt

- stourer une Stanchtirti parfaite au stockage

- no pas me d~grader au contact du combustible

- Stre apte A une pressurisation par de l'air chaud prilevi en amant des injecteurs (tempfirature
pouvant Stre mt-.pirieure A 400% sulvant la vitese. de vol)

S----------...--Vers
Fig. 7 t Alizmntaticin par turto-.poupe et vessie Scuple .96ow Iccvblo~ -

/ . Prise

Turbinei Var donn

air sguldt 'on

-avoirune bonne tanue micanique pour r~sister aux efforts dus aux facteurs de charge at nfianmoins .,

Itre suffisaaument souple pour expuleer presque tout le combustible (par exemple rendement
d'expulsion suprireur A 96%).

CWest pourquoi dea vessies mitalliques sont parfois utilisfies.

4.2. Principe de Ia r~itulaticen

LA systame do rigulation dolt fournir I chaque instant le dfibit de combustible. PoScesaaire au

programme do vol (rW. 4). 11 doiit donc, dona tout le domain. (caracthirioi par la variation du nombre de
Mach, de l'altitude, de l'incidanc~e, de la tempfirature de l'atmosph~re, ..
*assurer un fonctionnement correct du moteur ; pour cel& ie d~bit de combustible dolt toujourm Itre
coupris entre deux valeurs Unmitas
- une valeur maxirsale correspondent A la poussie maximale compatible avec la limtite d'extinction riche
at 1. fonctionnemont stable des entries d'air,

- une velour minimal. correspondant A la poussie minimale compatible avec Is limits d'extinction
pauvre. W

*persettra de rigler la vitesse du missile aux valoujrs demandiAe par le programme do vol, en fonction du
temps, do l'sltitude,

Un. chatne de r~gulation comprend donc une vanno do r~glage du dfibit, un ensemble de capteurs, unH
calculateur at un syst~m. dlasservisaement (figure 8).

La position do la vanno de r~gulation ost asseervie de fagon A co que la valeur d'un paramotro
caractiristique de ls poussie, calculie A partir de mesuros faites aur I. statoniacteur en fonctionnement,
soit &gale A Is,valour do commands.

11 faut noter quo Ia r~gulation d'un combustible liquid. pout me faire avec un temps do riponse
beaucoup plus bref quo celut d'un combustible gazeux (pour un atatofusie avec ginfirateur do gaz shpard, 11 -
faut on ef fet tenir compta de Is riponse du propergol et du volume du ginirateur) at donc pout Itre
utilisfie mnetsau cours do manoeuvres tras rapid., (mises on incidence brutale, par exemple). la moteur
pout alore fonctionner avoc une marge supercritique plus niduito at donc des performances optimales.

+ig dowa Stow Asevs 8 Schnna e ~glt

4.3. Inissile
Techol'ledsin
4.3.1. ote rs atu -i-.
 5-12

Celledc set fonction de la mission duamissile et de son domaine de vol. Par exemple. pour un
missile air-aol capable de voler A tras haute et tr~s bases altitude, ou d'un missile mer-mer avec croisiAro
A soyenne altitude, et acceleration au niveau de la mer pour des 6vasives finales, le ddbit de combustible
a~cessaire peut varier au moins dans ianrapport 1 A 15

Dan@ ce cas, la diffhrence de preasion A trevers ianinjecteur A section d'injection fixe devra
varier dane un rapport 1 A 225 (*). Or, pour bien pulv~riser iancombustible liquide, surtout s'il eat ianpen
visqueux (pour mernoire certains combustibles deases cant visqueux A froid'l, il faut une diffArence de
prossion minimale A tr~vers l'injec'.eur. Supposons que celle-ci colt de 1 bar A haute altitude, cette
ditffirence de pression devrait varier alors de I A 225 bars, ce qui correspond A basse altitude a une
pression d'injection trop Iflev~e, do l'ordre de 235 bars.

Cleat pourquoi, pour do telles missions, on utilise

soit plusisure circuits 6quip~s d'injecteure A section constants, coupfs suc-essivement lorsque le debit
diminue (figure 9).

-soit un soul circuit 6quipA d'injecteurs A section, ou A coefficient de debit, variable permettant
d'exploiter iaumieux lee possibilitfis du systAme d'alimentation.

Usa technologies mises en oeuvre sont trbs diverses : injecteurs A boisseau tournent, A boissesau
coulissant, A membrane, ...(figure 10).

En debars des problames classiques, de micanique, de lubrification et dlusinsge, I& misc saupoint de
tels injecteurs consiste a vfirifier quo l'homogdneiite duapanache et la qualitfi de la pulvfirisation se
v~uiservent malgrg ln variation

- de la section d'injection
- do 1a pression d'injection
- do 1iipression du courant d'air.
lnjsteuors i 2circudts

ksMrdCliamborl de combusta

let?2 Vannes de laminege A positionnement proportonel - )hcee oie __

3 Electrovanne do barrage
Circuits Iet 2 basse A moyenne altitude3)honsdorbl
Circuits 1 vanne 3 fermdfe :rnoyenne 6 haute altitude l
1) A ~JWisseau tournant
Fig. 9 z Injecticn icircuits s~ParS P'ig. 10 Injecteurs h sectioni variable

()Rappel : q - CD SV2A

q : debit massique
CD: coefficient de dfibit caracthrisant la geor-Atrie
D de l'injecteur
Sc:section d'injection
e : masse volumique diacombustible
4P : difference de pression A travers l'injecteu.e

4.3.2. Locelisation do 1'inJlection

Ua localisation precisc de cheque injecLeur set ianprobldae difficile et qui reste px beblement pour
de nombreuses annges, essentiellement empirigue . Cleat pourquoi la mise iau point complate dlun moteur
paa~rsd ~lg anndes
plusteurs s au banc, surtout si le domeine de vol eat Ctendu. En effe, le nombre de
lvnecessite
d'eoseis

- type d'injecteurs
- nombre et position
- orientation duajet do combustible : dens le sene duacourant, A contre-courant, etc.,..
Pour cheque configuration d'injection, 11 faut optiminer, dens toutes les conditions du vol (Mach,
altitude, incidenco/d~frapage, ..

- le rendemont do combustion
- lee limitec d'extinction riche et pauvre
-les portes de ch&rge internee
- lea instabilitge de combustion
 -' ~. -, - -. ,..L

5-13

5 CRAi(BRES DE COMBUSTION

La chambre de combustion d'un statorgacteur eat un des 6i6ments du missile doat lt dOveloppement eat
le plus long. En effet, tile doit assur2r un grand n-abre de fonctions, principalement our lea versions
moderns* oO l'accgihrateur eat intdgrG. Ces fonctions sont lee suivar.tea
"* assurer une combustion avec un bon rendement, tout en minimisant Ia trainfie adrod-Aamique duimissile
ceci conduit I une rfiduction des sections frontales, donc A des -itesses internee tras Glev6es
(fr~quemnent 100 A 150 matres/sec dens Ie? entrdes d'air et 300 matres/aec. dens Is chambre). CWest
pourquoi la fiamme doit Atre stabilisde par des isoyene perticuliers, mdcaniques ou ad):odynamiques.

"* contenir le propergol de l'accdigrateur ir.,40rS et par consfiquent lee contraintes qui r~sultent de son
fonctionnement :trappes assurant 1'6tanch6it6 A l'amont en phase acc~lfrde, tuy~re maintenant ls pression
do llacctitrateur (70 A 140 bars) puis P'6jectant en fin de s~quence.

"* contribuer I l'endurance duimoteur, aurtout pour lee vola de longue durne avec des condttions do
teupfrature at de preasion interne tr~e variables.

fl faut noter que ces fonctions ainsi que lea contraintes qui en risultent, sont pratiquement
Ind~pendantes de Is nature physique du combustible, qulil soit liquide ou gazeux (statofua~e).

5.1. Conception

La conception d'une chembre de combustion dfipend de la configuration adrodynamique duimissile et de

I& mission qui liiieat 6.saignge. D'une faqon gdn~rale, on peut distinguer

5.1.1. Les configurations avec accfl~rateur sdparA , celui-ci pouvant Itre montS en tandem, latfralement
ou coulissant dens la. chambre de combustion. Dena ce cas, Is chambre n'a pas A subir its contraintes d'un
acclfrateur inittgrfi :

-tile peut contenir des eccroche-flammes mdtalliques pour faciliter Isacomibustion i

- eli. fonctionne A basu. pression (inffrieure de 10 bars environ) at ne comports pas d'obturateura
Ajectables
-sa structure peet Stre optimiage par lee conditions d'endurance en Vol de croisiare, avec un choix
beaucoup plus large de technologies. Par exemple, on peut envisager des solutions oailee parois sont
refvoidies par film dPair prhlevg A l'amont, comme pour lee post-combustions des turbortacteurs

- la s~quence d'ellumage du statorgacti~ur n'est pas critique, puisque celui-ci peut Itre aliumE y compria
pendint I. fonctionnement de l'acc~lgrateitr

- ]a longueur de la chambre eat dfiter'.infe par l'obtention d'un boo rendement de .ombuction. Elle peut -

Stre plus coairte que pour lee versions A acc~idrateur intfigr6 (soit I matre au mAlimum).

Example do caicul d-! is longueur minimale d'une chambre de combustion (figure 11)

Considlrons un mfilange prdcarbur4 p~niftrant dens une chambre de combustion. LA longueut minimale do
la chambre eat calle A pertir de laquelie Ia combustion peut Stre consid~r~e comme compldte, avant que
l'jicoulement ne pin~tre done ia tuyare d'Ejection.

A partir du point 15,zone o1 Isaflaimme s'accroche, me d~tach. uL. front de flemma BA A la travers~e
duquel la combustion as produit ; si eli. 6tait instantanle, ii eat cicir que la combustion strait campl~te
I partir do A, point oO se encontrent ies fronts de flamme. Mais en rfialitg, ii faut tenir compte d'une
distance stippi~mentaire qui set fonction d'un temps caract~risant Ia dur~e des rfactions chimiques.

pint doflaMMO

d
S 20 Mach 3

0 Fig. 11 : ieigneur udiinizle d'une chenire de cvnbis-

LMach 2 L min (m) t

D.0,40m

La longueur minimale eat alora la somwe de ces deux termes

V.,,et V Etant lea vitessae dana la chembre, avant et apras combustion* Ua vitesse de
progressign du fdnt de flamme fitant

Vb -r V(P )('
 /t

5-14

Avon lam conditions de r~farences

Vr- 15 aSO pour Pr - 1 bar
attTr -288'K
tc, tempa caractaristique des rfactions chim~ques, eat donng par

tcX(P3 )0 ,25 x(T id)1,8 .50

La figure 11 montre que la longueur de is chambre de combustion eat d~ mace par l'altitude
maximal. de vol. Bien ontendu, d'autres configurations gdomatriques quo celles c. ctta figure permaettraient
do riduire cetto longueur si cela slavdrait nizessairo.
Cotte configuration eat calls de I* plupart des sisciles I statorhacteur concus dena le passa. Mais
il eat probable qu'olle roviendra dens l'avenir pour lea vole de longue durfie (plus de 10 minutes par
example).

5.1.2. Usa configurationc avec accilgrateur intagrf dnsn ia chambro, qut mont actuellemant lea plus
Itudifies pour lea missions de courte ou mayenne portge. an raison de leur excallente counscit6. Les
contraintes qui guident lour conception sont lee suivantes

Isllongueur est gAnAralement daterminae par 1. volume occupg par le bloc do propergol, cette condition
6tant dimensionnante, maine dons le cas oO le missile eat tirg depuis un avion rapide
I&lflas eat stabilis~e princtpalement par des moyens purement sarodynamiques, car 11 faut 6liminer toult
dispceitif accroche-flammes au contact avec lea gez du propergol. Macnaomi, at lee volumes de
recirculation sont insufftsants, i seatpossible d'amaliorrr la combustion par de petits accroche-flammes
situfm en amont des obturateure fjectables.

I&1technologio do I& protection thermigue get un 6l6ment daiterminant car lea parois sont soumises A des
* Icoulements de gaz A forte tempdrature et grande vitesse, entratnant des 6changes thermiques tmportants
de plus elle doit r~sister A des conditions trAs dtfffarentes

Done Is phase d'accalgration, lee gsa du propergol mont r~ducteurs, A des tempfiratures alevaes,
jusqu'a 3 600*K, I des pressions do l'ordre do 100 bare, mats avec do faibles viteeses A la parot et
des durfies do fonctionnement courtes (4 A 10 secondes environ).

Les gar du statordacteur sont' plut8t oxydants at A beaucoup plus faible temp~rature (2 000 K
au maximum, eauf dens les noysux de recirculation) at preasJon (inf~rieure ou 6galo A 10 bars
environ), mats les vitesses peuvont attoindre 300 A 350 aS at lee durdes do fonctionnement
pluseiurr centaines do secondes.
Bien ontendu, la protection thermique doit etre do faible Spaisseur af in de no pas raduire la
pouesee du mtatorfiacteur (en effet, A tuyare donnge, is poucsaie mnaximale diminue avec l'augmentation du
nombre do Mach dans le chambre).

Ces probl~mes techniques peuvont 8tre aaseez facilement raisolus pour lee missions de courte durde
avec des conditions do fonctionneinent rilativement peu variables (par example mission aol-air de dur~e
infarieuro ou 6gale 1 30 secondes). Par contre, at le vol dure plueteurs centainem do secondes, avec des
conditions tram variables (trajectoiies diversifi~es A basso ot haute altitude) Ie problAme deviant plus
difficile, d'autant plus qu'il eat parfois accentua par l& prdsence d'instabilit~s do combustiomn (voir
5.3.).

Lee solutions envisageables sont diversea, suivant 1a duras do Ia mission

5.1.2.1. Protections thermiques fonctionnant on r~gime tranuitotro il e'agit do mat~riaux

composites peu conducteurs qui pyrolisent et so caramisent progresuivement en surface. linefaible
conductivitf therinique permet non seulemont do conserver la paroi externe "froide", mais aussi de
riduire lo flux do chaleur qui panatre dans '.omat~riau, grAce A uno 6l6vation rapida de la
temp~raturo do paroi Tpiau contact avec la flammo

Q Ch tVC P(Ta -Tpi)

Ccoefficient do transfert de chaleur (do 0,001 A 0,005 suivant ia zonra de la

chambro do comtustion)

CV,Cp, Ta, caract~riatiques des gaz de omsin

Do plus, lee raactions d'ablation ottde pyrolyse qui se prodaisent on surface
refroidissent la paroi. Mais loreque le mat~riau eat enti~rement ablat6 ou clramisfi, 18
conductivitA du rfasdu est la seule propriat6 qui agisse sur le gradient thermique dans la paroi.
C'ost pourquoi, un tel procad6 eat limit6 en durfie tde i'ordre do 10 ou 15 minutes pour des
6paissoure inffirieures ou agales A 15 mm.

Deux types do produits ont donnA satisfaction

-rigidimArasA base de rfisines thermodurcissables (phfinolique) chargas de produits rifractaires et

atructuras avec un tras fort poujrcentago de tissue de ailice disposgs non parallAlement A la
parot. Cette atriicturation doit Stre telle qu'olle permetto lee dafor-mations radiales et
 777-----7~.---- M --. -,--

5-15

GlastozAres A base de r~sine silicone chargas en matfiriaux et fibres rfifractaires qui peuvent
Itre soulds in situ dana is chambre de combustion. Cette solution eat performante (car faible
conductivitt) et de mise en oeuvre en principe simple. Mais 1a couche pyroiysae 6tant plus
fragile, 1 eat parfois n6cesnaire 0 'armer Is protection thermique au moyen d'une armature
whtaliique ou en matariau composite (r~f~rence 5 at figure 12).
ArotondoJE dj*PS 14 Para-
mtattlqu (MM)
fS

-. *-%.~jj.
--- 7<~.- Protection
.- ~:--
thrniqu.
~tc*
(r~ne ig. 12 Gradient thermique dans 1' dpaisseur dI'
.

DC 92-104; protection thrniqi. isolante

r(IC) 5

5.1.2.2. Protections thermiques fonctionnant en rhgime permanent *Pour des tampa de vol encore plus
longs ii faudrait faire appal A des techniques de refroidissement, par convection forcfie do i'air, par .

rayonnement des parois ou ,localement, par circulation du combustible. Tous ces proc~d~o sont complexesa
ou co~teu~.

Un autre concept peut At re propos6 ii s'agit de rhalise7 is protection theratique an

Nat~riaux "thermostructurauc", c'eat A dire qui ne perdent pas leurs proprifitds oficaniques, m~ime A
haute temp~rature :ceci peut 6tre obtenu en associant une structure rdaistante (fibres de carbons, de ..

carbure de silicium, d'alumine, ... ) avec une matrice stable A haute tempfirature (cerbone, carbure de
siliciuA,Ialumine, ... ). t.avantage eac que 1'6paisseur du matfiriau eat faible et pratiquement
ind~pendante de 1a durfie de vol. Par contre, ia temp~rature 6levfe de Is paroi externe at is n~cessitA
d'6vacuer lee calories par tonvection ou par rayonnement compliquent l'amhnsgement du missile. A titre
d'exeuple. lo figure 13 indique le gradient thermique dana un matfiriau thermostructural, d'6patsseur 8 mm
at de conductivitg k - 10 W/m.!( ; deux missions sout stuvisaghes croisiAres A Mach 3,5 a1una altitude de
25 ks, at A Mach 2,6 A une altitude de 10 km.
Flu~xde Convection
2
ars mbwstion

Fig. 13 Gradient t1muiju dans la paroi d u-e chant-re OM Z.OKM

dobsinen inet~riaL thernostructural ~

R*Arodiwenent Refrokdisssmsnttpir
par convection setv es raymonnmuust(haute altitudte)

Uas hypothasee de calcul sont ies suivantes

Coefficient de transfert de chaleur

Ch - 0,0010 A i'extdrieur du moteur

- 0,0025 a l'lntdrieur du moteur

4
Emissivit6 de ia paroi externe 0,7
Richesse d~i foteur :0,6

5.2. Performances et mod~lisation

Uas perforinancso recherch~es sont lea suivantes

"abon fonctionnement dana une large plage de richesse ; celle-ci peut slier d'une valeur minimale de 0,2
(uanoeuvres A haute altitude, dicfil~ration) jusqu'A une limite riche souvent supdrieure au
atoechiom~trique (forte pouss~e pour accdlfrer ou monter A haute altitude, manoeuvres,..)

"* randeventde combustion (rapport du ddbit de combustible br-019 au dfibit injact6) suparieur A 0,9 en
phase de croisiA-re.

"* effica'-itSen pression (rapport de is pression totals au col A la preasion totale en fin de diffuseur)
sup~rixure A 0,85 malgr6 lea pertes de charge Ze combustion ou d'origine adrodynamique (obstacles divers
JI
dana leas manches Aa ir :injecteurs, stabilisateurs de flamina, .)

Cea performances ne sont acquises qu'au prix de nombreux essais effectuea our des maquettes en vraie
grandeur. Un tei d~veloppement eat coliteux en raison du grand nombre de paramAtres A faire varier. C'est
pourquoi ii eat prdf Arabia d'optimiaer A priori is gfiomtrie et le fonctionnement du fotaur grAce A
i'exploitation de sodales mathdmatiques approch~s. La dfimarzhe eat alors is suivante (rdf 6 at 7)
 5-16 8

5.2.1. Des essais do.,

visualisation, en 6coulement froid, sant r~alieaes au moyen d'une maquette
transparento ('Fig.14). Deux mfithodes sont employ~es

* mthodedu tunnel hydrodynamique : par injection de microbulles, on 3bservs l'6coulement dane la

chambre de combustion, afin de aituer lea zones de recirculation, lea d~collements, lea tourbillona.
L'utiiisation d'un traceur color6 et de techniques viddo permet d'obten1't des r~sultats quantitatife
volume des zones de recirculation, carburatio'n locale pour chaque position d'injection, etc...

authode en veins forc~e agrodynamique :grAce A des mesures diff4rentes (vitesse locale de
l'Icoulement d'air, concentration locale d'un traceur gezeux, ... ), ii eat possible de compldter le.
description qualitative et quantitative de i'Scouiement dana la chambre de combustion.

Fig. 14 Visuialisation hydrodynaniiqu de 1'6coule-

tnent dans tine chantre de conbustion

5.2.2. A pertit des r~sultats exp~rimentaux obtenus prAc~demment, une mod~lisation simplifide de la
combustion set effectufie, en consid~rant globalement lee zones principales de is chambre, par exemple
lee r~acteurs RI, R2 et la zone de jet J (Fig. 15). Pour chacune de cos zones, lea essais ont permis de
determiner le volume occup6 doen la chambre et les fractions de debit d'air et de dfibit de combustible
qut lts treversent.

Fa1 assisilant lea riectours Rl at R2 A des foyers homoganes, ii tat possible de prfivoir lea
limites de stsbilitb de ls combustior. done tout ls domains de vol (Fig. 16).e,

Bien quo Is m~thode ci-dessus soit sommaire et ns rends sans doute compte qu'imparfaitement des
ph~nou~nes, silo permet d'effectuer une optimisation paramAtrique de la configuration, ce qui facilite
et rlduit 1e coOt de is mnise au point de la chambre de combuction.

%30- Z(kin) m avec combustion

20 m

zone do riaction a

,zone do riaction porimaire R, 0 12 3

Fig. 15 bd~lisati-An d'une chanbre de contustion Fig. 16 Pr~1icticn du donnaine de stabilit6

5.3. InatabiiitLSo de combustion

Tout actsur thermique auquel on demtnde d'exceiiontes performances dane un domainete fonctionnement
trAs ftendu risque do rencontrer en queiques points do cs domains des ph~nomanes vibratoires gananta. Ce
n'est pas une particularitS du statordacteur, et V'on sait que do graves d~boires peuvent apparaftre sous
forms d'inatabilitfa de combustion, aussi bien our lea fusfiss A propergol solids que aur lea fusAes A
liquids. das quo l'on recherche des performances 6levesa (exemples :instabilitas du ler Atage de SATURNE V,
instabilitfis tangentieilss du moteur ARIANE, etc...)

Tours bonns combustion Atant assortie de vibrations, ii faut donc e'aaaurer quo colles-ci no
smpiifisnt peas de tolls sorte qu'slles aient des consfiquences Whastes, soit aur l'intigrit6 du moteur, on
'a
particulior do as protection thsrmique, soit sur is fonctionnement des prises d'air, soit sur is
comportement dss fiquipements.

Enfin, I'ecuitA de ces problames n'apparait parfois qu'au coura d'un d~veloppement, c'est A dire -
~
lorsaue io motaur sat test6 dane touten son eonndirinna do fnn.'tinnn~v%^ni nn5ratrninna1 .. earI. nArhmoA^
 7 7% -. rn -. 7 - - - ~ -

5-17

5.3.1. Tjeti d'instabilitis et consadquences

Sch~matiquement, il exist deux types d'instsbilitas de combustion dont lee spectres sont bien
d~finis ; elles produisent

Sdes vibrations de basse fr~iuence (100 ou quelques centaines de Hertz). 11 s'agit d'instabilitits quiL .
apparsissent sux limites thermodynamiques de la combustion et qui sont entretenues par des r~sonances,
moit des jets pAnatrant dans la chambre, soit de l'ensemble chambre de combustion + prises d'air. Elles
sant plus sensibles aux basses tempfiratures ganfiratrices.

Si leur niveau eat 61ev6, cOB instabilitfis de basse frfiquence risquent

- de fstiguer et de d6tfiriorer les strucruz.as et lea 6quipements du missile

- do provoquer des phinomanes de coupisge aSrodynamique avec ia prime d'sir at sinsi de limiter Is&-
pounsse usiimale (Fig. 17).
*des vibrations de haute fr~quence (1 000 et queiques mIlliers de Hertz). Ceile@-ci me. dicienchent pour ~
certaines piages ds richesse. Ces vibrations risquent de ddtiariorer is protection theruique et de faire
subir do fortes acc~lirations aux 6quipements embarquls.

prizeCrair(Pi2
Pic
Pi.g. 17 Liinitation de la pouss~e naxiniale par
suite dlinstabiliths de conbasticn -3e
teasse fr~qiunce

Mach

Application calcul des fr4quences des principaux modes acoustiques (figure 18)

La friquence F de chaque mode peut Itre estimfie A partir de la relation suivante, valable pour une
cavitAS cylindrique
F ~ 2 +(n2

a. vitesse du son dens la chambre de combustion

D. diamatre intdrieur
L, longueur
ktet n, coefficients dont lea valeurs pour lee principaux modes sont .ndiquges ei-dessous

kin

ler mode longitudinal 0 1

ler mode tangentiel 0,586 0
2A mode tangentiel 0,972 0
ler mode radial 1,220 0

LI L

----- 4

Modes tangentiels Fig. 18 Mxdes acaustiques transversaux

1"mods 2*mode 1"mods radial

Pressions Pp I

i.. t T - \s IsV

~V
 A titre d'exemple, syoc leg velours numlriques,
-I
a -750 uS- D - 0.40 a L3 - 1,203a L 1,10 a

leg frlquences des principaux modes mont alors

ler mode longitudinal (chambre seule) : F -340 Hz

"Basmen" frdquences( ler mode longitudinal (chombre + prises d'air):
( ~F - 160 lie
lot mod tangential F - 1100 Rez
"lisutte" fi'rquences( 24 mode tangeantial F - 1820 lRz
( br mode radial P - 2290 Hz
5.3.2. Origine-et remAdes

Comm la plupart dss phfnomanes inmtattonnairem, les tnmtabiiit~s de comubution des

- . tator~acteurs fonr. apel I des phdnomenes phymiques c.mplexes. Bien que des trav&ux approtondis sient
ftf entrepris d.'puis dlx ans, en particuiisr aux Etats Unis et en France, tim restent encore Mal
6lucidtm.

Les instabilirfs de combustion peuvent Atre engendr~es par diffirents mdcanismes agissant corn
source excitatrice at qui peuvent Itre amplififs par apport dl6nergie lif A la combustion et coup16
avec lem modem acoustiquem du moteur. Si com derniers peuvent Atro acme: facilement identifils, ii n'en
eat pas de mgme pour lee mh~anwases d'excitation et d'amplification :dfitachement Pdriodique dec
tourbillons do Von KOrmhii exist ant done lea prices dPair et I l'entrhse de la chambre, dietorsions
instables dans le diffusear ceaaprieso d'air, et aurtout comportement dynamique des zones de
recirculption.
1A complexitf de cev pLInomlnos rend Jiffl.ile le travail du mcientifique qui cherche 1
;* co-srendre et A prlvoir lee 1.nstabilitgo, ... mate ii favorise Ia oultipiicitt deis rex.~dcs que
i'inglnieur pout apporter expfrimentaleaont.

Munst, queiques mxemples concrete peuvent illumtrer l'efficacitG do cec remldem

1cr exemple : i'injection d'air auxtmommets d'un accroche-tlauae permet de repousmer is valour de la
richeese limit* pour laquelie les instabilitdm de haute frlquenco (ou "screech") apparaissont (r~t 8). Enm
effet, cette Injection dPair portor'~e Ia couch. cimaiill&. oa me forment lea tourbillons qui so dltschent
de laccroche-flaso. phlrcom~ne strodynamique pirio-dique qui, done ce coo, 6tait ia cause des
* Instebilitfs (Fig. 19).

I-
InjoclonI daw aux wwronotsI dun

Fig. 19 :Variatimr de ia richesse lintite d'ap-

ii - - ___ arition des instabilitds en fonctiori
* ~du ddeit Weair inJect6 (r6f. 8)
- 190-K

Usm. etemple i Ia manila dont Is combustibl, met injectio vaporieS et carbure lca diffirentem zones de
Is chambre do combustion, a un aM.e trim important. Munst I& figure 20 montre quo muivant le dlbit de
chaque injecteur, on pout faire apperattre ouadisparattre doemIntabilitfie de haute frlquence.

L'oxplication physique do coo phinceinee *it tanprobllme difficile at controvereg, bien que tras
Itudid 1ii somble qu'on agismant our *Linjection,on modifim Ia combustion done lea zones do
recir,:ulatinn, principal*@s ources d'dnergie acotastique (rft 9 at 10).
 N., 7

5-19

A pRipartlon do I'njection on % Woweedes

on A on 0 Instabilitisdo
haute friqumnce
I AP/P (on%)
830 70 0
40 60 062
A 50 50 8610
60 40 12630
Fig. 20 Apparitioni d'instabiiiths de conb~stion en
fancticn de la r~partiticei du &~bit W in-
* jection

Umie exempie %d'autres remades peuvent agir, non pe a en au~pprimant Is cause des instabilitfis, rnais en
modifiant la structure spatiale du champ de pression et les fraquences. V'est le cas des cloisons
longitudinales dispos~es A Ia paroi de ia chambre (Fig. 21) et qui sont susceptibles d'amortir certains
modes de vibration (modes tangentiels)-r~f 9. Ce remade est d'ailleurs bien connu des utilisateura de
J soteurs fus~es A ergols liquides et A propergol solids.

Cloiso,,s fongitudi e

(bar)
APh~

1- ~Same
dclsonjs

* Fig. 21 Ancrtisseirent des ixistabilit6.s de haute

A. frdquence par cloiscris longitxxinales

01000 2000

5.3.3.Quiusrsre

11 est indAniable que, dens Is cas de missions de longue dur~e, is problAme du risque
*d'apparition d'instabilit~s de combustion doit Stre aArieusement prig en compte. Clest une fitape
*importante done un d~veloppernent.

Chambre pe~r~e N.e

Fig. 22 Cons~quences des instabilit~s de haute

fr~quence stirune chaxrfire d3econrbustion
destatofusde

Toutefois, quelques rfiserves doivent inciter A la prudence le responsable d'un programme

a) il n'y a pas, A priori, de configuration "miracle" permettant A coup s~r de s'affranchir de ces
difficult~s. Des dchanges franco-smAricains faita en 1982 our des g~omfitries tr~o diverse. (une, deux,
trots et quatre entresa d'air) l'ont sontr6.
 5-20

noteur 4 entrfes d'air alimentf nsec du klrosAne, un combustible liquide dense, du propane 1 50%C et

also type d'instabiliti, avec des nivesux 6quivalents. Comme le montre I& figure 22, cee
instabilitAs peuvent parfoiB avoir des consiquences catastrophiques I

6 - INSTALIATIONS D'ESSAIS ET PRINCIPALES ETAPES D'UN DEVELOPPEMENr

Comme pour tout propulseur, l'Alaboration d'un moteur A statordacteur passe par des phases
successives, de mise au point dfitaillfie, puis de recette dans i.outes les conditions du vol, qui
nhcessitent des travaux exp~rimentaux importants.

Do is agme fagon que pour lea avions, Us tendance actuelle consists A qualifier le missile au
sol, done un environnement le plus rialiste possible, afin d'entreprendre les essais en vol, toujours
tram onireux, avec urisforte probabiliti de succis.

N C'est pourquoi la France s'est 6quipie depuis dix an. d'instsllations moderns. bien adaptfies
ce. diveloppements, I savoir
A

* * Centre d'essais de l'ONERA A Palaiseau et.a Modeanpour ise essais en conduits forceg et en jet
semi-libre de statoriacteur A combustible liquide ou de statofushe (Fig. 23). On y traits dea
% ~rechercheR, des diveloppements exploratoires et des diveloppements opfirationnels avec simulation de
toutes lee conditions de vol (vitesse, altitude, tempirature, conditionnement du combustible.
94 ~incidence. 6chauffement cinfitique, etc... ) (rif 11).
. * Installations de l'Afirospatiale (Bourges Le Subdray). De criation plus ricente, ce centre a des
-. capacitis voisines de ceux de l'ONERA. Bien 6quipfi pour lea problames de sficuritfi, il est plut8t ax6
Vers lee essais de qualification industrielle et de recette (r~f 6).
aCentre d'essais des propulseurs A Saclay * Cette installation tras puissante permet des essais en jet
libre d'un missile complet (slimlation de haute altitude).

Fig. 23 Essai d' ii stator~acteiur avec acc~lirateur

intkr6 en soufflerie (CNERA/M:ane)

-. LA d6la de diveloppement d'un missile opirationnel I statoriacteur n'est pee supirisur A celui
~.'d'un autre type do missile (I propulseur fusfie ou A turborgacteur). A titre d'exemple, voici ie
calendrier do diveloppement d'un mtissile franqais de ce type

to t Lancement du programm

to + I an I Dibut de le miss au point du moteur

*to + 2 an@ : ir essai en vol depuis une ramps terrestre

(*an*saccilirateur intigrA)

to + 3 An. : br eassi au banc du systime propulsif couplet

to + 5 an. o esusi anvol depuis avion

to + 8 ano : Miss an service opirationnel.

I FS Pour un tel programome (domaine de vol Atendu), is. principals. qualitis demandies aux
installations sont la disponibilitf et Is fiabilitfi . A titre indicatif, pour saettre au point le
* statorfiateur et le qualifier avec sea 6quipements dans touts. lee conditions de vol, il faudra 600
rafales par an pedn n. (902 des refales durent environ 30 secondes at 102 ont une dunfe
su ~insure) et 1'tiistin d'environ 80 tonnes de combustible liquids, tindiveloppement analogue avec
un statafungs A eindnataur cagarA devrait Stre nettemsnt nlua lourd (alffrfi des oerformances momns
 5-21

7 -CONCLUSION LIAVEKIR DlESSTATOREACTEURS A COMBUSTIBLE LIQUIDE

Laprogression de I'efficacitfi des d~fenses adverses at l~esouci de maintenir la plate-forme do

lancement A une 1Mine distance de s~curit6 de sa cible, laissant A penser que le stator~acteur A
combustiblc liquide a un avenir assurA (Fig. 24). Mats le spicialiste salt tras bien que ce n'est pas une
machiine aussi simple qua l~elaisserait supposer son schfima de Principe I Cleat pourquol. son utilisation
rester& tourn~e vers d d missiles A~ hautes performances et done relativement complexes.

IV*

i 4Fig. 24 MWssile Air-Sol sous Mirage

11 faut noter que ces techniques et technologies sont de mieux en mieux mattrisdes par lea
industrials :par example, en France, l'Agrospatiale qui a fait d'importants efforts intellectuals et
financiers A l'occasion du d~veloppement du missile ASMP, bient8t op~rationnel. dispose de nombreux
&touts pour entrcprendre avec le minimum de risques la rfialiastion des missiles future do ce type.

Or cette expArience acquise semble indiquer qua ce mode de propulsion ii'spas encore donno toutes
use possibilitfis. Cleat pouLt ol des recherches et ft-ides sont menfes pour mieux r~pondre aux exigences
du futur.

Ainsi d'ici la fin du sl~cle, on 'rerra apparattre des am~lioratlons des techniques et
technologies dlcrites dana le prfisent document. Parmi celles-ci

*nouveaux combustibles de plus en plus fnergftiques,

*amllloration de l'endurance des moteurs at des chambres de combustion (structures, protections

therisioues, systlmes de refroidissement), b~
*miss au point de modiles permettant de mieux prfivoir le fonctionnement des chambres de combustion, en
rfigime stationnaire at instationnaire, at ainsi de r~duire le noabre des essais au bsnc,
*am~lloration des performances at de la pr~cision des systames de rfigulation, compte tenu des program
des calculateurs at des aenseurs.
*g~om~trie variable des entr~es P'air et Eiventuellement de la tuy~re de sortie.
La figure 25 illustre ces variations g~omlitriques, en prenant pour example un missile capable d'une
croimllre A Mach 4 Ah~utealtitude. Pour aclr fnhentat ainsi r~duire la dur~e de la
pur ~duie l co:01tion en cos~e
n~cesaie 'piiaindspromne odi lr
A uilier
es ntres 'ai etuncol variables, ca qul correspond A une technologie 6ophistlquEe
at coilteuse. C'est pourquoi des solutions Lnterm~dialres, telles qua Ia variation de section des
entr~es d'air ceules, m~ritent d'Atre 6tudlles.

Bien sar, un changement trls important des techniques at technologies d~crites dans le pr~sent
document se produirait avec l'utilisetlon de la combustion supersonique pour propulser lee missiles A des
*viteases hypersoniques (mach 6 at plus). Mai* ce sulet, passionnant, eat trait6 par aillaurs (F. BILLIG).

A,
*CROISIEA
Q7
// 2
Limfitation
combustron
46 (Mach Mf3 %
Fig. 25 t Variations de g~cu~trie pernrettar.,. d 'optirdser
q5- ACCELERATIO les phases d'acc6l~ration et de cr.ois..bre
% ~Accilration
d4
0, s o,,6 07
AAct1D
 ~7 .. :",/ 7, 7

5-22

REFERENCES

(1) P. CARRIERE et R. MARGUET

Airodynamique interne des r~acteurs - 3ame partie : stator~acteurs
Cours A l'ENSAE (1975)

(2) CW. BURDETTE, HR. LANDER et J.R. Mc COY

High energy fuels for cruise missiles - AIAA 78-267

(3) I. PELEG et 1.M. TIMNAT

Investigation of slurry fuel performance for use in a ramjet combustor.
Israel Annual Conference on Aviation and Astronautics (Tel Aviv et Haifa 1982)

(4) M. RAVEL
Conception d'un statorgacteur A accAlArateur int~grA
AGARD CP 307 (Octobre 1981)

(5) W.E. ROBERTS, E.R. BOWMAN et S. WASSERBERG

Test and analysis of the ASALM-PTV insulated combustion chamber
ASME Paper 79 - ENAs - 21

"(6) P. BERTON et Ph CAZIN

Mhtbodes et moyens d'essai des stator6acteurs et statofus6es.
Comparaison des r6sultats obtenus au sol et en vol
AGARD CP 307 (1981)

(7) P. CALVET, A. GIOVANNINIT, P. HEBRARD, G. TOULOUSE

Quantitative interpretation of recirculated flow visualization by the analysis of video pictures.
International Symposium on flow visualization - Bochum (September 1980)

(8) C.L. BARKER

Experiment concerning the occuring and mecan.sm of high frequency combustion instability
Ph.D. Thesis - California Inst. of Tech. 1958
(9) A. LAVERDANT
Instabilitfs de combustion de haute fr~quence dana lea foyers de stator~acteurs.
Etudes documentaire ot expgrimentale. Document ONERA non publi6

(10) R.C. WAUGH, R.S. BROWN, T.S. HOOD, G.A. FLANDRO, G.C. OATES,
F.H. REARDON
Ramjet combustor instability investigation literature survey and preliminary design study
UTC/CSD - 2770 - IR - 1 (1982)

(11) P. BERTON et D. REGARD

Installations d'essais de statorGacteurs A I'ONERA
T.P. ONERA n*1980-132.
 6-1

SPECIAL PROBLEMS OF RAMJET WITH SOLID FUEL

by
T D Myers
United Technologies/Chemical Systems Division
P.O. Box S001S
San Jose, California 95150-0015

SUMMARY
Ramjet propulsion systems that use a solid fuel have been of interest to missile .
design engineers since the lq30's. The operational simplicity of the solid fuel ramjet
is of great interest since it does not require fuel tanks, fuel pumping devices or fuel
controls to operate. The potential for a highly relible and storable ramjet propulsion
system exists at a cost only slightly higher than a solid rocket motor. The solid fuel
ramjet can provide specific impulse in the 900 - 1000 second range, resulting in typical-
ly a 200 - 400 percent range increase over a comparable siue and weight solid rocket
motor.

The "self-throttling" characteristics of the solid fuel ramjet permits high perfor-
mance operation from sea level to high altitude conditions. Thus, early beliefs that the
solid fuel ramjet would only be used for constant altitude and speed missions have proved
to be erroneous as a greater understanding of the solid fuel ramjet engine operation was
developed.

Another unanticipated benefit of the solid fuel ramjet engine that evolved during 10
ears of development tests at the Chemical Systems Division was the high degree of com-
ustion stability experienced with a wide variety of combustion types and sizes. The
basic diffusion-controlled solid fuel ramjet combustion process results in a distributed
energy release throughout the combustion. This uniform energy release in the solid fuel
ramjet engine is believed to be the basic reason why no combustion instability problems
have been encountered in over 2500 combustion test firings with many different combustion
configurations.

Given all of these attributes of the solid fuel ramjet it is interesting to consider -.-
why the solid fuel ramjet engine has not yet been selected for an operational missile.
The probable reason is a lack of understanding of the special problems of ramjets with
solid fuel that cause the potential missile developer to have concerns about its satis-
4.:
factory operation over a full flight envelope and with the "off-design" effects that
result from enviro7.mental conditions and manufacturing tolerances. Unfortunately, the
basic operational simplicity of the solid fuel ramjet conversely requires a relatively
complex analysis to select a grain design for a given set of mission requirements.
The spjcial problems that are associated with ramjets using a solid fuel include: - -

a Selection of a fuel type.

o Flameholding requirements that limit maximum fuel lcading
* Fuel regression rate behavior as a function of flight speed and altitude
o Diffusion controlled combustion process that requires special mixing section
* Inlet/combustor matching.

Considerable insight has been acquired with these special solid fuel ramjet design .,-
problems over the past 15 years. Over 2500 ground test firings have been successfully
conducted that have included direct connect, semi-freejet and freejet tests. Moreover,
in the last few years several hundred experimental flight tests have been suzcessfully
conducted that confirm the operational flight characteristics of the solid fuel ramjet.
In this lecture we will briefly examine each of these special design problems associated
with the solid fuel ramjet.

SOLID FUEL RAMJET TYPES

Two types of solid fuel ramjet engines have evolved in the course of recent develop-
ment testing. These two basic types are shown in figure 1. The zop schematic shows the
non-bypass combustor type with a special vane mixing device installed aft of the solid
fuel grain. In this solid fuel engine all of the inlet air flows through the solid fuel
grain. The lower schematic nhows a bypass combustor where a fraction of the inlet air
is bypassed to the aft mixing section where combustion of the fuel-rich combustion gases
from the fuel grain section is completed. Bypass air ratios from 2S% to 801 are typi-
cally used to match a specific set of mission requirements.

Both of these solid fuel ramjet covibustor types can be used with single or multiple
inlets, using an air mixing manifold to provide a uniform air flow to the combustor.

Because of the special operating featu es of the solid fuel ramjet, additional
station locations have been added to the standard ramjet stations as shown in figure 2.
These special solid fuel ramjet engine station designations include:

o 2a -entrance to inlet air manifold

a 2b - entrance to air flow injector
 r7-

6-2

ThLese special station designations are used to help analyze the solid fuel ramjet engine
combustion efficiency and pressure losses.

There are a wide va-iety of solid fuel binder materials that have been investigated
to date. Depending on the mission requirements, the best gravimetric or volumetric heat
release fuel binder can be selected. One of the best performing solid fuels developed
to date consists of a blend of PB and polystyrene, providing high gravimetric heat release,
good mechanical properties, good regression characteristics, and high combustion efficiency
over a wide range of conditions.

It is possible to increase both the gravimetric heat release and the volumetric heat
release of the solid fuels by use of metal additives. The boron family of metal additives
offer the highest potential increases in both gravimetric heat release and vclumetric
heat release. Aluminum, boron and boron carbide are all attractive additives for solid
fuels.

With the solid fuel ramjet casting the metal fuel dir.-tly in the combustor case
eliminates the difficult problen associated with storing, pumping, and injecting liquid
slurry fuels. Attaining high combustion efficiency with the boron fuels, however, presents
a challenging design problem.

SOLID FUEL REGRESSION RATE

The fuel flow rate in the solid fuel ramjet engine is a function of the surface
Prea of the grain, the regression rate of the particular solid fuel formulation, an! the
:light speed and altitude of the missile. Basically, heat from the combustion gases cause
the temperature of the fuel grain to increase to a point where the fuel is vaporized. The
hot, vaporized fuel then diffuses in the boundary layer until a combustible mixture of
oxygen from the air stream is established. A diffusion flame within the boundary layer is
established and steady state combustion is released.

The basi' model of the regression rate for solid fuels is shown in figure 3. The
"solid fuel grain is heated by convective and radiative heat transfer from the diffusion
flame. The vaporized fuel diffuses from the grain surface into the boundary layer. The
convective heat transfer component is dependent on the air mass flux through the fuel
grain port, the air temperature, and air pressure. Thus at low altitude where the air
mass flux is high, a large convective heat transfer takes place, causing a high fuel re-
gression rate. Conversely, at high altitudes the air mass flux is low and the solid fuel
regression rate is lower. While the resulting fuel-to-air ratio is not perfect, this
self-throttling feature of the solid fuel permits good performance over a wide range of
altitudes and flight speeds.
The radiative heat transfer function is typically lower than the convective heat
transfer, but is significant. Radiation from carbon particles, water vapor, and carbon
monoxide from the combustion products are the primary contributors to the radiative heat
flux.
An approximate expression for the convective heat transfer function is shown in
figure 3 where:
x Y Z
"convective - G T D"
The strong dependence of convective heat flux to the fuel port mast flux (G) is the major
contributor to the self-throttling characteristic of the solid fuel ramjet.

The resulting solid fuel regression rate is basically then a balance between the heat
transfer to the solid fuel grain and the heat required to vaporize the fuel:

rfuelG,AQ, hv, P4

here: G - fuel port mass flux

AQ - heat of combustion
hv . fuel. heat of vaporization
Pf - fuel density

A typical effect of flight altitude on the fuel flow in a solid fuel ramjet engine is
shown in figure 4. Here the resulting equivalence ratio (*) that occurs as a function of
flight altitude at a constant speed climb is shown. The inlet air flow rate drops and
the air temperature increases as the missile climbs. The equivalence ratio varies from
.a - 0.95 at 30K ft. to * - 1.15 at 40K ft. While the fuel flow is still higher than
desired at high altitude, the "self-throttling" characteristic of the solid fuel ramjet
reduced the fuel flow rate to a sufficiently low value that a significant range was
achieved.

The relative etfect of port size and axial position on fuel regression rate is seen
in figure 5. In this figure the relative fuel regression rate along the length of the
grain is plotted for a 2.5-in.-diameter motor. The experimental values were then com-
pared to theoretical regression rate, with and without considering air acceleration
effects;. The experimental data fell approximatelv eouallv between the theoretically

* . . . . *, /
 6-3
2SO tests with engine sizes, ranging from 2.S in. to -in.-diameter..-..

SOLID FUEL RAMJET PLAEHOLDING/COMBUSTION EFFICIENCY

The solid fuel ramjet engine uses a rearward facing step to produce a recirculation
zone that forms the basic flame stabilization region. This flame stabilization region
is depicted in figure 6. A critical step height, h, is required for combustion to occur
in a solid fuel ramjet engine. The required value of the step height is a function of ,-.
the inlet air mass flow and temperature. Each particular solid fuel formulation requires
a minimum step height for sustained combustion to occur. The importance of the minimum
step height limitation ultimately limits the maximum fuel grain loading and thus range
in a solid fuel ramjet engine. Thus it is highly desirable to minimize the required A'.
step height.
Studies at CSD have shown that combustion in the SFRJ is 1overned by the degree of
mixing between the fuel and air. Within the combustion boundary layer along the grain,
the mixing rate is controlled by turbulent diffusion and there exists a natural degrtZ of
separation between the fuel and air. Calculations of mixing within the fuel port based
on turbulent boundary layer theory showed that only about 5Su of the fuel had mixed and
burned at the chamber exit plane. The calculations indicated the importance of additional
mixing aft of the grain and are in general agreement with combustion efficiencies
observed in motors with low L/D mixers.

Analysis showed that the parameters controlling combustion efficiency are the mixer i-q
L/D, equivalence ratio, and port-to-injector area ratio. Equivalence ratio is important
because it determines the relative thickness of the fuel-rich layer which must be mixed;
and port-to-injector area ratio controls the turbulence level introduced at the injector, .,-
thereby influencing mixing throughout the combustor. Correlations of combustion efficiency
in terms of these three parameters has been successful.

Early attempts to evaluate SFRJ combustion efficiency at low pressure indicated that
a reduction in efficiency could be expected at pressures below 25 psia. This conclusion h..
was based on seven tests at pressures from 7.5 to 16.4 psia. More recently, CSD has
co nducted a series of 13 tests at pressures ranging from 7.6 to 35 psia and found that
combustion efficiency does not degrade at pressures down to 12 psia. At pressures below
12 psia, some of the individual test points indicate a lower combustion efficiency.
The combustion efficiency of the solid fuel ramjet engine has been correlated for
both circular and spoked type grains, ranging from 2.5-in.-diameter to 10-in.-diameter.
The correlating factor, designated burned factor (BF) correlates the combustion efficiency
as a function of the equivalence ratio (C), L/D of the engine, and fuel port-to-air
injector (A3/Ai) ratio. Good agreement between combustion efficiencies predicted by
the burner factor (BF) and a large number of combustion test firings and configurations .. _
has been found.

The effect of combustor size on combustion efficiency has been studied. The highest
efficiencies are achieved with the smaller 2.5-in.-diameter combustors with decreasing
efficiency levels at the higher motor diameters. The combustion efficiency increases
from approximately 701 for a * - 1 to a value of approximately 85% for a value of 4 - 0.5.
This trend is typical for all motor sizes and is probably due to mixing limitations since
perfect mixing becomes more important as * approaches a value of

Thus combustor designs and devices that promote mixing tend to improve combustion
efficiency in the solid fuel ramjet. in nonbypass combustors the use of special vaned
mixers at the aft end of the solid fuel grain are effective in increasing overall engine F
performance. In the bypass combustor radial injection of the bypass air in the secondary
combustor promotes improved mixing and thus increases combustion efficiency.

SOLID FUEL RAN&JET INLET-COMBUSTOR MATCHING

Matching the inlet and combustor in a solid fuel ramjet engine requires special
attention since not only inlet/combustor pressure matching must be analyzed but inlet
flow distortion effects could potentially result in nonuniform fuel regression. To study
these effects the inlet-combustor simulator shown in figure 7 was fabricated. The
objective was to establish a baseline engine performance base for a well-stirred
combustor. The design permitted installation of turbulence screens at the inlet-com-
bustor dump to also evaluate the potential effect of local turbulence on the SFRJ
combustor performance. Having established the baseline SFRJ performance, test firings "
were conducted with: (1) a single side-mounted inlet dumping into the plenum chamber,
(2) dual side-mounted inlets at 1800 dumping into the plenum chamber, (3) the single
side-mounted inlet in combination with a special tube-in-hole injector in the dump plane
and (4) a screen at the inlet dump to increase the turbulence level. The effects of the
various inlet-combustor combinations on SFRJ operating limits are shown in figure 8.
The single inlet required a much larger forward dump area and lower combustor Mach number
that would limit maximum range and engine thrust, respectively. The dual inlet results
showed that only small maximum range and thrust penalties would result. Incorporation
of the tube-in-hole air injector reduced the effect of the single inlet on performance,
resulting in only a small range and thrust degradation. The effect of increasing
turbulence level on fuel regression characteristics was also very small, causing a -
minor change in the flameholding limits.
 /

%777

6-4

The results of these tests showed that with careful design of the inlet air dump to
the combustor, using devices such as the tube-in-hole injector that satisfactory inlet-
combustor matching was possible. Thus the inlet .ould be designed to best fit the missile
envelope and performance requirements.

The pressure loss characteristics of various inlet air injectors is in Higure 9,

using the dual inlet configuration as a baseline. Total pressure ratio through the inlet-
combustor section is plotted as a function of the inlet dump Mach number. The tube-in-
hole injector shown schematically in the left hand figure has the smallest pressure loss
of all the types tested. Over the typical range of inlet dump Mach numbers between 0.2
and 0.3 the tube-in-hole injector provides inlet total pressure ratios in excess of
90%. Fortunately, the device that proved best for uniform fuel regression in the solid
fuel ramjet also provides the lowest pressure loss, superior even to a straight orifice.

It is generally desirable to design the ramjet engine so that the inlet is operating
with a small supercritical pressure margin to ensure stable inlet operation aad avoid
inlet air spillage drag losses. The liquid fuel ramjet engine can operate with a very
close supercritical margin control (3-5%) by varying the fuel flow rate around a closed
loop inlet pressure controller. The SFRJ, however, has a fixed grain geometry and the
fuel flow rate isdetermined uniquely by the missile flight conditions, i.e., altitude,
speed, and angle of attack. Therefore, the SFRJ must use either a higher supercritical
inlet margin and/or select an inlet design that will operate stably at inlet subcritical
flow conditions. Both of these choices result in some range penalty. The repeatability
of both the SFRJ fuel regression iate and combustion efficiency determine the required
inlet matching condition. Based on current state-of-the art it is recommended to either
design the solid fuel ramjet engine with a 10% supercritical inlet margin or use an inlet
with a stable subcritical operating capability.

The bypass SFRJ combustor presents a special problem for matching the inlet to the
combustor. This occurs because ei.ther separate inlets or a diffuser bleed system is
used to introduce the inlet air both upstream and downstream of the SFRJ fuel grain. The
inlet pressure upstream of fuel grain is higher than the downstream pressure since the
section. Thus speciul attention must be given to match both the forward and aft inlets

to avoid coupling as well as to match the inlet air pressure with the SFRJ primary and
secondary combustor sections.

A typical solution to the forward and aft inlet design matching is as follows.
In this application the forward inlet used more inlet flow turning compared to the aft
inlet to achieve the desired pressure match. The forward inlet in this manner provided
a higher pressure recoveiy to compensate for the higher dump and Rayleigh losses that
occurred in the fuel grain sectionc. There was no flow coupling between the inlets in
any of the installed inlet wind tunnel or freejet tests of the solid fuel ramjet engine.
CONCLUSIONS
The functional simplicity of the solid fuel ramjet combined with high performance
makes this engine type quite attractive for many future tactical missions. Conversely,
the complexity of the analysis of the fuel regression behavior, combustion efficiency,
and inlet-combustion matching has caused vehicle designers concern over the ability
of the solid fuel ramjet to be able to operate satisfactorily at all the off design
conditions required of a tactical missile propulsion system.

A large engineering data base has been developed for the solid fuel ramjet over the
past 15 years that provides answers to all of the special design problems of the solid
fuel ramjet. This technology base includes ever 2500 ground test firings as iw.ll as
several hundred successful flight tests. This broad technology base provides confirmed
solutions to the special design problems of the solid fuel ramjet.

\ %
 6-5

Nonbypass with vane mixing device

Aft bypass Inlet

r1

Forward Inlet Mixer plate

Bypass with simple orifice mixing device

Figu~re 1. SFRJ Comibustor Config-iratiofls V8

V08113

oFLAME STABILIZA~TION

'FUEL REGRESS ION

*COMBIUSTOR PERFORMNCE
*COMBUSTION EFFICIENCY
*FUELS
2c0

Section '-Cham'ber

Figure 2. Solid Fuel Ramjet Combustor Phenomena

6-6

-- Engine~j-____

Air Port diameter

Boundary layer edge I

w - f l a m e~ f

Fuel regression rate

6cenv ` convective heat transfer
0,con +0ra
0
rad = radilative heat transier
G fuel port mass flux
Ocnoc xTYD'L TS air temperature
D =port diameter
Orado:a,.P, 5T a =constant of proportionality
P =combustor pressure
T, flame temperature
6=combustion zone thickness
Figure 3. R~egression Rate Model for SFRJ Fuel
33210

60~

404

204

80

0. 1.0 1.5 2.0L

Equivalence ratio, 0

Fig~ir4.Effect of Flight on Fuel Flow

33211
 6-7
2.5

2.0 Theoiy (including

air acceleration) ~

.3 *%%

C0

'0' 0 Theory (neglecting air acceleration)

cc

0.5

0I I I
0 1 2 3 4 5 6 7 8 9
xID)
P~
Figure S. Regression Rate Distribution in 2 1/2-in.-Diameter Motor
33212

-sp Inict air velocity profile

Rearward-facing ,.~Shear layer

an regi~b ton Combustion gas

Fuel plus products -velocity profile
ep Recirculation zone

Vaporized fuel
Solid fuel grain

Schematic representation of SFRJ combustionL

Photograph of two-dimensional Plexilgas combustion

Figure 6. Flow Field in the Flame Stabilization Region h.

V06146-Rl
6-8

Igniterlinjector assembly
Flow settling chamberiin scto

Fuel

ScreenTurbulence screen

Stainless steel honeycomb

h~
Figure 7. Inlet-Combustor Simulator
V08563

0 Combustlon sustoined
0 0 Not sustained
Flow A3AIa *\ o
vilaQCW lchamberH-----
A3IA&(M3)
Tests 1 through 9, 19 through 21

A.-JAG (M3) A31A. (M3)

Tests 10 through 12 - Single Inlet Tests 16 through 18 - Tube-In-hole Injector

S..dflowI trbulence
A31AI chamber njector A3IA.

A31As (M3) A31 AG (M3)

Tests 13 through 15 - Dual Inlet Tests 19 through 21

Figure 8. Inlet-Ccuibustor Flaiieholding Studies

V08441
 6-9 IL
0

1.0 ~

0.9

0.8
A-

0.7

*.A&
/'A 0.6
12b 3b
11Symbol Injoctor

0 ~ Tube in-hole orifice

0.3 0 Sleeve
0 Orif ice
0.2

0.1F
0 0.1 0.2 0.3 0.4 0.5
Inlet dump Mach number, M2.

Figure 9 .Effect of Inlet Configuration on Pressure Loss .. -

33216 *.
 S7-1
.1

SSOLID PROPELLANT RAMROCKETS 7"

Hans-L. Besser
Messerschmitt-Bolkow-Blohm GmbH
Unternehmensbereich Apparate
Postfach 80 11 49
8000 MUnchen 80

SUMMARY
The solid propellant ramrocket (SPR) is very suitable for military applications because of Its high values
uf performance and reliability. The capability of fuel flow modulation, however, is a precondition for
various applications. The subject of burn rate modulation has been intensively investigated for SPR prop-
ellants and presently can be considered to be state of the art.

The lecture introduces basic design principles of the SPR. The main components of the SPR propulsion system
as propellant, fuel flow control, control valve and ram-combustor are discussed in detail.
Finally the lecture reviews the presiht-state of technology and dicusses possiblefuture applications for
the SPR.
LIST OF SYMBOLS

A Area r Burn rate

c* Characteristic velocity R Gas constant

AH Heat of Combustion T Temperature

6Hf Heat of Formation Ratio of specific heats
% K K-ratio Efficiency

n Pressure exponent 15? Constant pressure coefficient for temperature

p Pressure sensitivity of burn rate

Dntnsity
SUBSCr!PTS

b Burning

c Combustion
t Throat

.1 INTRODUCTION
The solid propellant ramrocket (SPR), also knuwn as ducted rocket, is an airbreathing propulsion system
which belongs to the family of ramjets. The origins of the SPR date back to the 1950's. A considerable
effort was made in the USA between the mid 1950's and the late 1960's to investigate this propulsion system.
Even flight demonstration programs were included in these activities. This was followed b a phase of lower
activities in the 1970's.
In the USSR, SPR development was probably carried out since the late 1950's leading to the propulsion module
of the SAM 6 "Gainful* antiaircraft missile which became operational in 1967. Technological and application
orientated development work in the field of solid propellant raiarockets has been carried out in France and
especially in Germany since the end of the 1960's.

Since several years SPR development activities are stimulated in the USA and Europe due to the growing inter-
est for the application of this propulsion system to tactical missiles of the next generation. Its high
volumetric impulse and thrust density make the SPR a preferable candidate to meet high velocity, range and
maneuverability requirements even at low altitude flight as well as volume limitations being frequently
imposed on missile design.
During recent development programs, flight demonstrations of SPR propulsions systems were carried out by
the French ONERA in 1976 (missile outer diameter 0.4 m/16 in) and the German aerospace company Messersrhmitt-
Bdlkow-Blotim GmbH (MBB) in 1981 (missile outer dianeter 0.24 m/9.5 in). By now, a high-energy SPR propuision
system is under development at MBB for application in a joint German-French anti-ship missile project. In
the USA, the DRED-program (Ducted Rocket Engine Development) has been conducted since several years, A
flight demonstration, originally scheduled for 1981 has been delayed. Nev!ertheiess, a follow-on program
has already been initiated in 1582.
 7-2

* 2. CHARACTERISTICS OF THE SOLID PROPELLANT RAMROCKET

The basic design of a solid propellant ramrocket is illustrated by fig. 2.1. For comparison, sketches of a
conventional liquid fuel ramjet (CRJ) and a solid fuel ramjet (SFRJ) are included in this figure. An oxygen
deficient solid propellant burns within the primary combustion chamber (gasgenerator). The fuel-rich
combustion products are exhausted into the secondary combustion chamber (rcncombustor). There they mix and
afterburn with the ram-air supplied by the air intakes.

The precombustion of the fuel is the main characteristic of the SPR. The oxidizer incorporated in the
/propellant of course reduces the specific impulse compared with a system using the corresponding pure
fuel. This somewhat hypothetical disadvantage is more than compensated by various advantageous character-
istics which are highly favorable to the application of a SPR.

Performance:

o The SPR allows the incorporation of great portions of high energy/high density ingredients such as
carbon, aluminum, magnesium, boron etc. in the propellant, the application of which is not practicable
in CRJ's and SFRJ's at the present state of technology.
CFR with slurries: pumping and combustion performance problems
"SFRJ with great portions of high energy ingredients: combustion performance problems
o The high temperature of the combustible products injected into the ram-com.bustor
- allows to achieve a high combustion efficiency for high energy fuel ingredients which are difficult
to burn, e.g. boron.
- reduces or eliminates flame-out problems and may eliminate the need for an additional device to
ignite ram-combustion
- makes the SPR fairly insensitive to cold environment operation (where problems may be anticipated for
CRJ's using high-density hydrocarbon fuels)

o The SPR has a potential to allow higher maximum thrust levels than it is possible using CRJ's or SFRJ's.
This is due to the higher temperatures achieved at stoichiometric combustion as it is illustrated by
fig. 2.2, which compares curves of combustion temperature vs. equivalence ratio (air : fuel) for
different fuels and SPR propellants.
o The SPR has a fuel flow modulation capability by using a solid propeliant with a pressure sensitive
"burn rate. This is very important to fulfil the requirements of a wide mission envelope (e.g. flight
at different altitudes, different flight velocities, maneuvering). Obviously achievable turndown ratios
(TDR) are lower than using a CRJ. Nevertheless, TDR's of SPR propellants up to 18 were demonstrated,
* while values between 5 and 10 should be considered state of the art for medium to hie-h energy gasgenerator
propellants.

Design:
o The SPR is very suitable for volume limited propulsion modules, due to it. capability to incorporate
ingredients of high density (i.e. high volumetric heating value) in the solid propellant.

o Especially the SPR with fixed fuel flow offers a very simple design without movable parts. The simplicity
of design is comparable to a solid rocket or a SFRJ.
The SPR with variable fuel flow customarily uses a pressure sensitive propellant and a valve for the
variation of the gasgenerator throat. The basic design complexity of this control valve, being the only
movable part of the variable flow SPR, is more adequately comparable to a fin actuator being present in
every modern missile, than to the fuel-feeding devices of a CRJ. Nevertheless, this should mainly be
understood to be an argument for the favorable storage characteristics of a SPR system. High gasgenerator
temperatures and particle-laden gas flow may induce sophisticated design and the use of exotic (thermally
resistant) materifls for the control valve.

o The SPR has the capability to accommodate an integrated booster in th? ram-combustor in the same way Ps
the CRJ.

Logistics:

o Storage and maintenance requirements and long duration storage characteristics of the SPR are corresponding
to the solid propellant rockets being presently operational.
"o The SPR induces no additional t ire hazard for storage and handling by leaking fuel. This is especially
"important for naval weapon systems.
o SPR systems having a low optical signature (a requirement sometimes stated for air-to-air missiles) are
feasible. This requirement, however, puts some limits to the performance level that can be attained with
* such systens.

3. DESIGN PRINCIPIES OF THE SPR

All major components of a SPR are indicated in the schematic view given by fig. 3.1.
"Main design aspects for these components
- propellant formulation
 7-3

are discussed in detail in the following paragraphs. The subjects of air intakes and structural design are
not specially treated in this lecture since both are corresponding to customary missile or ramjet design,
respectively. Where the design of a SPR component requires the application of special material, this item
will be included in the discussion of that component. The integrated booster and the boost sustain transi-
tion phase are treated very shortly, because the inherent problems are commrion to all ram-type propulsion
systems.

3.1 PROPELLANT FORMULATION [ 2, 3, 4, 5]

A complex interrelationship exists between the propellant formulatio, and the performance and characteristics
of the SPR propulsion system. Fig. 3.2 illustrates the influence of the propellant formulation on parameters
which directly or indirectly affect the figures of merit usually required for a propulsion system.

Main performance related parameters directly depending on the propellant formulation will be discussed in
detail. These parameters are indicated in fig. 3.2 by bold frame lines. Additionally, indications will be
given where the selection of propellant ingredients has a major influence on cost, aging or handling charact-
eristics and optical signature-of the propulsion system. Two further parameters are discussed which are not
performance related, but which are indispensable to consider in propellant design. These are
- processability which must allow a reproducible industrial product
- mechanical properties which must be compatible with the selected grain configuration and anticipated
environmental loads.

3.1.1 Basic Aspects

Usually the main requirements established for a propulsion module to be designed for a certain missile
application are concerned with
- range
- mission envelope (flight trajectories, maneuverability)
- missile handling
- missile signature
- design restrictions (length, weight)
- system cost

The range requirement essentially means a required energy content of the propellant while the mission
envelope determines the required burn-rate level and turndown ratio for a chosen propellant having a fixed
heating value. Besides these two main items, various system related aspects have to be considered in prop-
ellant formulation selection to achieve an overall optimum SPR propulsion system [ 2 ).

A high SPR performance requires:

o A volumetric heating value sufficient to meet possible volume restrictions for the propulsion module,
but always going along with a high gravimetric heating value.
The required total energy to be contained in the gasgenerator should mainly be attained by means of a
high gravimetric heating value with secondary importance laid on density with respect to the propulsion
module weight. This is very important because the propulsion-module weight influences
- the required total impulse and volume of the boost motor
- missile maneuverability and

may be a restricted figure for airborne missile applications.

o A high expulsion efficiency to make use of the 'ropellants energy content. This parameter refers to the
fact that usually a small portion of the combustible is not injected into the ram-combustor but is re-
tained in the gasgenerator. Expulsion efficiency depends on the burn-rate level, the primary combustion
temperature, the portion of condensed phases in the primary combustion products, the geometry of the gas-
generator outlet, and the volume of the empty gasgenerator at burnout.

o A high secondary (ram)-combustion efficiency to get a high degree of conversion of the propellants
energy content into heat release in the secondary combustion chamber. Secondary combustion problems are
frequently inherent in the use of high energy/high density propellant ingredients.

o An adequate turndown ratio, with an upper limit determined by the maximum thrust level required by the
mission envelope and with a desirable lower limit according to economic fuel consumption at the minimum
required thrust level. A low figure is desirable for the upper limit of the pressure bracket used to
achieve the fuel flow modulation. A high primary combustion pressure leads to the need of heavy gas-
generator structure and increases propulsion -module weight.

Good handling and storage properties of a SPR require to exclude as far as possible propellant
ingredients whch are"
L...
* - toxic materials
- explosive materials leading to a high degree of propellant hazard classification (e.g. greater portions
of high energy oxidizers like HMXand RDX) 'e..
- materials undergoing a temperature dependent change of their properties (e.g. the phase change of
, aammonium-nitrate)
 7.-4

A low signature of a SPR propulsion module requires:

- A low metal content of the propellent due to the visibility of condensed metallic oxide exhaust.
- A low content of otner fuel ingredients which are difficult to burn under ram-combustor operating con-
ditions (e.g. graphite or carbon black) and consequently may induce soot formation in the engine ex-
haust.
- A sufficient turndown ratio to avoid excess fuel flow at any operating condition.
- The exclusion of ammonium-perchlorate (AP) as oxidizer if minimum signature is a strict requirement.
(A visible contrail of the missile is formed as a result of the hygroscopic reaction of hydrogen
chloride with water vapor in the exhaust plume.) This argument to exclude AP is only valid for low to
medium altitude flight conditions since a visible contrail of water crystals is inevitable at high
altitudes greater than 9000 m/27000 ft even without AP.
I
The compatibility of a SPR system with possible design restrictions essentially requires:

- The favoring of propellant ingredients with high volumetric heating value if range and low propulsion
module length are the predominant requirements.
- The favoringof propellant ingredients with high gravimetric heating value if the propulsion module
weight is restricted.

Desired moderate cost of the SPR system induce the following considerations for the propellant formulation:

- Some applicable high energy fuel components (mainly metal or metal compounds) are relatively expensive
and may boost propellant cost if major portions are included in the formulation.
- Physical and chemical properties of the ingredients insluded in the propellant may require sophisticated
or time consuming and consequently expensive processing procedures for propellant mixing and grain
manufacturing. Here the use of ingredients with a very fine particle size or arrangements to avoid migration
effects may be quoted as examples.
- Characteristics of the propellant like very high combustion temperature or very high particle loading
of the combustion products may indirectly contribute to an increase of system cost by inducing the need
for sophiiticated design or exotic materials.

The processability of a SPR propellant

- is mainly depending on the balance of liquid and solid components in the formulation and
- is strongly influenced by the particle size distribution of the solid propellant ingredients.

All processability problems inherent in the manufacturing of SPR gasgenerator propellant are similar to
those of conventional solid composite propellants. Basically the processability features of gasgenerator
propellants tend to be inferior compared with conventional solid propellants.

The aspects stated concerning the processability are also true for the mechanical properties of a SPR prop-
ellant. Usually, depending on the grain configuration, required figures for mechanical propertins of gas-
gnerator propellant are significantly lower than for conventional solid propelldnts. Since an endburning
grain and a cartridge loaded configuration are preferable for a SPR (cf. para. 3.2) only minor stress loads
may arise due to
- the propellants own weight 4
- pressurizing during ignition
- mismatch of the cartridge properties (thermal expansion etc.)

3.1.2 Ingredient Selection Trade

In the following, major aspects for the selection of the different type of propellant ingredients
- oxidizer
- binder
- carbon and hydrocarbon fuels
- metallic fuels
- catalysts
are reviewed.
Oxidizer -

Selection criteria for a SPR propellant oxidizer may be summarized as follows:

o High gas output in order to achieve good expulsion effi~iency of the solid organic/inorgan'c particles
from the gas generator
o Highgas output in order to minimize depos;Ls within valve and injector orifices as well as to minimize
expansion losses in the thrust nozzle due to particle flow
A
o High combustion temperature, high oxygen content, and high oxygen balance L
o High density in order to minimize the oxidizer volume content in the propellant, since heating values of
oxidizers are negligible unless they are monopropellants like cyclotetramethylene tetranitramine (HMX)
 chlorides (primary smoke), and hydrogen chloride (secondary smoke enhancement)
o Compatibility with other propellant ingredients
o Low hazard classification
o Minimization of risk and cost of the oxidizer as an ingredient as well as overall propellent manu-
facturing cost
o Availability. -
Due to general considerations to be discussed in para. 3.2, the selection of an endburning grain configur-
ation appears to be most favorable for a SPR. Conseouently, required burn rates (influencing oxidizer
selection) may be localized between 3 mm/s and 30 mm/s (G.15 inis to 1.2 in/s). Twelve oxidizer candidates
are listed along with their pertineiit properties in the table of figure 3.3a. Figure 3.3b illustrates the
selection trade for these oxidizers.

Nitronium perchlorate and nitrosyl perchlorate cannot be regarded to be feasible oxidizer candidates 0
because of their extremely severe hygroscopicity and their incompatibility with common binders.
Sodium perchlorate and potassium perchlorate are unfavorable because of their insufficient gas production
and condensed reaction products (low expulsion efficiency). The same argument applies to lithium nitrate,
sodium nitrate and potassium nitrate, since they form a high portion of condensed products. Ammonium
nitrate in turn yields only marginal burn rates, is hygroscopic, and undergoes a phase change at 305 K/
550 R. The latter does not apply to the expensive phase-stabilized ammonium nitrate.

As far ascombustion characteristics are concerned, 114X and cyclotrimethylene trinitramine (RDX) are camp- ,. ,
arable. However, HMX has a higher density. The advantage of HIX is that it has a moderate heating value,
lacks hydrogen chloride, and thus exhibits minimum smoke. The heating value improves the propellant energy
level by 80 KJ/dm3 (1.24 Btu/in3) when 1 % ammonium perchlorate (AP) is replaced with 1 %HMX. Less advant-
ageous is the negative oxygen balance of H1U, which tends to inhibit satisfactory decomposition and
combustion of any kind of binder (marginal expulsion).
Further drawbacks of a SPR propellant containing HX as theprincipal oxidizer can be summarized as follows:
o The burn rate is low. Compared to AP there is no known burn rate catalyst that can significantly increase
the marginal burn rate. Also, unlike AP, the particle size of HMXdoes not have a major inf'uence on burn -,-.:
rate.
o Raw material costs are about four to itxtimes higher than those for AP.
o A desired low hazard classification may prohibit the use of ROX or HMX in greater amounts in the
formulation.
Thus, of the 12 oxidizers considered, AP and lithium perchlorate (LP) appear to be favorable as principal
oxidizers with HMX as a potential co-oxidizer. Whereas LP has the higher density, AP has the advantage of
being the most frequently used oxidizer in composite propellants. Since LP has a more favorable oxygen
balance compared to AP, one may be tempted to conclude that less LP produces the same total propellant
energy. However, the amount of condensed lithium chloride formed by the decomposition of LP (according to
calculations with NASA's Chemical Equilibrium Code (6 ]) is high even at tempe'atureb up to 1860 K/3350 R.
The liquid lithium chloride may pose problems for the valve flow passages and injector orifices. In addition,
the hygroscopic character of LP poses problems in both propellant manufacturing and aging. Finally, LP is
about seven times more expensive than AP.

To conclude, AP may be recommended as the principal oxidizer for a SPR gasgenerator propellant. Ho ever, the
addition of lesser amounts of HMX, chlorates, or nitrates as co-oxidizers in AP-oxidized formulatio,!s should N
be considered if necessary as a means of tailoring the burn rate and pressure exponent. .. '."
Se*~

At the present state of technology, there are two binder systems currently used for castable composite
propellants. These are polyurethenes and polybutadienes. The more modern polybutadiens include

- Carboxy-terminated polybutadiene (CTPB)

- Hydroxy-tentinated polybutadiene (HTPB)
- polybutadiene-acrylic acid (PdAA)
- polybutadiene-acrylic nitrile-polymer (PBAN)
While PBAA and PBAI4 certainly will remain in use for the next decade, future systems will mainly apply CTPlB .
or HTPB. The preferred application of these two binders Is based on the filivwinq characteristics:
o Low viscosity of the prepolymer in order to facilitate loading with a high content of solid particles V ,
while maintaining processability and castability

o Good mechanical properties over a wide temp,r.ture -snge in oder to ensure the mechanical integrity of
the propellant
o Low oxygen content
o Secured availability.
 T7 77. 7 .- T 77 .

7-6
Carbon and Hydrocarbon Fuels
S..

A selection of liquid and solid carbon and hydrocarbon fuel candidates is listed in the table of fig. 3.4.
Desired properties for the fuel components are:
o High heat of combustion
o High density
o Combustion products that are gaseous (or at least in a state to facilitate expulsion).

The binder polymers (HTPB and CTPB) are exhibiting favorable figures in terms of grcvimetric heating
value. Due to the low densityhowever, there is a relatively poor performance on a volumetric basis. Liquid
plasticizers like IDP and Oppanol may replace part of the binder to reduce the viscosity of thc propel-
lant slurry during mixing, or to allow higher portions of solid ingredients to be incorporated. The use of
certain plasticizers (e.g. Oppanal) maybe interestingalso from the energetic view, since their energy
content is superior to that of the binder. Liquid burn rate catalysts (like organic iron compounds) also
act as a plasticizer and exhibit high volumetric heating values.

Regarding the solid non-metallic fuel candidates, elementary carbon (as carbon black or graphite) has the
highest volumetric heating value while the hydrocarbons (Zecorez, CLPS, PANS) are superior with respect
to the gravimetric heating value.

Two additional aspects should be considered for the selection of non-metallic solid fuel components.
o the use of carbon blcL and especially of graphite may impose problems for ram-combustion efficiency and
consequently contr'abute to soot formation and visible signature of the exhaust ccntrail.
o Lower expulsion efficiency can be anticipated for the application of elementary carbon fuels compared
with the use of "olid hydrocarbonswhere portions of the decomposition products are gaseous.

"Metallic Fuels
Metallic fuels may be incorporated in a SPR gasgenerator propellant mainly for two reasons:
o to increase the volumetric heating value
o to increase primary cowbustion temperature in order to achieve autoignition of the ram-combustion
and possibly to enhance combustion efficiency in the ram-combustor.

On the other hand, the following problemsinherent in the use of metallic'fuels have to be considered:
o High primary combustion temperatures may pose problems to valve and injector design
o Metallic fuels contribute to the particle loading of the gasgenerator combustion products which may be
detrimental to expulsion efficiency and enhance the risk of deposits being formed in valve passages or
injectors.
o High ram-combustion efficiencies a-" not easily attained using some highly cnergetlc metal-compounds
like boron.
o Metallic oxide exhaust contributes Lo the optical signature of the engine contrail.
o Some metallic fuel components (e.g. boron) are very expensive compared with other, more customary fuel
1igredients.

Gravimetric and volumetric heating values of some interesting metallic fuels are presented in a bar-chart .-0
by figure 3.5.The corresponding values of carbon and Kercsene are included foricomparison. The table of
figure 3.6 gives data about density and heating value of various metallic fuel components.

Boron has most favorable energetic features but is difficult to burn under ram-sombustor conditions. This
is due to the high temperatures required to evaporate boron-oxide which otherwise forms a layer around
tht boron particles and terminates combustion. Boron combustion has been intensively investigated and high
combustion efficiencies can be attained at present state of technology.

Comparatively little experience exists concerning the application of boron carbide which is not included
in fig. 3.5. Boron carbide has a gravimetric heating value of 52 lJ/kg (22350 Btu/in3 ) and a density of
3
2,5 g/cm (0.09 lb/in3)leading to a volumetric heating value insignificantly lower than boron. Combustibility
behavior can be considered similar or worse than for elementary boron. Alumintm provides a reasonably high
volumetric and gravimetric heating value and yields less combustion problems. Hover, the use of this metal
is limited to small quantities because of its strong tendency to form deposits.

Magnesium provides no significant improvement over carbon or hydrocarbon fuels from the energetic point of
view. On the other hand, magnesium is easily burned with high efficiency and yields an increase in primary
combustion temperature. Other metals not included in figure 3.5. like zirconium or titanium. are sometimes
discussed as fuel ingredients. These are exhibiting a very high density and a volumetric heating value
comparable to aluminum, but consequently a low gravimetric heating value which is not fivorable for overall
performance as outlined before.

The use of metal compounds (e.g. borides) or metal alloys have been also proposed to increase the heat
content of SPR gas generator propellants. Since there is very little experience with the application of
these ingredients a valuation can be hardly given.

Ikiiillwv tha ann14alitnun ri a htbn Patio ratialwat it Pansired4 MAafli4,tt the Nair ratea n# an nvv&,l&4.4e&nf",.-
 V- _77777 7777 7"7-_.

7-7

of the propellant).
There exists a vast number of catalysts used in conventional composite propellants which are also applicable
for gas generator propellants. Only one type of catalysts being frequently used in gas g~nerator propellants
(with AP as oxidizer) shall be presented here. These cdtalysts effect an increase in burn rate by fine iron
oxide particles owhich either are incorporated in the propellant as an ingredient or are formed during comb- L
ustion of an iro, containing catalyst.- '
Iron containing - alysts are .'

- ferrocene (sol. ,, rarely used)

- n-butylferrocene
e(liquid; strong catalyst; frequently used)
- catocene J trn
The use of liquid burn rate catalysts is desirable with respect to the processability of the propellant. The
liquid catalysts are exhibiting a considerable heat content as mentioned before at the discussion of hydro-
carbon fuels. Iron-oxide (Fe 2 O3 ) itself is also frequently used as burn rate catalyst. Its catalyzing prop-
erties are relatively weak and depending on the applied particle size. Iron oxide, of course, has a zero
heating value.

Iron-oxide is a cheap commercial product which is commonly used as a pigment. The aforementioned liquid
catalysts are produced in small quantities and consequently are very expensive.

3.1.3 Examples for Formulation Screening

To complete the consideration of propellant formulation some examples for formulation screening are added
below.

Figure 3.7 compares density and gravimetric and volumetric heating values of five hypothetical formulations
with different amounts and types of fuel components. These formulations are

1. 25% AP + 75% HTPB

2. 25% AP + 25% HTPB + 50% PAMS
3. 25% AP + 25% HTPB + 50% carbon black
4. 25% AP + 25% HTPB + 50% Mg W-.1 1
5. 25% AP + 25% HTPB + 40% B + IO% Al
The portion of binder in these formulations should be understood to include approximately 5% of curing and ..
burn rate additives in a real propellant.
The bar chart of fig. 3.7 illustrates the stepwise increase of energy content when hydrocarbon, carbon, .*-

and metallic fuels are introduced into the propellant formulation.

To compose an appropriate propellant formulation for any given application, firstly a decision must be made
to use a metallized or non-metallized propellant. This decision has to be based on the requirements for
energy content, gas generator temperature and possibly optical signature. Additionally, a proper balance has "
to be found between fuel components with lowor high energy content and good or marginal combustibility
properties to get an overall good performance.
Figure 3.8 presents an example for the screening of a non-metallized propellant formulation. The
volumetric heating value is plotted versus the AP content for the following variable formulation:
@
AP variable (10 to 40%)
binder : 25 %
carbon black 0 to 40 %
PAMS 75 % -% AP - % CB

Lines of constant primary combustion temperature (according to chemical equilibrium calculations) are also
indicated in fig. 3.8. The predominant influence of the oxidizer content on the heating value is quite evident.
However, burn rate requirements impose a lower limit for the oxidizer content, rarely allowing oxidizer
portions lying significantly below 25 percent.

Furthermore, fig. 3.8 shows that the increasing of carbon black by forty percent and reducing PAMS by the
same amount results in an increase of the volumetric heating value by about six percent. Only an increase .'... .s.
of the density is effected by this variation of the formulation while the gravimetric heating value even
decreases when PAMS is replaced by carbon black. Gasgenerator temperatures for a fixed AP content are dk- r
creasing when the portion of carbon black is increased. Besides the theoretical consideration of energy
content it must be reminded that good afterburning efficiency is presumably much easier to achieve using
a hydrocarbon than a carbon fuel. Consequently, only lower portions of carhon fuel can be recommended Iff.. . ..
volumetric restrictions are given for the propulsion module and severe requirements for the optical signature
do not allow the application of metallic fuels.

Gas generator temperatures which are calculated for the propellants of fig. 3.8 will be probably insufficient
to achieve autoignition of ram-combustion.

I
 //

7-8
thermo chemistry are again indicated in the figure. The propellants considered in fig. 3.9 are based on the

following formulation:

AP variable 10 to 32 %
binder 25 %
metal 7.5 to 15 %
PAMS : 75 % -% AP -% metal

The oortion of metal included in the figure has the following composition:
boron : 80 %
aluminum 20 %

Three major aspects can be concluded from the comparison of figures 3.8 and 3.9:
o Replacing part of the solid hydrocarbon fuel by metals leads to a moderate increase of the gravimetric
heating value (while carbon fuels reduce the gravimetric heating value)
o The incorporation of the mentioned metaliic fuel effects a significant increase of the volumetric heat-
ing value (which only could be achieved by much higher quantities of carbon fuels)
o Primary combustion temperatures are much higher than for the propellants of figure 3.8 and by far
sufficient to attain autoignition of ram-combustion. It has to be added that measured gasgenerator
temperatures are below the calculated equilibrium temperatures indicating a reaction rate of the
metallic ingredients below equilibrium. However, this does not affect the aforementioned statement
concerning autoignition.

( Afterburning of metallic fuels is not easily done but poses no greater problems than for the compared carbon
fuels. This is especially true since much more experience exists for the application of metallized than
for carbon loaded propellants.

To summarize, the application of me'tal fuel can be recommended to increase the energy content of the prop-
ellant if optical signature is no severe requirement or is inevitable anyway due to the characteristics of
other propellant ingredients as for example AP. State of the art propellant formulations and overall
performances will be discussed later on in par&. 4.

3.2 GRAIN DESIGN [ 3, 57

The gas generator propellant grain configuration must be selected early in SPR design since it drives such
key propellant characteristi':s as burn rate as well as other system aspects such as valve sizing. Experience
gained in fixed and variable flow SPR programs leads to the conclusion that An endburner grai., configuration
is optimum for the following reasons:
o Any internal burning configuration (star, rod and tube et. al.) requires very low regression rateswhich
usually can be achieved only by pyrolysis rather than by flame front combustion. Pyrolysis in turn imolies the
danger of unacceptably high residues in the gas genLrator as well as a gas temperature that is too low
for autoignitlon in the ram combustor.

o Depending on the internal-burning configuration, there is a more or less pronounced variation in burning
surface which has a direct impact on the variable flow valve size as well as on the maximum/minimum gas
generator pressure range.

o The volumetric propellant loading of an internal-burning grain Is inferior to that of an endburner.

A cartridge configuration is the appropriate design for an end-burning grain especially for long burning
times. Severe problems are inherent in a case-bonded configuration of a long end-burning grain concerning
the considerable axial elongation of the gra.in between commonly required low and high operational temperature
limits. Using a cartridge design, a radial Vap between grain liner and gas generator insulation may occur
atlow temperature which has to be sealed or filled by elastic material since a variable empty volume of the
gas generator will pose ignition problems.

Different from customary solid rockets, aerodynamic heating occrrring during long flight duration at high
Mach numbers may warm up the propellant and at least may provoke conical burning. Therefore, not only heat
flux to the gas generator wall downstream of the burning surface but also heat flux from the structure into
the propellant has to be considered fbr insuiriion design of a SPR gas generator. Aspects of liner formulation
concerning adhesion or migration problems Lre similar for SPR and conventional composite propellants.

3.3 FUEL FLUW MODULATION

Fuel flow modulation for a SPR is -isually achieved by using a propellant with a pressure sensitive burn rate
and a valve to vary the gas generator throat. This concept and the inherent problems are discussed in detail.
Some remarks are added concerning alternative, more exotic concepts for fuel flow modulation.

3.3.1 Fuel Flow Modulation Using Pressure Sensitive Propellants

The characteristics of fuel flow control using a pressure sensitive propellant are discussed below. Addition- t.:
ally, basic aspects of valve design are reviewed.
 7-9-L

for SPR gas generator propellants is corresponding to conventional composite propellants and may be described
by the following formula

r a p (3.1)

Where a and n are empirical parameters which are not necessarily independent from pressure over a utilized
wide pressure bracket. Consequently %ss production may be written as

mprop = prop b a pn (3.2)a

The mass flow ejected through the gas generator throat, a critical pressure ratio being provided, is given by
-Aej Pc (3.3)..'...-
'ee c*

with
c , 1 . R T. (3.4)
and
and o (3.5)

according to fundamental fluid dynamics. .-----.

Equilibrium exists when the produced mass flow equals the ejected mass flow. Consequently, the equilibrium
pressure is given by

Pc " (a .%prop c* K) (3.6)1-n

where K is defined as the area ratio

Pressure and burn rate variations are depending on a variation of the K ratio as follows:

dPc I dK

and ~m
dr n dK -- - ~~~~(3.8)
(3.8 ?.-''.

Fig. 3.10 illustrates the influence of n on the ratio of burn rate variation and K variation. It is quite
evident from fig. 3.10 that the application of propellants exhibiting very high pressure exponents above
.75 may be very hazardeous from the stability point of view.
Fuel flow control using pressure sensitive solid propellants is especially complicated by the inverse response
of the fuel flow to the actuation of the control valve. This characteristic Is illustrated by figure 3.11
showing traces of the valve throat area, pressure and fuel flow for a control action to increase or decrease
fuel flow, respectively. To increase fuel flow means to reduce the valve throat area and consequently to
increase pressure to the values which stationarily correspond to the desired fuel flow. When the valve closes,
first the ejected mass flow decreases corresponding tothereduced throat area. Now the pressure rises since
the produced mass flow is greater thin the ejected mass flow. Mass production and ejected mass flow increase
corresponding to the pressure rise until the new equilibrium is reached. The inverse occurs when the valve
opens to decrease fuel flow.
The magnitude of the fuel flow over-or undershoot depends on
- the turndown ratio of the control action
- logic cf the control loup and velocity of valve displacement

The time span required to change from one fuel flow rate to another depends on
- gas generator empty volume
- the turndown ratio of the control action
- logic of the control loop
. pressure exponent within the pressurelbracket used for the control action
the achievable turndown ratio depends on various factors which are related to the propellant characteristics
as well as to system aspects. One main limiting parameter for the TDR i.,the pressure bracket allowed for the
burn rate modulation. The upper pressure limit is imposed by the design of the gas generator structure. The
lower pressure limit depends on the pressure level in the ram-combustor and on system requirements to
include or not a second sonic nozzle downstream of the control valve to evaluate the fuel flow. Evidently,
the other main parameter affecting TDR is the pressure exponent of the propellant within the allowed pressure ...- ?..'.
bracket. Since the burn rate of solid propellants is more or less temperature sensitive, the maximum achiev-
able TDR is also limited by the upper and lower operational temperature required for the propuiion system. !
7-10

Figure 3.13 summarizes the parameters limiting the maximum achievable TDR. Lines of constant TDR are indicated
as a function of pressure exponent n and temperature sensitivity - [% / " K]. The left figure is related
to a configuration where the control valve fcrms the only sonic thrBat whereas a second sonic nozzle is added
in the right figure which increases the lower pressure limit. The calculations for figure 3.13 are based on
a 103 bars/1500 psia pressure bracket and a 128 K/2301F temperature bracket.

Essentially two types of control logic may be applied to solid propellant burn rate control.
o A sonic nozzle is installed downstream of the control valve. Pressure and possibly temperature are measured
upstream of this nozzle allowing to evaluate the actual fuel flow if gas properties are known with suff-
icient accuracy. The control valve is adjusted to a position where the desired fuel flow is stationarily
evaluated at the second sonic throat. The applied control algorithrrmust taka into consideration the mass
flow overshoot during the control action, the pressure sensitivity of the propellant (for stability reasons)
and a possible dependancy of the c* figure on pressure (which is mainly according to a variation of gas
temperature)especially if no temperature measurement is provided.
This control logic which appears to be simple and evident at the first look has several disadvantages.
The installation of a second sonic nozzle downstream of the control valve and possibly still upstream
of the injector (a simple sonic throat configuration for the fuel injector may not be compatible with
combustion efficiency requirements) is space-consuming and may not be compatible with design restrict-
ions given for a modern SPR system.
Pressure measurement shortly downstream of the irregularly shaped valve cross section, a tube turn etc.
may be of poor accuracy.
- The dicussed control logic is restricted to propellants which predominantly produce gaseous combustion
products with 6 minimum of condensed species, since two phase flow effects and deposits possibly formed
in the second tnroat are highly detrimental to control accuracy.

o Another control logic which may also beapplied to solid propellants forming combustion products with high
loading of condensed phases only relates to the predetermined ballistic data (including c variability)
of the propellant and the calibrated throat area variation of the valve. The control algorithmoadjusts
the valve to a position where the K ratio corresponds to the desired burn rate. A comparison of the actual
and the expected pressure allows to correct disturbances caused by grain temperature, coning effects or
cloggingof the valve. The essential control variable for this kind of control logic is the required
thrust level or flight velocity. The fuel flow not being explicitly known is adjusted until the desired
performance of the propulsion system is achieved.
This control logic requires amorecomplexaild skilful control algorithm but is capable to handle highly
particle-laden flows with moderate depositing characteristics.

Any control logic must take into consideration that the overshoot of the fuel flow during control action may
induce operational conditions for the ram engine which may be off-limits for air intake operation. A sudden
increase in fuel flow and heat release in the ram-combustor will cause a pressure ri-e which may blow off
the final shock forcing the air intake into subcritical operation. If the air intake exhibits no stable
subcritical operation the air intake will go into inverse flow leading to t'e destruction of the missile.
On the other hand, a long lasting undershoot of the fuel flow (occurring for high modulation ratios, big
gas generator empty volume, low n) may lead to thrust deficiency and deceleration of the missile and con-
sequently also may pose problems concerning air intake operation.

3.3.1.2 Control Valve Design 1 3, 9, 10, 12 .

The subject of control valve design, due to its complexity cannot be treated completely in this context. Kajor
aspects of valve design and some basic valve design principles will be discussed below.

All customary valve design principlesare applicable to a gas generator control valve as long as primary com-
bustion temperature is low (e.g. below 1000 K/1800 R) and the portion of condensed species in the primary
combustion products is low, too. Problems arise for high temperatures and highly particle-laden flows. Problems
may also arise from high pressure ratios across the valve if the configuration is not adequate.
The following aspects have to be considered for gas generator valve design:

(I) space requirement for valve and actuator -,"-"

(2) compatibility with other components, e.g. injectors --
(3) sealing problems
(4) actuation loads (determining type and size of the actuator)
(5) mechanical and thermal stress arising for the components
(6) heat flux into the control valve and to the actuator
(7) clogging risk
(8) overall functional safety
(9) cost
The significance of several aforementioned items dependsof course,on the required endurance time. The consider-
ation of the items (I) to (5) and (8) leads to the conclusion that a rotating motion of the control valve is
preferable to a translatory motion. In most cases a rotating valve and its actuator may be accommodated in a
smaller volume than a translatory valve. Sealing problems of the rotating valve are inferior to that of
translatory valves. Rotary sealing surfaces usually stay clean of deposits while any translatory valve com-
ponent may draw deposits up into the bore on which it seals. Rotary seals may be composed more easily of
heat resistant materials like graphite foils or wool. Translatory seals on the other hand usually apply
elastomeric materials which must be thermally protected. Actuation loads tend to be greater for translatory
4
.
-e -Ar -a *+k A lA k.1 .+ *k. - -1 .1a. ij
 V.7

7-11

actuator are low. Problems may arise frma dividing the fuel flow to be fed into the ram-concustor by several
injectors. Where narrow bending of the flow is prohibitf.d due to the clogging risk, a valve configuration
with several throats actuated by one unit may be the only adequate solution. This, however, is not feasible
with every valve type.
Referring to the items (5) and (6), probably less problems arise if the valve is installed in an environment
with low flow velocity. If the valve is installed in a region of higher flow velocity, e.g. in a gas pipe,
thermal stress will be great and possibly non-uniform around the control element. Consequently, heat flux
into the valve and the actuating unit will be greater for a high flow velocity than a low flow velocity
environment.
A clogging risk (item (7)) only exists for particle-laden flows. Deposits are predominantly formed at
- stagnation points
- deedwater regions
which cannot be completely avoided by valve design. If depositing is expected, any valve configuration ex-
hibiting a small valve lift or narrow slots for the valve throat siould be discarded. Further on, it must
be considered that additional valve actuation loads may arise from stripping off deposited material.

The overall functional safety (item (8)) summarizes the remaining risks of the items (3) to (7). Cost of
the valve may be boosted by the need for sophisticated design and complex manufacturing procedures, by
using exotic temperatureresistant material,and by type and size of the actuation unit.

Figure 3.14 pres.2nts six basic valve configurations which will be shortly discussed below, referring to the
aforementioncd design aspects. A temperature bracket of 1000 K to 1700 K (1800 R to 3100 R) an upper
pressure liitit of 100 bars (1500 psia) and particle loadings from 0 to 50 percent are implicitly assumed
as possible operational conditions for the gas generator control valve. The qualitative estimation of Le4
characteristics and the applicability of the presented valve design variants are b.sed on available su-te
of the art experience.

(1) Rotary Blade / Butterfly Valve

- Space requirement issmall for valve and actuator
- Feasible but infavorable, if more than one throat is desired
- Rotary sealing poses no major problems
- Low actuation loads due to ideal force balance
- Moderate mechanical load but high thermal load due to high flow vel.:ity around the valve
element
. High heat flux to the actuator, thermal insulation of the actuator may be problematic
. Unfavorable for use with highlyparticle-laden flow due to the depositing risk
- Excellent functional safety for gas flow with moderate temperature and low particle loading;
hardly applicable to highly particle-laden flow.
- Moderate cost, if applied to an adequate fuel flow Pnvironment
Expensive wmterial may be required for the blade if the valve is used in high temperature flow.

(2) Translatory Plunger

- Little space is needed for the valve but onsiderable volume may be required for the actuator
- Feasible but unfavorable if more than one throat is desired
- Translatory seal may pose severe problems when the valve is applied to high temperature particle-
laden fuel flow
- High actuation loads
. High .ei.hanical loads due to a non-uniform force distribution around the control element; high
thermal load due to high gas flow velocity around the control element
- High heat flux into valve and actuator; thermal insulation of the actuator may be problematic
- High depositing risk when used in particl.--laden flow
- Medim functional safety for the application to gas flow with moderate temperature and low
particle loading, due to the &echanical and thermal loads at the control element as well as to
the high actuation loads. Low functional--afety for the application to high temperature or
highly particle-laden flow
- Costly because of the power requirement of the actuator and the need for sophiticated design
and the use of expensive materials for the control element if the valve is used in high
teperature flow.

(3) Needle Valve - Opposite to Flow Direction

. Highly space-consuming for the uni-axial configuration of valve and actuator. Alternative config-
urations with lateral actuation are feasible but exhibit a complex design. High space require- .. .. ,*-.
ment of the actuator.
- Feasible but unfavorable if more than one throat is desired
- Extrewesealing problems since the rod housing inevitably forms a stagnation point for the flow
- High actuation loads
- Moderate mechanical load but nigh thermal load due to high flow velocity around the valve -
- High heat flux into valve and actuator; thermal insulation of the actuator is highly problematic
- High risk thet deposits are formed at the rod and the orifice when used in particle-laden flow
- Medium functional safety for the application to gas flow with moderate temperature and low-to-
zero particle loading, hardly applicable to high temperature, highly particle-laden flow
- Costly, because of the power requirement of the actuator and the complex design of rod guide
and sealing. Expensive material is required for the rod, if the valve is used in high temperature '..
flow.
 7-12

(4) Needle Valve - Parallel to Flow Direction

- Reasonable configuration only, when accommodated in the gas generator.

Medium to high spte requirement for the valve installation; high volume requirement for the
actua'ar; lateral actuation is mandatory and feasible but includes complex design using lever
bars. Even a rotary actuation is feasible.
- Configuration with several throats are feasible
- Sealing is difficult but feasible L
- High actuation loads
- Moderate mechanical load but partially high thermal load due to high flow velocity around the
control element
- Putentially high heat flux into valve installation and the actuator
- Thermal insulation of the actuator may be more easily accomplished than for configuration (3)
- Deposits may be formed at the orifice and less likely at the rod. Deposits are not necessarily
detrimental to valve function if the throat gap is wide and sufficient valve lift is provided.
- Medium to high functional safety even for the application to high temperature, highly particle-
laden flow.
- Costly, because of the high power requirement of the actuator and the complex design of the act-
uating l:ver system. The use of expensive material may be mandatory for the control element and
the actuating lever system.

(5) Gate Valve

The gate valve in some kind is a variant of the translatory plunger accomnodated In the gas generator. The
translatory gate covers part of the gas generator outlet forming a variable throat. The tr3nslatory motion
has usually to be transformed into a rotary motion to accomplish a feasible ,nd compact actuation. Alter-
nativety, the translation of the gate may be accomplished more favorably by rotatinq an eccentrically
shaped disk..
- Medium space requirement for valve installation and actuator
- Configurations with several throats are feasible
- Sealing is feasible for rotary actuation
- Comparatively low actuation loads
- Moderate mechanical and thermal load. Partially low flow velocity around the control element
. Medium heat flux into valve and actuator. Thermal insulation of the actuator is feasible.
- Deposits may be formed at the deadwater region of the downstream side of the gate. Stripping
off deposits when the valve moves may increase actuating loads.
- Medium to high functional safety even for the application to high temperature, highly particle-
laden flow.
- Moderate cost referring to actuator and design complexity. The use of expensive thermally
resistant material may be mandatory for the gate element

(6) Disk Valve

The basic arrangement of the disk valve is similar to the needle valve parallel to flow direction (4). The
valve is accommodated in the gas generator and laterally activated by means of a lever system.
- Medium to high space requirement for the valve installation; high space requirement for the
actuator
- Configurations with several throats are feasible
- Sealing Is difficult but feasible
- High actuation loads
- High mechanical load; partially high thermal load due to high flow velocity at the downstream
side of the control disk.
- Medium heat flux into valve and actuator; thermal insulation of the actuator is feasible E
- Comparatively small valve lift (depending on the diameter of the gas generator outlet) leads
to a narrow throat gap of the valve. High clogging risk at narrow gaps excludes the disk valve
from the application to gas flow with considerable particle loading.
- Medium to high functional safety for the application to gas flow with low-to-zero particle
loadlnC (wOere another Yalve design like (1) may be the better choice). Hardly applicable to
highly particle-laden flow.
- Contly, because of the high power requirement of the actuator, the complex design of the actuating
lever system. The use of expensive material may be mandatory for the disk element.

The selection of an appropriate actuator for the control valve depends strongly an the actuation load of the
valve. An electrical actuation is frequentlydesirble due to aspects of the missile system. However, electrical
actuation is restricted to moderate actuation load; because of
- limited power supply of the missile system
- volume requirement of high power electromotors
- thermal problems.

High actuation loads may be handled using pneumatical or hydraulical actuators. If a pressurizing system does
not exist in the missileanair turbine fed by tap air from the intakes may be applied. However, a pneumatical
or hydraulical actuation is space-consuming due to th4 required auxiliary valves and power supply.

Several types of thermally resistant materials may be applied to valve components:

- high melting metals like molybdenum or tungsten
- ceramic material
 7-13

The application of high melttngmetals is a customary and safe approach but It includes a considerable weight
penalty. Ceramic materials are exhibiting a lower density and are highly temperature resistent. However,
some ceramics are sensitive to the thermal shock occurring at the ignition of the propellant. The manufact-
uring of ceramic components of complex shape is very expensive. Components made of graphite or CFC exhibit
medium to low densities und high thermal resistance. If such components will be exposed to payticle-laden
flow, usually an impregnation or coating of ceramic material at the surface is needed to avoid erosion.

Surface coating may also be considered to protect thermally stressed valve components. Spinels and zirconium- l
oxide can be quoted as examples of coating materials. --

3.3.2 Alternative Concepts for Fuel Flow Modulation

Some alternative, somewhat exotic concepts for fuel flow modulation shall be presented below.
Retractable Silver Wires

This concept Is illustrated by figure 3.15. It is based on a burn rate increase around a silver wire embedded
in the propellant which is due to increasedconductive heat flux into the propellant effocted by the wire.
A local increase in burn ratc leads to a conical shape of the combustion surface. When several silver wires
are axially embEdded at regular intervals in an endburning propellant (as shown by fig. 3.15), a cone will
be foried around each silver wire leading to an increasedburning surface dnd an increased fuel flow. When
the slyer wires are retracted behind the burning surface by a skilful mechanism and a locally increased heat r
flux into the propellant does no more exist, the burning surface equalizes after some time due to the uniform, 0
burn rate. Consequently, this concept for fuel flow modulation means a controlled variation of the burning
surface instead of a variation of the burn rate. The logic of the control loop may be bdsed on the evaluation
of the ejected mass flow through a constant sonic throat (measurement of gas generator pressure and temp-
erature).

Various disadvantages and hardly estimable risks .re inherent in this concept:
- The achievable burn rate increase by means of silver wires is limited to a ratio of 3 to 4, con-
sequently (according to geometrical considerations), the ratio 0f surface increase being effected." .
by the coning is limited to about 4. The resulting narrow cones (cone angle - 30 degrees) will pose
problems concerning burning characteristics.
- The response time of the system may be poor depending en the time needed for the formation of the
cones and the equalization of the burning surface, respectively. The response time will improveif
the number of silver wires and cones is increased.
- The response time problem of pressurizing or depressurizing the empty gas generator volume when gas
production changes and has to be fed through a constant throat, is also inherent for this control
concept. However, no fuel flow overshoot will occur during a control action.
- The axial bores in which the wires are drawn , probably reed to be insulated to avoid conical burning
induced by entering combustion gases when the silver wire is withdrawn.

- Since pressure is a fairly insensitive input parameter for the control loop Lo adjust wire drawing to inL@
the actual burn rate, control accuracy and control stability are questionable.
- The overall functional safety of a wire drawing mechanism is questionable either. The space needed
for the accommodation of this mechanism mvy be considerably greater than for a control valve.
- The manufacturing of the prope lant and the embedded wire system will be highly complex and expensive. . ,
The concept of retractable silver wires for fuel flow modulation exhibits no favorable features. Its realiza-
tion seems to be hardly feasible.

An lacrease of the burnlnU surface by coning is also effected by a concept proposed by Thokol. The THERMA-
TROL (Thiokol Heat Exchange Rocket Motor Augmented Rate Control) concept employs small diameter metal tubes
embedded longitudinally in an endburning grain. The forward tube ends (at the bottom of the grain) are
connected to a flow rate control system. Hence for this concept the heat flux into the propellant (to effect
coning) is delivered by the combustion as flowing through the longitudinal pipes.
Most statements brought up concerning the silver wire system do also apply to this concept. Feasibility of
the THERMATROL concept may be somewhat superior to the silver wire concept.
Matrix Propellant .,.

Another concept (cf. fig. 3.16) proposes to compose two different propellants to a so-cailed matrix grain.
According to this concept, preproduced granules of one propellant would be added to the mix of the other
propellant. By this way, a high energy propellant which is not pressure sensitive could be embedded in a
highly pressure sensitive propellant exhibiting a Inw energy content. Fuel flow modm~lation would be effected
by burn rate variation of the pressure sensitive propellant according to para. 3.1.3. The granules of the
high energy propellant would be ejected from the burning surface by the combustion gases of the pressure
sensitive propellant, somewhat similar to erosive burning.

Basically, this alternative concept for fuel flow modulation may be feasible. Nevertheless, two objections S

.1)

I I I I iI I
 '.''7-14

will not reach very high values.

-"A possibly low deqree of decompositon of the ejected high energy propellant granules may be highly
unfavorable for afterburnirng efficiency.

- Propellant processing will be highly cumplex and expensive. Poor mechanical and agine, properties of
the matrix propellant can be anticipated.

This concept (cf. fig. 3.17) acts somewhat similar to the application of silver wires. Here, a central rod
made of a pressure sensitive low energy propellant is cmbedded In a tubular high energy propellant which
is not pressure sensitive (a configuration similar to a candle). The core propel lant is embedded in the
tubular propellant without insulation. The gas gemerator pressure is controlled by a valve. The core prop-
ellant needs to exhibit the same burn rate as the tubular propellant at the lowest required burn rate level.
When the valve closes and pressure rises, the burn rate of the core propellant increases and effects a
more or less significant coning depending on the burn rate ratio rrod : rtube. Fuel flow modulation again
is done by a variation of the burning surface.
Of course, several axial propellant rods embedded at regular intervals may be used 'nstead of one core
propellant. Reviewing this concept, the following critical remarks shall be added:

- The loss of energy content of the overall propellant is lower than for the matrix propellant but is
not negligible, especially if more than one low energy rod is used.
- The achievable TDR is limited since the area ratio of a plane and conically shaped burning surface equals
to sin (c/2) where a is the cone angle. Consequently, in area ratio of 5 corresponds to a cone angle of
23 degrees.Obviously, the accurate handling of such narrow cone angles can hardly be estimated to be
feasible (erosion, mechanical properties of the propellant).
- The control logic corresponds to the one applied for pressure sensitive propellants (cf. para. 3.3.1)
while the control accuracy may be inferior.
-"The system response to a control action and inherent problems correspond to the silver wire system.
"Complex and costly manufacturing procedures can be anticipatcd for the propellant, mechanical and aging
' properties are questionable.
No advantages can be stated for this concept concerning simplicity, control accuracy or volume requirement
(since a ccntrol valve is mandatory). The feasibility may be estimated to be superior to the silver wire
*. system. The concept of coaxial propellant may be applied when pressure sensitivity of the high energy prop-
ellent cannnt be effested by variation of the propellant formulation and the requirements stated for TDR,
control response and control accuracy are moderate.

If only two (or a maximum of three) fixed levels are required for fuel flow and if sequence and duration are
predetermined for these mass flow levels, the application of propellants with different burn rate levels
offers a simple and reliable concept.
Different propellant types may be cast one on another or the gas generator may be divided into different
S."chambers which are ignited and operated independently. The use of several primary combustion chambers leads
to a low volume efficiency and an increase in structure weight. Usually, mission requirements are prohibitive
to the application of this simple concept.

3.4 RM4-COMBUSTOR
Three topics have to be discussed concerning ram-combustor design:
- optimization of the geometry to achieve high combustion efficiency
"- thermal protection of the combustion chamber
- thrust nozzle design and thermal protection
Especially, the first item is reviewed in some detail in the next paragraph, since there are characteristic
aspects inherent in the SPR system. Insulation problems and thrust nozzle design are similar to other ram
type engines and are only shortly discussed.

3.4.1 RAM-COMBUSTOR CONFIGURATION [ 4, 15, 16, 17, 18, 19J

The basic design of the ram-combustor (secondary combustion chamber) is mainly determined by the missile
configuration. Different air intake configurations are presented by figure 3.18. A lateral arrangement of
two or more air intakes usually leads to a side dump configuration of the ram-combustor. A chin ar ventral
intake or a non-integrated design of the propulsion module are favorable to the application of a coaxial
ram-combustor configuration.
Possible ram-combustor configurations are s$hewmatically schown by fig. 3.19, along with some examples of
detailed design patterns.
 7-15 .. 4

are predetermined by the required mission envelope of the propulsion system. These are:
- equivalence ratio
- chamber pressure
- thrust nozzle throat area (i.e. combustion chamber contract!-in ratio)
- chamber length

ratiofigures
-he fromtotal
result for the mass flowthrust
required the ram-combustor,the
throughlevel. All three parameters togetherratio
equivalence determine chambercontraction
and thethechamber pressure. .]
* The chamberlength usually corresponds to the required length of an integrated boost motor.

Thequivalence
W ratio (determining combustion temperature) and the chamber pressure have a strong influence
on the combustibility characteristics nf most types of fuel components. The chamber length and the chamber
contraction ratio determine the residence time of the combustion products in the ram-combustor which is a
very important parameter for the efficiency of particle combustion.

As a consequence, the optimization of the ram-combustor geometry is restricted to the following subjects:
* " To provide proper mixing of the fuel-rich primary combustion products and the ram air (possibly for a
-' wide range of operating conditions)
, To provide local variations of the equivalence ratio different from the overall figure which may be -,
favorable to the combustion of high energy fuel components exhibiting marginal combustibility character-
istics. Variations of the equivalence ratio are effected by special flow patterns like recirculation
zones.
"- Toprovide a uniform pressure and temperature profile upstream of the thrust nozzle to get optimum use
of the expansion flow in the divergent part of the nozzle.

Beside a high combustion efficiencyit is a main objective of the ram-combustor optimization to provide
autoignition of the secondary combustion (if the temperature of the primary combustion products is adequate).
Since the usually required accommodation of an integrated booster is prohibitive to any installation (e.g.
flameholders) in the ram-combuster, the aforementioned optimization goals must be achieved by proper design
of
. the air injection into the ram-combustor
*. the fuel injection.

%-It is obvious that there is no general approach to the optimization of these geometrical parameters. Optimiz-
ation depends on the predetermined overall ram-combustor configuration (coaxial or lateral air injection)
and the applied fuel components (particle loading, combustibility). Conseauently, only some major aspects '-
can be discussed in this place.

Cd4,y
Generally, three major parameters are available to optimize air injection:
i- injection angle
ar
- air injection velucity (i.e. area of air inlet into the ram-combustor)
- shape of the air inlet
-. Additionally, vanes or similar installation may be inserted in the air inlet in order to effect flow
.. deviation or swirl.

. A high angle of air injection applied to a lateral combustor configuration or swirl applied to a coaxial
geometry may be favorable to combustion efficiency due to efficient mixing of air and fuel within a short
distance. This is especially true for gaseous primary combustion products. If major portions of condensed
"pha~es are included in the primary combustion products and the equivalence ratio is on the lean side
however, rapid and complete mixing of air and fuel-rich gas and the resulting cooldown of t7, condensed
ph.3es may be detrimental to combustion efficiency. The last statement also applies to autoignition. A
penalty in terms of an impulse loss is inherent in any flow deviation of the ram air. The magnitude of
"the impulse loss is proportional to the sine of the flow deviation angle.
A high air injection velocity may improve mixing and combustion characteristics for lateral combustor ."

p configurations. It may be helpful
However, the objection stated for
"fuel applies to the air-injection
Intake port to the ram combustor.
pre;sure loss being a performance
to break up highly particle-laden jets of primary combustion products.
the air injection angle concerning rapid and complete mixing of air and F-
velocity. A high air injection velocity is effected by throttling at the
Throttling of the air flow leads to a loss of total pressure. Total
related parameter should be kept low. '4

"Aparticular shape of the air inlet(s) may be designed to induce high local turbulence or to adjust the circum
ferential or axial extension of air injection (e.g. using rectangular instead of circular intake ports). An 0
* appropriate geometry of the air inlet may be important for the combustion of fuel gas containing high
portions of condensed phases. A sophisticated shape of the intake ports poses severe problems to the design -
"of the port covers which seal the air ducts during the operation of the integrated booster.

The axial position of the air injection is a geometrical parameter which applies only to lateral air intake
* configurations. A downstream positioned side dump configuration of air injection has to use a high injection
* ]angle to get advantage of a recirculation to be formed in the combustor dome.

u ' , ,, , ,tSI m
 7-16

duct near the intake port of the combustor by upstream or downstream orientated injectors similar to
customary configurations of CRJ's. Primary combustion products of high temperature or with significant part-
icle loading have to be fed into the dome of the ram-combustor. This being the *:ase, the following geometrical
parameters are available to optimize fuel injection:
- fuel injection angle relative to air injection
- nunmer of fuel injectors
- angular and radial position of the fuel injectors
- fuel injection velocity
- shape of the fuel injector

A widn angle included by air and fuel gas injectinn (e.g. a perpendicular configuration) is desirable to
provide proper mixing. On the other hand, a high amount of particle loading in the primary combustion products
may be prohibitive to wide angles included by fuel injectG:,s and ram-combustor axis. Significant erosion
can be effected by particle-laden fuel flow directed to the ram-combustor wall. As a consequence, proper
mixing of air and particle-laden fuel gas is not easily achieved for a coaxial ram-combustor configuration.
The lateral combustor desigo usually allows an almost perpendicular injection of air and fuel gas.

The number of fuel injectors depends on the overall combustor configuration. One central fuel injector is
frequently applied to a coaxial ram-combustor. This injector , however, may feed through several nozzles or
ports ejecting to different directions. An impinging injector consisting of several injector nozzles
installed at regular inteivAls in the air intake port is also applicible to a coaxial combustor configuration
if particle loading is low for the primary combustion products.

For lateral configurations the number of fuel injectors usually is related to the number of air ;nlets into
the ram-combustor. The angular and radial position of the fuel injectors are parameters mainly applying to
the lateral ram-combustor configuration. Corresponding angular positionsof air and fuel injection are
desirable to ensure break up of particle-laden fuel flow. Furthermore, it is evident that fuel injection
at anouter radial position is more favoratle to mixing than a central fuel injection.

The optimization of fuel injection velocity depends on the composition of the primary combustion products,
and the configuration selected for fuel injectio!1. Plume expansion downstream of a sonic nozzle may provide
sufficient distribution of gaseous primary combustion products. Particles needing time and distance for
velocity relaxation possibly are not properly distributed by a sonic injector. Any sophisticated shape of the
fuel injectors designed to induce particular floo patterns for mixing will only be effectivefor gaseous
primary combustion products. A rapid deviation of particle-laden flow (e.g. by a vortex) will more likely
"result in a separation of gaseous and condensed phases than in an improved distribution of the particles
due to the aforementioned two phase flow problems.
I ~ One important aspect has to be added concerning fuel injector optimization which is not related to combustion
efficiencv. While no major design restrictionare imposed when the primary combustion products are pre-
dominant~y gaseous and have a low temperature, any sophisticated design of the fuel injectors will be
highly complex if not infeasible for high temperature particle-laden fuel flow. Remembering that depositing
and erosion risks areinherent in particle-laden flow and high heat flux exists in narrow gas pipes and bends,
there should be two additional objectives for fuel injector design:
- To minimize the duct length between gas generator and ram-combustor
- To avoid complex flow patterns

* The use of thermally resistant materials may be req'iired for fuel injector components. For this purpose, all
materials quoted in para. 3.3.1.2 referring to control valve components can be applied.

3.4.2 Thermal Insulation 1 10, 11 .

Obviously, an insulation of the ram-combustor is required to protect the structure from the hot combustion
gas. The application of an ablative layer is customary for the SPR system as well as for CRJ's. This material
usually consists of a silicon rubber filled with different portions of carbonaceous or ceramic fibers or
powder. Fillers are added to improve insulating properties to adjust ablation rate and to ensure sufficient
mechanical properties of the pyrolized material. The ablative layer may be applied to the combustor by centri-
fugal action or by pressing depending on the viscosity of the uncured material.

Insulation thickness should be kept low since it determines the diameter being allowed for the integrated
booster. An Increase in insulation thickness results in an increase of booster length referring to a fixed
booster mass and volume. The required thickness of the insulation depends on

- heat flux into insulation (depending on ram-combustor operating conditions)

- maximum temperature allowed for the structure
- operation time of the engine
. ablation rate
- insulating properties of the pyrolized matrix
. rigidity of the pyrolized matrix

. Problems have been experienced concerning the last item with CRJ's. The pyrolized insulator was damaged by
* combustion instabilities occuring for equivalence ratios near the stoichiometric figure. Although the risk
of combustion instabilities appears to be inferior for the SPR (according to available experience), the
mentioned problem has to be considered for the design of the ram-combustor insulation.
 7-17
L, j
3.4.3 Thrust Nozzle! 10, 11.,
All main aspects of thrust nozzle design for a SPR correspond to CRJ design. However, two problems have to
be considered additionally which may arise when the combustion products of the ram-combustor include a
significant portion of condensed material. This will occur when metals are used as high energy fuel components.
In this case
- there will be a moderate risk of erosion or depositing which both may degrade the contour of the thrust
nozzle
- two phase flow effects will increase expansion losses.

Both effects are not very detrimental to engine performance and may be minimized by adequate design of the
convergent and divergent nozzle contours. Tite thrust nozzle is exposed to high thenial stress for a long
duration. Consequently, the application of temperature resistant materials is mandatory. High melting metals
cannot be used for the thrust nozzle due to the unacceptable weight penalty. Since large ceramic parts are
extremely expensive and still imply a considerable manufacturing risk these are net exhibiting a proper
solution either. Consequently,
- steel nozzles with thermally resistent coatings (spinels, zirconium-oxide)
- graphite nozzles i with ceramic coating or impregnation
- CFC nozzles
should be applied.

3.5 INTEGRATED BOOSTER AND TRANSITION

The subject and the inherent problem of integrated booster and boost-sustain transition phase are common
to all ram-type engines. The ignition of the sustain propellant and the subsequent pressurizing of the gas
generator are the only particular characteristics of the SPR transition. Since a short time span (usually
below 200 milliseconds) is needed for this action it has to be included adequately in the transition sequence
of
- booster burn-out
- booster nozzle separation
- opening of the port covers
. ignition of ram-combustion.

An additional device to ignite ram-combustion will not e needed for the high energy SPR, due to its auto-
ignition feature. If an ignition device is required for a low or medium energy SPR system the same install-
ation may be applied as for the CRJ.

4. DEVELOPMENT HISTORY AND STATE OF THE ART [ 1, 2, 4, 12, 13, 14, 20,
At the end of this lecture, some historical highlights of SPR development shall be presented along with a
short summary of the present state of the a,'t and anticipated future applications. Several SPR systems of
different size were flight tested since 1955. One SPR system is operational in the Soviet SAM 6 missile.
These flight vehicles are shown by figure 4.1. Available technical data are summarized below.

o 1954/1955: Four SPR fliGht tests carried out by Thiokol

- flight vehicle: diam: 2 in (50 am); length: approx. 28 in (710 mm)
- SPR engine: annular scoop air intake; fixed flow gas generator
- tardem booster; ground launched

o 1958: Three flight tests of FDR3 by Thiokol

- flight vehicle: diam: 6,62 in (168 mm); length: approx. 4 ft 7 in (1,4 m)
- SPR engine: conical nose air intake; coaxial ram-combustor configuration, fixed flow gas
generator
- tandem booster; ground launched

o 1966: Two flight tests of SPARN (solid propellant augmented rocket motor)
- flight vehicle: modified airframe of the AQM37A drone
dtam.: 13 in (330 mm), length approx. 12 ft (3,65 m)
(length without pitot probe and booster nozzle)
- SPR engine: 2-dim ventral air intake; side dump ram-combustor; staged propellant to
demonstrate two level sustained thrust.
. integrated booster, air launched
- flight envelope: Mach 2.5; flight altitude 45000 ft (13 km); burn time 182 sec; range 74 nm (135 km)

o 1967: Soviet SAM 6/Gainful becomes operational

- type:
"- configuration:
- length:
- diameter:
- weight:
- SPR engine:
Surface-to-air tactical anti-aircraft weapon
Slender cylindrical body with pointed nose, small cruciform wings near mid-
length and cruciform tail fins, four air inlet ducts between wings
20 ft 4 in (6.2 m)
I ft (33 cm), span (max): 4 ft I in (1.24 m)
1212 lb (550 kg)
four lateral trailing shock air intakes; side dump ram-combustor;
r fixed
 7-18

a 1976: Two flight tests of an SPR experimental missile carried out by

the French ONERA
- flight vehicle: diam.: 15,75 in (400 mm), length 18 ft (5.5 m), weight 14301b (650 kg)
- SPR engine: four lateral bi-conic air intakes; side- dump ram-combustor, rod and tube
grain configuration; medium energy self-pyrolizing propellant, fixed flow
gas generator w
tandem booster, ground launched
- flight envelope: ballistic flight; culmination 8600 ft (3 km) with Mach 2.1, approx. 50 sec of
sustained flight; range 19 nm (35 km)

o 1982: Two test flights of the EFA experimental missile carried out by the
German MBB (Messerschmitt-Bdlkow-Blohm GmbH) company
- flight vehicle: diam.: 9.45 in (240 man), length 12 ft (4.2 m)
SPR engine: four lateral half-axisymmetric air intakes, side dump ram-combustor configuration;
endburning grain; high energy propellant (40 % boron loading); fixed flow
gas generator
. tandem booster, ground launched
. flight envelope: one ballistic and one guided flight; altitude approx. 4300 ft (1500 M);
Mach 2.5;h1gh gside maneuvers during second flight

o scheduled: DR-PTV (ducted rocket propulsion test vehicle) by Hughes

- flight vehicle: diam.: 7 in (178 mm)
- SPR engine: two 2-dim lateral air intakes; side dump ram-combustor configuration;
endburning grain, medium energy propellant; fixed flow gas generator
- integral nozzleless booster; air-launched
- in the VFDR (variable flow ducted rocket) program, initiated in 1982, a throttleable propellant
of increased energy content will be developed for use in the DR-PTV system.

The present state of SPR technology shall be illustrated by some figures referring to
- propellant energy content
- propellant throtteability
- ram-combustion efficiency

Concerning the energy "ontent of the propellant, a hypothetical highly metallized formulation containing
50 % boron
5 % aluminum
20 to 25 % AP
25 to 20 % binder

and exhibiting heating values of B u.

1700 - 1800 -Iu 40 - 42

- ooTn"
1050 Btu / 68 BP
NJ %.

may be consider?d to be the present feasibility limit referring to processability, burning characteristics
and expulsion efficiency. Assuming reasonable parameters for performance calculation, a flight Mach number
of 2.5 and sea-level flight, a specific impulse of 14000 m/s can be attributed to a SPR using the afore-
mentioned propellant. This figure ;hould only be understood as a rough hint for performance level.

As already mentioned in para. 2, turndown ratios up to 18 have been demonstrated for SPR gas generator
propellants.Nevertheless, referring to high energy propellant formulations, available information indicates
that turndown ratios between 5 and 10 more likely present the current state of the art.

Referring to afterburning, there exists considerable experience with low energy propellantswlthout solid
fuel additives and with highly metallized high energy propellants (up to 40 percent boron loading). Less
experience is available for medium energy propellants with solid carbon or hydrocarbon additives and low
to zero metallic loading. However, increasing emphasis is placed on the investigation of this type of SPR
propellants. Summarizing the available experience, the following figures may be indicated as state of the
art (temperature rise) afterburning efficiencies:

- low energy propellants : > 90 %

- medium energy propellants
(carbon or hydrocarbon fuel additives : 85 to 95 X (depending on ram-combustor operating conditions)
low to zero metal content)
- high energy propellant : 80 to 95 % (depending on ram-combustor operating conditions)

Problems concerning the design of control valves, injectors and thrust nozzles that arise from high temperature .. ,
particle-laden combustion products can be handled for state of the art propellants. However, sophisticated
design and ultimate use of modern materials and manufacturing technology are required for the construction
of these components.

Finally, some possible future applications of the solid propellant ram rocket shall be reviewed. As already
indicated in the introduction, the SPR is a favorable candidate for the propulsion system of medium to high
ranan tactirAl mitti1o nf thp noit nanaratinn In thie fiald n# riols +h& CDO r&+t 4,I.k rO.110 .. A Cr,P.
 7-19
0
The SPR provides the general advantages of ram-type propulsion referring to range, sustained supersonic .
speed and the capability to perform high g maneuvers even in the terminal flight phase. While the CRJ may
be preferred for extreme range requirements and for reusable vehicles and the SFRJ may be preferably
app ied where simplicity, small size and low cost are the predominant requirements, there is a considerable.
domain of mission profiles where the SPR features are equivalent or superior to those of its competitors.
These features are:

direct control capability along with relative simplicity

autoignition capability - reduced or eliminated flame out problems
- all solldifuels
high volumetric heating value, application of high density fuels . -

The SPR may replace the solid rocket propulsion unit of operational missiles to upgrade their performance.
Here the SPR is preferable when tne CRJ is incompatible with logistic considerations (e.g. liquid fuel) and .-
the performance and system flexibility of the SFRJ are insufficient.

By now, SPR systems are in discussion or under development for the following applicaticns:

Surface-to-surface/antiship: A high energy SPR propulsion unit (40 percent boron loading) is under develop- -.
ment at MBB/FRG for use in the ANS (anti-navire-supersonique) missile. ANS .
cf. fig. 4.2) is a Frennh-German cooperative project carried out by SNIPS
socteta national tndustrtelle abrospatiale) and MBB. ANS shall replace ___-__-_

missiles 1ike Exocet, Otomat orHarpoon in the 19901s.

Surface-to-air: SPR application to future medium range SAM systems is discussed in the
European part of NATO. Even for the short range low altitude surface-to-air "
mission, the application of the SPR provides advantages. A study considering " -x.'.9.
a SPR Roland missile (cf. fig. 4.3) provides a significant potential for range ,.,**
increase and a tailchase capability which does not exist for the rocket-powered .
missile.

Air-to-surface: In this field, SPR application is considered in the USA. An air-to-surface

mission with cruise flight at high altitude and terminal dive to the target may
be accomplishei using a simple fixed flow SPR system.

Air-to-air: Medium range air-to-air applicati1s for the FR are discussed in Europe and in
the USA. Possibly, a throtteable variant of the DR-PTV may lead to a future
AMRAAN (advanced medium range air-to-air missile) propulsion unit.

5. CONCLUDING REMARKS

This lecture reviewed design and performance characteristics of the solid propellant ram rocket. The advanced
technical standard that has been presently achieved for this propulsion system allows its application with
moderate Ievelopment expense and relatively low technical risk.

While no general design or selection criteria can be provided for a missile propulsion system, the lecture
tried to indicate major advantages as well as problematic aspects of the SPR system to be considered for
propulsion system selection and to present a rough guide to performance tailoring and component design.

The authorwishes to express his particular thanks to All -,ers of the MBB-UA airbreathtng propulsion group ....-
for their advice and assistance to compile all the inforrr-.1on that formed the basis of this lecture.

REFERENCES
1i Anonymous
Presentation to FMOD/MB .'
United Technologies - Chemical Systems Division, Sunnyvale Calif. Juli 1980 "'

[2] Besser, H.L.; Strecker, R; Weinreich, H.-L.

Advanced Variable Flow Ducted Rocket Prupellart Development
Final Report - MBB TN-RT31-7/83, Nov. 1983 V,

13 ] Anonymous
Advanced Ducted Rocket Gas Generator ... ',
Volume I, Technical Proposal
United Technologies - Chemical Systems Division, Sunnyvale Calif. May 1982
[4 1 Besser, H.-L.
Investigation of Boron Solid Propellant Combustion in Ducted Rockets
AGARD-CP-307, Ramjets and Ramrockets fUr Military Applications
London, Oct. 1981

5 Strecker, R.
Private communication about gas generator propellants r -,.

[6] Gordor., S; McBride, B. .. ,,

Computer Program for Calculation of Complex Chemical Equilibrium Compositions, -
O..b*D- -Tn.. 7 i Ant ani Rflarta1 .*hndke and r.hanmin-,jnucumt-rsetnnations
 /p
7-20

' Thomaler, 0.
Untersuchung eines nicht-iinearen Regelkonzepts fUr den Gasgenerator des Feststau-Antrlebs
MBB-TN-RT31-18/78, June 1978
I B] Thomriaer, 0.
Private conmmunication about fuel flow control concepts for the solid propellant ramirocket
191 Engi, E.; Engel, H. .+ o

Private coammunication about control valve design

1 101 M4enkel, H.
Private coammunication about temaperature resirtant materials and ablators
(i Crispin, B. et aI
Feststau-Ra keteiiantri eb
IUB-Pr~sentation vor dem BM~g und BW8, June 1976 *

1121 Weinreich, H.-L.

Skrlptum der Vorlesung ORaketen- und Stauantrlebe"
Hochschule der Bundeswehr, 8014 Neublberg, 19P.4
113./ Anonymous
History of the Ducted Rocket
Thiokol

*1141 Marguet, R; Ecary, C; Cazin, P.

Studies and Tests of Rocket Ramnjets for Missile Propulsion
4th Imt. Symp. on Air-Breathing Engines
Lake Buena Vista, Orlando, Florida, April 1 -6, 1979

15] Henkel, H.
Experimentel1e Untersuchung von Gasvertei1em und Brennkat mererorobung plasmabeschichteter
SchubdUsen fUr die Feststoff-Staustrahlrakete als Vorbereixung der Funktionsuntersuchungen
des Windkanal-Versuchstriebwerks
MBB TN-RCs-35/71, Nov. 1971

( 161 Engl, E.
Gasverteiler und Mischsysteme bei Staustrahlraketen mit vier Diffusoren
MhuB T r-RT34-6/76,
May 1976

1 171 nThonayer, 0.
Feststau-Brennversuche mit 226 Triebwerken im "Connected-pipe3 - und Freistrahi-Betrieb .
MBB Th-RT31-20/76, Nok. 1)76

1 181 Besser H.-L.

;
Untersuchungen zur Optimierung der Luft- und BrennstoffzufUhrung am Feststau-Antrieb bei
Verwendung von Bor-Treibstoffen
MBB TN-RT31-35/87, Nov. 1980

1 191 Jell, C.S.

Air Intake Aerodynamics and 18Operational ] and Installation
esser H.-4.-.i.
Effects on Missile Powerplant
Performance
AGARD LS-98/14 Supplement Missile Aerodynamics, March 12-16, 1979
[201] Jane's Untesucungn
derLuf-zu ud
Opimirun
Bennsoffuf~rungam esttauAntreb ei :'r
Pocket Bookrn 10 Missiles
Verwedungyon or-Teibsoffe

MBBTN-T31ZO/0, ov.198
 7-21

Fuel Tank Rom Cowoustor IConventional

Liquid Fuel
~ amist Fig. 2.1: Comparisonl
(CRJ) of Different Ram-Type
Propulsion Systems
Air Intake .f 6Fuel Flow Control and Injection

Primary Comb. Chamber

(Gas Generator) I Secondary Comnb.Chambe
(Ram Combustor)
Solid *

pDucted
Propellant

Rocket .
Airinlke Oxygen lieficient Fuel Flow Control
Solid Propellant and Injection4

Rom Combustor

Solid
uci

Air intake Solid Fuel Mixer Vane Aft Mixing

Chamber

hi eergy SPR
MbC. -
meimeerg SP ig. 2.2: Coimbust onl
2000. Temperatures of Dferent
2000

1500.

Equivalence Ratio (Air :Fuel)

- Propellant,
- Grain Configuration
Rom Combustor

'4

Fig. 3.1: Components of a Solid

Propellant Ramrocket (SPR)

Air Injection Thrust Nozzle

Thermal Protection

- Fuel1Flow Control
- Fuel Injection
7-22

PI

*F* .

SPR~W Prpusin.yse

0~ m~ mtnQvt,
-
~
Ualtki~Ow"/
B~4Pit.
2
Usd"PeuaPoucsga~Cnmt Rst~ rdc
Niroiw OOOr222s.

SPR ropulsioSyte

mmnw N1Chemca Nd,~~c

N 2,tk 53.5
Ce %Pamo
perxidbat Plr 11PodcsCnetS RatonPout0
Sizodiwi NCIaO4 e 2.22 52.20 /3

Perchlorate
Lithsium LIC104 LICI 2.53 60.1 1 M10231
pardhlorate

Ammonium NHNO03 N2 , H1,2 1.93 60.0

necitrate Hj, .2

Sodsium KNO10
3 NI20 2.34 32.2 1933603
pnitrate

Lihium LHN 0)2 3 L10 .39 63.231

Amoniu NfCH 02
2 %N 32 20 1.73 33.2

en rnitrate n)
 7-23

Amount of suma Mamdfac-

Utsndamedchtarac. turft
Oxidizer species Perforrmwne tefistics Allini Experience Overall
Nitroniwn .
Perchlorate
Nitrosyl -- -- Fig. 3.3b: Oxidizer Se~lection Trade

Lithium -. 0 0
perchiorste
Ammonium .

sodium,. 0
perchloroate
potassium .- 0 .0
perchlorate
Lithium * .. *
nitrate
Ammoniumiw .. * 0
nitrate
Sod ium -- 0 - 0-
nitrate

Potasslium -. 0 -0
nitrate
HMX .. 0 -. 0 0
RDX . i 0 -- 0
Legends .very good,.good, 0fair,-marginal, - bad
Notat HMX . cyciotatrametttylene tetranitramine, ROX *cycionite (cyclotrimethyl-
ene trinitramlne)

Ingredient Oxygen Requiremet Density Heat of Combustion State at

gram oxygen/lb/Ibn3 g/cc 18tu/Ib kJ/kg 181ln kJ/dM3' mt~ent
_______fuel_ condition

Hydroxyl-terminaetd polybutaditnt 3.19. 0.0339 0.930 1104S. 61972. 606.3 39036. I

jHTP8)
Carboxyl-tarmlnstied polybutsditne 3.20 0.0347 0.960 179''. 41666. 621.5 40014. 1

Isodecylpelargonat (lOP) 1.99 0.0311 0.660 16126. 37506. 501.9 32288. I

Polylsobutylene (Oppanvol) 3.62 0.0329 0.910 10713. 636. 1.6 93.
Catocane 2.56 0.04666 1.29 16297. 33161. 664.6 62776. 1
Carbon amorphous 2.66 0.0117 1.62S 14095. 32786. 627.6 S32)1. a
Graphitt 2.65 0.0813 2.2S0 14095. 32784. 1165.9 73777. $
Polydlcyclopentadiene 3.1s 0.0391 1.100 17533S. 40786. 696.1 64817. a
(Zecor*2)
Cross-linked polystyrene (CLPS) 3.07 0.0331 1.046 17139. 19869. 69. 620612.
Poly-..Ialpha-unethylstyroen) (PAMS) 3.11 0.0357 1.07 1731, 0279. 670.2 631530. s

Fig. 3.4: Properties of Carbon and Hydrocarbon Fuels

At

C
RP-1
04 80120160 Fig. 3.5: Heating Values of Different
x101 CKJ/Kg I CrsvImeirIc ,.,
1
7-24

Fue Ad~tve ODns ity heat ofCoftustlon

lb/in'
3
9/cm
p Btu/1b 6.3/kg Btu/in- k{ O
a (amorphous) 0.0802 2.22 25 486. 59 280. 2044. 131602.

Al 0.0975 2.70 13 356. 31 066. 1302. 83 878.

Hg 0.0629 1.74 10 636. 24 739. 669. 43 046. .

Zr 0.2345 6.49 5 177. 12 043. 1214. I 78 162. *~

Ti 0.1626 4.5 8 487. 19 744. 1380. 88 849.

4C0.0903 2.5 22 159. 5132. 200. 118831.

Fig. 3.6: rr~operties of Different Metallic Fuels

40

2, 4 Fig. 3.7: Density and Heating .

l~T. Voaluesof Different Metallized

35 . and Non-Metallized Propellant
Formul~ations

30 t0

688 7.

% -,

530

%U

20.... 10 2f

Neo.orl 1AP 2on3en:Ammonium~o

44 680
 7-25.

C 870 T O

60 K1700

320 1 is
3000
20 324030

Amonus pe6hort co1900 Knt

Fig.~3
ofaLeal
3.0 dPoeln
Sceein

- -60 ulFo eraeI K Fe lwices

>/d .........

0*3

Expnen on9threnVaiatio of K-Reatlioe Sepel

nsiiePoplat

5 F~~~~~~~~surelFoDeraeFe FlwIc as

Pressureur

~~ ~
Fig. ~ ~ ~ ~ ~3threenecofthacivbltO
nPrsuead ~ ~ ~ ~ ~
eprtreBakt ~ ov

Iare
1-010
7-26

8,7

0.7

0,6

0.6 ..

a's

a.~

0,6

0.10 0.1$ 0.20 0.25 (%/iFj 0.10 0.15 0.20 0.25

0,1$ 0.27 0.36 0.65 !%IKJ 0.16 0.27 0. 3b 0.45

Propellant Temperature Sensitivity arp

a9nle control valve control valve *second sonic nozzle

Fig. 3.13: Parameters Limiting the Achievable TDR *-

Fig. 3.14: (1)Rotating Blade/Butterfly Valve (2) Translating Plunger

 7-27

Fig. 3.14. (5)Gate Valve (6) Disk Valve

A0 .

Fig. 3.15: Retractable Siher 0,' Fig. 3.17: Coaxial Propellants

Wiresf

Fig. 3.16: Matrix Grain

~~0c

0
 7-2)t

Four Inlet Side Dump~

Design Study)
(P485

~ Two Inlet Side Dump

Gas generator device

BoosIerlgaa generator arm fire

(OR PTV Ramcoetustor

*Inlector Propellant

-- Coaxial

* (143Winidtunnel-tested
~- ~ ~
_ __22A- ~ __ IEngine; Half-Integrated
~ Configuration)

Fig. 3.19: Ramconbustor (c'nf igurations

 7-29

Air ItakeInjector

Gas Generator Ran.-Combustor

THIOKOL -FDR3 (1958) [ 13]

BOOSTER HEAD-END DOME -

SUSTAINER GAS GENERATOR- BOOSTER NOZZLE

BOOSTER- AFTERBURNER CHAMBE

.
AIR INLET

THIOKOL -SPARI4 (1966) [13]

*USSR -SAM 6 (1967) [20]

Guidance and Control Air intake Injectors

Gas Cenerator Rom-Combustor

Flight Test
-.4 instrumentation
ONERA - Experimental Missile (1976) C 14 ]

Fig. 4.1a: Operational or Flight Tested SPR Systems

7_30)

Nose Cone Electronics Modlule Ramrocket Propulsion LUnit

Recovery System (Guidance an
IControl, Telenietry)i

MBB -EFA Experimental Missile (1981) [ 4]

Gas generator Right air Induction unit

Gas Port cover

Booster
combustor
* ~-Saet and arm

Pnasue fittings Booster igniter l

Inlet turn Pressure

*and dump transducers

Left air Induction unit

Hughes - DR PTV (scheduled) [ 1

Fig. A.1b: Operational or Flight Tested SPR Systemt (no uniform scale)

Fig. 4.2: SNIAS/MBB Anti-Navire-Supersonic (ANS).

Anti-Ship Missile Project

Fig. 4.3: SPR Powered Roland Missile; MBB Design Study

 8-I

RANJETS WITH SUP'ERSON4IC COMBUSTION

Frederick S. Billig
The Johns Hopkins University -
Applied Physics Laboratory
Johns Hopkins Road
Laurel, Maryland 20707 ,j.

SUMMARY

A terse look at the requirements for high speed-long endurance missiles and the virtues of airbreathing
propulsion are presented to support arguments for developing ramjet* that employ supersonic combustion. A
review of generel classes of inlet configurations leads to the selection of multiple inward turning scoop
(ITS) configirations to serve as the basis for engine cycle analysis. A fixed geometry ITS inlet is shown not
only to have an attrcctive performance potential, but to be quite versatile for use in a variety of engine
configurations. When it is designed to have full air capture at a flight Mach number H somewhat below the
cruise Mach number, it can produce high pressure recovery over a wide M0 range with high air capture during
climbout at low angle of attack a as well as dur.ng cruise at near optimal cz. The geometry of a family of ITS
inlets is examined to determine optimal configurations for various constraints and to provide quantitative
values for the flow properties for entering the combustor analysis.

A flow model for the combustion process that includes the major features that have been observed in
experiments serves as the basis for determining the pressure rise in the shock train that precedes combustion
and the conditions at the combustor exit. Calculationo for simplifiPd frictionless flow with constant specific
heats show that 1) the combustor area ratio is set by the acceleration requirements at the end of boost; 2)
the combustor operates in a subsonic mode at low M and in a supersonic mode at high M0 ; and 3) optimal per-
0o
formance occurs when the combustor inlet conditions correspond to moderate inlet contraction ratios. Computed
values for the net engine force coefficient show that forebody drag must also be considered to determine the
optimal overall engine configuration.

NOMENCLATURE

A cross-sectional area
A effective flow area

A physical area
A projected area of inlet
Ai surface area
w
CD forebody drag coefficient - D/qoA Rd,.
CDADD additive drag coefficient - DADD/qoAR 4.

C net force coefficient

Cf mean skin friction coefficient
CT thrust coefficient
D drag
DADD additive drag
D4 diameter or hydraulic diameter at station 4

ER equivalence ratio
E~eff eofective equivalence ratio - ER*-c-
f fuel/air weight ratio
f stoichiomatric fuel/air weight ratio
F net forne - T-D-DDADD
t
stream thrust - pA + pAu
kI constant in Eqns. 10, 11 -

L lift
X ?4ch number
MFC ~Machnu..,ber at which inlet obtains full capture at m -0*
M molecular weight , -
p pressure
q dynamic pressure
R gas constant
s axial distance from origin oZ shock train
axial distanca between fuel ports and end of shock train
T temoerotura
 /

8-2 r

a
a inner of
angle attack
scoop angle with respect to missile axis

y ratio of specific heats

6s boundary layer thickness
- boundary layer dioplacement thickness
6c half-angle of conical compression surface
C Crocco parameter for pA processes
Crocco parameter for processes between stations s and a
ic combustion efficiency *

inlet kinctic energy efficiency

nozzle efficiency
0 boundary layer momentum thickness
0 flight path angle
angle between cone axis and ray passing through scoop
8w leading edge
P density
wall shearing stress
V
* roll angle
*b smile angle of ITS scoop
sector angle of scoop duct
ow flow angle along conical ray
Subscripts

0 free stream conditions

1 conditions ahead of scoop lip shock
2 conditions in plane of crotch in side plate
3 conditions at end of supersonic compression processes
conditions corresponding to downstream of a normal shock
at station 3
4 conditions ahead of shock train
5 conditions at combustor exit
6 conditions ct nozzle exit
c conditions on cone surface
CR flight conditions corresponding to cruise
a effective; conditions at origin of - constant heat addition
ROB flight conditions corresponding to end of boost
f fuel
s conditions at downstream end of shock train
w wall

Superscript

mean value
conditions downstream of normal shock
conditions at the sonic point .

INTRODUCTION

The attractiveness of the supersonic combustion ramjet as the powesrplant fur a high-speed missile for a
has been recognized for more than twenty-five years. Arguments supporting the use of the scramJet cycle for a
missile and conparimon of performance with engines using subsonic combuition were presentee in Ref. 1.
The effective defense against future air-to-surface missiles, on-the-de k attacks, and IRBW attacks will
require major reduction in time-to-intercept of the defense missiles, compared with presently available
surface-to-air missiles. For the foreseeable future, effective intercept of ballistic missile warheads,
as well as air-supported attacks, will occur within the atmosphere, becatse the atmosphere provides the
best means for distinguishing between the ballistic warhead which must be destroyed and extraneous re-
entering material accompanying the warhead to which firepower must not be diverted. The margin of superiority
of ramjet missiles to rockets increases rapidly as missile speed increases. To provide equivalent perfor-
mance to a Mach 7 supersonic combustion ramjet vehicle at sea level, a rocket would have to have roughly
three times as much weight. Included in the open literature (see e.g. o Ref. 2-4) are a number of super-
sonic combustion ramjets that have progressed through the stage of free-jet engines tests.

Before proceeding into the main discussion of this paper, some clarification in semantics ti needed to
differentiate between the conventional subsonic combustion ramjet (CRJ), the scramjet and a hybrid cycle, %,.'.
the dual-combustion ramjet (DCR). As used herein, the CRJ is an engine cycle wherein there is a normal Y
V.
shock structure in the inlet compression system for a11 operating conditions and all of the heat addition '.
takes place In subsonic flow. All of these ramjets must be boosted to some supersonic end-of-boost Mach
 8-3'

number (MOg) by a rocket engine. ScramJet will refer to an engine cycle which over at least soum :,rt of '.
the Mach (HO) altitude (Z) equivalence ratio (ER) operating envelope, heat is released in supersi, iLlow.
Host practical scrsaJet engines for missile applications must operate at high ER upon the ramjet tai.- -!-r
at
ad Y',, <- 4, and at this modest speed a normal shock structure will be present and some, if not all, ! A
heat will be released in subsonic flow. This cycle is often referred to as a dual-mode engine, but ", t -:.
paper no distinction wvll be made from the scraujet. Dual-mode should not be confused with the UCR, 1'.,
is a hybrid engine cycle that combines a CRJ with a scranjet (Ref. 5). The moderate H0p for the scra.. 7A
or DCR is a consequence of the optimization of a tandem or integral-rocket-boosted, fixed-geometry scrmj-.
system which dictates that 0.55 MCR < MO 1 < 0.6 H, where MN is the Hach number during high-altitude
cruise. Typically, 6 < 14R
< 8; thus MgOB 0 3-4. Since cost and complexity are key considerations in
determining missile feasibilty, variable geometry and active cooling of the airframe are not practical.

Schematic llustrationa of the ecrmajet and DCR are shown in Fig. 1. Nose entry configurations are .
shown, but aft entry configurations such as the so-called underslung engine (Ref. 6) have also received
considerable attenticn. Inlets for these engines are of the external-internal compreasion type wherein the
pressure rise it obtained via a series of weak shocks and regions of isentropic compreosion, a portion of
which occurs in the internal ducting. Typical inlet contraction ratios, A1 /A 3 , vary from 4 to 8 (see Rcf.

7). Whereas the inlet flow is fully contracted at station 3. additional length must be added in an isolator
section to prevent combustion-ind.ced disturbances from causing adverse interactions in the inlet (Ref. 8).
The combustor generally has a diverging area to permit high heat release at low H0. In some configurations,
a step increase in cross-sectional area at the combustor entrance has beeT, used (see e.g., sfef. 9, 10) to
help mitigate combustor-inlet interactions and to enhance flame-holding. '.he exhaust nozzle has either sa- .
straight or a contoured shape and the demarcation between cWabustor end nozzle is the change in shape rather..
than the point of completion of heat release. All of the fuel can be added at the combustor entrance or
fuel injectors can be staged at se,-eral axial stations. The latter method can be used as a means of adjusting
thu effective area ratio of t1 combustor thereby enhancing p. 'formance, but at the expense of weight,
complexity and cost. Instream injectors have been given consideration, buit with some trepidation due to the
difficulty in providing adequate structural integrity. Injection ftom ports in the will is preferred. 4

If the scramjet must operate at H0 < 4.5, which is generally the case, the static temperature in the
combustor is below the autoignition point of conventional hydrocarbon fuels such as JP-5 and RJ-5. For .. '
these, some type of piloting system is needed end piloting has proven difficult. Even when ignition has
been obtained, flame spreading has been minimal and reaction rates slow compared with typical residence times
of 1-2 ms. Highly reactive fuels such as HiCal 3-D (a blend of mono and higher athyldecaboranes) and penta- - *

borane-hydrocarbon blends autoignite and burn rapidly at static temperatures at least as low as that corres-
ponding to H0 - 3.5 and have been used in frea-jet engine test demonstrations of the ecramjtt (Ref. 2).
Unfortunately, these fuels are expensive, toxic, pyrophoric at room conditions, and no longer in production.
It was these limitations which led to the concept of the DCR. '

The DCR is conceptually a scramjet with a large pilot burner. The Inlet airflow is split so that 20-
25 percent is directed into a "dump-type" subsonic combustor, herein called the gas generator. Because the
flow is subsonic, the static temperatures are higher and ignition and combusticn can be obtained as in any
well designed CRJ. From both an engine cycle efficiency standpoint and a design simplicity standpoint, it is
advantageous to add all of the fuel in the gas generator and eliminate the need for a fuel injection system "70
for the main supersonic combustor. This moans that for all overall engine iuel/.air equivalence ratio (ER)
requirements X:eater than the value corresponding to the fractional air flow split (gas generator/total), the
gee generator operates in excess of stoichiometric. The bulk of the air captured by the inlet is ducted
supersonically around the gas generator to the main combustor entrance. Refer to Fig. lb. The downstream
portion of this duct serves as a combustor-inlet isolator. The hot, genervlly fuel-rich effluent from the gas
generator mixes and reacts with the main flow before passing through the exhaust nozzle. Presuming that the
gas generator can operate efficiently over a wide ER range, the designer can tailor the air split and the area S
of the discharge port of the gas generator to obtain optimal performeaice for a given application.

INLET DESICN

before presenting in detail the current understanding of the physical processes and the state-of-the-art
of the analytical tools that are used to analyze superuoni: combustion, various concepts for scrsmjet and DCR
inlet designs will be discussed. The literature is rich (see e.g., Ref. 11) witt the results of engine cycle -. .,
calculations that give the reqtirements foe ,,ir capture, inlet compression ratio and efficiency that result in
a n.ompatitive propulsion cycle. Some genera! guidelines are: the inlet contraction ratio should be sufficient
to p twide 0.35 < 133 /M0 . 0.50; at EOB ,;he Air cApture ratio A0 /Ai should be no less than 0.6 to 0.7 of the
maximua air capture ratio; the inlet kinetic energy efficiency prior to any shock structure in the isolator
should be no less than 0.95; the cowl Lip aon waie drag must be low; and operation at angles-of-att,-ck, U, of
12" or greater should be tolerated without c usi'l engine unstart. These requirements have been paramount in
the design strategies employed in the develo mant of the three types of inlets shown in Fig. 2.

Figure 2a is a modular or Busemann-type inlet wherein the flowfield into each inlet module is in effect a
section of a fully internally contracting flowfield that has been cut along srrsamlines and.turned inside out.
The swept "cowl" leading edges are shaped to follow the intersection of the first compression wave with the
captured free sLreastube when the inlet is operating tt its design H and at a w 0. The sweep angles ate
typically greater than 60, therefore the leading edge drag and heat transfer rates are minimal. Theoretical -" '
tools have been developed (see e.g., Ref. 12) which can provide contours for compression surfaces for any
desired contraction ratio up to that corresponding to H3 - 1. Practical designs are limited to contraction

I I---II-II I I-II--lI IS
8-4

SUPERSONIC
COMBUSTOR
OBLIQUE FUEL INJECTION
SHOCKS
mz

>~3PRECOMBUSTION
SHOCK STRUCTURE

a) Supersonic combustion ram jet (scram jet)

COOMBUSTOR
AIIR INLET
(TTYP.)

COWLBCROTR

ABLEE FLOWT

Mo

a) SECIDEAVIEW OFMHOCKL TRAPSEATYPE INLETe

SCOOP SIOEPLATE

M.

0 st e*s,.a n. . ...- M3l

 8-5S:_=_'' .J

contraction than would be allowable in the fully internally contracting counterpart because the controlling
area to meet the starting criterion is the cross-sectional area of the auct at the crotch station rathe.r than
the cross-sectional area at the cowl leading edge station.

Designs can be developed to bring a cylindrical capture tube into any number of modules, but most atten-
tion has been given to three and four module designs, as typified by the photographs shown in Fig. 3 of inlets
tested over the range of Mach 4-10. The modular design brings the flow to the periphery of the engine and

SCRAM INLETS

3-MODULE INLET MODEL 4-MODULE INLET QUADRANT MODEL

3-MODULE INLET EVALUATED IN 7 WIND TUNNEL TESTS AT Mo - 4.0,15.3,10.0,74. AND 8.1

4-MODULE INLET EVALUATED IN 6 WIND TUNNEL TESTS AT Mo - 4.0, 5.3, 7.8, 8.1 AND 10.0

Fig. 3 Photographs of Busemann type hypersonic iniets.

cannot in supersonic flow practically be brought back to a central duct combustor. Thus, the designer has two
choices: either each inlet can be used to supply a separate modular scramJet combustor, or the flow from each
module can be spread circumferentially in a contoured duct to bring the flow to a common annular combustor. %
In the free jet test previously cited (Ref. 2), a three-module inlet was used to supply three separate comn-
bustors, and the exhaust nozzle was shaped as the inverse of the inlet to bring the three streams together in
the nozzle exit plane. For such a design the modules function separately as three independent engines and by
using differential fuel-flows it becomes possible to obtain thrust vector control.
Figure 2b shows the so-called "shock trap" inlet that has been examined for use in both the CRJ and the
scraujet. Annular and chin-type nose and half annular aft designs have been examined. In these configura-
tions, the flow is first turned outward on the external compression surface and then turned toward tha axis
through a relatively strong shock emanating from the cowl lip. Ideally, the external surface of the cowl lip
is aligned with the body axis to eliminate cowl wave drag at a - 0. To mitigate the adverse effects that are
caised when the strong shock strikes the boundary layer that has built up on the compression ramp, a large
bleed slot is added. When used in the CRJ cycle, the bleed acts as an "educated" &lct to provide additional O
inlet margin. The slot's intelligence is ascribed to its self-regulating flow characteristics. When the
disturbance in the region of the slot is weak, the local pressure is low, and the bleed flow is minimal,
thereby matching a low mass flow requirement. When the disturbance is strong, as is the case when the inlet
approaches critical operation with the normal shock positioned in the region of the cowl lip plane, the pres-
sure is high and the bleed flow is correspondingly high as required. Whereas bleed can increase the maximum
pressure recovery and inlet margin, it generally reduces the effective air capture of the inlet, requires
additional ducting, and incurs a loss in momentum (bleed drag). Various bleed configurations, including
slots, holes and scoops, have also been used to vent low energy air in other inlet configurations. An alter-
native technique for boundary layer control is the diverter. Here, the inboard surface of the inlet duct Is
raised above the boundary layer on external compression surface, or above the body boundary layer in an aft
entry configuration, and the low energy air is diverted laterally. This method is restricted to segmented or
fractional entry inlets such as the chin, where lateral spill is possible. In all cases, a detailed composite
design and performance analysis is needed to assess whether the benefits derived outweigh the penalties in-
curred.

Figure 2c shows an innoxative inlet concept that has been given the name "inward turning scoop" (ITS).
Multiple scoops, each capturing a sector of the flow are used and the area between the scoops is used for
venting during starting and for diverting boundary layer air when appropriate. Three or more scoops can be ' - .
arranged symmetrically to produce overa.l inlet characteristics that are quite insensitive to roll angle,
th-reby permitting the use of skid-to-turn control. On titeother hand, two or more scoops located on the
windward side provide an alternative configuration to the chin-type inlet for a bank-to-turn vehicle. In the
latter case, maneuvering is made in the meridional plane of the inlet and the designer can take advantage of
the additional compression on the windward surface. As in the shock-trap design, the external cowl lip
.. - - -,.1. .. 4.t. , .. 4.e4 1. 4.. t'nt,.,^nrlv. In confisurations with the cowl located in the fore- S
to permit starting is moved from the plane of the cowl lip to the end of the ventin; slot and lateral spill is ,. .
prevented at M0 > with a - 0*. Within limits, the designer can change the circumferential "smile" angle
Sof the scoops to accommodatc different maximum air capture ratio requirements. For nose entry configura-
tions, such as shown in Fig. 2c, the forebody angle can be reduced aft of the sideplate to reduce drag.
Although a conical forebody is shown, other shapes can be used, albeit to a'commodate a more optimal radome
shape or to improve the inle." efficiency by including some Isentropic turning.

Aft of the plane of the end of the sidcplate vent, the designer has a number of options for ducting the
flow. They are typified in Fig. 4 for the particular case of a 4-scoop symmetric inlet in a conical fore-
body flowfield. In this figure, CRJ, scramjet and DCR engine options are considered. Figure 4a depicts the

e
A

?_ _ _ _-:.:,:
a) Annular or modular scramiet. b) 4 duct dump type ram jet

c) Center dump ramjet d) Center duc scramjet e) Dual combustor ramjet v, .....
- '.".et

Fig. 4 Engine options with inward turning scoop inlets.

option for an annular or modular scramjet. In the annular design, the flowfield from the truncated sectors at
station A-A is spread circumferentially to form a complr:e annulus at the scramjet combustor inlet. Alter-
natively, the flow can either be distributed by shaped transition sections into 4 circular modular ducts or
"dumped" into 4 circular scramijet combustors, the . vsumption being that circular cross-sections in the high
pressure-thermal loading combustor are desirable, I iot mandatory.

In configuration b, the quadrant flowfields are itended into circular cross-sections and brought par-
tially inboard after the flow has "shocked down" to subsonic. As shown, the four ducts supply a subsonic
"dump type" CRJ. Similarly, thi flow could simultaneously be spread circumferentially and be directed to the
missile axis to provide a central duct dump combustor as shown in Fig. 4c, or a central duct scramjet as shown
in Fig. 4d. Finally, the flow from the inlets can be split to supply two %ombustors as would be needed for
4%
the DCR. Here, two inlets supply the subsonic dimp combustor as depicted in the lower half of Fig. 4e, and
the flow from the other two inlets is spread circumferentially to form an annular concentric supersonic flow ,
in the plane of the gas generator e:cit. In practice, it has been shown that more than two scoops are required - "'. '
to provide a reasonably uniform flow in the supersonic annulus, thus the configur-tion shown is only illus-
trative.

Given that the ITS inlet is potentially attractive for a variety of engine concepts, it is an appropriate
choice for describing a strategy for designing the inlet for a scramjet engine. The first step in the design ...
process is to exa,,tne ilieeffects that changes in geometry have on inlet performance using simplified inviscid
lowfield solutions. Then for a few geometries, perhaps only one, the inviscid flow solutions are used as -
edge conditions for a boundary layer analysis to yield the displacement thickness, 6*, and the surface geo- " " "
metry is accordingly adjusted by 6h. With the revised geometry, fully viscous solutions of the fljwfield are
obtained either using the parabolized approximation of the Hevier-S-.okes ouations (PNS) or. if muhanM..
 DATAUELEMENTECTD
INPUTS IODEL
) GOER EUTDOT
Z_19.. WE FLZIRPEOTt FLWPOPRYPRF
PARAMETERS
"KE "Pt

OFSTATE
INITIAL. I ' LNTODF
ETH'ODOF i -"TAVALU
PRlqOPEFITIESATSELEC,'.O.!70 " -.
CODITLONS, NAERASTE
M O D E L S E IR
,IS
I
-'1
30 M O,
E IQ
T=ACERALL
TECNN UE CO2 LXNT If '.'E
SECTI
YS HEA R H'"","
r .' '
S,
EQUATIONTERAIO
TRAN SPO RT V- E rNA PROPERTI- -
-ER INTTORATI
PRlOPERTIES CODES FLOW..O T - - - -_ -. - - - -

EDDY =INTEGRAL CODE~

LNECROSS
SELECTED ~~~~
I TCHI NLT EFFIEN OF

FIV. 5 IEND
E.0FNU
TEE.NOO
Blckdaam COWL LiP I,',.TE INTEGRATION I I" N
SI EXT.~LANE TEC.N.OUE LANCE"'''',LO. '"''""i
,'HIOU
LOE CNOElT, AIRON.N
CAPWIL3OAO
URE
REATFXN.....
EWIRICPRFILE ATLWCOE
; I ULLNAVER-TOK$IFURSERTE L- INTEGRATION

MODEL
ES E% LN /"-'-"."-
TEC'HNIQUE
VAISCOS IT

t. . . . . . ST E MLN r._'- - .- '

In this example it La assumed that HEB-4 and the missile will Climb and accelerate to cruise at HC =.. " .- .-
7. To minimize fuel consumption during the climb phase, it is necessary to cong;rol both the rate of accelera-'',. -,
tion and the flight path angle , hiaving entered the climb phase at an optimal *~B The determination of the :;:::.
optinal climb-out trajectory is a f~ormidable task but for a r~elatively high thrus[-tO-weight ratio missile,
optimal climb-outs generally require ci 0 until an altitude Z 0. 7 ZCR is reached, at which time M0has .. .
nealy
eaced CR.Thus, inlet opration ant angle of attack and low M0 is not a significant design con- ''....-'
aideration. On the other hand, during cruise at high altitude, maximum range results when the missile ic -.-:.
flying Just below the ci correspornding to maximum LID. Typically, ci at (L/D)MAX lies between 7 and 12" for .--..
aissiles with MH - 6-8. So, inlet operation at ci, particularly with rs.spect to air capture ratio, is iI!por- '-...'.
tant for H0 MCR. Arguments are presented in Ref. 13 showing that for syrsietrical inlet geometries it is
advantageous to design MFC - MCR - AM, where AM Z 0.5-1.0 to obain op~timal performance. Consaquently, in this . S e
xarple MFC - 6.--" - "

The conical shock lies ahead of the cowl lip, since M0 < C and the air entering the inlet is reptesented ".... ..
by A0 and the last captured streamline. A ist the ER_.T.
projected area of the enti..e sector of the circle bounded """:
by the simile anglue of the scoop.
All
The conical compression field is
TS:G
ATIO
E
turned toward the missile axis by the
cowl lip ,hock. Internal cowl lip ,anges 0 of 0" to 20" will be exam 4 .ned. Inlet sideplates follow the '----'-
cowl shock to prevent lateral spiliage at ci - 0, and no further turning of the flow is introd.ced ahead of:.'''
the plane of the aideplate crotch (Station 2). Thus, the innerbody is turned approximately parallel to the :".
inner cowl surface along the line of the shock intersection with the conical surf acu. Downstream of the
station of the sideplate crotch a "boundary laye.r recovery" section is s'hown. Unless the boundary layer has ,...,
been re~noved from the inlet innerbooy, the shock from the cowl lip distorts the boundary layer to a profile
that is nearing separation. Length at constant pressure is required to restore the profile to a shape more i ''
tolerable to txposure to the adverse pressure gradient in the isentropic compression region 2-3. In region 2-
3 the flow is compressed by gentle turning of the cowl side of the internal duct. The innerbody also curves .-
inwardly tO provide cancellation of the weak compression daves. This section can hiave an arbitrary lengt:h :.i:"':""
greater than a minimal vaiue corresponding to a centered wave compression focused on the Innerbody. Engine
packaging considerations and perhaps viscid-inviscid interactions such as those Suit sentioned would govern,.....,
the lengthn selection. In this example, the length was chosen su'ch that the inner duct cc~ntour is the middle ..
streamline in a centered compression field. The amount of allowable internal contraction is governed by the .''
starting area ratio criterion, i.e. that is, A3 /A 2 can be no rmaller than the ratio corresponding to sonic r. "
flow at station 3 when a normal shock is positioned in the plane of the aideplate crotch. The internal c'~m- .,.-..
pression is assumed to be isentropic. For the engine configurations shown in Fig. 4 that require transition :.:'..-.-
from a sector of having a smile eangle I to, say, a quadrrnt of a circle or annulus, circumferential spreading -'..'.
of the duct aft of sideplate crotch can be incorporated. Providing circumferential spreading while main- ""." '."."
taining either constant pressure as in the "boundary layer recovery" section and downstream of Section 3, or "". *" """
providing a "shock-free expansion-wave-free" compression as in Section 2-3, is a challenging design problem
that is beyond the scope of this paper. Suffice it to say that generally the desired longitudinal variation " "."
---- I ...
------.... 4o .e,4.
.. ,I.,,.. hk, r'vefr11ln,e tha ahace of the inner and outer duct walls 0
 .... ..... .". ~.
V
LAST CAPTURED STREAMLINE
b
...
.......

Y ON . .
0

. .
A

............... X :..... .,.

LAST CAPTURED STREAMLINE . ... RECOVERY:: .......

a) DESIGNED FOR
4/Ai 1AT M 6
/ !i i iiiii i i Eii!i iil
i~i=i=i==i~ii=i=:iii==i==i!
........ .. i iiiiiiii==!i==!:i=i===ii ON E iii E i
.. ........................" :'" : " :'

Rei
EdDo Fig 6Ge of tRo I inetE

Teowrelce shchaa
X)
sl ight
inwr iiiiiii
iliiiiiiicratr DESGNE
decreasing gradiiet
idue toi thei FO ...
in
A -
Mach umbefro
TMFX6X ..

LATbATUESTEMI E SGF FO o//IAT -

the c.owl lip to the conical surface and due to flow convergence downstream of the shock. The waves in the
isentropic compression region are simnilarly cu.rved due to the persistence of the outward to inward decreasing
Mach number gradient downstream of the cowl reflected shock as well as flow convergence. In the numerical
examples of the inviscid flowfield solutions that fo'low, two simplifying assumptions were made aft of the
cowl lip station: 1) a mean Mach number that satisfied conservation of mass was used to represent the flow at
major stations and 2) the pressure rise and total pressure loss across the cowl reflected shock were taken
M0 for
equal to those -4
C a planar flow with the same9Pc0 t3 t
turning angle.

Other scoop geometries could have been considered. For example, the external compression surface could
have comprised a cone followed by an outward turning isentropic compression surface. For the same amount of
external compression, viz Nop
0 the efficiency would be slightly greater, but the Inlet would be longer.
Presuming t'hat the innerbody houses an RF seeker, this shape would be less favorable from the point of view
of minimization of radome boresight error. Likewise, the cowl inner surface ahead of the sideplate crotch
plane could be shaped to provide isentropic compression for cases corresponding to 0 > 0. For the same com-
pression (p2 ip1 ), the efficiency would be higher but the subsequent amount of allowable internal contraction,
A3/
*would be slightly less. For most designs, the curved inner cowl lip gives better inviscid performance
but the issue would be moot unless the inlet incorporates boundary layer control. If no boundary layer con-.
trol is provided, the subsequent discussion will shiow that 80 is optimal.
Figure 6b is used for exemplary purposes to show the general utility of the Inviscid flowfield analysis.
This Inlet has the same cone angle 6c and cowl angle B-0", and it Is operating at the same M0 as the
inlet shown in Fig. 6a but it is designed for M~c - 4. All of the other c,.netrainta regarding Internal con- -

traction, etc. remain the same. It can be shown that the static pressure increaseb and total1 prcssure losses
through the shock and isentropic compressions are nearly identical for the two cases (Figs. 6a and 6b). This
is because the average Mach number of the flow ahead of the cowl reflected s~aock is only weakly dependent on
,F.e.g. for d~c. 15" R, - 3.230 for MFC - 6 and R1 - 3.255 for MF 4 The internal duct is, of course,
larger for the M~-4 case, but A /A0 9 A /A2 /c'Pand p /p~ are nearly the same for both cases. It is

also apparent that the inlet contraction ratio Ai/A is smaller for
T4the case which means that for-M
> 4 the NyC - 4 case will have lower P3 /po and therefore the cycle performance will be lower. Thus, as in a
number of other situations in well deaigned fixed geometry inlets, a loss in air capture Is, in part, com-
pensated by increaeed compression and vice versa. Thr loss in air capture at n - 0 can be obtained from
solutions of the conical flowfild to yield flow angle and M , the Mach number along the conical ray
wthe
w w
that strikes the cowl lip and the total pressure loss across the conical shock pt /p t
together with the
relationship w t
+ 1w
 8-9

MFC'- 5 6 7

1.0

0
cc

-17.5

0.7- / - -- -20.0

0.6 I
4 5 ft7
MACH NUMBER M0

for inets with conical forebodies.

Fig. 7 Air capture ratio

1.0

0.9 d

0.8

/ dc, l7.5 20

0A20
16-

0.4

3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0 5.2 5.4
MAXIMUM INLET CONTRACTION RATIO A0 /A3
Fig. 8 Total pressure recovery for family of inward turning sconp inlets at M0 4.
 / - " / -

S o.

8-lI0

CONE ANGLE b DEG.

25 22.5 20 17.5 15 12.5 10
. ............ '

I. .. :. :
0 . ........
0=20 O

o -.-. %.'oO,.

L" 15
k-

w
-- - -S.:::::
Lc
I5 A A-.
00

cc4

SEPARTED WITHOUT BLEED 1

ATTACHED .S

i.
2.5 26 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6
MACH NUMBER ON CONICAL SURFACE Mc

Fig. 9 Maxiumn contraction ratio contours on grid of shock pressure ratio at M0 4 for inward turning scoop inlets.

Figure 7 shows A0 /Ai for KMC 5, 6 ar4d 7 inlets with 6c - 12.5, 15, 17.5 and 20* aver the range 4 < H0 <7

Air capture is only weakly dependent on 6c. At M0 .,4.M - 6, AOIA i =0.73, so the inlet contraction ratio
can be about 37% greater for a H C- 6 versus a MHiC- 4 design.

Figures 8 and 9 show the method for determining the preferred geometry of the ITS inlet for this example.
Total pressure recovery pt /p t.C is shown as a function of A0/A 3 for cone angles S from 10 to 20 and cowl
3

angles 8 of 0* to 20' in Fig. 8. Three important pointb can be learned from theae results.

1) There is a maximum allowable inlet contraction ratio A0/A 3that is considerably smaller than would
occur if the compression wee isentropic and the internal contraction was not limited for starting, i.e..
- 5.4 vs. 10.7 for M0 - 4, the Hach number at whicl, the scoop lip shock is designed to strike the knee in the
innerbody surface.

2) For a constant 6c and increasing 8, or with 8 constant and increasing 6c results in a monotonically
decreasing ptt3
/pt and a corresponding monotonically decreasing value of H.
3 , but a maximum occurs in A0 /A3 -
to
3 for a given A0 /A3, so only the uprer brsnches of the
Scramjet cycle performance increases with increasing MH
curves are of interest.

3) Small or negligible internal cowl lip angles result in the best performance, i.e., maeximum p /P~
3 3t 0
for a given A0 /A3. Indeed, for any value of Ao/A 3 < 5.3,. 0* is optimal.

CA 0 are mapped in a plane of cowl shock pressure rise, vs. the

Mach number on the conical surface, H , to examine the problem of boundary layer separation on the innerbody
surface in rig. 9. The separation criterion used to establish the demarcation between separated and attached ....

flow is in accordance with the values suggested in Ref. 16 for turbulent boundary layers. Additional ex-
perimental data are needed to establish the veracity of the values so derived. Nonetheless, this figure shows
that geometries having A0 /A3 values greater than 4.6 would be expected to have separated flow at the side3.late
crotch station even with -d 0. A local separation would cause unwanted compre/sion and expansion waves and I - "
 At H0 4 for most scramjet engiac operating conditions there will be a norAsIl shock preced' ig the comn-
iation (dual mode operation) and since the ITS inlet is also attractive for use in the CR.], It is4useful to
gamin* conditions corresponding to those with a normal shock located dowrratream of scsatlon 3 with no addi-
Lonal diffusion. Figure 10 shows the total pressure recovery pt' /p fut hsfml iilt tA-4
3 00
Rximum values occur for a given 6 when the total turning of the flov ir the comprcssion procens is appicxi-
Iatly 40%. If there Vere nlodesign consideraLio..s other than in.]et recovery and drag on the portien of the
orebody not wetted by the air captured in thb. inlet, a "esign with 6 - 17.5* end 3 - 3' would be optimum.
oreover, since the optimu~m occurs with inward tuining at the scoop lip. somewhat higher recovery could be . .,

btained by curving this surface throigh the 5 degree turn. c'r. the other hand, if the constraint to avoid ..

oundary layer separation is operative and 6 < 15% a 102 lvwer inviscid rec~overy would have to be accepted.

At this point in the discussion all but a riarrou tange ot geometries can be eliminated frcm further______
ensideration. Four cases have been selected. Case A with 86 - 12.5% 0 - 0 is a conservative geometry, and
asa 3 with 86 15.0', B - 0* Is the optimal geometry for inlets without boundary layer control at a - 0.
asas C end D with 86 . 17.5* and B - 0 and 5o are near optimal if boundary layer sep~iration is not a problem.
able 1 lists inlet parameters and flow properties for these MF 6 desi.gn inlets at M0 - 4. 5, 6 and 7. In
11 designs the scoop lip shock strikes the knee on the innerbody at a - 0, MH
0 4. For !0> 4, the lip shockp
alls downstream of the knee and a far more compler wave structure exists in the duct. Whereas the-te flow-
ields can be obtained using the computer routines shown in Fig. 5, a good approximation of iverege properties
t station 3 can be obtained with the assumption that the total pressure loss in the sum of the loss across

TOTAL PRESSURE RECOVERY OF INWARD TURNING

SCOOP INLETS AT M 0 =4

0.52* - P=

10 12 14 i 8 2 2 2

-- 2

0.52

> P=22.5
 8-12

Table 1

Inlet Parameters and Flow Properties for Four ITS M4FC 6 Inlet Geometries

Inviscid Flow

Case 6c Ai/A 3 M0 Ao/A3 M3 Pt3/Pt P 3 /P 0 M31 Pt3,/Pto P3'/PP

Deg. Deg.

A 12.5 0 4.934 4 3.682 2.522 0.9245 7.938 0.511 0.4531 57.6

5 4.305 3.140 0.8346 9.750 0.467 0.2427 110.6
6 4.934 3.670 0.7373 12.019 0.445 0.1356 1.86.9
7 4.934 4.232 0.6230 12.532 0.429 0.0712 259.7

B 15.0 0 5.881 4 4.297 2.280 0.8632 10.813 0.537 0.5113 63.8

5 5.040 2.828 0.7335 13.692 0.486 0.2789 125.5
6 5.811 3.272 0.6069 17.445 0.461 0.1570 215.0
7 5.811 3.744 0.4797 18.519 0.443 0.0828 299.7

C 17.5 0 6.667 4 4.842 2.045 0.7886 14.598 0.570 0.5516 68.8

5 5.787 2.523 0.6235 18.632 0.511 0.3055 135.3
6 6.667 2.909 0.4866 24.002 0.481 0.1728 232.9
7 6.667 3.314 0.3650 25.888 0.459 0.0913 327.3

D 17.5 5 7.160 4 5.200 1.806 0.6983 18.284 0.615 0.5656 66.5

5 6.215 2.259 0.5253 23.702 0.539 0.3160 137.2
6 7.160 2.596 0.3883 30.929 0.504 0.1793 237.9
7 7.160 2.945 0.2762 33.823 0.479 0.0951 336.5

the conical shock and the loss across a scoop reflected shock that would turn all of the flow to an angle
With the air capture characteristics given in Fig. 7 and the contraction ratio Ai/A 3 set by MFC 6, all
properties at station 3 can be obtained. The trend of improving performance with increased compression ti
was previously shown at M0 - 4, is present for M0 - 5-7.

Whereas these results provide a basi3 for selecting the compression field, viscous effects must be it
troduced to define the geometry and provide a basis for proceeding into the combustor and overall engine c,
analysis.

Using flow properties from the inviscid flow solution as input for conditions on the edge of the bouni
layer, the displacement thickness 6* is obtained by one of several possible techniques (Ref. 17). All con-
tou:rs are adjusted by the local value of 6* and fully viscous solutions of the flowfield are made, 6* is
adjupted and the process repeated until the core of the flowfield is approximately the same as that of the
initial inviscid analysis. For the asymmetric portion of the flow, e.g., in the internal ducts, 6* varies
circumferentially as well as longitudinally. The resulting cross-sectional area Ag, the physical area, is
therefore defined. Additionally, the local static pressure in the boundary layer at a given location is ti
acme as in the inviscid flow. Superficially, it would appear that Ag 3 and p3' together with the continuit
equation, could be used to define a set of mean flow properties that would be appropriate for a cycle anal
sis. If this is done, however, the deduced mean value of the total pressure is artificially low. In effe
the mean value of the momentum, which is directly related to the thrust, would not be well represented.

A more appropriate method, albeit more rigorous to obtain, is to calculate the mass averaged value of
total pressure Pt , and with the local static pressure, P 3 . define an effective flow area A3 e that also

satisfies the continuity equation. The resulting area lies between A3 and Ag3 A numerical example will
to clarify this point. Let us suppose that the viscous solution of the flo,-':eld for Case A at M0 . 4 sho
that (A /A) 3 - 1.2 and that the profile in the turbulent boundary layer csaL be represented by u/u 3 = (y/6)
For an adiabatic wall the mass averaged total pressure in the viscous layer is 59% of the value in the in-
viscid core. The resulting mass averaged total pressure for the entire flow is 83% of the inviscid value,
which leads tc Pt'Pt0 - 0.7689, (A/A) 3 = 1.079, and a corresponding effective Mach number M 3 - 2.404.
30e
These effective flow properties now become the values to be used in an simplified one-dimensional a-alysis
Accordingly, the conditions corresponding to those down3tream of a normal shock at station 3 are M3 ,e = 0.

Pte, /Pt 0 0.414.

Since the boundary layer parameters are dependent on the geometry, M , the altitude, and the wall
9
cooling of each case, each flight condition would produce a different Ag 3 A3 . At first glance it would ap
that the geometry selected for one flight condition would not yield the compression field previously compu
at other flight conditions. However, it turns out that for a given geometry and altitude, 6* is only weak
dependent on MO, changing only about 10% from M0 - 4 to M0 - 7. Moreover, changes with altitude at a gi',e
are also not too great, e.g., a 30,000-ft increase in altitude Z increases 6* by about 20%. Mach number a
altitude also have only a small effect on the ratio of mass averaged total pressure to core flow total pre
sure. Thus, even though the boundary layer thickness, 6, can vary by more than P factor of 2 over a typic
flight envelope, the changes in Ae/A are quite small. Thus, the design strategy is to base the geometry c

the conditions at the end of boost and accept the relatively minor changes in the flowfield
at M0 > MEOB.
 Table 2
Inlet Parameters for Four ITS " 6 Inlets
-C

Viscous Losses Based on (A /A) - 1.2 at M0 - 4

Case a MO M,3 P te33/o Ao/Ae 3 %E e3 Pe3/ 0 Pto3

Deg. Deg. te3' 0

A 12.5 0 4 2.404 7.938 0.7689 3.414 0.9756 0.523 52.2 0.4140

5 2.999 9.750 0.6771 3.992 0.9764 0.475 100.7 0.2224
6 3.512 12.019 0.5905 4.574 0.9774 0.451 170.9 0.1244
7 4.055 12.532 0.4943 4.574 0.9772 0.434 238.4 0.0655
B 15.0 0 4 2.159 10.813 0.7148 3.960 0.9685 0.553 57.0 0.4624
5 2.688 13.692 0.5921 4.645 0.9684 0 497 113.1 0.2533
6 3.116 17.445 0.4825 5.356 0.9679 0.469 194.7 0.1433
7 3.571 18.519 0.3770 5.356 0.9672 0.449 277.4 0.0755

C 17.5 0 4 1.921 14.598 0.6493 4.425 0.9590 0.592 60.4 0.4920

5 2.380 18.632 0.4989 5.289 0.9560 0.525 120.0 0.2737
6 2.751 24.002 0.3825 6.093 0.9561 0.492 207.9 0.1553
7 3.141 25.888 0.2831 6.093 0.9557 0.467 293.8 0.0823
D 17.5 5 4 1.652 18.284 0.5531 4.604 0.9424 0.653 55.2 0.4840
5 2.081 23.702 0.3976 5.503 0.9397 0.564 115.8 0.2716
6 2.400 30.929 0.2865 6.340 0.9404 0.523 202.7 0.1547
7 2.731 33.823 0.1996 6.340 0.9403 0.493 288.3 0.0824

Moreover, it is possible to adjust the inviscid flow results to values suitable for cycle calculations by
obtaining only a few viscous flow solutious and applying a constant value of Ae/A over a suitable range of
conditions. The value of 1.2 for Ag3 /A 3 would be representative for engines of about two foot diameter for
Cases B-D, As well as Case A at HEOB - " Z - 5,000 ft. Ae3/A3 would be different for each case, increasina
from 1.079 for Case A to 1.129 for Case D. Maintaining these values of Ae3 /A3 constant at higher M0 results
in effective flow properties at station 3, as shown in Table 2. The kinetic energy efficiency of the inlet,
NE. at station 3 has been added to the Table since this parameter, rather than Pt 3/Pto is generally used as

the basis of empirical correlations. For a calorically perfect gas. nKE and Pt /Pt0 are related by the
simple expression 3 0
Y-1
2
1 ( y-1)M1 [
[KE
- l
(2)

The same trends with M0 that were present in the inviscid flow results are maintained, but a few addi-
tional points can be made. Total pressure recovery is lower but the effective inlet contraction ratio is
higher. Moreover, nKE3 is nearly invariant with M0 for a given case. This result lends credence to the

results of simplified analyses of performance that are based on holding nKE constant.
There are many situations when the rigor of the foregoing approach is not warranted, i.e. for preliminary
engine performance evaluations. Reference 7 gives the details of a method that uses a strategy similar to
the one just described, viz. it is based on the premise of a uniquely defined effective inlet contraction
ratio. The method is based on an empirical relationship

-KE31 - 0.4(1-MH3/Mo) (3)

and an inlot air capture ratio characteristic similar to that shown in Fig. 7. It is of interest to add the
"data" from Table 2 to the Figures from Ref. 7 as a further test of the adequacy of Eqn. 3. Figure 11 shows
that for inlets with high values of Me3 /M., e.g. Case A, Eqn. 3 yields iKE3 values that are somewhat high, but

the agreement improves at low MH3 /M.. Note also the near independence of Me3 /No with M0 for each of the
cases. Figure 12 from Ref. 7 shows the corresponding values of it /P versus ADA aid to complete the
t 0 A e/3 ,
picture, Fig. 13 shows p3 /po versus M3e/Mo. For clarity, the values from Table 2 have not been added to
these figures.

The design feature of the ITS inlet that remains to be substantiated is the favorable inlet air capture
characteristics at angle of attack at M0 ' MCR. Figure 14 shows this effect for the 6. . 15* conical fort-
body, symmetrical ITS inlet with four scoops, each having a smile angle * of 45. Curves of air capture are
shown for three values of the roll orientation angle f of 0e, 22.5' and 45'. The decrease in air capture of
< 9% at a - I0 in the least favorable roll orientation is remarkably small.

Io
 1.00 1

0.99
4
nKE * 1 -0.4 (1-Me 3 /M0 )
w 0.98 -
Case A/ %

0.97 B

p0.96 c\
Sym. Inlet type RjL
0 APL modular 2
- .Marquardt scoop 28 S0.95
NASA 2-0 integrated
w/o forebody
2 II
0.94 - NASA 2-D Integrated 29 D

A NASA annular (Hre) 30

0.93 -
1.0 0.9 0C8 0.7 0.6 0.5 0.4 0.3 0.2 0.1
Diffuser exit-to-free stream Mech number ratio, Me3 /M0

Fig. 11 Kinetic energy efficiency as a function of diffuser exit-to-free stream Mach number ratio.

1.0

0.9 -
0.8.
M 0.86-

0.7 -

. -Max
Ao/Ae 3

0.3 - 8_

0.21
2 3 4 5 6 7 8 9 10

Effective inlet contraction ratio, AO/Ae 3

Fig. 12 Inlet total pressure recovery as a function of effective contraction ratio.

COMBUSTOR ANALYSIS

Treatment of the combustor of the scramJet varies in complexity from relatively simple, one-dimensional
integral solutions to relatively complex finite difference solutions that include wall shear and non equili-
brium chemistry. The key to the successful adaptaticn of any of the theo'etical descriptions of the combusto
flowfield is the adoption of models that take into consideration the ever growing base of experimental obser-
vations. The schematic illustration of the combustor flowfield shown in Fig. 15 will help to develop this
point. The combustor is a simple diverging cone frustum. The qualitative features of the flow for other
diverging shapes such as those containing step increases in area just downstream of the fuel injection static
are the same, but the details of the flow are different and must be considered even in a simplified integral
analysis. The subsequent discussion vll show that constant-area or converging shapes are not practical for
the types of missile applications that have been postulated.

For all but the highest Mach number and lowest equivalence ratio conditions, the blockage due to the
combined effects of the injection and heat release generates a "shock train" disturbance that originates in
the isolator section and extends downstream of the injection ports into the combustor. In general, the pres-
saure rise associated with the shock train is sufficiently great to separate the boundary layer. In a well
designed engine, the isolator is of sufficient length to prevent the combustion-induced disturbancts from
 601 1 1 1 1 o

0/
w8 /

UE30-

20

z 10

0
1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2
M3/MO
Fig. 13 Inlet pressure ratios based on data correlation.

1.0

MFC =6
.98 -M 0 ='7
*=45'
.-97
C 45 (TYP)
.96
S.95
ujKW
~.94
-

< .93 = 22.5*

U
.92
.91,
.90

0 1 2 3 4 5 6 7 8 9 1C
ANGLE OF ATTACK a (DEG)
Fig. 14 Effect of angle of attack and roll angle on
the air capture of 4 ITS inlets.

FLOW
STATION 4 se 5

M.1IXING AND APPROACH

ISOLATOR' COMBt~jT R- 4
Fig.~
Scheati ~ ~~-OB
~ ~ ilutaZno 15 lo cse niONEto
and cmbustr
secOons
 disrupting the flow in the inlet. Generally, combustors with step increases in area require less isolator
length because the strength of the shock train is diminished. Downstream of the shock train mixing and
combustion are intensive, with large radial, axial and, perhaps, circumferential gradiets in flow properties
and chemical composition. Further downstream, the mixing and combustion is less intense and the gradients are
considerably weaker, and the flow can be reasonably approximated by one-dimensional mean flow properties at
each axial station.
A similar shock train structure can be generated in a duct of this shape by p'.acing a physical blockage,
such as a partially closed valve, downstream of the duct, or by simply increasing t:he back pressure in a
direct-connect testing apparatus. Several investigators have exploited this fact to study the shock train in
an unheated, nonreacting air stream that is more conducive to measurement. In par~icular, Waltrup, et al.
(Ref. 18) made sufficient measurements to model the pressure rise over a range of conditions typifying those
in the entrance of the scramjet combustor. Figure 16 from Ref. 18 is a data correlation that shows that the
shape of the pressure rise curves for various p /P4 values of 1.6 to 6.9 with initial Mach numbers of 1.53 to

2
8.0 $0M
4 1)Re0' /I I /p 1\2
7.0 -- 501- -1/ +170 ---- 1| -

)6.0
6. p4 --
5.0 00

P4 4.0

2.0
I fJ 1. __________.0
"0 1.0 2.0 3.0 4.0 5.0
'M4- 1) Reao
(x 10-)

Fig. 16 Normalized pressure distributions in shock train

flow structures.

2.72 can be collapoed to a single correlating curve. Here, p5 is the pressure downstream of the shock train
and P4 is the initial pressure. The correlating equation is

s(M4 - l)Re 6 1/
D/= 4'/ 50(-p 1) + 170(4")()

4 4 1p / p /
Equation (4) can be used to obtain rhe total length of the shock train s shown in Fig. 15 for given shock
a
pressure rises and then used to find p - f(s). The nondimensionalizing length scale of (D/ 2 61/ 2 )4 was based
~\ on a regression analysis of data taken ovar a limited range of these parameters and could, therefore, be
subject to revision with the compilation of additional data. Until additional experiments are conducted, it
is neceseary to use the hydraulic diameter in place of D4 for isolators having non circular cross-sections.

Experiments are currently underway (Ref. 19) to develop similar data correlations in co-annular isolators
typified by Fig. 4e. The overall pressure ratios in Fig. 14 correspond to a range of shock strengths equiva-
lent to either oblique or normal shocks. It has been found that the mean value of other flow properties down-
stream of shock trains correspond to conditions downstream of the equivalent single shock. This suggests tha
wall shear contributes only a small part to the total pressure loss in this strongly coupled viscous-inviscid
interaction zone. The static pressure distribution downstream of a shock train depends on the durnt shape and
on whether the flow is separated at the end of the shock train. In separated zones the pressure would be
nearly constant. For attached flow in constant area or diverging sections and subsonic in'tial conditions,
the pressure decreases monotonically. For supersonic initial conditions and either constant area or very
small divergence the static pressure rises sltgl.tly, but in the geometries of most interest that have moderat,
divergence, i.e., area ratios of about two, the pressure decreases monotonically.

Combustor analysis based on integral sclutions of the x momentum equation require that the wall stress b,
well posed. Therefore, it is evident that not only must the shape of the pressure rise in the shock train be
known, but its position relative to the fuel ports must be defined.

In Ref. 10, the relationship

ad/D 4 / 4 0.5 ( 1-) (5)

was derived from hueristic arguments and limited data to define the distance, 5 do that the shock train extend
downstream of the fuel ports as shown in Fig. 15. Thus, for a given set of initial conditions M4 , 94- Ree4

and the shock train pressure rise ps/P 4 , the pressure distribution to the end of the shock train or the re-
o.A. ,
 attachment point is defined. No ambiguity exists if a step increase in area is present because the pressure
is given as a function of x rather than A. Downstream of the shock train, it turns out that the pressure
distribution can be well represented by a pressure-area relatiohship first suggested by Crocco (Ref. 20) if
the area is taken as the local cross-sectional area of the duct. Crocco noted that the relationship
pA '/-1- constant, where C is an arbitrary constant, embraced the particular solutions: 1) c = 1 corres-
ponding to the Rayleigh process, i.e., one-dimensional constant area heat addition, and 2) c - 0 corresponding
to constant pressure process. Reference 21 added to the attributes of this formulation by noting that C - -yM2
corresponds to e one-dimensional heat addition with constant Mach number.

Note that the assumption of the vail pressure distribution does not require unidimensionality of the
flow, but the association with the corresponding one-dimensional flow having the same value of E can be in-
formative. It is particularly important to grasp this point if one is to use integral techniques effectively.
For a given set of conditions (not necessarily uniform) in a starting plane and heat release, there is a
unique set of unidimensional properties in the exit plane for a given total force and heat flux on the boun-
dary (the combustor wall in this case), regardless of the path of the process. The total axial force on the
wall is the difference between the integrated pressure on the projected area and the wall shear. Thus, a
given solution of the integral equations is valid for an infinite number of combinations of axial pressure
force and wall shear. To %xploitthis reality, one needs a shear model for reacting ducted flows. Although
considerable progress has been made in the development of analytical models (see e.g. Ref. 22), these efforts
have not progressed to the point of practical application; consequently empirical correlations still must form
the besis of scramJet combustor analyses.

In Refs. 9 and 23 heat flux data from six different supersonic combustor geometries, with both gaseous
and liquid fuels and H4 from 1.6 to 3.2 were correlated and used to obtain a deduced combustor shear para-
meter. Figure 17 shows this shear parameter Cf - 2'wIP4u4' 2 (A 4/iau4) versus EReffR where ii the mean

EQUIVALENCE RATIO, ER (TYP. LIQUID FUEL)

0.005 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80.005

DEDUCED WALL
0.004 -SHEAR PARAMETER -0.004

0.00 - -0.003 c

_ / WALLHEATFLUX
,o 0.002 PARAMETER -0.002

0.001 - 0.001 -

0 10
0 400 800 1200 1600
(Btu/Ib)

HEAT RELEASE PARAMETER (f- aHf -n)

Fig. 17 Combustor wall heat transfer and skin friction coefficient

as a function of equivelence ratio.

shearing stress, ;a the weight flow of air entering the combustor. Here, EReff - ER'nc for typical liquid
a ef c
hydrocarbon fuels and n c is the combustion efficiency. The form of the parameter was intentionally selezted
to enhance combustor calculations by giving the mean value of shear in terms of combustor inlet conditions.
Preliminary results from the aforementioned theoretical analysis confirm the qualitative trend of increasing
shear with increasing heat release.

With this modeling of the wall stresses the solution of the integral form of the conservation equations
can be carried out. In particular, the x mnmentum equation can be expressed as

St 2
-+
,,x p d* + ( 1-A) A A] - Td A P 5A 5A
- 5 5 p4 4 " p4 u 4 4 - Pf u f af

J ePPna 5 (Pe P4 f twdw 5 + - P44- 4 -4 U

x-0 xx- x=0 (6)

where the subscript 5 refers to conditions at the combustor exit, subscript e to conditions corresponding to
the point where the pressure variation with area can be represented by the Crocco parameter and subscript f
refers to th: fuel. For normal injection, the f.jel momentum term is zero and for other injection angles the
term is small and usually is neglected. In all of the data that have been examined to date, either xa - xe,
i.e., immediately following the shock train the pressure-area distribution can be represented by - c process
or A - A , i.e., the downstream end of the shock train is in a constant area section. Thus, the second
intetral Ras been set - 0 in the analysis of data currently available. In the latter case, it still would be
necessary to relate pa to Pe to obtain solutions rigorously, but considering the approximate nature of the
shear modelling, pe is frequently taken equal to Pa" With these assimptions, a set of initial conditions,
including the fuel specification (atom balance, flow rate and enthalpy) and a value for T and a procedure for
 relating the state variables, the integral form of the conservation equations can be solved for selected
values of ps/p 4. Generally, local thermodynamic equilibrium is assumed at the combustor exit, where the
composition is assumed to be products of combustion at EReff - nc ER and unreacted fuel in the amount
(1 - ric)ER. References 24 and 25 provide the basis for computing the state relationships. Of course, the
application of this analysis is limited to either the determination of the global heat release and combustor
exit conditions given a wall pressure distribution or the prediction of the wall pressure and exit conditions
given an assumed heat release. If neither the heat release nor the pressure distribution is known a priori,
then suitable kinetic rate equations must be introduced and solved simultaneously with the aforementioned
conservation equations. A discussion of this subject is beyond the scope of this paper. Conversely, if both
the pressure distribution and the global heat release are known from experiments, the veracity of this analysis
can be examined.

Equation (6) can be simplified to

(I- C:
[P5A5

2A (L_" )
-
(Pe Ae]
ee-
P5 A5 1+
_
- P4 u4 2 Z(7)

where
fuff
P4A4
2
+
/ e-- X~x
P4A4 )

-- d
57A
4

--
(
xx
p5
a
d
( 4-
p4

- Zf
-

pu
u
-_ (8)
P4APfu
4
A f P4 AA )
4 f P4 (A4 2C P44
4 A4
x-O x=x

and the procedure for obtaining all terms in Eq. (8) has been described.

"For a given
combustor and shock train pressure ratio, t',re are two solutions for every fuel flow rate,
"one of which canbe eliminated because either the pressure increases downstream of the shock or the second law
of thermodynamics is violated, or both. In effect then, fc',a given heat release there is one candidate
solution for each value of p/p 4, and an additional constraint must be imposed to yield a unique solution.

Reference 26 introduced the constraint (ap/aA).C

6 - (ap/aA)sB- as A - A5 and (dTt/Tt) 0 which simply
0,
states that the slope of the pressure area relationship for a constant E process at the combustor exit, where
the derivative of the total temperature approaches zero, should be equal to the slope of an isentropic pro-
cess. More rigorously, the slope should match that of an adiabatic process that includes wall shear. How-
ever, if the shear terms are acicad, no simple analytic relationship has been found, and indeed, when typical
experimental results are examined, the effects do not significantly affect this constraint.

With this constraint the additional condition

5 +y51_C71/2
M5 + Y5 (1-)J or [I + (y-l)M-lJ (9)
"must alsobe met, and a unique solution for each heat release can be obtained. That is, there is a specified
shock train pressure rise for a given heat release. This result represents an analogou~s situation to the well
recognized thermal choking point in a Rayleigh heating process. Reference 21 gave the name "entropy limit" to
solutions of the conservation equations that were so constrained. One caveat remains, viz for low area ratio
combustors and high heat release rates, the solutions to the equations would yield shock train pressure rises
which would exceed that of a normal shock. For those cases, the slope constraint is relaxed, the normal shock
* pressure ratio is held, and the combustor exit Mach number, H5 , is permitted to decrease. The limiting heat
'5
release for a combustor operating in this mode corresponds to M5 - 1. Analysis of free-jet tests of scramjet
engines verifies this mode of engine operation.

The steps in the procedure for determining the wallpressure distribution, wall shear and flow properties
"IJn the exit plane for a given combustor and heat release rate are as follows:

1) The shear term in Eq. (8) is computed directly using the relationship from Fig. 17, A, and the
initial conditions.
2) An initial value of p /P is estimated from previous calculations or from a calorically perfect
s 4
solution, which will subsequently be discussed.

3) Equation (5) is solved for Sd; then Eq. (4) is solved for a. and values of P/P 4 vs x in the region
x - 0 to x - sa - sd' These values, together with the geometry of the combustor, permit the evalua-

tion ot the left hand integral in Eq. (8) and define As/A 4 .

4) The energy, continuity and state equations are combined to obtain P5 A5 /P4A4 which, for fluids that
can be treated as calorically perfect at a given flow station, yields

2
T 1/
P5 A5 (1+f 4 IY 4I, 5 124r (Y4 -l)M 4 J s 1 (10)
p4 A4 (14 f) g ) 2+ -M'
5 J \Tt)/" M5 2 (y5 5 2]1/ 0l)M

5) Equation (9) is substituted into Eq. (7) and set equal to Eq. (10) tc yield

2
e 4 + (Yb-)Mzl (1 + y
iA4 4 Z)
ZM- k (11)
2
P4 As(1-M5 ) [2 + (y I)M5 ]
J--e
. 5' . . . .- 5.. 5 .. . i . .,. .
 8-19

For cases where PS Pe' as has been observed in many experimtntal results, the computed value of .
pe/P4 is compared to the initially selected value of p s/P4, a new value assumed and steps 2 through
5 repeated until the suitable agreement is reached. When p5 0 pe and the values are known a priori
and A - Ae, the above matching is carried out for the known value of ps/P , which could simply be
aa e
an iterant in a ps . Pe solution. When p a Pe and the respective values are not known a priori,
or if As 0 A., or both, a joining process is required. The joining process necessitates the adop-
tion of an additional constraint, viz, that the flow at stations s and e can be represented as
unidimensional at those stations. Aga'n, the integral method does noc require one-dimensional flow
between these stations. Between stations s and e it is again assumed that the pressure distribution
can be represented by another value of the Crocco parameter, i.e., E. Solutions can now be found
for any other imposed constraint within limits. For example, it could be assumed that the variation .8
in heat release Is known as a function of distance or residence time, thereby the values of Tt /Tte
t t
would be known. Solutions for the upstream portion, i.e., from 4 to a would be joined to the down-
stream portion, i.e., from a to e by an E - constant process, and from e to 5 by an C - constant
process.
6) For the converged value of p/P 4 and the corresponding values of M5 and y5 , Eq. (7) is solved for C
and a 4.5

p5 _p 5 e
8 p5 1 \ (Ae e e) (4
p pp
4 T4 14 5 ~ ()l(12)

7) For cases where the converged solution for ps/P 4 yields a value greater than that correspondIng to a
normal shock at station 4, the normal shock pc/P 4 value is held, and the constraint imposed by Eq.
(7) is relaxed. A value for M5 less than that corresponding to the entropy limit case is assumed,
and Eq. (10) is solved for A5 P5 /A4 P 4 , which is substituted into Eq. (7) to find t and into Eq.
(12) to obtain a calculated value of A5 /A 4, which is compared to the actual A5 /A4 . The process is
repeated until convergence is obtained or the limiting value of M5 - 1 is re-zhed.

Results from this analysis are peculiar to a particular combustor and operating conditions. On the other
hand, a number of important qualitative features of the analysis can be examined by considering the idealized
case of a calorically perfect gas with y - 1.4, and neglecting shear and the mass and momentum contributions
of the fuel. Figure 18 shows a typical set of solutions for t.he case of M4 - 2.5, so - sd - 0 and p e - pt.
The solid lines show the p5 /p 4 values as a function of the combustor area ratio for the entropy limit solu-
tions at selected shock train pressure ratios of 1, 2, 4, 6 and 7.125. The dashed curves for M5 values of
1.0, 1.2, 1.4, 1.6 and 1.8 are solutions that correspond to the normal shock pressure rise, ps/p 4 - 7.125 but

M4 - 2.50
.=1.4

4.0 1

3.0 .

-Q 2.0 SHOCK 1.0,

2. .SOLUTIONS 8
0 4 "1.2,

Lu 0.8 - I
ENTROPY 1I
.8 "
. SOLUTIONS

U) . t1t6.

0.21 1 I- I - I

1.0 1.5 2.0 2.5 3.0 3.5 4.0

COMBUSTOR AREA RATIO A5 /A4

Fig. 18 Combustor exit pressure ratios for selected heat release rates.
 8-20

with the slope constraint at the combustor exit relaxed. Superposed on these solutions are curves corres-not
pondiug to total temperature ratios of 1.5, 2, 3, 4, 5 and 6. For all As/A4 > 1, p5 < P5, but this does
preclude the pos3ibility of pressures within the combustor being greater than ps" In fact, if the combustor
comprised a cylindrical or slowly diverging area section followed by a more rapidly diverging section,
p > ps would be expected.
The distinction between solutions resulting from this modeling and those for flows that are constrained
to be unidimensional throughout is clearly evident for the particular case of A5 /A4 - 1. As shown in Fig.
18, the p /p4 - 1, 2 and 4 curves intersect the A5 /A4 - 1 axis at p5 /p 4 - 1, 2 and 4, respectively. Upon
closer examination it can be shown that for all values of pa/p 4 < 4.0625, p5 - Pa. whereas for all values of
ps/P4 > 4.0625 there are no solutions at As/A 4 - 1.It can be rigorously proved that every combustor exit
solution of p5 /p 4 < 4.0625 corresoonds to an end point state on the supersonic leg of a Rayleigh heating
process, i.e., a one-dimensional frictionless heat addition process in a constant area duct. Foreover, the M5
values are identical to those of the corresponding Rayleign heating process. However, from Eq. (9) the slope
constraint at station 5 requires e > 1 for M5 > 1. From a phy-!ical standpoint, this would imply that for all
T t/T < 1.4083, the value corresponding to p5 /p 4 - 4.0625, M5 - 1, shock trains having strengths corres-
514
ponding to single oblique waves would be present and not the shock-free all supersonic Rayleigh process. Note,
too, that these results would also hold for cases where the shock train extends into the duct as long as wall
friction is negligible. When wall friction is included, the same qualitative feati.res are present, but the
values of p5 /p 4 and ps/p 4 shift for given values of It/T. Also note that as A5 /A4 is increased slightly
5t 4
above 1 in the wall-shear-frae case, solutions for p/p 4 up to the normal shock value at M4 are possible, but
pS/p4 is always < 4.0625. This can also be seen in Fig. 19 where pa/p 4 instead of p5 /p 4 is the ordinate.

M4 2.5," 1.4 , w =0

1.1 1.2 1.4 1.8 Ms , 2.0

5S6 t
. / ' ' - 2.5-
I

2 31
4V t

4.4

StJ 4

1.
St 4 4

intersect the As/A 4 ., 1.0 axis and, indeed, there are valid solutions for As/A 4 < 1, i.e., for combustors
with converging areas. The curves for Tts/Tt > 1.4083 are cusped at As/A 4 + 1, i.e., there are also

solutions for As/A by

iondition streained d andslopermonitteda
1,
4 < the those Tt5/Tt th
curves also become
eqcombusltor asympotic
he xitu.
For to the
of As/A
/values 1 axis. Overlaid on
.2 4 - <.403tecre

S....
this-. grid are-dashed curves for constant H5 values of 1.1 to 2.5.
condition5
isrelaxedte ndo
maH5 ismp et eqoaustys
rmitst ios themaximh
The entropy limit solutions for ps/p 4 -
sh/Ic isa6.293.
 1.0 I II I I

0.8 M4 - 2.50
1.w-O
,y 1.4
-
1'.
-0.6

S0.5 T /T:

"tu 1.0

04 1.5 NORMALSHOCK
SUOLUTIONS
-2.0
o
j-0.3-"

LIMIT 4 4.0
SOLUTIONS "-2.'

0.2 I I
1 2 3 4 5 6
COMBUSTOR AREA RATIO AS/A 4
'
Fig. 20 Total pressure ratio for selected heat release rates.

Figure 20 shows the important effect of combustor area ratio on the total pressure recovery in the heat
addition process, Pt /Pt Curves for total temperature ratios it5/Tt of 1, 1.5, 2, 3 and 4 are shown.
Points to the left of the dashed line correspond to entropy limit solutions and thoce to the right of the
dashed line are for shock trains with normal shock pressure ratios. The T /Tt curves terminate at H5 - 1.
5t 4
For T /T > 1, the P /P curves decrease rapidly with increasing A5 /A', but reach a minimum at M5 slightly
t t t t 545
5 4 5 4
> 1. 4hereas larger values of _^/A4 permit more heat to be released, the expense in total pressure loss is
insignificant. For the same Tt5/Tt4 more heat can be added in a given A5 /A4 combustor as K4 increases. As

shown in Table 2, M4 - Me3 increases with Mo, but Tt /T decreases with H0 for a given ER, which is, un-
54
fortunately, the opposite of the desired trend. For this reason the designer generally selects the smallest
possible A5 /A4 that provides sufficient thrust to climb out at lLOB and then accepts the loss in pt /p that
5t 4
occurs at M0 H>EOB where As/A 4 is greater than the optimal value.

Similar curves to Figs. 18-20 can be generated for values of Z 0 0 (Eq. 8) to provide a means for
assessing the effects of shear, combustor shape and of the shock train extending into the combustor. Of
course, for a precise calculation, the actual properties of the flow at the combustor entrance, the fuel type
and injector configuration, etc., must be known, together with the appropriate state equation and solutions
must be done on a computer.
As previously stated, when the bulk value of the heat release is known, experimental measurements can be
compared to the results of this analysis. Over a period of about twenty years at the Johns Hopkins University 47
Applied Physics Laboratory. bulk heat release rates have been obtained from a steam calorimeter attached to
the exit of supersonic combustors. One cf the first reported experiments (Ref. 21) remains as a benchmark to
index the veracity of analytical solutions due primarily to the atypical test configuration. In this test,
the boundary layer of the incoming air was removed just ahead of the fuel injector ports (see Fig. 21a), thus
providing the unique situation where the shock train can collapse to a simple ncrmal wave as sa,. 0. For the
ERff - 0.424 with hydrogen fuel and combustor entry conditions of M4 - 1.95 and Tt - 2188*R, the foregoing
eff 4
analysis would predict a normal shock-entropy limit solution. For this A5 /A4 - 2.28 combustor using a value
for Z in Eq. (8) deduced from the measured heat transfer and a modified Reynold's analogy (see Ref. 23), the
entropy limit solution corresponds to C - 2.004, p/p 4 " 4.28, p5 /p 4 - 0.835 and M5 - 1.676. Figures 21b and
c show the excellent agreement between theory and measurement of the wall static and combustor exit pitot
pressures Pts, and exit Mach number deduced from P5 and pt,
Iv all other available experimental data there has been a boundary layer of finite thickness at the
combustor entrance, thus it has been necessary to compute aa d and p - f(x) from x - 0 to x - s. - ad.
Figure 22 compares measured and computed wall pressure distributions for three different combustor geometries
(see Refs. 9 and 10 for details). In each case H4 , 84, D4 and A5 /A4 are 3.23, 0.0146 in, 2.74 in, and 2.0,

"M 4*.

I II II I I I I I I
 8-22

COMBUSTION AND FLOW

CALORIMETER
STATIC PRESSURE SPRAY
FUEL TAPS

STORAGE
PDM COMBUSTOR CALORIMETER
HEATE PLAINUM GAUZE =
CAALYST
.- 24 INCHES.---

0 2 NOZZLE PITOT
.- PROBE RAKE
TO EYHAUST

UNDARY LAYER BLEED

SONI5 1'0HOLMS 0,0911-1CHDIAMETER

LAYER BLEED '42

a) Schematic of test apparatus.

0 CENTERLINE
195
-. EREFF 0.424
4 .0 * T t4 2 18 5 R I g
S _ P4 =17.4 PSIA
2.004 Ttf= 1715"R #rI0
II
S-+0.5
-0 MEASURED PRESSURES 0
0 3.
3.

w I I SOLID POINTS ARE I

- I I I R 1.0 * BELOW CENTERLINE
- -2.00 " OPEN POINTS ARE
wI-
cc :I1 * 0 aTHEORYc
2.0041.
ABOVE CENTERLINE I

"
I!t
lot
I -- 1w I
-.- S,=
go* =
l~ i~
1.5 :.TKEORY1
0

WALL
I0.2 2.0 0.4 0.6 1.4 1.6 1.B

0.0
8,0 1.2 1.4 1.6 1.3 2.22.4 EXIT PITOT/INLET TOTAL, EXIT MACH
A/A 4 AREA RATIO P,/Pt4 NUMBER, 5
c) Experimental and theoretical cornbustor-exit profiles.
b) Experimental and theometical combustor pressure distribution.

Fig. 21 Comparison of experimental measurements with theoretical

analysis for combustor with normal shock at entropy limit.

respectively. Since these data were used to obtain the empirical relationship for ad given in Eq. (5), the
positioning of the shock train with respect to the fuel injectors is reasonably well represented. Some
changes in the theoretical pressure distributions from those presented in Ref. 10 have been made due to the
improved modeling of the "Z" term in Eq. (8). For the conical cc-ibustor, no joining process is warranted
since it appears that station s and e are about coincidental. It the short cylinder-cone and the step cylin-
der cone, the constant area sections necessitated a joining procedure to complete the theoretical analysis.
In the absence of any well posed theory or experimental correlation, the simple assumption of a linearly
increasing total temperature with distance was assumed to hold between the shock train and the - c region.
This zone is so short in the short cylinder case that no judgment on the appropriateness of the modeling can
be made. In the step cylinder cone, both the analysis and the experimental pressure distributions show that
the flow is "overcomoressed" with respect to the duct area by the shock train, i.e., the effective cross-
sectional area of the flow is smaller than the duct area at staion a. The theoretical pressure distribution
in the joining region was based on the assumption that the total heat release wa, evenly split between the
cylindrical and conical stations and station e was located at the entrance to the 0.7* half-angle cone. With
these assumptions, the pressure is predicted to decrease by about 10-15% in the constant area section.

A much larger data base will have to be analyzed before any conclusive arguments can be made to improve
the modeling. Nonetheless, the general agreement between theory and experiment is considered to be quite
good.
 8-23

FUEL 4
INJECTOR

I-10.37- 24.00CONE0-... J
CYLINDER 2.67
CYLINDER

-12-8 4 0 4 8 12 18 20 24 28

DISTANCE FROM FUEL INJECTOR (IN.)J-.

Tt 4 P4 T4 P4 Tf ER "'

R pea R psia R.+-

04135 458 1581 7.52 1158 0.78 0.81
:4150 458 1590 7.82 1160 0.49 0.92

a) CONICAL COMBUSTOR

FUEL
IN.,cCTOR
AIRFLW 2.74 I.D. -1.3 3.876 I.D.
FLOW __ CYLINDER CONE "
2 4.0 0 CO NE - 4 5 - -1 2.9
CYLINDER 14
2i1*19CONE "
0 .32 - - - 0.3

0 .10 - I - : 74 I .D
. '3. 7
.-0.7 *ZL 5 " K--3
3.28 I .D DO
. '.
0 FUEL
.oe - S-- INJECTOR .08 ,- a,n
1-i

.02.0 00

0-84 04 8 2 20 242 32

DISTANCE FROM FUEL INsJECT)R (IN.) DISTANCE FROM fUEL INJECTOR (IN.)+,
Tt2 T4 P4 Ttf ERl7 .Tt Tt4 Pt 4 P4 ER0 J

"*R pu *R pI.l *R "R psis R Isla R"."

O 4000 449 1515 7.38 516 (180 0.572 O-875 448 1452 7.32 112 0.93
4 0.68

03970 450 1500 7.43 1185 0.51 0.89.

b) SHORT CYLINDER CONE COMBUSTOR a) STEP CYLINDER CONE COMBUSTOR

Fig. 22 Comparison of theoretical and experimental wall pressure distributions in supersonic combustors.

Since in the design of a fixed geometry engine A5 /A4 must be specified, a procedure for selection must be
developed. Each application imposes different constraints, e.g., the available length will directly affect
the combustor wall shear, whereas a change in combustor area ratio may or may not cause a change in the
allowable fuel volume, therefore there is no one design strategy. Nonetheless, some of the steps in the
process that can be general'.zed are as follows.

Generally, as mentioned, the maximum thrust requirement occurs at NEW, %ndA5 /A4 is selected to meet the
requirement and provide some margin to assure that the missile will climb out and accelerate when anomalous
conditions arise. However, an overly conservative value for A5 /A4 is not a Judicious choice, because pro-
vision for additional thrust margin invariably leads to engine inefficiency during crui3e. The following
example should help clarify these points. Consistent. with the general nature of this discussion, the sim-
plified assumptions that were adopte') to develop Figs. 18-20 are the basis for the curves shown in Fig. 23.
Here, the maximum values of T /Tt for the normal shock, HM- 1 limit are shown as a function of Hf 4 for
5 4
values of A5 /A4 ZrL 1 to 3. The aforementioned trend of increasing (Tt /Tt ) with Increasing H4 is
5 14 max
characteristic of all of the A5 !/A4 curves. Overlaid on these curves are curves drawn through points corres-
ponding to the combustor requirements fol the four ITS inlets (Cases A through D) that were previously dis-
cussed (Table 2). Here, H4 - H 3 and4 the vai;:As of the required T /T assume that the engines are opera-
e3 ~t t: 54'.
 0*4*'

SAsjA 4.0 4 I I I IJ I

"03.00 2.75 2.50 _2.25

w" xs, 1 .50

2.1.0

I- . I I I I I 1.2

.. 1.6 1.8 2,0 2.2 2.4 2.8 2.8 3.0 3.2 3.4 3.6 3.8 4.0
SCOMBUSTOR ENTRANCE MACH NUMBER M4

VFig. 23 Maximum total tempeature in dlverwng combustors normal shock pressure ri.e. M. - 1.
5425

'..'-' ting in the tropopause with TO - 390"R and that the fuel ii Shelidyne H (R.J-5) with a heating value of 1
I gtu/lbm and a atoichiometric fuel/air ratio (ER - I) of 0.07284. The required Tt /Tt4 values are based
0 0
real gas in thermodynamic equilibrium, but all other parameter. are consistent with the simplifying assus
tions of a calori.cally perfect gas with Yr4 - -5 1.4 and no mass or momentum of the fuel.

The open symbols are for Tt /Tt values that correspond to ER - 1.0. The required A5 /A 4 vary from 2
"2.5at H0 - 4 to b... 545 The solid .symbols are for ER - 0.5 and the required0. 5
at H0 = 7 for this ER.

Sare correspondingly reduced, varying from 1.58 to 1.91. To ..ize the combustor, the thrust requirements
would have to be known, but in this example it is assum~ed that at H0- HEOB - 4, Tt /Tt4 - 2.5. The cort

P.. ponding ER would be 0.639 and the values of A5 /A4 for the four engine designs are listed in Table 3. Fox
i"." maximum acceleration in the climbout, the fuel control could be prograrmed to increase ER to the allowabi
." limit as H0 increased. This would occur at H0 from 4.3 to 4.4 for the four cases. Zn this example, fLt
-". Is also assumed that the engines would accelerate at ER - 1 until Mo - MCR - 7 at which point they would
throttlMB beck to ER - 0.5 for cruise.Cbe

n With A/A andT W specified, solutions can be found for each set of initial conditions. Result

* from these calculations are suearized in Table 3. At H0 - 4 all cases correspond to normal shock preass
es with H5
reas 1
em Asi 0 increases, H5 also increases until a conditisn is rench here it is possible

'% obtain an entropy limit solution. Normal shock and entropy limit processes are labeled N.S. and S.L., re
*. spectively, under "Type' in Table 3. For all points other than Case D at H0 > 6, the static pressure at
combuetor exit is higher than it is prior to the shock train. In the absence of a more detailed analysit
th5 viscous layer at sto atio, i io us assuTe that (A/A 5l a(A/Afor ER h-.5assumptionres made t

"obtainthe i of A0 /As
values shown in the Table. Total pressure recovery in the combustor, pt /p , incre

MO.effoctl tninial Hachin number more witk

the
i inlet orcompression
asicompensate
thx increases
the dMtried
because
talwlffdoc the beneficial
of larger A5 /A 4 . Co4bining the redoced
op pressure recovery the inlet

combustor
s rasulte in the values of Ptc/pt0 shown in the last column. An extremely important trend is shc

in the values of pt /pO. C.is. C yields higher total pressure recERery than Case D, even though there is

"W"less
compreson than for Case D. Tsuf, sltire would be no reason to select D and exacerbate the possible
problems of boundary laterasmsearis due to reflections from the inward turning scoop inlet.
3h.ck Thi
rawking of Cases 1A, and C minas,also take into consideration forebody forces and additive drag before a
final ranking can be established.
 Table 3

Combusror Parameters for Four Engines with ITS MFc m 6 Inlets

Case H0 A5/A4 H4 ER Types Tt /T t4 H p/ P5t 'Pt0

* ~ ~~
0 4 4 ~ ~ 5 ~~ 4 sP A 5 ~4 5 0
A 4 1.717 2.404 0.639 N.S. 2.500 1.000 6.58 3.12 2.968 1.989 0.3C19 0.2936
5 " 2.999 1.000 S.L. 2.241 1.479 9.33 1.63 2.467 2.329 0.2396 0.1622
6 " 3.512 1.0C'` S.L. 1.805 1.799 6.64 1.97 2.215 2.665 0.1638 0.0967
7 " 4.055 1.000 S.L. 1.463 2.342 4.40 2.40 1.744 2.665 0.1427 0.0705
7 " 4.055 0.500 S.L. 1.290 2.773 3.00 2.64 1.258 2.665 0.2005 0.0W1

B 4 1.822 2.159 0.639 N.S. 2.500 1.000 5.27 4.05 2.378 2.174 0.4489 0.3209
5 " 2.688 1.000 N.S. 2.241 1.493 8.26 1.70 1.925 2.580 0.3059 0.1811
6 3.116 1.000 S.L. 1.805 1.793 5.80 1.97 1.715 2.941 0.2230 0.1077
7 " 3.571 1.000 S.L. 1.463 2.292 3.84 2.37 1.361 2.941 0.1994 0.0752
7 " 3.571 0.500 S.L. 1.290 2.675 2.67 2.59 1.006 2.941 0.2672 0.1007

C 4 1.954 1.921 0.639 N.S. 2.500 1.000 4..4 6.39 1.871 2.255 0.5117 0.3321
5 " 2.380 1.000 N.S. 2.241 1.478 6.44 1.85 1.503 2.707 0.3773 0.1882
6 " 2.751 1.000 S.L. 1.805 1.800 5.04 1.98 1.298 3.119 0.2961 0.1133
7 " 3.141 1.000 S.L. 1.463 2.249 .'.38 2.34 1.051 3.119 0.2677 0.0758
* 7 " 3.141 0.500 S.L. 1.290 2.583 2.40 2.55 0.798 3.119 0.3423 0.0969

"D 4 2.154 1.652 0.639 N.S. 2.500 1.000 3.02 44.56 1.377 2.129 0.5674 0.3138
5 " 2.081 1.000 N.S. 2.241 1.477 4.89 2.08 1.117 2.555 0.4467 0.1776
6 " 2.400 1.000 S.L. 1.805 1.827 4.31 2.00 0.931 2.940 0.3814 0.1093
7 " 2.731 1.000 S.L. 1.463 2.223 2.97 2.32 0.774 2.940 0.3511 0.0701
7 " 2.731 0.500 S.L. 1.290 2.509 2.16 2.51 0.603 2.940 0.4279 0.0854

N.S. - normal-shock processes; S.L. - entropy-limit processes

" Solutions for the flowfield in the combustor based on the use of finite difference solutions of the
differential form of the conservation equations are also in an active state of development. The deficiencies
in these techniques are in the formulations for modeling turbulence and kinetics and in incorporating boundary
constraints that are not in opposition to experimental observations. indeed, it has been found that the most
successful approach has been to impose a wall pressurs distribution governed by the foregoing integral analv-
"sis. References 13 and 27 discuss this methodology for the particular case of the DCR engine cycle. The
practical limitation of the cost of the comrutstions still precludes the use of the finite difference approach
- as an effective design tool. but the value of having the capability to examine the entire flovfield in detail
"* for a few selected cases cannot be underestimated. As reductions in the cost of CPU time accrue and better
techniques for adaptive grid point spacing evolve, these techniques will play an important role in design.

NOZZLE ANALYSIS
The analytical tools that are required for the design of the exit nozzle and the analysis of the flow-
"fields for the supersonic combustion engine are not significantly different than those for the CRJ. The
principal difference is that the throat station in the converging-diverging nozzle provides a reference plane
in which sufficient constraints are present to describe the flow properties, albeit with certain assumptions
regarding the chemical composition of the gas. In the scramjet, the distinction between nozzle and combustor
can be obscure and as the previous discussion suggested, station 5 can, in fact, change with changing con-
ditions at the combustor entrance. Moreover, significant gradients in flow properties in the initial plane
are much more likely to be present in the scramjet.

"As in the case in the combustor, both integral analyses and finite difference techniques have been
developed. Emphasis shifts from properly modeling the mechanisms that contribute to the losses in total
pressure to adequate representation of flow angularity, especially in the nozzle exit plane. The greatest
deficiency in nozzle flow analysis, as iv the combustor, is in the modeling of the chemistry. In the com-
p bustor the processes during the early phases of reactions, e.g., chain breaking,cause the greatest diffi-
culties. In the nozzle, modeling of the recombination reactions and, in the case of some specialized fuels,
phase changes of metal oxides are problems.

Much of the work in the development of useful design approaches based on finite difference techniques is
still in progress, but Refs. 31 and 32 give some insight into these methods. The integral analyses are at
present quite rud4imentary. Given a unidimensional representation of the flow at station 5, calculations of
isentropic expansion corresponding to the effective nozzle area ratio A6 /Ase era made for two cases: one
3 assuming thermodynamic equilibrium and the other assuming a constant chemical composition "frozen" at station
S..' 5. Two empirical constants, r and f, are then introduced, the first to account for the loss in nozzle exit
stream thrust that is attributable to finite rate reactions (and phase changes) and the second to account for
friction, divergence and non-uniforaity. Thus,
" flNr N
p6 + (l - r) 6 ] (13)
EQ FZ
where

a6 P6 A6 + 6uA(14)
 and typical values are r - 1/3, T) - 0.97-0.99. An adequate data base to substantiate the validity of this
modeling has not yet been forthcoming.

With I6 defined, the thrust and net engine force can be determined for specific values of Ai/AR, whet
AR is the frontal area of the engine. In Fig. 14, ' - 45' for each of four inward turning scoop inlets,
thus A /AR - 0.5. By -hanging either or the number of scoops, or both, Ai/AR can be varied with an upper
bound that is governed by starting. The relative performance of engines corresponding to Cases A-D can be
obtained by examining a single value of A/AR, e.g. 0.5, and then the effects of changing A/AR can be
"studied independently.
The thrust coefficient is given as

CT -OA 6 - -= 2 [p6 (1 + M6)- - 2 AO(0

"and thedrag coefficient due to the pressure acting on that portion of the conical forebody not wetted by
the air captured in the inlet is

The ddtv for 0 < MFC is given as

2.- w [1[+ YM-(sin(ew-1)cosN A_

0 2 YMP0
DD sin 0w1 2

Therefore, the resulting net force coefficient is

CF - C T --DC - C DADD

Table 4 lists the Mach number and pressure ratio at the nozzle exit plane and the calculated force co
efficients for Cases A-D. Again, for clarity, the simplifying assumption of y - 1.4 was adopted, thus r
and nN wa3 taken as 0.985. At M0 - 4, the exit pressure is somewhat below the ambient pressure, but the
values are not low enough to suggest separation in the nozzle. The gross thrust coefficients, CT, exhibit
the same general trend as the pt/Pt values discussed earlier. However, when additive drag is subtracted,
t5 t0
it turns out that Case B yields the highest values, and if the forebody drag is subtracted, Case A is best.
If other values of Ai/AR between the limits 0.4 < Ai/AR < 0.8 are examined, the trend is toward higher re-
lative values of CF with greater inlet contraction as Ai/AR increases. Nonetheless, over the entire range
6
reasonable values for Ai/AR, the optimal conical forebody angle is c < 12.5". These results clearly show
"need to consider additive drag and forebody drag if a true optimum is to be identified.

"Typical values of the ratio of missile weight to maximum cross-sectional area are from 4 to 7 lbs/in2,
For a value of 5 lb/in2 and flight at M0 - 4, Z - 5000 ft, a value of CF - 0.26 would correspond to 7.1 "g'
of accelerative capability.

It is of interest to compare the results obtained from these dual mode engines with those from a set
engines having the s~ne inlets but with dump-typo subsonic combustors. As in the case of the scramjet,
acceleration requirements at the end of boost prescribe the combustor-nozzle area ratio. To obtain maximut
performance in the CRJ, the nozzle throat is sized to produce critical flow in the inlet at this condition,
In an inlet having internal contraction such as the ITS, the critical condition corresponds to locating a
normal shock (in actuality, a shock-train having a pressure rise equal to a normal shock) at station 3.
* These conditions are shown in Table 3 and are labelled 3'. As M increases and/or ER decreases, the total
temperature ratio in the combusto- decreases and the normal shoci seeks a location where A > A3 e,. Para-
. doxically, the inlet must throw away pressure recovery in order to satisfy the sonic flow conditions imposo
at the nozzle throat. For the example cases, this area is set by the M0 - 7, ER - 0.5 operating condition
"Larger ductareas would be possible, which would lead to smaller duct and dump losses but with a loss in ti
internal volume available for packaging. The minimal duct area was used to obtain conditions at the com-
bustor dump plane. Total pressure losses during heat addition vary inversely with combustor entrance Mach
number, so to obtain minimal loss the combustor area was taken as AR. Again, to simplfy the calculations,
y - 1.4 was assumed, the duct flow was frictionless, the dump was considered as an isobaric area expansion,
and the heating was taken as a Rayleigh process. With these assumptions, the total press-re loss from
station 3e' to the nozzle throat and the throat size for the M0 - 4, ER - 0.639 operatlng conditions can bo
calculated for each of the four engines. These losses vary from 16% for Case A to 21.1% for Case D which,
%
%
when coupled with the inlet losses (Table 3) give the values of p t*/Pt at M0 - 4 shown in Table 5 and pre-
0
scribe the nozzle throat size A /A that is also listed in this table. For the M0 - 5, 6 and 7 conditiont
pte/Pt 0 is deduced from the continuity equation using the lower values of total temperature rise in the cot

bustor given in Table 3. Proceeding backwards through the engine cycle computing first the total pressure
loss in the Payleigh heat addition an i the isobaric dump lost provides a means of verifying that the
inlet must indeed operate supercritically at all but the H0 - 4 condition. Had total temperature ratios
below 2.5 (ER - J.639) been prescribed at the end of boost, A /A e would have been set by the H0 = 5 con-
ditions with critical inlet operation. The static pressure ratios p /P 0 at the sonic point listed in
Table 5 are about one-half of the maximum value in the combustor. When compared with the maximum pressure
ratios (pp/Po) that occur in the scramjot at the same operating point, a significant difference is present
at high H0 . For example, taking case B at H0 = 7, ER - 1.0, the maximum pressure in the subsonic combustor
is 103.79 0.5302 - 195.8, whereas in the scremjet combustor it is 3.84 x 17.445 - 67.0. The implications
with respect to the structural design of the combustor and nozzle are obvious.

With A*/A prescribed and A6 - AR the nozzle exit conditions p6/PO and H6 and the force coefficients can
be calculated by the same method used in the scramjet analysis. The values of CDADD and CD are the same as
for the comparable scramJet at the same operating conditions and therefore are not re-listed from Table 4.
Comparing CF values from Table 5 for the CEJ with those for the scrsmjet from Table 4 shows that th% CRJ has
somewhat higher performance at M0 - 4 and 5, about equal performance at M0 - 6, and significantly lower per-
formance at M0 - 7. For most applications of engines operating over this M0 range, overall performance
would favor the scramjet. A superficial examination of the listed -esults will help to support this argu-
ment. For example, if the CLJ engine for Case B was required to produce the same CF as its scramjet counter-
part, the ER, or fuel flow rate, would be reduced by about 9% at H0 - 4, 5% at M0 - 5, but would have to be
increased by 50% during cruise at M0 - 7. Both cycles exhibit the same trends in performance with inlet
contraction ratio and the conclusions previously stated for the scramjet therefore hold.

Table 4

Exit Parameters and Force Coefficients for Four Engines

with ITS MFC - 6 Inlets
(Supersonic Combustion)
Case H0 ER H6 P 6 /Po CT CDADD CT - CDADD CD CF

A 4 0.639 3.246 0.848 0.3370 0.0124 0.3246 0.0584 0.2663

5 1.000 3.402 1.293 0.3219 0.0059 0.3160 0.0555 0.2605
6 1.000 3.631 1.668 0.2291 0 0.2291 0.0538 0.1753
7 1.000 4.141 1.257 0.1173 0 0.1173 0.0527 0.0647
7 0.500 4.611 1.237 0.0631 0 0.0631 0.0527 0.0105

B 4 0.639 3.347 0.795 0.3530 0.0188 0.3342 0.0804 0.2538

5 1.000 3.515 1.228 0.3402 0.0089 0.3313 0.0771 0.2542
E 1.000 3.733 1.610 0.2500 0 0.2500 0.0752 0.1748
7 1.000 4.202 1.573 0.1328 0 0.1328 0.0739 0.0588
7 0.500 4.617 1.247 0.0756 0 0.0756 0.0739 0.0016

C 4 0.639 3.409 0.772 0.3565 0.0257 0.3307 0.1057 0.2250

5 1.000 3.570 1.182 0.3283 0.0120 0.3163 0.1021 0.2142
6 1.000 3.802 1.538 0.2373 0 0.2373 0.0999 0.1374
7 1.000 4.224 1.542 0.1209 0 0.1209 0.0986 0.0223
7 0.500 4.584 1.250 0.0623 0 0.0623 0.0986 -0.0363

D 4 0.639 3.348 0.777 0.3248 0.0257 0.2991 0.1057 0.1934

5 1.000 3.507 1.220 0.3250 0.0120 0.3130 0.1021 0.2109
6 1.000 3.763 1.567 0.2354 0 0.7354 0.0999 0.1355
7 1.000 4.132 1.605 0.1185 0 0.1185 0.0986 0.0199
7 0.500 4.434 1.327 0.0575 0 0.0575 0.0986 -0.0412

0.985 r - 1
 Table 5

Nozzle Throat and Exit Parameters and Force Coefficients for Four Engines
with ITS " 6 Inlets
-C

(Subsonic Combustion)

Case M0 ER A*/A3e Pt*/Pt0 P/ 0 CT CT - CDADD CF

A 4 0.639 1.293 0.3898 31.27 3.544 0.729 0.3562 0.3438 0.2855

5 1.000 " 0.1853 51.79 " 1.207 0.3366 0.3307 0.2752
6 1.000 " 0.0894 74.53 " 1.737 0.2225 0.2225 0.1687
7 1.000 " 0.0411 89.84 " 2.094 0.0886 0.0886 L1.0360
7 0.500 " 0.0386 84.35 " 1.966 0.0203 0.0203 -u.0323

B 4 0.639 1.325 0.4412 35.39 3.689 0.674 0.3742 0.3554 0.2750

5 1.000 " 0.2126 59.43 " 1.131 0.3528 0.3439 0.2668
6 1.000 " 0.1035 86.57 " 1.648 0.2518 0.2518 0.1767
7 1.000 " 0.04'/5 103.79 " 1.976 0.1085 0.1085 0.0346
7 0.500 " 0.0446 97.47 " 1.855 0.0392 0.0392 -0.0347
C 4 0.639 1.408 0.4612 36.99 3.764 0.635 0.3480 0.3223 0.2165
5 1.000 1 0.2249 62.87 " 1.080 0.3406 0.3286 0.2265
6 1.000 " 0.1093 91.15 " 1.565 0 2347 0.2347 0.1348
7 1.000 " 0.0502 109.87 " 1.887 0.0995 0.0995 0.0008
7 0.500 " 0.0472 103.18 " 1.772 0.0307 0.0307 -0.0679

D 4 0.639 1.496 0.4518 36.24 3.740 0.643 0.3477 0.3220 0.2163

5 1.000 " 0.2202 61.55 " 1.093 0.3402 0.3286 0.2262
6 1.000 " 0.1071 89.31 " 1.585 0.2355 0.2355 0.1355
7 1.000 " 0.0492 107.69 " 1.912 0.0838 0.0838 -0.0148
7 0.500 " 0.0462 101.12 " 1.795 0.0305 0.0305 -0.0681

fl o0.985 r- 1

Al/AR - 0.5

CONCLUSIONS

A aummary of the contemporary approaches to the design and analysis of scramjet engines has been pre-
sented. Where possible, comparisons with data have been made and judgments made regarding the adequacy of the
modeling. In many instances the treatment herein is superficial and only in conjunction with a thorough
knowledge of the ever-growing base of reference material can the designer properly assess the state-of-the-
art.

REFERENCES

1. G. L. Dugger, "Comparison of Hypersonic Ramjet Engines with Subsonic and Supersonic Combustion,"
Fourth AGARD Colloquium, Milan, Italy; also Combustion and Propulsion-High Mach Number Air Breathing
Engines, Edited by Jaumotte, Rothrock, and Le Febvre, Pergamon Press, New York, 1960.

2. P.J. Waltrup, G.Y. Anderson, F.D. Stull, "Supersonic Combustion Ramjet (Scramjet) Engine Development in
the United States," Proceedings of 3rd ISABE, Deutsch Gesellechaft fUr Luft und Raumfahrt, Cologne, West
Germany, Rept. No. A77-17266-05-07, pp. 835-861, 1976.

3. "Hypersonic Research Engine Project - Phase II, Aerothermodynamic Integration Model Development - Final
Technical Data Report," (AP 75-11133, AiResearch Manufacturing Company of California; NASl-6666) NASA CR-
132654, 1975.

4. T.D. Burnette, "Dual Mode Scramjet," The Marquardt Corp., Van Nuys, CA, AFAPL-TR-67-132, June 1968.

5. F.S. Billig, P.J. Waltrup and R.D. Stockb:idge, "Integral-Rocket Dual-Combustion Ramjets: A New Pro-
pulsion Concept," Journal of Spacecraft and kockets, Vol. 17, No. 5, Sept.-Oct. 1980, p. 416.

6. J.L. Hunt, P.J. Johnston, J.M. Cubbage, J.L. Dillion, C.B. Richie, and D.C. Marcum, Jr., "Hypersonic
Airbreathing Missile Concep'ts Under Study at Langley", presented at AIAA 20th Aerospace Sciences Meeting,
Jan. 11-14, 1982, Orlando, FL, AIAA paper No. 82-0316.

7. P.J. Waltrup, F.S. Billig and R.D. Stockbridge, "Engine Sizing and Integration Requirements for Hyper-
ionic Airbreathing Missile Applications," AGARD Propulsion and Energetics Panel, 58th Symposium, Ramjets
and Ramrockets for Military Applications, London, England, Oct. 26-30, 1981. Also, The Johns Hopkins
University Applied Physics Laboratory Preprint PEP-58-8, Oct. 1981.

8. F.S. Billig, G.'.. tugger and P.J. Waltrup, "Inlet-Ce-bustor Interface Problems in Scramjet Engines," let
International Symposium on Air Breathing Engines, Juae 19-23, 1972, Marseille, France, Johns Hopkins
University Applied Physics Laboratory preprint, June 1972.

9. F.S. Billig, R.C. Orth and J.A. Funk, "Direct-Connect Tests of a Hydrogen-Fueled Supersonic Combustor,"
NASA-CR-1904, August 1971.
 10. P.J. Waltrup and F.S. Billig, "Prediction of Precombustion Wall Pressure Distributions in Scramjet
Engines," Journal of Spacecraft and Rockets, Vol. 10, No. 9, September 1973, pp. 620-622.

11. P.J. Waltrup, F.S. Billig, R.D. Stockbridge, "Procedure for Optimizing the Design of Scramjet Engines."
Journal of Spacecraft and Rockets, Vol. 16, No. 3, May-June 1979, p. 163.

12. S. older and E.J. Szpiro, "Busemann Inlet for Hypersonic Speeds," J. of Spacecraft and Rockets, Vol. 3,
No. 8, 1966.

13. M.D. Griffin, F.S. Eillug and M.E. White, "Applic~tions of Computational Techniques in the Design of
Ramjet Enginec," 6th ISABE, Paris, France, June 1983. :

14. F.S. Billig, M.E. White and D.M. Van Wie, "Application of CAE and CFD Techniques to a Complete Tactical
Missile Design," AIAA 22nd Aerospace Sciences Meeting, January 9-12, 1984, Reno, NV, ALAA paper No. 84-
0387.

15. A. Kumar, "Numerical Analysis of the Scramjet-Inlet by Using Two-Dimensional Navier-Stokes Equations,"
NASA TP-1940, Dec. 1981.

16. A. Mager, "On the Model of the Free, Shock-Separated Turbulent Boundary Layer," Journal of the Aero-
nautical Sciences, Vol. 23, Fo. 2, Feb. 1956.

17. J. Persh, "A Theoretical Investigation of Turbulent Boundary Layer Flow with Heat Transfer at Supersonic
and Hypersonic Speeds," U.S. Naval Ordnance Laboratory, White Oak. MD. NAVORD Report 3854, May 1955.

18. P.J. Waltrup and FS. Billig. "The Structure of Shock Waves in Cylindrical Ducts," AIAA Journal, Vol. 2,
No. 10, Oct. 1973, pp. 1404-1408.

19. R.D. Stockbridge, "Experimental Results from the HDCR Combustor/Inlei. Interaction Test Program," The
Johns Hopkins University Applied Physics Laboratory. Laurel, MD, June 1982.

20. L. Crocco, "One-Dimensional Treatment of Steady Gas Dynamics, Fundamentals of Gas Dynamics, Vol. III of
High Speed Aerodynamics and Jet Propulsion, Princeton University Press, 1958.

21. F.S. Billig, "Design of Supersonic Combustors Based on Pressure-Area Fields," Eleventh Symposium (Inter..
national) on Combustion, The Combustion Institute, 1967, pp. 755-769.

22. E. C. Anderson and C. H. Lewis, "Laminar or Turbulent Boundary-Layer Flows of Perfect Gases or Reacting
Gas Mixtures in Chemical Equilibrium," NASA CR-1893, Oct. 1971.

23. F.S. Billig and S.E. Grenleski, "Heat Transfer in Supersonic Combustion Processes," presented at Fourth
International Heat Transfer Conference, Versailles, France, August 1971.

24. S. Gordon and B. McBride, "Computer Program for Calculation of Complex Equilibrium Compositions," NASA
SP-273, 1971.

25. D.R. Cruise, Notes on the Rapid Computation of Chemical Lquilibria, J. Phys. Chem. 68, 3797 (1964).

26. F.S. Billig and G.L. Dugger, "The Interaction of Shcck Waves and Heat Addition in th3 Design of Super-
sonic Combustors," Twelfth Symposium (International) on Combustion, The Combustion InstiLute, Pittsburgh,
PA, 1969, pp. 1125-1134.

27. J.A. Schetz, F.S. Billig and S. Favin, "F'owfield Analysis of e Scramjet Combustor with a Coaxial Fuel
Jet," AIAA Journal, Vol. 20, No. 9, September 1982, pp. 1268-1274.

28. L.C. Dunsworth *nd K.E. Woodgrift, "Dual Mode Scramjet, Part I: Inlet Desisu and Performance Charac-
teristics," The Marquardt Corp., AFAPLTR-67-132, Part I, December 1967.

29. C.A. Trexler, "Inlet Pcrformance of the Integrated Laugley Scramjer Module (Mach 2.3 to 7.6)," presented
at the AIAA/SAE 11th Propulsion Conference, Anah,-im, CA, Sept. 29-Oct. 1, 1975, AIAA Preprint 75-1212.

3' L.F. Jilly (Ed.), "Hypersonic Reseir ch Engine Project - Phase I1: AM.rnthermodynamic Integration Model
Develolment Final Technical Data Report," Garrett Airesearch Manufacturina Company of California, NASA-
CR-132654, Mey 1975.

31. ". Agosta and S. Hamer, "3Sciazjet Nozzle Aita.lysia, PEI Report 7%-1, Propu.sion Sciences, Inc., Melville,
N.Y., Februaty 1970.

32. D. VISdal, "Supersonic Annular Nozzlel." AIAA 9th Aerospace Sciences Meeting, January 1971,Preprint 71-
43.
BIBLIOGRAPHY
 N V GJM di Gi ap iA -a . vo cu -i w Z; 0 N 0= W a M 0 a, V, vi
Zc
a aa 0 Ww 0 M v z r C; u .0 CRI x C
W4.1.0 m m C ww z cu
4. _j a z 14
> 4,
D .0 Z " 2: m
P . ': =3 c 0z 4-C,
M_ m V N m 9: 4, .0 0 'p z wj 0Z (P
LO z m > c
7" -P 4. 0 C, z 0 0 z 0 9 u - v 0
!2 C, 0
0)
M Lr7 w
.. . m a
1* -0 - . In - x a) CL m Pc
cc
w a
-1
. ' x (nd a m 0 0, v o a -, Q.= 3 c z 'm a) q
4, A 0) 0) M.C a al 0 aj z x cu , ri 0 C3= a z
(D Ic. u u M
z ... p m p 7 _j 3 41 4, z CL 41 w
6j z 0 rl 0 _j '0 m '0 10 z
MCI.PZ Go 0 14. Z .. 0 41 a,m 0 _Wl a, a
017 N 0 'r f 0. o o z z 0 9L - ru x W a -W 4.
v A < u 0 En 0 E 0) C 4 Dwo J1. 0_4 w
4 3 ciZ CL U0 'j a w 1; C , 4, w M ki 0 ao)
lu 4 c _j C o- w ana)
3
W,,r OZX
kA4 u CU
tn CX
o a) M. 0
4, Q, _j Lu Li UJ
z T r E r x m 'o
r
w 5_ . 0 1 11 m< Z 0. ` 8 N
0. 0 0
3 cn _J :3 i W
Z a) 4 to a, Gi 41 ol Z ce > a
0
a- 'c CL
m Ic
0. m r 0 0) r_
do 0
Ift 4%
Ic L En 0 a cE 0) 1 lc al u i fd
w ul (Oz u
u 61 GO 41 > m 41
Z N 41 C LLI w -C
IA0 (n-x z r m 0
mm u do 0 W0 v - r4 u 0 w .0 %, :c rj U - M p (o V m
19 L - 4J (d M 1, Z 3 a: 4, le .0 u z 0 ,.a 41. 0 ri
8 a.. M 46 a
cu 0 u M cn w ti)
_j w -0 u .0 re ws u C, m w z Ift C cn Lu w
M m- 0)-6 0 m c - LU In 0 r) u 0 11 > 11r_ ci z 6< -
0 1 a, 41 m cuL') c u 0 > al W N 0: LUZ u t= 0 U-0 Z. 0
:1 ern A 'A4 w (n 0 C, (n ! u k" 0 Im 0 11 41
cn
'z C3'x I w
c ci w N mU fn w ev En N a, a - 0 1-0 u u Cc Ln
al a uc .0 E! u La M
z w tn T a,
C, 0
4,
4j E 0 E
3
L x
0
U) x > M x v =Zzz u a
o u v 0 m OL 40 L0
01 0W. u 0 C,
0 ob 0 r. 0 0 0 0 "a u 14. . c uj m m 71 u fro m 41
fl 4 C Lu
Id 40 0 L w z z c 00 CL4) j 4 M p Z lp U) 9) 0
at c 41 L v zlw m m w 4, z .0 0 owo u
m cl w ID 'L '3 w u CLQ) m -1 00 10, M
1 .1 CI_ >'a It. _j W w 0
'a m z 5 o a z N 11 z _ M a I Ch cc z 4 cu - ul 0.
.0 4,
%.xa t' 3 a w 14. 11 to (Dw x = a N_ 0 0 v 11 'L 41
0 0 u M z In r 3 0 -0 x N 0 c 4 -P 0 - 4, C -Z ow 0
C. I. c 91 .4 (1 to 4. Go = a E u In 1. C 14 C -P 41 4 4, 41 z C9 1, L M . u (n -
c .3 %.4, -A r - CL 0 - c - - 0 0 _ >. .4 w W 0 4. S. r. Qw 't, > 2 OLx 0 u ag a
4, 0- 0, 6 > lb 0) 0 N 4) Qj Z 3 (n 00;w M 41 - LA U 0, '- fa z D 3 al
go
-0 CL.;
v 91 Q. 06 > w J al 'i -'o u r_ .2 Ch0 u S. L. el 0 Q0 lov
ai CL 3 N a u -1 Z 'r . I u ij Z cL.6 s.0 cc in 5 .0 z -Z
0- N Id ra
.0 4, k. 6 t;hc 41 -0 0 WZN 3 Ala) q. 4C do 4 3a, ol C
1.- 0, tj _ t :3 w0 ki r
N W 0 41 4 M,4 W 0 C En L _j to CL fl 0 w
r
4
ww 03 N > p
gr c 14.
a w 'n r 4) J m WLn rdW fa ea
aM IC CP "
3 r E 3.0
W -P0 Qo Za cc Z W
a cn 4. U C2 -, 0 > w0 _j LP z
>. M 0..v r C i.. a z ol w 1 CL 0 6 co al fa o o M m , I . z_ 0, w 4 m
L. 0 a .. - 'I .' , I., a, 'n r 0. w U) 0 M J) r 0 4p "1 0 '. M S.. V 6 1 n ui . )- 1. - :3 =) m c CD
m 1 41 fa - a N 1. 0 0) > _j (n W (n CL Im I LA 0- 0 C 06 C 30 1! 0 U U) M 0 .0 O m
-PN a w a a I - Q I x z - or uj (n L) 'r 4 6 Z L 4 4.0 4 00 x w U) 4) E
- w It 2 "'o 4, CL
go0 0 4,
CL'o 13 1 3'A CL > 13 .0 cu u do -0 K
> al u . Ca li w 0w w A
CJ..
K j RIO 40 4-0 41 N 4. a C4. C-m 0 - 4. at 0 fd
N r
T u ccal T. -
O_ -'fl z'p .0 L
1' 0 1.
m t 41 M 0 (A W' (n
* c m 0
(n 0, m-u a
W V.- D c x1 2 41
1 0 .0
ID r. fd CL CK a r 1 0
0 0- = 5. 't a. %' 10 .0 to tft v w r 4
N a, .00 C a. 0 0 0 Z! 'x
0. 'o to 4 I z w
b, -P u 02 T3 m 0 0 0 ct CL u tv - 40 w u 0 c 0 u r. ds C'J
al a 4t U, . o w 0 z - - _I * I'll P4 0 'P _.z 0
ev a, > U;
'0 0 10 U .1 4 0 A- - w U" w c C 30. ull U0 0 Q. c m 4 3 . 41 c r L ol A;ow 0 41
'a 0 m a; a 0) - Go C v 4. 4. 4 w 01 4 m OC 8 'NI:; U. 4.1 z m 'A LA0,
0 cm 0 w
- w a 0=0 to u
2 u c CD0
m a j. w Q m -p 5 a, GA 0 CDw - I m.cc cn 0 a. :
0 In
c Ift z _j w 0 0 c fli L. < .00 't Qj 9 L' j
0 W 09 C J
u
M to Go
30.. a, C
S 06 lp 1) 4._ w li L' Z w z ul 17 % d' 4 06. >
ffl
0 Z W W m >
w a, 3.0 W CD 4- c :I to 1 0 11 9 dl I- a. 4 (A 0 c

OR"o U ta > 40 M0414-q-4, N

A.
0 m c 0 0 14 J 0 > 0 14 4 r'a
v D. > u '1 4. - ul *1 0, -m 0
o 0.- 5 - i. A 3 .. to 0 0)
to
0 Im.0 a) 0 4 av S. t 0 vi 0, L., $. co
9 M -P ko I u x w z 0 x w CIO>
u Ifflc die do r
> a 0 CL V 0 C 0 aj E 0 Ir 0 06 (3
X M *1 '4. U 41 1- <1 U 41 z
U 0 D M W 'u 11 9 In W
v w tr -V P ftj 0 r u ml u
C,4)
S.. In a a W u .1 a, w X x 6 T c w 0 'v 3 do 2 n 41t W
ru a P 6 0 m Z. 0 E 11 cu rd t k" ffl a -1 4 " u Cil
C, 01 n j
rd La 'm 1! 0 . N z C_W -P 0 a,
14 * 41 41
A 40 a,
L. 4, 0 Ir
0 -1 al CL I14 _A . m In . $. 11 - 0 (x Z 0I M u -P r = m CL- 4, CK C
N 0 al 91 004 O-P C 'j U - - K al r m ol to w
Go c a CO a fn 0 1 IV u -P
v > 4. w W 0 ff.. Qj 3 4 1 41 1 a .0
u CII.P a c 4 3U Cn w
Z M 0 41
II w
_j
N Z 0 4a ri IV
g I It '.f u w z fo c 0 al a . 0 a z
ch U) %. m 41_0 a u ke (a Go 0 L ed.0 x L CLcv a LLJ
L 4J A 4J X CO. W D 4J 41 3 0 En > 'n . . > u CL
0) 40 r 40 1 41 0 0 Id X tn., 4, A, 0.0 d' 0 N 'u 0 2:
w z
rd -C
0 D, 4 4. V r. n 0 1. C
.4. . CL m 0 c 41 :3 < . 0 a, do . 4 to - r 'I
2, a 0 > 0 0 w ix U 40 .. Q CAto u 4 L'306 -P
to z E w u al- v E W 0 = C" 0 a al . u if z
u _j do 10 d, M M 41 0 ft V (d U (a z ds -P 0 Ift N 4. 0 0) 14 .. 4 w
CA p 10 > a Z M v E w C) Z 3 u - CL 4 0:
0 0 1 0 6 uj
'A 0 4 IS z 3 J '- m CL
GI a l6 - 4
IAJ I IL 4- M 0 4, 1 m _r n - Lo 0 = o c q o 1, '_ w
.m tn CL '5 . 0 11 . M 0 _j w v U 'a V` , I , , 1 11
in 0 40 GO- m 4- 0 > 41 W G:D a CL 0 0 14. 3 w v Ix
W , m . a, 0 to a 0 0 r (d 0 010
Gi E _.P to w
a
u Q. > do a T .. 13 P a 40-0 w x C, u -P . 3 4p a$I (d ff
4141 x
41 a- 0 3 m a = - a w 'o v CL'v C' w w U..o
W 99 L 0 R d, 01. U M , .. 0 a C. w -a r
X (a
> x 41
. 0 0-10
41 0 -!c -W ty U 41
'3 0 U 0cm U. cu M - ,0 L
'a - 0.0
u a :" -'A Z, :
0 W02 C, 41
Nr Or)
I .'3 'c >. 'w a I In
-01 ft . 4J W W 0 E C z x
CL
lu c 5. V. ul
CL
a, P 3
'" '1 0 W
CDdo bl Go ul w cn w 0- r -P w to a 007
u
Mai _j 's A 'I o 2 w w _j v %ft " "a ..' Q
u Ul 0 5.16 0 (n -i m -P > V
Ar a 0- to -C 41 -Qc to a- w CA lz to r
v -P 0 > Co kj r- O c Of 0 X 0
V.It m ro 1- 0 0. , 0 M . m a
11 a CL cog '0.
w Z r .0 ff 3 up tn L. I. tj U) x gr 11 rm u r 'doin cm a c c S. Q T 0 ut
ic to so WE , . m & ' w u
m IR = a0 ..6 .S. u CS &. w -D to -.U WCL CL 0 ol
CL -P 3 6:.m . - (n -P
(n 0 U) M 5. 5 4 40 0 ul M(1)
, Z v0 0 CL
a CLCO u Lj
3 a, 0 -W 0 ; 4)
I fe -P 0 R a a w x w U) go x K In c u u A w a 4; al COL
0
" 'i to
. w 2 '4 .1
Q'x CL fa t ; > al 1 w
-0 0 4, & 1.. 3 0 .0 6 -P a, C C 41 u 0 41 X Li > 0 OC
3 A 0 . S.. 0 CL..0, 0 to 0 CL97 C; C 14.
c cL r ri 4, 0 U; 0 M
0, ju
j.- va ! 2 cK tn
z :; ZO 'fi a 0t
Pc . 4 ff . 0 Q $- 3 m Z t
0 z - LL -CIO 40 0 al Al 0W c)
IA 'j 19
4; 0. >. .0 a, u c 6 z to La
.00 _'I do In.A rc, -we . .0 %j 0 u0,60 41 4 0= d- a w
0w 0. 0
CU 0 0 di K al a 4 A
a .0
c no kj. .. to m tv L z _ x C" -0 a, t. _1140. a 3 0 7 111 1 -1 1
0 W_ , Z.C In a u .. ru -i 0 V, V 0 v 4 -1 to -P 0 QW
-P - 2 10,
U>O I r 1. L. 40 u 0 0 0 v c 0 tn ri)
U) 0 1, w V Q (u L :) 0
ru 0 c 0 . 'r S 0 CLI,., w cc 4' CL1 41 W, 0 en u 5-
.dr V W M WOO a 4 4j aU.P 0 a Zdo OE
0 -02 q. 41- tn 03 06 fe I Z< -a x IT
'A, 11, 10 cr 1. . N.
a10, OUolj UVarl al
w 41 CA u -m co 5. m 'i a4.0 4, s. 4- co
 M a, M 3v d,-.Cv 4. 0 3
'a Q
3 w am Ul
at ol w
> 0 I cu- 0 0
0 . C W cn Im
.C 2: Li 11 u o 0, 0 F
U,
I 1 0 a, Mm Is A: r- 11 L-1 . w . C,is
a, a V) E2 z CL 0 z MI 0, -Z 7- T w lb a, - , I
z z F 0 N r M > m d'
a < w 0
(D (K u CL dl- 0 0 1,10 1
-1 gr 0 ID 3 0 m a E
tf. 0 1 = a, c I . 'A 0
(4 V. m r I (D 0 =) 0i ul -T L c aj ri
-0 al a, 0 cu(V s A 0 -1 m o 0 do dA v r
C" S a 0 1 U 0 E 'n 0 4J 0
rg 10 1,
E -M C3"
Z N M z u :1 111ra 1R
41 C I GOJ W d) 41
Q,
rm tm In u W uj .6 > - I 'd a -0 aj -i U) ac
LL m 40 v 0)
0 M E a u x . - a r
ri m a a (4 a m W E 0 a m a- aj
03 ME X: -j cm E 5 r 13
C u 4j 3 do r m 40 U tm T D 9 C A ot 'n 10 m
. I C M M 2 . Q) U.0 I C
0 *1 Ul r I a, - - >' ej m rg C Q
z UJ
z wC 0
- . 4.
ul O.U aj 0) 3 u c
0 0 W E -P 0 a, W a' -0 -C > 00 u 0
0 X: m 41 D cl 0 c m u
Q) Lj 0- - tu 41 N IV 41 fe m CL. c - E? r 41 u 0 0..
w 4, 0- M 11 C I - - .1 = lu 11 c LL Id M II
a = - - w > > - 11 11 2 11 a, 0 -4 4) - * E -
41 1 " Li In W I 'm 4. u .0. 'A 4 CA L'i
W (n 1li C
03 0
I> 0 a 41 1- = 0 ? 11 0 u 0
M- 0 uz- 21 -0 0) % Ix . T" 2 O'
'0 o I w cr, .9 w a O>W3 0 1 ri u 0 0 1 ..P
0 3 u >
'd 0 m m I w -i
mI u .1 M 0 a Z 3 - %.0) u a ai 0 C a, a
U M 4- r Q - U; tm41 1 c c Z umi (n > 0. M3 0 11u M
a r a 0 '1 M 0, C 0, 3 1 :11 u ri 0 m I LO- 3 mv m I ol *j 0 Id
L. 14 Ln IV :1 Ln - u IL =m0 -r
m
0- oll 0 0 1 ".2 L' Z
a0
2 . . W L'i X 41
'n LL
L" -a 0 E
1 0 o 3 'd
0 a) m 9 1 (n L., (o 3 CIJ vi I m 0 u 0) a a$ =z E L D
IUa, E 10 U al x w L, 1 0 3 f 0
<C LU u Q- o o rl
a, w 0 0 M
1 t-u u I I v I Um I I Im, I lu, LL
LU (> WU >'v r 0 a 1 00 -0 C
a) 3
E in 0 M do
r caj.14
-1 0 0u U u 0 'C w 0 r z ; T w ol r 'n
0 -0 0 a 0 w Li w a . - 0 ' <1 0 m.4 a u o m 0 41 0 6 3 .3 0
_ 0 w -J Z O.m
41 D V c C I M M C
C z 0 1 E IOE z n 01 E E 3 W 0 'M
ci r z 1 0 -33 (V 'P D W _ I.- w w 1 0 . tn 11 1
.1 C' r) 'o 3.0 C. u m 0 cn fl E 0 " I > a, E Ir In
k) I C" -j " 'l w 41 m c .. m u cq c .. a, 3. m- Q, d,
p c a z 11 -
d' ri .. ai w m m . co ri 6 a, z 0 z CL 0 0 1 c v 0 M0 0 M 0", 0 0 1 a
16.
41 1 Go Ll 0 m '1 4, c
011 a 41 0 qr m z 0 4.0 a, C 0 u
.0, M- 0 03 m IJ W w > w
w w -j ol 11 0 3 u w z z w up (d m 41
1
M- 4, 07-
E z C" m 0 o' 14 t- a, 0 4' cu - -c E 0 (3 mr- .0 %. 0) W
a, tn (r - u- a 0 o w ci 0 1. 1 1 a m 43 0 0
In 0 w :3 a, m ra w Cl) a 1 0 4z u 0 c 0) C cu u z ui 41
no Lli CL 0 0 0=3a)- u all
in :> a,,.0 1
a) z C u I
w I <1 0 0 d -P a z CL *A
0, LL, C)
u 11 0 41 u IL
0 0 0 0 M
a, 0 u m 1 .0 :n w 4, r 0 0 a
d. u U, 0, 4..C 11 GoU) 0 z z ra a, fa
If ej 4'
> o 14. 1 0 1 U) z fd
u
0 z x m 0 d W
0 I c a IL 0) 41 u 13 W) al 0 -0, 0 E aj
11 co0
1 0 a d' 41 u 6 0 10 1 M do a v a .. 0 0 - 1 0 w
.0 M -C m M 11 1 a, 0, &- p > w m 0 al u tu I"
J V .1 41 O W dl rl .0 a, .1 0
d'
o mo V) cq u tm- 0 3' 0 ka, 'u
0 3 M
r C" C CD1 9:
ff > u > -T A E 14- 0
W <C
X 11 ml C
.- eg'a U, it
'a cn
u u 0, E 0 L. :
0 A
11 11
w
Z Ow
03
lczn11 u
W99 41
> w w M
u 'A. 0 fflv
mm

U W 0 .0 d' lb (b C, N al to W -P 4- C oc .4. 0 4 u 0) 0
rl r- Is 1gd 8 ; - cm 4,
10L.41a w
a =z D 0 z di
41 1 a q cn (f) 7. 4- 5. - ,, '. 1 .1 .00 1 A" 0 m 'm
-P "M 0 0 w 0 im 0 ow c O.C 3 r il
p L 0 u
(0 LU a 0 01- 0 1M
01 .. 1 11 CA I W
il W
U I. W U cn a -0 u x D -0 czm cn wM w w I aI a, a (m
a U n (n Ift % ai m a 4- 0 1 a m a
ma :1 11 0 11 = _.. 'A_ 0.. :3 0
il (n 4. z L
S,
0 - :1 u I fa a 11 0 M.0
Fs ul CL 0
41 41 a, w 0
w a a 0, lop 0 .1 30 OM .1 3 = 1- 1- 01. w I T 3
L: m w u M q m
CIO m 0 z C) 0, P.-
- w
In Ic CL 0 0 V.P v (n a
cn > 0 0 ol U
a, z En 0 0 fd ci a0 S 6
I
41 -C al C' 0- CL <Ec M
a, N u 0 u u
0 -P = a,-
3 Do u -. C 'm 2.
ca 0
U 'o 4p
0 'a -
1 . 11 CU
a, a It = %.. 41 .. to m a a u In E -P
I C LM M di 3 N - C w -7 N 0 M 4 0 M0 X C 1 41 ul ! fd a)
3 0 0 u > v u f. w (n 2 > a E u 0 w (3 0 a
-W 0 13 z r u 41
U:) w 11) . r. u o Go 'r w 0 a r m M
r 'a 11 W. a Im A u p - w w 0 0) r a,
0 4m-WIZO-m D I w 'c 41 u fada, -1 c 0 0 0 C
u - IV v 0 = E 0 w c ca v
0 0. u W u 0
E 0 E I M c a) & tm 3 (d v .-.P M.-T a zr 0 4. W 11)
4. M a Cn oj 4, 0 z 4 a - A al . 0 3 w < qj 3 0 41 00
Z 'm . ru m 0 a Lo I .. u a 4. 0, - 0 a 0 " m
-j 0 1 a -T . m m (u 41 41 al N m ri m I c 0) rd
c > N as E 'c .0 a, F N el z u Z.C 0
m c -0
CU1 0 In
M 0
iz a) 02 a
n M.C u 1 0 fq u -1 0, MX 1 0 E CL a (4
m La ZZm o 0 W N u
m I- ki Ir . I., a, .0 U) C a,
a, 10 c al a. U) U
0 > WIM >1 4.0 I'll
W 0 .1 0 :3 0 a 0
0 1 41 . m u 0
0- 4 4, '1 0 - 0 co - 'I 'v MWI , x to u (d -6
L-j K 0 0 kA ED in (I P 3 u L' I Q x 0 a S.
lp W En w 0 0 u m 41
Im -1 0 a. M u OL G,
0 a 0 'T .4 0 - = u 4. m L u mo . a
u I" I 'p U)
1 I .. . 11. n N= 0 'o 1ft 0 0 3 1 0 1'u 0 0 -P
m z L) z >x 3 .c Q) 0 a,
c
ai c I to 6 m 0 z 0 ;; 13 m I -T 1
co 41 a z 0 al a) 0) It c cm
0 IT
11 -0) W.0 0 N tmu Z. m u < Q 0cl ." L) 7 N a &0 c
= 0 1 r
%ft
ja .0
cn 00 , 0 1 z L.
z CUdo K M 6 a -0 Im1 10 0 m -X = k) -10) 0 me 41 I
U w l 0 U7.x . (d m
41 v --l 4, W
0 0 a
0 11 0 0 rL w t 0 .0 1 1* I I u 0 w 1,
N-0 0 0 .6 0 E I 'A
re 'o I C3
0 cu c m 4 C,
.A a 0 mlvz u ocnx N 0 m
0, 1- 11
r 17
w Z ': T
0-
w m Z 41 0 M W w (n 1 M :1 m 41
'n
MI -T 'I :> :3 Li x 01 m u D ej
E 0 a) >- 3 'Z -i w CU a a 0
u L
'4 14 4, C N(V I re do I z u I '0 ow I J- u M <c w 0 W al

-ml
C c r. a'I w
- I 'aal
0)
I lb
4.
p
iA I:3 p 'a E IT
w cr, u '10
z 4 6 0, 4' ft m w (n 03 0 'w In v
u 0
u a In - ' ; 0 10 v CLI w w 0
0 at 0) C u CL 0 0 CL > CL U w 0 N >
a$ 0 r 0 m
a U) rn 1 0 cu fd a 0 m 0 >.I 'I u N -C 0 0 O.C m 0
0 lb 0 to = C >. -UV'4 3 D I Im- .0 1 14 (n I Lz 1 03 4.
a I 1 : 3 co
0 I
(1) ' Z ov NCO I a
w (n fa -7 CL 0) m
r. IC .
Z to 0
14 oC :1
i 'o
W 0,
U 'c 00 IW =
Z a. .0 z 0
WV 0 u >- I
0 1 -W
C 4. Iz " -W -j 9 0 41 . M 11 C, 'I . I '. .6 - 14 . 41 a
c L tn.. D at 0) - -j . .. .. 0, z u a; 0 5 ex a 41 = - 4 S. 0 a r aj
c 0 a &I " M c 0. . M a 'z In EO a m u a c ul 0
a I -'. m do I CD d' 6 3 5. Li 0.- '0-01 D I CL 0
0 (a ai 0 1..
vj a, 3 I0 0 41 0
L (n
Lu to cn 1
Cl) 0 M 1:) ,z 0,
Q 0 rd I -j .7 "x u .. -
0 M u fa -1 Cr< " Wm 0 u
u z 4.
11
u 10
I U ..U I L 0 o 'r 4, z . 3 0 c o (3 .. w
C' v-1 - 6 dl 11 > U 0 C a, .0 :1 op 1 0 u 0 x c
0 ol 0E r m t_; C" z ZI 11co v am
Q fd- MI Q Z mi& Q tr. <c a (> m T c '>' Im c
Cl) Ln Q 0 T ;
S. i= .
I1; U Ic 4 En =
> t > -
-j a to v (w a, d' V al CL In LL G: 11 z C a m > I m 91 a (D u-WV -1 co a x U- 3 u
 z
,L
r- 26 .1 OW
.11 1, '0 1 0.1 w v 0 3 -r 4-
4A 3 o No z 6
41 0 0 r V. ce 0 C b, A (n
_ (3
0 It 0
3*
_Wj 3 -W 0 m at
5.o 0 '0. 1*, 1. .50 -.0 .11 ka
j. P 0. fu 0
CL D 0 0 ;;.C
0 9 :c V)
v J2 0 1
0. M.. o 01.4 pu
0 C to I a
Is o CIO W m 'ic A V 0 C _j c.N0, '01 1-1 Ira Ckbl, I. CC 0
C
u 6 o i =zw Z -01
r I m Ov 000
IN IM4- - M 0 %. 0 U. 0 u >. m q A.0 Aft
u It z m C 4. WCL
0 u o 41 0 0 4-0 1. do
i W 3 0 t- S. (a 0 u W
40 4 z0r 0 WU0 W xm 4, 0 - 1& 9: d, Z
6 -o A E " > c 11 a6 1,. ':
3 I to u 4 a. .. A .0
41 a 0 -0 3 10 W
0 a, a fe 0 .c al c w I0 Ial u
C u r a
Irm- 1 3 ul a, I r X-C a.. t- 0 LI a 100go
o a) 3 _ C m 0 c
CW S a .1 mc 0 W L" 0 to 0 Ct tj a Le 1 4. dor a, V W
C, V a3 0 .0 too>
C
4, - v .4
4. G. u ri
cn
X 0..In 0CLrv 0)zm u
a
%. lu.. .. m 0 0 tn 0 w a a C,134 a . L4 W
do > W Mw CL DQ to 0 c
L,
W o
z CL 0 =I u jV It I W 0 tn
(n W
Ot
u 4o in>. m Go'a m m q. 3
z
z IV .. a cu
C, 0, " u u 'dW :, c , X ISM OD 0 r >
U) Z k" '1 Is
41 do
cm Tj
r.c -0 5 0- 0.. o -0 C C
j _j 0: W 0 A
0 '1 a E W 0 3 (D
u .0 o 41 0 C, > CU 41-. 2 waz r *1 4; 11
L4U) : 0 Q - 11
,.r .. 0.. a in to
o ; W 'V ). 0 C36 6 P 1 44 IC
a 4 S. to N.C IR L cr r. 0 9 W 0 al 4 40 41
0 0 > - 3
.P S u '0 Op 0 0 z C u a z
wu. W 3,0
rd .0 t
r A go L. 5 4, V 10 to a 0 .0 do 0
r K 0 a
a Ol
C w w 0 N r 4, a
10 a 0 u 1; 43 1
a a m 0 u
'A v u
Cz j Co . . m in A
0,
z IV 1.
W
41 a
tn
.4 a a C > 0 0 (v co
A; 4,4. .0 0 x N a al a 0 0 0 0o 0 oa 0 d, V
.P 41 3 cli , W CKW
CL 4. 41
0 0 0 0 q- - 0 L, z zo W 0 W 0 0) N U 0
0.. 0, .0 '1 c, ..0 r.
, 6 U) CL 20 6 40 CD0 _j
0 0-
C (a cl
cv CLa 10o 1 oIV 41 6 .0 -J Aj CL lb - C' U a' 0 a 41
1 0" 0 >, 3 0. W W0, CL WK %. C D
. a in Ic s.
r m41 a on 0 CL 0 Ci.
4.1 C
6, w u to doGoW N S. Go coo I at . in
CL
0 * gK 11
1.
.1 RIO
a go a k" u 01 r .11 X IL
CCJ 0 W A -. 1 q o cIu r
9; V6 0 > 0 CL , U) 03 o 4,
Ol v now
100 1o . 3,.
at 6 r. o 0 o u 0 v to W (A - CL S. .0 so.-, la (n r It 43. S. S. W 0
W . o . x X a,, . k. . 0- L. A 10 0 V
LL 0 . 4- 43
C lu
u u Z= CL ko 0- v N m q 0 >
gL
> AP C C c t a $. 3 1 .9 > u u C L. 1.. 0 S.'. a 0 kA to u
0 OA.0 .0 1 t% 0 u CC I.C 0.0
0, 0 1 u 'A ; ci ffi
rd o U, 0 0,.r r atf
C Wo 4" z - 0 7 m N u 41 41 10 04,
C
0 A q. u 4 z z ko 40 >
to (L ,
0 4 4 0 0 0 Z i- tn - it
04- 0 a 41 13 0 d,
GA IE It (n tp 4 z 41 m Q.
14
0 o Is -
z r 1. V Oc c..q. 0 o a i m Mo W 0
" C& a W in W -0 0 cn (n m W a - 31
C. - '. >! gr 0.
. 611 a 11 0 C 11 0 & 16. $. 0 W V 41 L W _j
cm C; O-V
0- Cguqm C. r p - 'a'. u IN Im , - 6 ..u
4 ei L. : 0 0 4 Lj
1 '0 a
ev Ol > a, 2.
O.P 0 a 0 Q
U) 0 w
- C a< Q.
0 W > C 06 9L 0 a to 03
m
go 4 C 0 U. ucom 0 u0 X
C CC
0 . -I .. t 11 IM
%. X
U C, m 14
CL CD a W M
03 cv 5. 0.- 11 6.. 6 .00 P .4 W3 00

dt ts '0 4 0 CU W w a, L, 06 P,
oz 0 Z u W
$. C I X ff a W r 0 0
.2w W Go- 41.,
a I., z 3 fa
0 0 0 0 -
0 z
07 cn o (r.
N - -
0, .0 'a oto va 42 41
0 N ed =
O'W
W q U)
o 'b- P 41 0 0 C 2 m
LL m A v 3 0 W60M
at W do Mc IT x Ixw
Um X a ej
4
CC
L6 u C
77 -cu 0
v
rl .
0
1- .4.1
., 0 91u 0 a 3 0 r 41 C a
CC , z z M U 41 143 XW
r _n 1 9 0 14 u C u x _j
00 -C
4. i:.p r- z me, 0 Z (IIILL 01
u 1 2)
rn
V
_j I ri u Z
U, W.. a. p d, q. Al a CL LL W
a) W
cv Oj 0.0 b-
C oin t;
th o 0 vZ 0 W tn- 4.1 to al m u
Mo ft IE a, C
CL .0 C, r_ u cj CL
cr _j
(n 0 0 o 09 us 0 U, ZO
N
10 W 0-1 u to .0 a u m r 0
r1l (h fV u W r Of M
0 40 C
m 4. V W
N X -J 0 Wz 0 0) di 0 a 0 0 -1 (n
_im 0;.p > v 0 0 a u
141oil r o 0 Q
a,a m 0 U. (n a L U rl r. 40
v 0 0 v 0 a a. S. 0 0. ta 0CL W *:. (n C lbro 'a Wu .-. ! w
X 41 dl
CC It 0 44 >
La kt, 1. 0) z W410, 1 3 3 1-
:03 -IT
(n X z :) o 4 C, I-- o W M 14 Z -0 .8v m .1 0 ,; x
AA _j C w co in
4- A 4- a 4.1 .4 1 W
U, a U) >, m "C T 0 0 n 0 M
o W q. 0 g, a, Zx tn - .0 . :1 o'
> X o o 4. A " S.
4z= o La s..
Z u a > a, u W W
cz Z W z 41
rd _j
r- 0 a C . 0
C . r. 0 40 rol "I 'r. W OL)*;. . . ; . tn, 0 tc,
4- 9 - Q WWX 1 3 K u m 3 x
>wr 9) 0 o n > > 3 M dj 6 u M C 1. L'i In
S.. . r, 0 E 1, Ic le tn cr Z L'i L. M ch at .1 '0 1.104.1
W 0 Oll _j _j
0 1 m to S 0 ce _j CL 0 U. c C N a 1. 0 0 2 v m 0
C 0 0 .0 0. S.
10 L 01 2 10
S.. al .4 L. x z _ 7. in 0 . T . *4
1) m .0
41
CLM 'G 0) aj %"
P C, u 'a 0, al 4. uzww _j 3. - 1 3 N.
w - m 2ca re, CC
cm C) 0 a W 1 0 M a C6 C w u cL aj Mj o _JMl c
In a 0 CL" to "0'
p 3
0 n
or
0 41 m W_j tn
C0 o 0 a Al u ag
a o u 0 Go = 0
cm4 a C W - (A CL.0 U
4 - to 4,go
>.Z.0
: ..0.
= 0
'1 6
0 U
X
I
U LO
W 0 C
to v
0
a
j 'a
La (n w , 0 A 6 m CP4. a v ", w A, 0 In"W
_j v 6
CL :F 11 Q = W. a, 0 C 3 In CoxLIj_
A.,3 o Ir ON
c w0
_j X 00 O_ 0 0 6 4) to - 41 > W re > am 4 0)A
610 CL
61 4, c -
:1 w r
vt
in 0. W
41 0 u CC do di m I o ON 0 _0
:3 .0 _j Z
W 0 0 tn IMO 0
06 W o -0 C o 0) a 0 41 o 3 3 :$. o0
L CKz N
WL'i 0 > a 10 'L u 0
CL 0 4.1
I n 0 kj _j X
10.1.1 43, 4>1.
-a, v 1 0 P I Cqol:z Imn
z " L"
cr u u 0 C r a S. C4 CL W a,
c w a z 0. :) 41 ce CL 0
a 0 c CL5- a- 3 10 'D
0 U) m
U)tn X tn 4 L. 41 0 CLLao z o >40
W - 9 .0 Q > tu
1, CL_j 0 wom W u a .0 , 0,
10
:3 Ul En 3 3 L a a, 0 C a CX E
0
m CL 4. a, v m a U)
u
_j
C6 0 41 2 CM> 31
6 ca
4 LL u el
CL . I
P..
0
(A d- -. L -
w LaV
M 0)
0
.0 V
Z
W 0 fo.m _
a 4" "d '0
.
rn W
Zr,
* V)
< 06
6 -0 W u o0 '0'
u 0 0 a, , <
10 j 4 a to 49. 0 U. 0, q. r tq ali. 17 Z -0 COL o 1.0cl
:3 Z 1 :6 '" C v o -zj co ax: 0, CC u W 11 .0-0 10
o 0:< 4A CL
(, ".C
31 04,
" 0
2 =(11. L" .r. .0
0 ,W 0 O'C
11 "1
0. w -1
L, * M -3, -m L, . 1. 0 U C; g g 1. 0m cn
in
: _j 0 in a a 0. 0 0 N- 0. cz 00 V a,.0 9 tn L m IW3w
0 w c, wo, j'O W tn N
al"U'lo CL W
z
fm T 41 3
;c 5 1 0 00 Nr 0 . 00 v 0 0
>. ON
IRq. Z &.a
o v Ul O.P
M 4. a z
4 z c CCCC-5.0ca 0
W Cm
C a 0 a u 5
m 04. to on m
V, (1) - al r cv 0
0 C C3 & D x 1.
Ll Z 4 0 4 0. j (1) . ch. 6 In .41 *, "A U, z W c"r r-DV-) z u U. L) C co 0 u 0
> W z m . Go " 1. do 0 W 0 - a, a, wzwzwP- W (1) a, u r
QwW I.. o > o CO U W l 41 W 11 'C W 0
0 0 &.
B4

M4; 0
3, W ,;;acwlro
0 4, : 00 1 u
M., 'n z M m5 " > . . (r)V)- CL
CL
0 q. " '.0
Goc m a, zU. =
Ij
mov 0 0 >. 10 ko z 0 w , 'o 9 tm 0
0 :) '. 0 3 0 r.
4, 041 .1 0
0 0 c "o a 0 m = 0 0,
z -P v 71 o a, "o z a, lp z
u ul
71 A a CL 0 Q u 6
GoZ %P W u fs m m a a c CL m 4. j 0 c 1,
CLX . -1 a E 4. a) u 0 z W r ft m 0 C', 4, 0 w u
0 0 x 0 N z rd -P M
C.
F i 0 En 41 4 10 m 0 1& It r
0 04.
00 4, 4, M 0
Z cn co
CLw c
C,
m =3
0 w 0 Q.
0 w -P
0- 0 4, C 0 c >.
a,
m .. c 0 ap 3 0 w I u
0 OIL 1. C z 01
M.0 Lj s0 w L to a 0
U) to m . do w
cm -j
El ul
m Z -j 0 aa :1 1 1 1 In 'i 0
L c r a, 0 0 to u
S. 'o a 0 4m
IX I a el 0 m 0 1.,
Q 0 ft w .1 r, u 3 'r . 4 a -i U . Z ,
w 4, 'r E In
W0 r
x ci w ra ol
z EDm X
.10 0)
1-
0 a
.c
u , -j N m. W
0
- a-, 3,0
6-0 a, to 41 4 0 3 >. -0
31 'c
41 4,
.5 3 u
- -
0 3 0 Ir 0: W
0 0-- cr 0 40 Rm0 >. 4- O-P 19 LLIC
0 41 0, , , m '. C, vIn j u
r I- m r= >
(ao
0
0 u u
10
0 3 c t;
0 U. 4
a 0 w v 0 >.4. .0.00 do
m Q- 0 w
m a a fn 0 -
3 d) 0 a u .1 0 r 0) - 4,
0
0 . m a T :, M
, =aj u 5 z c
ai :)
4 U
a *I tr)
41:3 'A 4.1 a, V u n. v 'I ri
Qj U., 0 to aj
c u a.u rd al C -C 'OL c 'r ru , Wm0
0 Ic w E .
cx 0 0 'In In v M .. c x Ld W)
a) L- I U Cj 0 x
13 In
0
3 Q
a 0 w
V. 3.. w
=
W - * tv m a, C, cli
F ko > 0) 4' a r; M I.E" >. LA
1.- 41 0 C 91 & M 0 3 -Z , 4 X U
0 V L .- L . z ul
N D
0- Ln a w >. - q E tn
'm -P , > M v r a u
4. 0 E w L. m 0,
0 'o u M- j Z
u q u m v Q 0 oE
0 1, M 0 2
0 w 4, Im 41,10 m r
0 CL ol z Go C, a, W w I LU W W E I I
WW- - .) 0 m 3 -j
4 r In W
411, 0. , to p cq, S

ED (n z cv- I ,m cli
re 0 ol 0
0 Ln
2 n, 0ul
vW 4 0 -P 0 -a L, 14. z u
(n C 0 F
0 14 E
2M L, !: L
E CV
Ll C, ci
a.
tr
0, n m'o 41
u "1 0
r
0Q zW cn
-j 12
5 c E w L, x 0;
Z 5 U r 0'. 0 fc.. Ln 0 L'i a.
o > Q- >
u o > L 0 41 cr 0 M CCU Ir 1 -0 IV 1,00m, 00 -T
0 > x -P CL n 3 a I r > ac C2 (n 3 P'l S. a, c r to
'a fa 0 71 v to a <
03 U. U) 14 0 c I u vuz z .4 cu .!L 7. >
= 3
pi a: Q z 0 m 0 41W W 1 0 c
u 114, 0 ,--a , -
0 In C) 0 at
cm M
4. Q 0 cIJ Zoo 1- 11 0 .0 11 1 C; 0 taZ)OW Ul 0
a Ix ow 0 m 11 1, In 0 LAJ 4 u x
tn V . Ln 0. 1 (a 0. 'm > u .1 .0 & 4 4014. m
UlUl 0 E E In z -PIn
crluj U) L'i
L 6, a C6 rg a, T 0.
vwmw 0. a u m u z z
m > cv 0
> a c vi 0) rMc a, z V-
It CL
Ix cu 11
co aa 1,- m ff 0 m Uw a)

W 41 a k, 0 0 4. 41 $. N 1 kAV
u u M -P 41 c M 41 w 0 0 ru
z , , .. . 0 W
0 C. M.... 1 LLL)l
wn a, -ul :L:I,
oz 0
LLC ru 6 3 w 0 4 c a L z 1 > V 11 u (n t)
w a 1110 -Z 0 4 OD M 3 a) 0
10 m M u ED!;
0 -. 0 v 0 m W T
cc L 0 v 0.
I"
li u V, 'W' 0
Q -W ,.,w m 0 o un
r Lu
IU a w do-m .0 0 a 0"d z W (n
(n c a 14 N Z 4.1 .0
3 as 6-0 0 0 1. 0
to M 0 0 = 0 zzz
a, >
z W WX v to mu 0.
0 a Im 4o I M ci ! a w CL
go 1. , ti
do 'I ..0 c
, a. m a u m-o -i LL7)
u c- CL x U. W (L
>
u
In a m rd IftI z (n
ri = m a, 4, ms 0 X En W a 0,0 a
I (n t) CL -P a=,Z W'XL-n CL > 0.0 u 1 0
a rn x '00CC3 0 . n V,
w ru u - a, m w w 41 M I., I z 0 a. ru
ua a,w -4 Vc I10 :1 w w W 4- 14 LI;:n
L) U, cu
0 0 06 C
m w
I 0
ri z n
L3 01 3
.0 > E tn r
<or
Fj Z L'i Z . ; X m rz U) W t,6 4. Qj 0 .3 rl L,
CD
0 4; 8' 0 10 U r u
cc 0
0 0 L. 0 '1 z 4)
..
u 0u n fa 3 0) go -0 cu 5 CDX
X
C W ULc L
z <r at L. a, cu 00) m cu -1 0 6 0, m rl L, w (n fd
w L ig o M F4
x w oz W
U)
0 ta 4-1 C3 9: ru u
ma c, C o) to Id o T
0 c
kiGo C; a -1 1 0 0 -MV L '04 a rawCKW r. 0 00 4 v ozwm a CLZ
Z.C tu 2 0 3 2 w 0) . fa
umo z 91 0 mr IT
u
a..
PI a ri 5a c
co
0
ru -.
C 4A
c -A u cm a C11.0 M 0 U) ru
: i =
r or W A 4. 0.0.. w . .. = " - . 'A 0 0,6 ff c zwo
at at S.. Ul 0 0.. ma _j >cu
- - m a 0 In 11 a
w >1 In D m x N "
'n 0 w 3 T
0 in > I V fd 0 a CL 0 CK- C 0 2u x0 m a
c%
.5 Go.Va ao -j v 1 r3 1, 1-1
0) La
cc n 41 a m 0 0m 0u -T W (L p 3 ri 0 p
U rl 0 0 m
V7 ri
0 u. Iz go 11 o 'n a C. u
W (n U. 'm 0 M 0 u .0c w tp 00
w
Q 'I eL cn 0 oa
j p Ll 0 m
a w rd 01 T C.4 10V41 p 0 nz 7 v ru 0)
ro 0 0 0 w 0. 4
u QD
I al
r3 W C,W a 0 001- 0- W 0 a C 41C I > t: 4.
0 Lo-j - Z -T c, 0 a m a Z n .,o
_j Z
- :) u ': -a Xo
(n 4, 0 o u0 -0
.11 al , 0 W of "V' . .. c z
L 4
3
0 W
W
S. C 9L IC > a 0 v
kn
ka m z 10 0
Ic L.)Z W2. z -X
(L 0 cx'n 'c m u 4 m 0 Q P- W Mtn m3
u
0 0
a Z .5 x
.0 0 L * r.-- w"to Ir 0
-w "
W!2 w
x " >
'T tr
U) a, 0 Z
w IM
,c z- r ix
v L;
(AL.
Iz CL
90
:1 a ! >C!; '" '1 0, m at c
o .4 z 0 "0 .0
-rc ua offol-0 :: w U,
W -0 z 4 u orIV410 v11 u CV
0 orct, vl;o to
u, , 0 N It Ld 0
C 1 0120 A.0 faTL
> b, 0 w
"I o um 40 0. 0 8 L. 41 so- W 6 m -
v p W
!;'02 no 0 A z 4 W en cn
% 0 0- 0 n 1 4 a -j ! - C , to
Ll co mru I L L = Z, Ioz
Q %4-1 0 40 ; X m
-0 00 w ox
Wo-
cc Ln
Z 'r
cl)
0 w cv c CL W v 0 t , 2, b, 8, , ow ic c cn 6
L > r r Z(/).Zm
A > 6 CL >..; C a fa 0 3 :3 C', CU 0 co 0 04. 0 0
C6 cvx (1) u 0 a 0 Is .1 a z 10 k72 co q. CL CL4 'o u Go v co go u tj CD CL
 w w g 03 0 4. c
al 04 a) 0 0 0
lw 10 w 0 0 z 'I z
uox G 0 00 Ll'ZI -jID 0 E 2 2 LJ
V u -3 V La 40 a, 0 0 u
coo u W r-' 41 E 0 W 0 -
r LLI=
-j En
3 u 41 to aj z
14 0: Z a: 0 LD u cmL. J$ 417. 0 ; 0 M cuc
M >
rlI E 30
.0 T
ouo E U ro M 04,
-W- 01 M.C 3 W
>
41
0 C
Q j U .9 M ED cc w ..0 . '0 . a, z M Ld 0
u -j a: 4 It. a, M u M ai 0 to .. r -P r LL
0 0 aj c a, I W d co to 'a L4
4.
v
M 0. r 2 3
C 4. 0 w a.
>, U, g 0 c 11 E
41 0 0 0 ID 2.0. a UZ
1. u I tp w LLjC
3 L4 I M a 6 4. W
-P IV r L w 4 0. a NXW
to c) C. 0 . u 0 a 0 0 - : 0 W 4J OC
U- -j (o'n 4 to
w - wi
; 0
- u a, 3
412 0, . ,
L W E 41 41
0 u
. In 0 Z
>,41 t ai
z Lij - -j n w U. >
lp - 0 0 to " P- M Lf)
U 0 _j U,
0 4. 4, - tv a, M W
Ul I- -P E r M 0 31 U VI 41 Z Ld
z 14 0) r 0)
. M 3 a X'Z
r= 0 0
.. . 08 kft
3 v
= M 'a Q.-
u
> 0
4. an a,
u
W E P . lp 0 0
u w z U
Q. a CL
M 3 ). z 0
L'i x z
w
41 <
=0 '0 0 W 0 0 3
re -P 0
4, ZXX L. - E 0 4. 1
mW A 'a a Go 'a Mc C6 60 0 0 S
I woc 'd, o W La
r 3 0 -4, to (d 0 w 0 - c "4, M
U, 7--,, c- 0 M a a.. lu 4
0 'u 4-
U.0 rl N 11 . . W CL
cc,
c X 4 4 z a, M 41 c - ML4. a '0 6 a 4. u 1941
0 : M 0 w a:
Ix CK
a 9 il.
0, 0 S a le to
> .9 uj -i
Go 0 9 M cu v 0 41
LA
in cl) 4,Ni Q a V) gx
240- 0 al 0 4 u a
w ! 4 ", - 0 N al M 0
-j
- w L) do '0
0 40 0 10 Of Z(r
OZ I M 9 = Z '0I vM 0 =
43 0
-J Z In -0 X 'a 3-
04. 41 0 t 34-U
0 C, Cm M0
a
doI
lp
I C.0M ul >- cn
tn Cl Cd 4I W M 04- 61M 4 C, 3 0
0 0 -P u 43 fa
a 6 al Q =I C ru 4.
M
-u ICII) up w .2
0 0 0 -W
U7 Olw 4, 10 rd
A ap u 41 U IE 0 x -P u (n W w P i. Aj (a
M 6 lp - M a .0 0 w . '.4 u
> 0 v M Mu 0 wonz
0 1 4S 1 40 M M w (D w
a 0 M0
EX" 41 10 -P go 6
U) u V EDN M cu 19
r) 41 0 3 z
10, M M v ff W. 0 q. r z w N a ..N "a 40 C3 o 3 Lz, W
do N C%a CKW C I -.. Ln M 0
a u
-x 0 w cn w
1, N .. . ki a 0 -,- %
a, 9 r CL0 0 W u u
V
a
<c
XWW a r*'0 41
U.C w
w v
0. D 0 U) al N 4) 0 6 >'O U.
U 0. 3 r 0 al CL

-P 4 U)4 cn - 40CL C 4 'M o 34 C. V)" .3 r.

to 4, 0 WS. , Z T
I.- w w (1) 0,7 4. v a V. a 0 K w
9 0. 0 M 0 0 vp 10-P IM 3 1-
z w a L. w CL 6 4- 64.0
CL. 0 pi w M 10 W-M
Ll
we
, ; , , 5 U, w
03 > c I- En w
w 0-
> -P tn
U, 4- rj) 4) 0 ol a w z
z N 4 L. 0 u -P W CA Z 14 cu 6 3 3
9 0 u
CL4, W ;, z a a .0 Go 4E 0 's 'Z
z
a - " r 0 a
w 1: (n a ft L -0 6w 0 0 Z
0 v Go 0 " -A C GO 0 Z ,a 0. 0 w
MI C 41 a 0 0 ;
O>CW a u
A a , 41 40 2 -C U w i
W LL -C (1) 3 = do
11 cv z w
0 La 0 0 N 4
Cn 41 LA u r 0
p a, W r .0M 0 a
at L. 41 -1 6) j 9 ki w cu - 0 3 CRO M > .0 V -
CV
C, 40 -1 U 0 0 c ix 0 , ; Ioz u I
OIL Z 0 I.- u u
> 34
0. CL LJ 4 M 0 a a v
. X 0 a 3 u -j u z X.V 0 -0 w (1)
cu 0 Z -0: w
li w 11
wow woo M '00
CY li 41 t MWuo Mwozaa
=0 am, 3 v u
0 (9 w >
IL ac
L.

w 0 r -P di -0
.09.P.C a -0 0 0 ORi -1 z
w 0 '0
& 90 .40 a. 0 1
0 W w $- c
w 0 0
-M :-u c 2
$. W. L) 0 6 . OP,
M 1, U. (a 'a 0
63 IN -P M D 30-=A..20
2. 'M 0 3 0 34, 4 = IM a ; U.P Q.
La W m w A c 0 19
to 0 M -P w u
q 4 10 41
CL a 3 3 a a a 41 r 0 a 0
6 & 41 M a" Z a. I.. . . 0
cv G, In
a u .0
a a 3 u 0 x CDz 3 ... a P 61v 0 0 u a
& W 'L
2, 0GoAn v -T > Ll
u fe - u to -0 w c 040 0 r I., 's 3
cly . 6 a tj ae
u go u - z -P V N tj CA
a 4. a.. V . M CL-W C - rj) <c a 0 Q 0 M 0
-j p 'a -0 040 v - 0 u 40
0 34. t CL i 0 ft -P 0
a z M 0 0 .. 4. Z
q. 11 > ko . 4 .0
0 0 V.0 6 a
4' ' P -.. L4 W 'U.0 ul Z
-W 0 1. V > 14- 0 W 0 -1 13 CL a al 1
a 1 0 u 1 .0 CU
C
%.
X 0 0 Z 414 4 In " 4-1 in L. L%M I do
4
0 j Zw
U. eg
-j
W C 0 Q.
0)0 M 1 IL m 3 0 0 0
- a. 3 a N 46
-j I.. OD D q. -0 4 I. & .5 0 c ix 40
6 'a 31 fUl U I WD11 .0 41 r r
q. z 40
0 3 w 'D 0
C, -j 0 w
0 . CK one
CL
Z41 4. 'a 4
CP. S. to p .0 4 E C CM 41
CD 0 1 .
4 10Zo U. 10 -
a
3
S. w
ascn M.U 03ro ra-, Ij 0 0 -Poa, w 0
co q.
In
0 0 3, a 0
.. 41 u Or., 06 3 V) 0 4. =I 13 -P 0. U u
.7 0 " ! b, do so I a 0 M W LU u (V vi M M4 3
0 '1
q-
M-P -0 $- 4 M >-Z S.
M r.
0 a a M0.0
u -S
0
wx
Or - w a 41 .1 = 0 0) . p .. 0 W 4
0 M OwR ol a W) D Lq > tv CA0 =0 11
- 41 -4. 6 u w 0 ; L. 0. , u 'o GOV
ox<
C 19
4
zoo a N U
a r x 0 3
W
> 5 u -i 0
CL
0 0 0 -0 0 uc w '20 6 V' W 1 0 a a 9 w 6 x
& 0 u E
a 00
0,0 D 0, cwz 0 u >
do . u 1 U)
U,
-P
to a "
cu 0 A. 0,
4. x fe W0
Mo x
. -0 a 0 c Z "
V at C, Gov
w cy >. N cc 4 a
11
kj 41 V.
0 4 0 3u c 0
>.
>.. 11 W U-
0 k"P, .1 c
-A 0 u a a
11 w 0
w 0: 0 slo (4 0
u 5
u a al 41, v w
cxw 3 =I 0 z a, w ru M Am0
C, 0 u
41 10 N 3
0 0 ZU,:3 11 0 L. - .. 0 a. 0 0 U.0 M -W 'o
> A . - 0 C4 4 M - fn I.C ._4'V)al
p 0.. A C."a
w 0
0 ..
. t GA
u D w 4!
6
4j = Wz 6
u
a- 0 U) w r 0 .1 9w
14
uj - - .. -2 u a IE u . M Zx
z -P . >, cv cu J: 41
ww M z cu cu 6-0 11 4. 1 0 -1
a - 3 c S.. cu :5 X ift a a u = - I 1 5 . 'o r
* ! uj w w w IF 0 v
q. 0 C 0 0 r 1
w c , . p - 0
v at -j w 1 0 '0 0 U p - " 11 ft
li z
%..L. n 0 inv z 0
ORou Or- 0 4 '.6. Zww ol
en torM CA
cr 0) , . L, > o 0 wa En t -j .0 a
cn 3
Q. CL
W 3 9 IC
d'
a - 0
0 0,w
30 3 0 LL 4j M 4. a a 0 W- to
- 0 E . ; M C
a CK a A Low x w 2 > M 'C 41 W IFE EM-P d, -C M3 0 -P 4 a CL
a tj c > 4 W) c a a b 4. x A. a.
> 91 A to M a L, z E 4; 0 a do to N > 4 0
0 me 0 v c ol a
C
a 0
'T ". .
0
- j .0 -
C. -4.
0 0 U. -C
w agLa Wa 0 ! 0 c T6 q. 6 0 M.
a 4C -j tn 0 u Ll a Mu
0 0, q, X A 0 L. W c g, v 1 0 C,
4, all tA 'a z 0 14 4 C
c 1. r- x 0 U cu
a z 0 c z
0 -4 16L0 .6. U&I >. 0 1, 4 cc W- I in M a .. a L. a a w dx0 -0 Go 0 0. 11
: 7 M u al P
33 a 17 m
'i 0
I
4, Ov a 0 w x 1 007 -0 0' k.
, 0 u I, '.Do K-
>
a, .A,11 0 0 N
Cj CL , , 0 u "o 0 ON C, " 0
-0, M . . :5 Cu > - 07 Cd a. 1*
- 0 0 0 0 W-2
0 0 -0
a (n M -i C, (n
0 A-
4, Cl) 0---
A 14- '1 3 V . u
01 -P w '..
uj cu A. id
I x IN 0 u u lie) co m 3 - ;E M -j ON I Q. v
a u Mc% i 0 0 x w ul -a W
0 .1
z CA i6v w 0 tv u0 to x z M w w (V 6 0 01-
0 cn z 0 0 W U;: w 2 WE 0 5. co E =L C L. T
c1d ol
z 0 -P 41 a 0 9 010 11
04, CD .0 1 UO
RwIDL 4 CIKI a) z 104,10 M 0
 B-6

0
cy n a, o m w to a zd' 0
za' as W CD 4, 0 % 0
0 W, -i I cm rd 11 v 6 m 3 p 4.
W W 0 C - ft 3 m 10 a 0 , u 0) 0 41 2 W :1
'a u u. cn cn u v7 C 0 D Ir a r. z 'a am rA W M .0 1. it, X m
.0 u 0 c 1 41 .1 4A
'a Lft
v 41 W -i
z z :3 0m u
r LT, n.2 m
F, 0
3 ao W W a, 4 0 t c u r r
7 a a 0 $. 1 > In a W tm 31 d. A u- W -1 - % .=
1.0 0 C M 4- C 0 W W -1 En >1 a 1Av m
Ol.. L 0 En m bi %. > u W V W :1
4 z = I Lj W 41 M GA V In -0 a, 0 U) a -P C LA 0 m 3
- on V) 0 a to c0 3 a 41 -0 W
0- u > W 0) - u
W . >
Ij N u m U)
W a IT 40 -1 r. I l ml,
m wz I I- C,
f- * XZZu 0 % u .1.1. M 11 a-= 41 x 11 .1 u .0a, CO C3 11 t a, . . m ES11
u 4 W a 4 A, 0 19 tm: I& mA 4 0 C 0 kn
r 14. In c r- t4 M 0
U) z 0 0 V, a v m-P W so
CL =3
C .W twn
-j 0u 0 4 0 30 v(U = -P
= 4 .0 0 M
LO 0 W I <1 0 u "'0 0
V W- D'a 0 - 0 c 0 - r C; m Wm > 0 -, "
r D
CL - 0 4. 0 .1.1 1',,
a M A 0
X: L'i W 0 cy M M U " 0 U Uf W CC U CMcx a, u -U.
I u 03 1 1 1 -- m-.. - - c It > 4- L" * 00 41 2 u 0 a Q. 41 u 3.C
u I CA c C) M a- 'n 3 q, Ch u 41 0 (Do u S. O0 , ... 3 14. '1 11
0 0 uWx r) ru 0 0 0 -P 3 = 0 -A r- X a, Cl) 0 M 0 0 4 4. a, 6 0
W U) 14 0 1 u M . tu u 0 C a, 0- L, En n F 11.. 0 m u
L. u t" 0) k C 3 0 Wfi. u CL (n , U
10 0 to 0 (n CL Ic - U o a 0 CLkft cn W k., U. 0 . 0 M 0
- Z > W 4. u 0 . 41 = L. - -i W M S. z L. W
I.- LLJ(n -i = do u z - Im c a kt, m -P do Z ! I u u 0 0 c .
.00 s: 'A to 4! a 0 u 3 V! 0 W . 0.0 (n do 2.0410 ad d,
1 0 3 .. C W C
C, (a Co
(V bi
a m 00 M v1. -r. 3 llnl GIS WWI 0W
m
6 a 4 di
4. k" j 0 S. 41
. 4
Z Ift
a,;
in
4 11 T In W 0 a A 11 r - -1
4. 0 W (n ]C OC cm u cu 41 1 11 3 0 1.- c U a w 0 (n '1 0 0 al M .0 0 M M Ol 'A 4. :3
U7._j W 0 CL -- 44 .. 0 4. > C,
41 CL L, U)
W :3 .0 4
a, >.O t6 u
3 0 C zW 10
u W 0: 1 a C
D, CL 0u - 41 141o- - :)
Im
(L. -P .1 f.a 0, to
W 41 0 4.
1-
z u 1.
z =0 r cu 0 4,
" a)
> N $. 0Z 0 vc
4, 41 u V 0
0. 41 u I Q C) r E . c a)
'n Z a. L. M 0 0. do a M 11 1.
CL 'm 4 W0 41 0 do CL0 0 U U I 1 E CM
U) CK0 L M a 1- 3 11 0 c4. IL C 0 0 3M 0 1- CM tj W
a W 0)
440 a U Ol U U 4
a I4, Wn2 -m
U
W W 0 c 0 0 a . 41 0 . . - - z -1 1 0 a 0 0 -0 , 11 In W E
ta- 0 cv 'a ho a) - E u CA 0 -C S. a 0 fe 'm L.) W alz 0,0 4. = 4 0 14 . .,- .1
(1121 1. m r., W a, L. IV t" cu u
4.1 3 , 0a) 0.
u 11 11 -C
41 41 al 0 z Q.
D go W z cn 0 M 3 a, M 0
CL-7 = CWLn v 14. 0 ,4 M, 3 C
0) 4. =.. '0 3 uCU10 D.- z x0
C) vM 11 a) (L >- r0 r -P 11 4- 0 41
IA L CL
>. 0 rl W 0 -j z (z a) ON- 0 v L 0 gu a) Q 0 &. 0 D 0 u 41 z
L. riW -j W cu cu 000 31 31 1 %m - 7. W
0 IM -P do >0 Owc z Iru 0000 4T
C c 0 m 0 a, 0) 0 cu 43
45
> CLrd 1 W A. I CU a 0
M a m z D Iq c tm m- Irl a <x ul I >
V) I u En ta LL (x L 0 0 cm M E .14 a M k.; i W 0 0 L 0 c 3
0 C% - - - - - - . c 0. . a 0 W -j 0) 1: W 4*1'4 M 4 d,
co I M u u 3 (X 0 a,
bl Z ta N M 0 .1 3 u C, 1. 0 Cu ILW N W) 0 3 41 M 10 3 a L 4
a u u v a
4- W ri) < u a 0 U.. M M 1= a m
0 CA Al m x z > 0 44 -0 'I c 0 a, 0) .0 r
0 1 0 L W.
0, > W
a Vz V) c P -P a -A
0 L, -T C 'm 0 mZ u m M >, M a = , c t4 0 " " '. p .. c 11 Z L, 0.0 W 0) 4j Q. C, u
0 a 0 m U) x 0 c al u M M0 > M C n u I x - = 0 0 a, W
cc 2, 1 0, r
W z u 0 x In X
00, CL a a,
14 0 0-0 al C, z _,a 41 C a) I a M a, CL 0, tp -41 W 0En =4
M 14 01.1 11 r, "I
0 V.0 In
0)
a.- 'n
K 0 W:3
W W a 6
Ci a lm CL , 0.0 0 m U
40 u0 (D .! : do
0. ; m a() -4-
E C - 'L CU -Z 4 It
c 0 mOJ--41 4 11 4, M U
W ( 0 > 41 u ,
41 a :2
c 0 U M 0 . 0 41 a L 1 0, 0 0 -2 m.u W a
ks c 0 N O4- . 4 MIT a kA o En m 4
c", -j z W 0 11 . &U 1 11 li .1 to U!!?5 -> m C. 7.0 0 m
I cli* (D
- Uu
m --x c zMn-, Ia Q, 0> Ift n 1 7 1.1Z le <
"W-
W a . > D
Z 4, a *,-
E fv u M Oj a$ a$ E C, m Go " (L
Z 1, , r
0 c do a :) 0 a 4 . C. c -T rd a 0 1, m
on z W CK 00 u cl u ri u 10 u 14 0 U S. a .3 CO a 2 3 UO C".

4- LA c I -j al X
CD5 0, S. W <E u
e u 41 o En 3 cn Z U
al
41 t, 11 00, 41 >1 m a, zz auzm
9x c W
4,
0, Wa
m
0 L:) 3 11 r 11 x m 3 u IA D .0 L. c 0 Oz 13 LD lz I W ck
a M 0- do r, S r a,
4, 11 11 P- 1 3 41 3 a . cu M VI 11 >
11 U)
141. 1 C V 0_ . :) 0 . 03
.17 OCI., % :1 > <zr I I- 1 0 0 G n 0 0 'a 7.1 Z 0
LL 6 E 0: J do.. -,a u IP E U, 4. c a
a, >. 0 a x I I In m C Vk Qj 0
.0 In a . > AW 4. . Cj 16. C C U. W
a, M a, u vi x X 0 0 C, (D 1 a O 0
c: 41 C3 0 LD 0 D 0 T c ;E z
CL v = W OKz U. M 0 al (T C t 4A41 t- M W Q W (b c
.P M o D m LO Lf ff x 'm a L do 0) 0 In 3 Z 0. z 0
V ej a tm mI M a 00 Go u a u D m u LL U) z -j
u do 3 -as Z di ri 0 > W
W2, W W J)
0- W ru -0 4 0 0 at
'216 u 0-u 3
di 3 13 c E r E
m (n
LAu a U" a) Oh a a En W (n
-j r 0 U)
Im "o I U) e 41 Cl u 0: .0 W
0 c Ix 0 kn CL I c C 0 W u kft al v
CL 0 r a, a, a
a a , 0 cL a 'q Ln ra W Mm "WD m C, 0 z uO
(n x u a, 0 u 0 'A n T CL 22:mo
OV C6 9=.c Q W 0 > u 0 U) 0 t- :) 0 cn , Lu
-T k- 4- 0- C z m LL
0, L, <zcOW u n 9 a T a v 0
CU 1 r 06 cn I V) 0
ul 6 O T 0 M 4 Lj Z x cn
IV u 0 al 0 u X: U) -T cn c
u 0 C,
-7 .. ru u X, M 0, .W 2, .. 0 X: cc
-T z LlC3 414J 00 0 N Em z 10: C, N . cu
cv (n S. 4. V. =CWZ tv w) I U IM 'G a, E d, v W 01 m co v
N 41 -4 M > r u kd W o a a, 0r a
> u C.,U '1 0
C, CIJ En 0 N r 0N 11.0
% u In
ru " '0 0 M 0 al M z m M u 14 a S. -P 0" = N W a, Z
a Z .0 In x
IZ L'i
(1) U7 W Ix
ru aC- = > 41 z a W 0 M 0 0 W .1 Q z z Lij 0 7.. 0.
old t. m a, M 'I CA Crd> W > 414, 0 cm
C-6 CN r Co.. Jj 0 V r 4, Cl) c
- F- u
-0 00 a 4 u M U) V) U, 41 3 .
C
U
0 cu 4' -U m >.
v
lp 4.)
z 0 0
0 c W 0
0 Z I E
'A CL 0 U) Q I (n
Go a4, f,m 0 > I 41 CL
n l (D 0 " I a L 00 C. 91 E 0 0 V 40 a'
u tA 20 W =-X Q
w - N cl (I C
X,
CUcu 0 few 3 a 3
lp co
r 2 Z tn a cu cli A W 0
a. E V) u 6 0) 0 W C%ru 00
cu ac W CL ol A 0 u 0 W Oj 1. 0 - (V 4 4, S. u
U) $. A M a u ") z W cr 0 V ko N 3 L. u 11 C- 3 41 0 0, * W X
W .2 .0 09 'A 11 ID w a I W 4. Cf) >. cly a a$ in L. 0 C4
WW Q u a- V) :0 xW 00) Ln CL a z v a 2
tn it 4 q. a 0 C L. U cn D a ci tL u V,
MZ r In U, 0 1 (1) u ru
x -P Cu
0 W " u 00 '1 tA m cv .1 S. 3 m LO M Ir L, U, :> U- 0 0 X ODN
U, 4 Ir 0 o V6 11 0 02 x W !2 41 0 - z > -P - . c W 'n (n Ln M
0. >. a L z a- Ir W 0 M r 0 a, a W a. E
> a u c CAm d, 0 0 0 > DO,- uW a 0 U C. 4. M %. 014- r, a al W aC cm
(n C .0 0 LIJ C tft E 0 w > Im vl)
4L 0-0 >. C it ao = Wm (n Emf
* * D 0 a W u L.) W
Z 0 C a 41 di 4, 0 0 a N 0 u -i C L.
*l 0 '0 Ul u z'0 (D m
Z U .41 4.1 M I W m I 0
11 z 0. 0
<1 a 0 3 u 41:, Co 0 WU a a41 0 M do
va :1
cr
lu 3u c W c
m c -0 a0 1.
a 11 110z z av r-
a v z - .. u 'm m
0. a 0
11 0 m -P 0 m w c tn n 0 a .2 2 P 0 y - c
.m 0- L 6 0 M 1. Is 0 S. - 0 W 0 u U; vVi00 c C M - 41 a u, t-0 cn cn .0 4 to 0) .3 0
0 U IC 4J
C'd.Tzl 41 r) 01 rE U v u u 'cl 4j (n W 1 01 0 W Cu L, .. u m C, ru
CIA z no 11 > 0 1 ru aw-mo > (D - 0 a; 0 D T 3 al a cn W U fli . < li
- 0 0 1 0 0 m C) a > a rl ; ri .1 a, u Q .0 0 CL a CL 0 40 u
Lnco i. c u) m a. m 5. > E p u < (D Ln On I in b, 0 L 0 4, W E c E > 0 0 z It 0
cu a m >-== I : ZZ- cv CL >. ., CD M mo 0 0 0) 0 wwz W cli
Go 0 ra N CL CL cwr CD x 0 a a IA S. 0 u u v C, I 'm a.
 B-7
ZIN
w w al C:x 0)2v v
LU .0 .0. -po 31 w " a '0 m
u 4- ': a -P Go m 0 r 00- 06,, m a a M rf)
0 a ci. 41-1 Li 0 0 0 Go
z - to V -W uj
u -P at a. 0 .0.0 m 'r Ic In
a
0
Mu w
cc
x C.
& w
0 0
> CO 0 LAM4- 01 Al ca w 0 - '0 W 0 M W
(D ul A x c
c .0 al m CS z U) t a 0 a Ij a' a 'A w 0..(n 41 91 M M P
41 3
K p w z z
o W
it
M 4. >.0 M 134 3 :1 .0
.1 Z 3 at cn E C) .0 u S. c
V m 0 03, tj M 4 0 4 0
w = RZ 0) 0 -P C3 lb w 3
0. W Go 40 C Cb 11 CL 4 4- 3 u EmW, w u a Ow v Z 'i
cc a w In C, 3 .0 u a,. 1 0 r W U CL 0 0 L.
5.- 0 m 21 Cl) 0 a 0 0 U- C 'r 11 p
cq LL z ac > OA ca a CL 0 0 w
r-
u Q. CK -W 4 x W cx
= z 0
0- 0 u I r w Q (D 3
u -P a X0 E
9;d3 w 0z 0- 41 au M w a,
uj CI
0
C; m fd CA w OZ u 0 2 0 tj M
df r L N in c
r G 01- 4, a 0 Qj ". 3
o
41 to 0 =1 W w . . . 4,
0, 1. .4. w .1l C, <wzZ , p E L, V, m 0 1
4 4. 10 a w CL 41
z c
-P to a. m v m x cz w ..
u U,
a 0 q 0 N 7.
0 Ol P
-0 a, z 0 a-
'a r -j w -j U) w Z
'* " CL 15 13 0u c " LO
of 61 a 0 m 9 tn .0 r al
r7, m I- u 0c
00,
W) '40
'1 cm 3 z c
ce 3 Is O.U ONul w
w
Z
'4. E :) 0 1. .4 N (M
z 11 r a- m 'a . T ,
N LEI
xIn 0W a 3 -T 0 M
0 z4, 1 ... 0
C', cc gr
4". 4. w a. Z I
w Z CV u
uj EF< a'r_ .0 a .0 ft 7)
00 E 0 w - 3 3 a 41 8 V a 0 a
0 u am .0 m
rd 40 v0 u -0 Im =
- . C;-Oc !N a,.4 a 0
10 -j I r CD tr w
ru IIn _j W D 0 'A z tj 14. L. A V.
0 , . cu p a CL m
fa al- -poI'
M4
0 to
A I
U. w
0
lp >
11
0 2 w - *,
I "', IV
D, 0 4 4 alL
0 c
S. 6 W 0 5. -X 0 c
CA
0
u z
4-
" N
u n
6, CL
> 64 0
> 0
0
3'T"
0 1 w
z m m Iffi
0 -P
z '0 .0
10
U
Is a I a, MVml,
00"
I Z
CA
Nw m
0 0 5 .2,
,
'A q- 'i u c di - ki
Z m 4. u Ul c
0 0 a, - Oc 1
cli 00 >,.C a p 14 4 .0
m a- w u "a) I -j w to c 3 3
0L. ..%. p cn m '0 E W w Z
M kj - 4J 06
C, 3 41 a n di CLv M
ll tv 0. vi co 0-4. m u U. m w a 0 0
"I m U) U- -P lu
a - z m 6 w Go.0 1 m 0 z u c w - D Lj I L;03 0 11 C)
.7 , m A- m c 0 L w q. a = LU =
0 A 0) do 'A
-0 0 o- x 0 3 =0u
U. W LaJ 0 WCL A.0 0, w doJ)c m .0
T n o, , "' z cq a - - n U.Ix z 41 00 03
W Z 0. C, c 0 c
W 30 ej 0 to 4) APr -0$ U) 3 w 0, CL
uj 4 cu , 11
In If, 11 .
11 41
u 0 W W 0 a Z Go Q
ew w U, m 1 4. m cm u a
al C a a lb
9L CK p fd c u D 41 CL 1- 0- - a -- = c I- W
> - a L 0 (4 . Z CC .. 0
CL CO 0 r 41 CL- > w %j L. cft,
0 a 0 cc r v
o 0 c ri R
C, 0.
CL- >. 0 -11 -0 0 -T oft 'I cu a E
M-
0,
-j 0 .09 x a I-
4. > 00 u2w >- w 41
Z
0 '0
0 w 4. to x "9
4) -M 4.
010 0 a 00 vzw
CD 0. z 0. .0 m r d,
z r- 0 .<Go 03 V 4, 01 0. 0 6 . c Go0. x a:c 41 .0
> I'D L>JCz Z a Ol E 2 4L um- 0
MOK W_ 4t V 61 '0 0 3 w co 0- '0 1
M U LL 0
.r En
LL; zom 'a 0 21 mm;- i ca Go
. 'r a, to Z COM r, w -j 0 Z Im, -IMV 1*
m 0 > .0 4 00 1 wz W 91 L LL
3 , z 4, 0 ID V 3 0 M. 0 6 r u -w
rJ 0 w cm
1 06 0 -P OC A. O"Cl c do LN
I d,
I '4$ k. C
0 00. -1 00 w at 0 b
z LL tj ;Dcc . 4c a ap
cy ca CLto cu r M CA a U,
w z 0
3 0.. 0.0 03 0 w 1,
U-0. ri
(X) 001 0. 4 -P 1. W
c m 3C E 03 - 0 a CL40 c

0 N w V w
x ww, w
0 -.0 10 w 0 0 " 2 to 0
d, A - M ca 41 a- 6I a F U) Q
0) E 3 r. W to tv >OW w 0 U, W
I, L C6 T 0
I'l 1. V a w ci 0.- 0 3 Z: L 0 $.
V a w -j
0 tu -x , . i
r A., U C of N =
CL 0 j 0
'0- 0 14-
IS tm In 0 m 0
> M 0 u .1 a v d. r xzw
0 41 C,u w 0z a P 3 0 02
0 P 0 a A--M,0 -M -j CL
C6 21 Go. M w WW 3
u A 'n 4J '0 91 L 0. M Cd w 0,
C, 0 ar
IR
0 -W > IIn r 0 IL
11 W wmw
r a al 0 0i E Q.
L L.
OF> a C 3 L. - u 0 *L =L v 0
M'C 41 0 41 C W o 0 o 0 o
S. 0 u W u
1w. M-P p 400 to
(b u tz
I. w cr
n 0 Ln 0 L. 0 Z0 .1 4D 10 11n 4M Z C3 r cm u m & 3r -1 CL
Q. m
$. 1. = C C C q. 1 CK j
0 0 .41 > u CL C
CL 4,
0. 0 GEO 0 . a, I IWW L -0 a. " 4-.p Lj w
X
Al o w w 7.
cx -C M a ef a) 0 0 r1. p r A cc kp "a
t4 -PLu4 0 0)0 a do .4 a) 03 <C 1.- -1 L %6;*
-. 0 9L.0 '),do 3 0 (A 4) M
0,
S.C L. to go > L. W Z- 0 - -0 m - 2,
r- 'm fin 0 -P m .0 w 0 to 'a 41 w x w A. N r . 1 .5 -P. u 0 1
M a L, 0 D- L. fe CK
0 4, 14. kft - v 11 L19 j u No
Z" ta m C c Go..a 41
Ul 0a rx N a of i v 0 m
-j
w 3 u do0 X., c u
0Mp" 4 al (n
.0 0 u 0 -1 vi i. N CL fd 0 * , 0 0 0 0. w
do u 6 Go Np > z x 41

V0 V ` M UlOC dWt6- 1. C cLA U,

> -W G a -P a 6 Al..- 0 W
UC51) 0
0 . 3 M Cm-> 's CL L .1 Ov.
=p w
0, 0 .6v - 11
a 'a 3 0.
. ., I -j -i -
00 u 2 -M
0 I - U 3110 0--.6 O(M'A
mi N. a <c 14 -A
mwCW3
to 6 L. '0 A 41 4, 4, u - 0 0 Non 0 z IT Ol 00CLv efte) I 'a
N - '4 0. . 0 Imu cc tm- al 0 lp 10 410 a 0 w co ck
qj 4, z !! 0 06.2 Go WZ
10 - 'a , 0 Z lw " U
Va 0 (b ta Go -P 4 1, 4, cu I cv
.0 0 1 a w tj 4, 0 0 1.-
n Ld 4. . . . t 00:
0 m .1 of a 41 0 -19, a 14, a U'a V'a
CL a 0 0 wwo . 4 z a
11 - I- a
0 -P
0 u --
-,w , T ,
A L. 0 -P .. m w m w 1. (Ql _U
'a OLV 0 -j P-
0 U- txwv al z-rol-w amc V)L'i c C 0 10 cm
0 4)0 0 3
u 0 0 a m -P -wo w a 0 a.
qmn W M = IN a
S.. a q C (n w CyCL
-C m'a c I m
p " -P z
A S.
W
W 4, D
0 W 0043 01" 1 0' -P CL 'a A'
0 M a
U ;z V) CL 0 0 v >1 bi
0
u um vo L. _j
Z W o ko cm0,. id -P 0 z Im 0 a . 4 a
p 6 =1 IZ IAJ-
_j .0 R-W 9L -P a -j a al .. Em . 0 V
C 1 C mL
41 6 $.. 1, W -o.
Go (L P- u _j u 0 a u I
0 LLJ a Iff, u 41! p go c L'I
m 0 a a < C (D 'A W S tft 0.. L. 9 al U. m
a CD 1 U, 0 0 0, U M V tj M X Lj q. 3 'j
.,P 0.6, c I.- '0 u I m z r
MO cc C au -SA a u
.0 0 41 a vo*
az z _0,2, 3 u z LO n' - 0go a v
Aft 0
41 .0 az -j
0 (a 19 =
C CL 41 4J
0 M L. 0 UNNl* a a
a. w W > 11 W)
10 Ul G A =A, & 0 7
in k- IMO -7 a a -P 5 *.Q a 0.,20
CL a ac '0
0 0 a CL
C6 4, 0 1 U. 0.- a it W W 'u C 10 4 5 71 n n M 19 lp, v Lxll
0 fm M. In ON
2. . W
0 w 4, a u 0 0 U)
cm I CVm m Q > E a u 3.0 0 zw 0 02
C. a u c S z -0 = ki z 0fu a0 Z=
0 m w ru 4 0 0- 0 m . v 0
Q. - 0 0 02 m % CS CIO co
U 0 cc
w ru 0 S. um S. w M u
OZ- co z u >04. 0 It -1 C CL
1, -j
w In
 -u~~
0~ ~ 2 UJf 0
C
CCUIo
>C~1
-
0 Li~Z
1

.-''0'04
tM
m~~~dC
7144-
).'4.4
~ ~ ' 4.-
2444
0.
OCNLC
4' I4' J
420
z.~1.0'
D56'0I
4
4
-0L.
'o-
N,.'
0 '
4.l
.

C3 Z' 644 6o -
0
N >04441<
Q.. v44v j C 43 z(
CL c.0 J ItSC S
N 0. Z
g4545... (4 * 442 6, U 4 a
0'4n N4 a0..

S 6'. 0 ' aS'n4 U ' C4A I2-4

X) u('C 0S

) .c( ,L *0 m t 011 1 U
N

0230 . * t44-2w .- 0 4 'o >. S..'04 .0 0

0,.a m. .C
c5 4 4-0 0' W6~ 6 S.4.0 V .a
A.104
044*1 4 '

0 v'0 v' a4 U
oS C0 C w z
CL
us8
4 44UO
C41
1.1
'11
-1SC. 1S
01 ~0(4 ~
M
6-4
i4c.
05
: -
5
.0
6
S .
w
x>1/
4
.J .. 0
o41 S O C
a444 v 6. UN
No UDS
1 0 4W 41 N0b
aL).4 044 U.
' 0.).S toUY445
CLZ C 4 . I 0 44
W414 NM 0 ... U) N 04 0L vi 0' wIS x4 a42 N.a40
.11,goo
-0I.4S
<o~
z.'
It
cnI
N 0'
.C
0.
05..
C0
452
P. O' 0' S
.. IM'
'4.
6 a4U
02
-e
0(4
w
.4fU
0an)
a/ L.S a
a6
~C S ' '
X6Z
u0
-44 6m
C.-

0, v4 W-..V) C $a2 *.I 1.45 U1 I () "'0 C) r- 4 5

a-
IS a4.) .4
S(0 6- C!4

50 4 C1 4. *5 Z- 5 4Ja "S 3 0 4. & 'I

45 .4 .2'C 24
C)~U w aU 4' 64. *'0CO r * a- N35'-r
.r
4.
'0.
-P
V>
0-0
1
7'N .'
W41044
c
CO)
4002W
- :
U>
.100)
I.
4
0
N
..
r4 0
(n
v
v

''464
~ .0
a
C
0 of0
a.0
44 . C .4
a 42'4
0'aa0.0w
I -
40
cn 5MN

- 5- 4 .4 0 CL- ('
44. 0 f 4 ' 4 6 a'4
'0 ' . 0(
54Q..42
4 4 04 ' O
C)in
4,U40
) ( .w V 4C
4.14 .4.(v c.
0.'2 00 4 NO44 ~ 4'. ZWt

c 6S.c1~
-W 4
V,-0 04.45 >." 0' 45Z 42'.. 0 Z

oKJ.4 CC S410C4 U4'0461ZCWLf0 6c 4. 1. 444 1 C W-6 I -

44 k-.4

04N4 cm0a' -. a Im .. Z. V0 0'.'- C1 a ~ a - L44A

0 :

DUU"C4'S~ ~ ~ ~ U5.455-S ~ U C;2Z .54.

OD U U 7 16.

".44.5415"2' : 0'440 0USC5.bl.04 -1. 0 45"6 4.042'44Z

Zjr 4.5
)44' 41 1 wL . C 0
0!4 i-' 4 05 00 'C 4 -r4
w O .00 mE u'U .14l 0.4 41 m2 4.
IS *U. 0 6 0 5 4 64MC
4. 40 A-a~ 2a m v4a5mP C 4P44 00. E3.
... 4 1.'4, 0. 2u -m0. o6.0a 11O-~ 'o m.44 n
:1
a 6444 4, U .Oa4. '4
C'4040I54. W
* 5.4.43L)
*.0.0040 I N. -...
J CC('4 .0' 3 41 -4J4.
I 4 .~ .. 5.

to6'. - A -41(l c54 09 0. ab w.f CL 25x4.w 6.-0 50.4

*L

v
*'S z241 *4;0 4 * To
4'0. 0...4.44 t-4'.
0.0 11

C"
54 8 *:.
SEC 4 r'
-N N005 O34' 4 m.4 T. 01004.3

* 2
.505 '0 '0 4 . 5. . 064. 541 .444 v6.u. '4 :1 16.
* ~~~
~ 0406.5
*OS4
tm
NCLN
23 40 ,'
'04,4~5
5
53
a2
544
'.5
. w0
2. 04.to
24 5.6
4
4''0..c .
c6' 0..4

~ ~ 0i ~
s4.S.4... (fl ~ ~ ~ .zWU* ~ 44.44440 ( 4 4. 46C ~ a C 4 4. .3004474

4.6 4 lb.4 5'

* ~ ~~~~~~
0 2 .. 44 0
0.
NC
4.4 4''
64 5
.. sa a.-.
40-4 (- a )04
.44. . 0041S0I 4 'm 41 ~~~5
' S C' 4 .024
04 0 r.4
4. 2 1

444 U' ei '0. ON . 5 4 c..' 4' 4 5157

'o54
-.n04t.4 a46 0.
5, b4'45. 9 . v 6 4
.'~~~~U 0.4
5- 0
04.)
.7 0 5 .4 6 0 5 ) 6 44 44 C .

* ~* I a6S4.44
6.4 50.52
a'44 4Uac
*'W fta
44
I.
4' 5. 4'
4.60
64'
a C,.O ,''"Z
62 04J'
.OM
''.4'04 -
'-4.Z
46
4.'4
mm ". 4.0 40 X 61 4.u 40
w 6.. >S0C U,204 20 0 N464 C1 .0
a,45
* ~ r 11 '0 55 4(4 04 04.'
0 a &44)4 - C U w.' 44,f 3,6C" 04 2 X46.0'

c v
4.45 4,
2-. .4 N.05 I.- 4 4 03 'A
I0C4 6S. 40a c45 74
* .'54
o. I.4 m O 4laU -44 CI'.' W0 44 L 44.06 4 4.4
' CL '4 'n2 (4 'A 0!6. 3.4.4
* 444.45 6 'SV00A(44
64 C..' 4.v 5 .,o 0 - u 44aL'P .'bL' w4> .0. . a.)A'a
'5. 10 C1.6' 5 44 C'4L4.
0,.6 . 0.. &'
0 - 44 m 4. #0 .4 C. C- 0 4.
64 40 -uL w 52 .4x ) 5 4 . 4

4'~44'.4
5.''.~~~-~~4'
1
4.*0 06.4..
*0
44-'
0.4 a.. 5".'0
4.4..'4N
n44. -W5
' U"
4
.5.424
~0 5
.4.
4
4a5
4
6.5S'4L. t" I -j.2 042 5 454 2*
o Ca4. co
1
4
.'
-c4
O-~.Lo
C424--
O .444 .a4
"of v. ON.-
* 45C. 6 S ' a Ut . W.44j44
I44644. 7 44054o
* 25L"4-C.
CL0SS1'
V'4-a
U Z La4
56405.44-04.2~
-' 0
U4
~~~~~~~~~~~
4-'04"'4'246
64.4..CC..0
U44L.4l
-W.445-
-.
42
4. 0
4
4
4'.
.0
.44
440
V4'-
ra5'-445~
 I In a v aj 0 (n L w mo
W)4 r c0 u
cn 0 011 0 91 a la m CV
I=, w 0 04
_j LL
4,
cv (b CIO cn 0- uj rl I .1 Xce< C _I 03
m w .
wo
V U) I I m ,
.0 W.J w 0 0 c W u w -i w
W cli aj 1 0 0
x
C I M 41 U 0 - - 'A a 0. mz
14 A U I .. :5 CL a m ul
E a
cn 0
0 Al 0 0 Li CD
Z. " E
w Lu 0 A -411, 10 14, IC) 'El 10,O 'm U) z u u tci c r 'n z
U) cc X i
w 01
w k4 C MI
r w M m
0 > L m u 4. 1 Z 0 w (r
0- Z_
n 0
CL LA w ol
0)
I
w
m
2 - :, w " X
o CKZ a, 0 o0 = I X> -a
v 11 a aj c Lm 0 La . : 40 0 d, cu 4 w
> Ln
0 0 id ri 0 A 0 1 (n LLa 6, u
W. d. 1 .1 u W cr
cm
u a 0
a. 0 uj
u I o w 0 z c I u 3 w w LLJ 5 0 0 w (U
cu 0 (4,0 j . T cn <C w 91 z
z U. ci
06 ru ;, w Q. 0 0 z (1)
a t I ml _j - w
Xwmw . U) =,
T 41 C I

vw 0) '2 m w w a 0
0 , w q
IV
a
21
to a CL I Wu u Dm W.- 0) 0
En
d,
Go. 14
" ; I V (n u
w m
: 's c . M W Xw 3
zw mM
.0
0 L
ul I m 0
0 1a, a, W Q
La 0 fa Na w
C 4; 00
0, D 'o
I
0 . 0 19) rn 0
> m
CL W M :> >
M "I n '" -
W 4. 0
u W CO c -0 z !
U, lwnw OC
> lw v 'o
_j > 0 'j, Wa, 4. U 0 ri
.< -m 40 6 4,
%. V x
C) uj U,
M L'_C'
0 m 19 1 E U W x (n 'r 0 0 Go -67.,
> LI. m
Z z x E 3 C M
1 w 0 'm I x z tj o 0 2 In W W W
La K 0 1 (D a I 1 41
kA W c v CA1) 1 01 lz b m
14 1 a Wx. cu 0 M
En w Lo 41 3 rwo N <
C) m a to ol tft 0 0 a '3
Ij (n > . . _j
0 4) 0 10 0. '.1 Go
u I (1 1; 1 w
0 0 -m 7- Ll W 0. OWI=
a ': 1, & . > a:
(V - LA Im1
w I a 0 -of >. Im
010, n 11 C. ap ki _j 0 , 8 j C, >1 u C) 0
a, m j tm uj > >
a, a T u im m I I m.
3 m m 041 3 C
0, W cq Is aj 0,
4, c tA ccall Mtn
a (L r a Q. I cm- w 41 a w u I.. a u
lo.. .0 T C*
L) c -rd cc
1 0 .4
,v -W) a 11 m w
U) 4K CL : 'a
w 0, p D 4. 7)-
CL
w . . I 1 '04, ..11..> a z 0. U, 0 tn Op 01a (a N m
z w 9 11 W .. (n
- T 11) ,
I V to C . w 0. m r.C. Ia, (n En 51
m u M r
u = i,, -0 0 31
0 X - ,* ul I:3
wW .1 1)
Li Eo 0
L m
0.0 E 0 mdo41 m
C r. w W U, u S.
re u N 09 1 U A
0 u cn .. I * X Ln 0 0 Q,0 C3 W r r
W 0 a cc > > a a 0 u C. Lli > 0
C) 11 Dp o-
'L m clc Q:U)
.a :, 0 oj w 0 I Id
M a, 3 - w C3 -T t d MIX
41 01 'D (n w c M N 0
> 1 0 3 u = (1) .
W
4, u C gz
<r 00
X AlT z< :>
4T (n al <C
G. w a, c u 4.1 -10 1 D En
a, uj r (n u to 4ocq cn o
0 .,m uj W W U) Z
Z 6 Q-0 0 LJ U _J
CP. 'DC, -ON 0 ci 0 'U.0 lu 1, -0 'C, U_
0 0 n 0
uj Ul 1 0 C, .. I _ W 0 1 .. 1- 0 m 3 2 0 %.Zt zo
F A I z W m
a . .
17 51 'v' mo: ri T wo 0 0
J cr a. all T a, 0 LU 2,0 .0 G1 ru u) mc 0 of
C% a
5m m
0
L11
z i 4 Z M
co
I0 I U c (I a >0 u
0 a, 11 u 4' iu z
0) C -a
OM D MM 0..t m
Vp
3 a a
4wooc
V W. v U X
m 0 1101. 0 j 00 LOA
z w I u x r
4 x > ED

r woo A, 910 v T r
3
ILA c a J w
(n LL a Ly
0 ,1 = _j - IU 0
cn M
0 " 'A , , I L = - CV (n ! 0 In
cn w
-m C-1Loj 10 3
0 (d '3 Q xV) CR;: C -0
D
rv C,

>, 1..v0 u ..t, w .0 CL .0 0 0 z

a, 0. 0 0 a,
CK
C ;
0 0
CAO.
4, a, xJ
_j U. w
;t 0 o E
a 'I 'm
to 0 - S. z . m, , w ul a,
0 rx 5M 0- 3 1 tL . 3 I
a. 'D 0 41 th 4 A,
u I0 Gox U) z L.
w
x 0
c I : d,
0 > > r :1
41 0 rd 0 v a. W le
pi 3 L" u
I , 4 . 0 1 C,
Lx, 0 Lo
M V
0 z c b 'm a; 91 11 0 Q A
0 " W
11 00 1- -P :2 1 . a, 0
I UP 0 a 0 4 v u : . :1 - -
0 v >
z a
Czw= cc , z , 'A .r 2
z 0 M 0 0.11 :01 a, L>j, cc
3 ai 'o 0 3 aj r
u ; L" . 1
0 14 a) > CK ci
M 0 w t_- In Z
3 La
_j 0. 0 w
C. CL m al Go
C, L, m N CL 0 a
U) ra a, 4) cm v 0
-1 41 1 3 C w a - -D In
u ,J) ru Gi
tr
0 IVz T wo- ",,.o X
n 0 Q 0
41 N 0 ru kn 03 a.
ClJ to C w 0 > CK zco u
z im
w a.-z 4, W
0 m -0 T. In m 10 0 n 'A v 0 z
> W m
0 00 0 "U W w
40 . C 7 0 u
u al
w q. R<
.41
02 %'I `2'j > uj , 0, Z k. En -i
z I , .
0 u > In q c 0 , C, 21 a Z W 0
Qi i = 0) q, 41 0 a, I0
0 1 C I CL,cn 0 X 'T 1 0 1 1, - LL w
_T 6 0 E 0 C'
x W
ly Q
ff c a 0 M6- I .
5 :, Cl I
Go a, u z0 w ;
0 4Z
0. " m Ln W ON= x
LL (n CA - "
03 X v 'a .0 . " m a 10,
1w w >
41 10 CK
0 a.- FU c 0 fa w 0 w m
_. .as r. > I w fo j"
r 4-0 w > , _- .. ...
10 (D 0.. rl I C , = .. .. Q v a 3 u a, C.
0 .0 ORc u I La _j 0 1 1 4 CA= : w I) -
A m > z 0 , 0
0.. .0- ul C u a z u
0) .0 & > m Z AS z w
Aj 1 0 10 w 11 r k. 16. L 4'
iv - 14 u I L. *u Ic C 4, L, w 0 ru D (D
0 g r m t c W,
m w W2 0 en z w m 0 1 wo 10
0 v a 0 m 0 X m 0 T C, &
m 0 Oc 0' 0 'cc -i < M 04 0 u 0 0 4 " j " 2
w a 0 w a z 0 a 0.0 0) .0 ei
- 0 w , v , >
0 fflO__ .0
41 A
7r 0 C3 r I D
a C0 - A q 9 0 z w
U 10U 0 41 o"V u I :w, 0 0 a
0 0. Q. al M Z 0
0 MM
 N to w CL In 0 a, w 9 w 11 -1 -0 3 to o N Lft iv v F 4 c 4
-0 0 0 D C 7 c 0 0 a, 4, r m a, n p I r :, 4;
0 6 0 0 o c I a, 4r c, o m o 11 11 2 a co -j al 4; 41 3
u ! r < ,
c z c a 4, m c6 a rj V 0 w I p r D fA. c
3 s. c u c o - c > z. u " 0 '.1 %. > rj m a, a 3 ej a) u o o 0
a Ib a, 4 cc 2 o r x Q a, a 5 u 0 0 1- r al 0 cn "a
f 5 'i >
o 11 o 3 w z u to c. z 4j a E
(I p u a 3 > o c -w N m 11 41 1. M W 41 u r, o
u a > > C F
r o tu ej v, 4; 11 11 Ift r It 0 Uk
61 Cc M j3 w CI
., 11 m a, -P aj Q
to 4A 3 0 c o x: (L
o) M w E (n w cu L. 41 Id
m Ln o> I"
CL
v 61 z u (x o to
wa m 11 o 5 F o > ., c, u
m d, I In a z o c s al a o
'L z Ln c a, o "j CL 3 j a, cv x 3
(L m 11 l
E Q ,p m 4. a w o :- a, m o m6 o 1 13, o w o > o
o .1 4, is W1 t- a, 3 > 6, - I- !. 6 o
o a .4 E v n z o o l w . o I n (u m 14 o. z u 'C 0 c at
a > W c a. u o c 41
MW C
0 4.1
C
o
Q) -
>
m
cu
u
m T
r L. w 4 u
o >. 1 7 cx 4 D
c u m
g z
I m o 0 .2 = - n u W 41
11 *j m u c >m o m , v x D2 4,
o ft - p- 4 Q w mc . a -j o .0
a, w a cLa a, 0 43 c) > (n o Ia o > (D 4-
>o 'u. uLl> o !u: n m 1. v:Io
la > 1 1. ou a I-j I-j a. o a -a o oo x r.
Ift6 6 4 $ .. " . 'D "
"' ..o . T
o N c m o o A zC x z r o o Ift Imc
u .. o m 14, L M V Cn
w
to IL w
, o ce u, M 8
41 . 24. 0 1 X 4' o.
n ou a 6 lo w T Z "Iola
u L Z o' do o 0 u a r o . o t" ma ro In v bi
u v o a, q a, a 6 6 z X
u fj
: mx
41 mim m V
m - a & . I 'D K to > u >
V %. a G. .1 v
iv wt .. o E m .. m
w au 'I w zx 'o 4X I lo u w ff
a 61 1 A x 4) N 41 o (n z o m do rw c*
Q eL c
s. o a t w N :2 o S 6 11 L w o) -a n L - ac
. . a 4) m c aj I! 'a _j o 4 m
4-
o " p T - o ..o o m4 -11 w1 o c 9 q, 41 11 -a Op
6, CL4 c v, m (3 u c- Ur 41 w. w % a. o z, - c
w. >. 0 E 0 2, z to cx _j 84 r c, o p m %ft 11 r 6, m 'n o 4"4
,p w > a a a -. 0 u
. 41 Go
Q- I A, 2, m v wucw c
m o T D m
> $. a m o a cn -, r. x u" 0 -4 a Z w u-o o' o
m w
m If a, o - R 3 3 'm c) m 41 cma a M 4. a, I.. o cK -T w
q, fd CD o
c I a cj >. c - E c') fo o w.. -w > -.C 3 > F! V o cu 3 c, w o
u C- C u z Ip (n a . > c t a, D 1 m-o cn a.
o c
U,
a- I I .. l I o- 11 114, 11 r a, co 'T 4 a - e
a . a a 41 .. .. u w a. : =; u 4' -1 u m c a w% Iv a r o w r v, Q) m
'a M m t., G L U a (n c w m tt o c ci r
w lu u r1i
z co = 4.
u) o -A T
41 o' cl
'o m w a, 4
o 4. U CL a,, c r I > V > o M oww Do- v m r r
c w w kA m w 14 I a Ip -
(I a.-
m a
I L co
o M L 0 > & 4 a v,
o W Lo ,V o > m o Z Z z r '1 0
w a w c, oQ a a ul a
.1 m o aj c v D a, - > CO > ff & omo a' al v 7. 4 -0 4-1 CrLM 'ft D 6 kA o m ul
Q u fe L .1 - - x - . cc v, co r 41Gia c n c u. I S . 41 M -P 44 0
as, 4' -u w > w 4, I at w w a 0 4, 1 uj LA IL mt c IT M4. u 6
z o 3 WIn a 4, '1 l w al "41 u a
Io w o o o T z v -7 0 2 M P 4 C Q, 4, v La to c
6- C m -- ,DR . u 1) w o C > g o o r S Wu. i u
tx= 4 V u - u 11 'o w I o 1* (0
:1 U. T C
o..4J ..
ed I o w Ic a. u 41 a
ap o o o" a) w w do Id 0 x o I (c a c i n
rd x m r 4 o c, I a, o m m I a > u W m 41 W o
0 o tv co,
'. a, (L co a, 2
> 4
u - . , c L; uj I _j -a r. 'g. I v al
zcu 2 I =
4 oE ta) C
> m
IN = m
Gom
a 'o r_
q 41 'i
Q o. a
c. 'A 0 <
U 0 5,
U Z m Ip >,=
tft Lrj
uo>
o: Enr-.c
cz" x:::,I roll z
.23
ff c z o., -
o *E'V>. ICL4
o. u -co I. o

a, 'a q. >. w 1, a, wn D.. I" >. v o1a cc Iv aj,,c -P N a

-c o- -c .2 zx Ln o 41 -V fa W to A o
:3
4' a w (n N - u uj o,
4. D o N c Q- c w o A Go U u ID c M Iwo u y
-1 x" j
41 o o W o tA -C Lft l a, a" o
E I-
Ln x
N 4 -4 R
(b A Q. aj CLc
-0 M D u
.0
r a n Q, 4
o'z m 6 w DI 4. a C z c o
m W to o.. LU rj Qj
r
c -P u 41 a 4 41 (n 0 a, a) w c - a. 0 4 L
Im al 3 L c cm L L" 0 Ul u 0 IP 4, 3 4J 1 U V 1. En -
z c w o- tm ww >. al o t4 C) c
'0 41 a w , g , z LA z Go -a 3 I a, a 10 o o 2:
0 E L) u c, al c C w m 'a 11 . . um
k" m o w 4 o u co u Q 4j c Aj Q)4. It -u Lu 4) a u m 6
o el a. a, u > I, D Z cn lb
O.O.C l . . . . w o j . In
4.
Ic c N a Z c > . 6 l w 8 > - - a, u
>. a, .. w 4. CAo NI ! .. 3.. 1 w . I I 1w, z ol M 0
o 4j Q o v . a Qj w c z " al > ra Lft -P w c _j (n c a al
m k" 4 o 3 En (n r u 1.1 D 2= c r m w L. .. u
fo 4 a o :; . o M M 4.
a, c fjo I m c 0 r 11 11 1 . m M N IE 14 > 3 ru
o' > M L W' X a u a) 4" ma
m D w w M
a . a, Q)a c IE.1 c r- 0 'o 11 1 mC 6
at Q. u 0 2 Lj o- c n fo o r -v E,
o 1, 3 -.. ft cm r 11 r > I I l u L. " Iv 4,
I o 2 C I-
41 a a w a m a, I _j < = ol a) IU M m
o oz L 11 Lj w u a11 a.
; a w o & Lq w 1. m
I c V I - u I W P 11
u o u 0 N lE u A
x a, r cn o o E
D. 'D w T 0
w Lft Ij "P Ix o
do m q. 4, Z m 4) Qj In > r, zo
2.
4, o
2 cL r u m o u -a 41 a . m -1 o., .4 z u
r to c z u 4 11 lu o di tft I a w La m 0 14. In
ej N -a
o If (a w c -13 m -m am M
2) I c v cm4 V a M 0 0 u .1 Ic W u w V
cm 14 z o -T 0 4j 'A m -- o w a LI u w v a I I o
o c - - z c, a o w u o.l D.. a u u. -j 'n Z - 3 .. a) . I
41 o G r a) x _j w . 5 * u) u a u to
o a o lu u u=,x m T I I o v o 1. - mr u
o - 11 -r
u, m u c
0 Z u l 4, C, fi a ID > o o m w z< Q I a, 'u, m
oo
'd1a Qj . . G 'X W a,
o -P > cmu 4 a
m o r > L o 0 _ W a, Ln r Ln cj o 0 dI CL C C 'r (b - c c x w ..- P Q.- 5.
Q). N - c. x L'i(n o . u C,'P c I u 'D x . . A .. . u
cit w A o -w - .1 00.. . D .. W 'o oj
o c) 41 c 0 >n > T
*1 o
& p, - Lm
tr3
3 ..C M, . - >, %. 4, m m0 w o _j -A-a di c P
M 41 O Q U IZ M M U 41 Qj
> M LD 'T N N
o ac,'o, . I " " o-- W 11
D 41 v) z cz m o-
41 c Q z o' Lij r, a al w E o 6m
6 C IM 0-0 41V C 4 J 2 0 a, o mV w r 41 z c X Lj C a$ o
C, C LC '. _ Ew 1, w u) z a I E m A- o D
o a, >0 4,
:) 3: w I c o m c w w 0 o
61P CL IL w,
'c u w M.. m -j
*1 o U o 1 11 m w
c > 'm 06 .,n 6 Et a) fa
W a, 0 C&
IL w oc, an
11 ego w o v Z
o > xej w 0 > o- cL
N-Mr, IT d. o'
CLV
a- of > aj u, ON am c o z U)
w u) a UI
m W
o z.6
m 13 m IV I = v .u
4Zo "I U. -o
_j 4,.o vj 6 n v do o al W z a e. a a L -P to I c o a. u z .. 5 .. T a, c.
c z -w w
N z x(L.u M 4 61 m a 6 o -w 41 > z Lo Y) 'Z C) M 4J 0
m w 2, w oI u
a o
o >. 'o to
bi 16. > Ij u 8 owl: uzj r, u c"o I do
1 E W 4 W vt s-
x:
fu VO l ` 11 .1; vo .6 c. o ou a
o -0 4
ov CL tm-P %.4, m mc-
a a wa C 0 cr _j u 40z c a 13- 3 40tj o 11 cc;: ux w aa
LL, cv & - o a% 14 - 3 Ij c" > Q, x QcIL
cW o
w cq udi ral 4
u cr > n .16.
3 cm a, o o- r. 0 'Q r. v Im u -< PLw
I a 6 Nt& o c - " c t j
4z L, o 6 o o
06 CLU .1 t, u a co r cc M a T -o o cL.Q 0 r -. c L cL w co cL a. u) z Q
 -- ~~ 140'1.0~. a~..-4001
v.00 0 0 01.

0- -,'~4 z u
0', l0 M0 3- 0 1z0 -
.c - Z 3.a M, (4co.-).1 N >0 aj.'-0:.0 a.4&n0T
cm W l 31
04.1 .. -. O 000
C, L.0M1.-OUU'P40MP1W 0* 3

-
~ c% ~~~
11~..
-
o~~
0.
(1
'I'
M10k,0% 0. 4
CC
>
111. ~4 IN4--00'1.
(n0
0
*L ell
1.0
Z
t..0z ... 01.111MM
M' IC 0 4
O~
..
4'9
O U3-O
ar
Il 11(1 4. LIJW .14 - 0 .. 00.4034
a1 0- v f
2l..4' 06 0 00. - 04 'o10.1 0 N
.10
01 .131 N 1I 0 1 WI40 *o
1.0a~.41 a1
> ~ 01>I ~ c 00'~ 0 0'. 0 41.01. 0 V 0 zO
M0v
c, 0 04'11do -1W
00 E . MI4
oo. a ~ c0-
CDr~
~ C.', 1~
X:
111
0
* 0 '
.030W
0C01,
3>0 4'
.143
0 4
1. 0 >'0C.1 C. (d * .,I- >01a3 i o m W4 L a C. xII

0 -0.- M
r11>).1 4,. U-l I-WaC.. 01-.03 0l1 In01
0. 3. ', Z .
1..l 1 1030 41. (
00q.L1.a 411,09.4 cl x 3
41 01 U. 0 0 .l
a- 4
1. CL cc0 a-,Z014 W-04 J).0- z' 10 L
0~~~ 1 0 0 1.. 0 0E1 0 0 V,. 00
t11: 0. M. 01 .

1.~~ .o.L - 9L.ooa

00 .. 1 0 .E0 >.- 4114. x l 00
IX 4'.'.'1 1 1 .1 0r 14 . -11 "1* 00 ,01 1 0 0, ft 0 c

va
0 00 0 ]A r a
0 -191 Q110 *1 000 1
14 w .0>o 0
M 0) 0) .- 0.0 a M E 4, W c c CL u 0
r >103 -o 0..40 0 W01
a 90~'
z-' .l0
CD0 z- , 0 .5 " 4 fd
d..1- 0 41 W 0)1
I-LL

01 co 'd OE0. W a r. .0 .0 0 t , c Ua'

0-
0
.140 .1 M .'0
1
o '.u
0 0c aW
LL.
Z.
00.
CU
-
.-13
C
~41 C-Aa , . 0x413U.-> 0In*. 0
04
>-.
UZ
>
0t
.0 1
-0

3 . I'..MI -0'. z.)

0 . 10 10' VDO
0 0 2 1. 10 01, 14 3

N' 20 0f0. 3 03 XOrU- 30301 .3' t00000

*000 Z- 6
N c
-il0.1 0 0.'301a L. M-0.Z1. 0 0000: 4' .0 u N0 n1
0C'NW.CL') L u.01.0-N 0 a a.C 06 cm u u C-3 40)
W 0410. 0 4' p 1 U 3 In 4' ..
V* NWa co L

%..1' C&Z ( fl.>.oL,1 k. 3--' .0 c u L.

4'U at 1
0.0 C.1 C. P0C1
0.04) l.0' 04 >- 0 CU
. CM *O -CA'0 - M 0C. D 0 c

01. c u
M a 614'94 (M. o. A W
E0. )1.1. .19 0 09- 14
.0 -P 0..-- W W..- 01
ty .k(O30 11 0 1 -0001 . kn0 CK *- CL 0 z 0
C 001 0 3 0 414 3.00
U045C
0 ino a o14 a.1 z 14W,0 1. 11 Lo, . :0-0
4. r 41.0 to 0. Lo at 0f
kdIn 0311 o00 0 W0 E1-PZI.3.31*.. 0I0 L. co41-.05
4v% - CL .- 4 X . a,0N to4C0011 $,4'. 0.
UOL
mv0,.0.''. .0(0 uW0W0

I Z11>110
(1UO000 11 1. En .1a) 00 3~ 4x
z0.0v P10.1 If03. E
4VU3 .0vi 0
w(n040,1-4W
w In - 1 a.. 41 .'a 4 M'0W0 ' 0411
U*500 3S 0 1u.1 .. 10. '
('50.0>00W000.-.'1 0001
4:,0 >::10
(0040 M11W11r1S4 0*0Y)1111>
O CO0 4 4101 3-40 x.W (Un- %. 140 .00O6IM 1.UE4a Ur0to-01W14
n 0q.00 0 <015.
0 M A ) - W.0 1 1 3- 1 1 . 4 3 .B 01
20.
0 1- U 01 0. 33.3-'1 11'
01 In 0 U..1 U 0 0 3Z.10

00 3 4
1 CL U M -05 10V 3 0 v 1 -. 01 (n 3
1 . ' -1. -.. 0>'-'
11.C'-- SW.4 .0 WII U' 'a 40-Ud,
>1.1 CO . .0
3- U k" . U Ifl p r CdoK) 4'.0 Rt 4.'0 MIC 0 05 3UU. w-- 0 4' . 1

W11
4' o . .9 Wa,00 ZCO 41p
. 0.
4r 00 9 0 ,W
a1. I--
3-, 0 1 0.4
ZO*.3 >'1101 3 -a0 o11 (..U.0d ,uc 41
4' 0 0K111
. 1 w 0 : >. 0 _j
g 1u: 1uv1 >4'.' o '.I Lo
0441 000
006 c9 54 0 -j1 I4 a, u4 1 - w.1S.
cc.. c1
11 111139 340 C v 4'I 40- z1 U .0 ud ,4 1:
0 * a- ,b u a, (
. a M.. a, 4.4WW0 0M0

4' (D a. co 0 N0
9.'- 4J Ji JWZ
4' ZV
11 MV L ZO4 - 0a 0 a .u a) 0 u 0.0*1 0M
1
11 M30 0 ol 0M 030:1 OI. W1 1 000 10* ON

ap 04 .
Lh-.00041 !.p0 4 0'.
MZ= 0 40 61. - 1
al go 0 a1 I-

V'..
0101...34C-1 04 - 0' Cie U0 11019' 0301' 4
M
cr11110 I-1 41 0 . 00.- 4' 0 M4 0d a911 f 411
.41-40 0 IM0
0. 1 0
11 Mu0. Ow0 0. a
1 0.->~ 0 - '
a,00 4 * 0.. Z0 1.
>' 0 X01fl-N
.01:0 4 411 11 MW4.3 . . 3-. 19 010U
0
0013 U UO13 -. 0 10 3 4 11. 04' 0 '-
r..- OE11 1. '0 In
.0>4 V 31 14Z
C0 01 11 .1 .
'AU * 0.1 x1 14010.
C10. 1. 01
1.>L
N
-- 3 an Nr- M o.
-3'0.01 -l 0A 04'01
<7 4'4'..l4W ''> 40.3 013 4*..9
410o01. a)
-'01 I. Uu . a..0' f '

7-..'-11.01-4
-13 1in' . M130
a2 to-. .4 0U '.in 4' 41 q.=00 No
0
1
u11
1 10 0
4, :.-
* 5.
... 110
1
~ l-011.. 1 00 .0.
40'' -.v 41:1
U 4 M1IU
>>
U4, 'a 0
~
D
01 _.. 0:
411

" "I10 3
343-.7. 1.3 410
00:1 a. 3 0os. -P0. -1 w3 4' -3.
0.
1 14.,
0.1 4' 4WN .
0.0
Z0.4'0.
S.1Z2; 1 W
. 0

a0.r a11 0W U.30of 0 v410 0 41 1. 1.'3 .>14 al*0

lp.l. W4

W C1.0 CD 00441. >-

330 .
a, L-91
-"'51
1.~
10..,
0.21 m
I.' 0
.00 1101: 0
33
UC'
.1
0 *0
~ .1 .04'0S01
1.
34..'1'I
7c1Z.
01
Il0,gU5: U
41
.1
111.
. U r0 a 401
to 01
'
01
00
11032 .4 .10 33)- 0. 14 0 * 3 * 0 0 1) 0 1:410 3p 0
0 3- I9 Ll4 4 14 -'0- 41W 4A 301 0 '..0:41. cm1. 1.
401.0.
0 00.U 0.00W 00. 0 -.0 C 0 CW -4- 00.I 1.1
0L 1 11Ir0 M1 . 41 .1 1 0 . 11 1.0 0 U 0 101.'1 4-100
a.
c101 011 . 4 1 0. a.43 -I *1 000 0n 41 I.Ir0,0-.40
0 .. 311 0 9
CO.. 3,010 31. 1 143 GEz10 I 001 3 11
01 1. 110. 0.
0. 11 . 1 4 to) 4
G 1. W3 0 l 3 a 1 NI 4 .' $.-. 41. >-0 4 JX
1.4 1
0 4'9 1. 11a9ia 4, x 0a 0 W a, 4' 1 0. 4-,
> a 0 04 *060
o0 0 1
C) U, OU Z- 01. 1 111xP--.0 0C-
01 0M0 30- I.W- 111. w- 0 I- *._j. *- It. 3.-
11- 0 0 -0-
in. 00La .1 o
di40 03 0 0 0 2 01 . do W0 u.0"1 .0
1
0 0 u- 4 4
Q1 o s -T11WaC
wxW 0.A -0 MM.aI zaa 40.
C19-.4 11
3 Ww0Z. a '. 0 06 M 0 0-P'0V1
2d 04411 9
4' 00 30
3>3-1 3 C 30 111-0- C-. 4a I.. U v 0 > 4- . (5
>4 01 30 3.' 014
00 4 4 1 -0 03 v.0 . 3
L 0 4 ' ' > 4
e r. . 04.3 110 4- 24.
..
0U#.:
41.0~ ~~~~~ 14 (34.
(11 14.1
0 0
>.0.goW
3.' .. W
1.p
U10 m0 . U
*1--11.
N-..31 U13 010,1 a
U. M1.l.100 .1 0 1-

cli 10191.
00~ (10. 10 OD -- 0141M M0.0 05.0401 .o3.'-1 v~.0 1:.. 0 0>1

O0*.O~ O o0. 1 . LL 0x O .1.'':'

r.13.l001 W0.In 0 .01 111 1
' .4 . U *E.1>U, 01 0. . 0 3 a' 1 .41 >00 -
-: g 31 .0 3- 1 v *0
x01-

7u0 a, z. 0
-0 u
5---
M
Goq
,r 4
01
I -I,
a . Wc
-% 6 10-
4 - .
o-
-

W02 M-m - .,>

-P c
a a 'In
 u 41
Z

C x
N

u ru
1.:'0 0,7
: - 2,
N I-
U,W-i N .1 w
t '*I
, . !
u
v 41 -j V) i j M r C Go w U LL:
_j J Ir 0 a, u 0. o z d, 1. Oj 3 W -0 0
'o (n 5 w (n
> > a o 2
19 o 4
m.. o o Zn lu-P 41 a c > ru o cc 4, "
at . - - .. u u m
r. u M lo Q. 4L z r- a Ir 6 11 . 11I I V 3 1 uo to In = C Z . -P 41 u
4114. cK j z , o q. A M 0 4 0 r- M
8- -P n rx a to o
0 C o CL _j - r A I A C j
cw Im Qu cr w w 3 o al -o CoL
.21
MM I G .0 c M u (u. m m m m E rA U M 4
r4'..3 -P u A La o
z 0 m
ul v v u m r A 4, w
uwz o
w v v. U = w 0 A a C u a o o u a,
u
OD * 11 A w x 2 A zoo o- ,
=
(1) . . .. zcl . , -T
n rU a, o I A m,
0 - 4 41 C C vo) oo cz - (wn u u N -m o (o tv '* al Qj m
14 m o 0 Al U n LL -J LU w A o
A aj ft w L"
a, o u d, U
cn A oa u,
- w w x: LL Z C , z z
16, m. 10
6 P, ozww m
IIK I > a u z
u o cw> o A s- w > (u u 0 4
GoZ -w 0 o o- 4J o al A 41 41
o
L) -j oo 410 w m
IT .0 on, o WW
's
z o > 6, w im a a o cc I. m u
1 w C;u,> A .. c o
(n -, D 1 96 CL v o a a' cc "o o CI-L
4. 4z' 'U'l
Z U V > 0 le a a o- A z L. & L 01
Ao a w (1)2: (n 1 '10 o N A.. A
V
w v 41 a, E 0
o r o il n 3 >' a qa
ac > . cn o > " z 6 ol 6 o w. _j c C
a - u. m 'o o
a . f, m 4) E 4
w w z 64 1, A
cL w Ol Znc _j
41
u
v V o LL w 41 -c u
'm m 0 z
r- N 'L r- lo
ww -j o 64 -v %.v m Cc
w
_j ; cL In C
0 o
m
r, r
a r
m :q,(b
W 014. re U
A 43 0 o' d, o C
M
cl A ro It a ("n,
1w,a
o 'a 14.6, L. u :) (n c= 0
-w IN 0
C 4j
E z
> x -j * w - o W, a. o I$ 1 0 u c C wo o 3
o a o &. v m cx W (D via. C o cL 41 m C w fl 3
N w o o N 0 a r, A
m C. am4 42, C all a, w L, o u L
41 0 a o to L. 0 -j m Aj L. u
E 4
ul
_j
n wI
I
o
f I
w x
u N u A m o w oA a Iv'a zox W)L o Q ;n C Mo U ? rj
* , ru z o
m V .1 w T ou -7 3 A
Q6
0 fe v N u Act D AA a 2 o o (4 3 o o
Q 4, m
11 o.
o I
D a A C W, z w 41 cv A In u o 4, E
a, w 's
o oft .1
11 8 p
ac IAQ4 'n C 41 o 11 ol=1 u m 4,
IL 0) Zn w w u c 94, o z (n a c&L
LOI a U) dj oa, w o
U, 'o o o 4 a a (n uj (n w (j) w , < o L" cm> 3 m a o w W m o N E Lr. Qj a, u
cL LL M 0 a a. -. Ps o r WV u (n to a. - _j K ooo a a 0 o C
> 'o A > MO I L) M G 11 D T - M W
z -;.u u Ul W o.
Ip 2. M
I- z
r " > w A , I = a
m > . a. w 0
ct) z o 3 - c r. x c , . . a C a al 4.
(m o -
w Z A u
z v) z a 3 v ww 0 wo IN 1111 114r tn (n CA.: M -0 C 41 M .1
c " w le cu v o .a M o Z
D- 0 fl U V 0 &
x sv o m A Ln(n zz u '4 r
ol . cvw
L f, -jvu u 0
w -cu al A
In m 'o m In V
cho) ff 'I a Z o
tj C > u
41 a = L - m 14 x u.
K La w) - 31 a - z w z (n 0 A o I..
w N O0 WN H 0 C W -zoc v >. C a o C Loc o 6 11 A m
o, z o o 2.x a o o cK o
t S
N a a tA
41 j r o' w
fu Lu

0
ca Lj
oc o z
M 0
c
Ift p P L -w
4j w jj
r
o a a o
o o a ru
m
o In 0 -A M q .9
A > c- A
V >1
v
o , 4) N
21 1 "' af 5. 41 1
r0 w I wx z a w o
6, --.A m
Lj a,
z ul z (n c rd <r o u L'I c o m M c 16 a E u c x cmq;<o
k"u K p u L. Z, m w m 41 o
a- a 0 a) o
m
41 = = o o fe A o m w- o 11 a o C z - a I o
r oo c 5 ID m 4L ul oD 'L - -w w2 j co 4 Q cL au to v o a a a m

cNw U) w 'a f 41 V u 'a 4.

a, luZ w x w w - = .1, A a, a Z z
a. 6 Ln w a
- tl Aw I, uof'. ! u 113 1 'or A co
Li 0 a u j, m >. z w <
z r (v
n n r c, m u
a I c . 1, - cq > eu Dv $) u - C a) i-
a, m 13 Q3 a c2 o o p C&
L.
En- oD m a r- m A r
I 1- 61 a 0
u 3 al
o G o v m a, Z Z m v m w z
c w w o o
fo 3 A
11 o a, w
o fa o 1 0 41 W c
lu IT V a, w W F cu ej C9 z _j w
11 u u V ru y o Iz 41
o M IQ o 3 c
U M 4
v A -P z cL &
o p a,
A a w 0 06 M o o
o o ..o .. w N Ij a, 11
Ll
x 0 Q 0I C > a, fn o o a o 'A o
4, > C -c o n _j
a, u - o Li > %..n.>a, a ED cm Ea L'i <1
ff d, o 0 a- m o 6, o - = _j z *' - m rd
4. u
oA 'a, V,I. Ol v
o o0)'1 03 D w - w w
D o ar.
c M.- 7 04 LLLU d, E
'o
w
41our c 0 -o
a In Lj
o 5 It
4 0 m C w
41 3 r A r (a w
a, u u w A qj
'o
u c C . z I a,
(n I v lo ffl
o m 0 u rj C v) o' K IT M M o :> w -o
o N u
ru C 4, a, > n M 3 V P to 41 j- 4- E
o C
:J >1
'p
u L; mu) o knj z
u 1 11 41
u o A 'a o A
-1 - r o a w w 'j, c -dj.2 d, w i 4
41 'z m cn a) w <2
z u
m >, C, > A, o p A 0 c E o) x cr)
w 0 a
o (C A U M 4 'D u u
"D w ;
o rj
u i a lm If o E
E a' Z E
A o-- u a m V a D z a m CN= w ch L.
Q 41 41 0 6. -x cl 14. a c"m ,.- 0 4, o
N
z -1 ca r I -- w u
a. 3 -j Lu u o A, u u
K a. R c .1 > m
re di '0 w L3 -T epo' C o "o a m o w 0
ul o o 11 c) 11 v I vj w it fn w) N a vj c a rj a
') - - , u -j
co .1 1 U.... .. . I "' w o -1 J- x a'! > > a, *-I, is m 6 m
a. 4. 1 o C A c u. -7 Lu m tj o L A 'D a) u 'd v m o a-j 3 a, L,
lc"j ow 0 o 'd
-a 6 a. m o
=zv a a
C FA
o E
rj o
u
-,
o .
W fs
> A inm 8 N Q.
o o- A
IK cn Z 0 w 'm o
a' a a ul
u U u) w 0 6, 41 Ma 'n A "L M'a'O L
In w) c a.. w cu
-D .10
m m4l ml
a a m m IF
u. M w
w
oo . o- =a -
" m > . _j
AZ x A mw Ir
c o aj cn 41W r
v, o umm A L
M a a
m m a)
(n
r- 0 .11 4j 0 M 61 o x: m > .1 - .. w C', u
rd o v o Qv a < A Z - A v, , > - (A L rj W
cv z o,44ol.7 4 .1 a M _j C >, a
14V 0 4J a C to u. ruv o > z L:
r m N 0 w a a, fa g L) m 6b z u m ua, a A a u 1 4
A 1-112 C .. > oz w
N wc m u U lc'- A 41 Ima > 1411w t. o o -j o o 4 o ol o-.. o r e $. C a,, o
cL. c o o - 1x ujcn ru c 0oc Q a w !n c o oN m a %. a I
06 r
4, o m z a, v A 4u 4 =)
,A A %. A o a,
c% 6 2 L2 1 a <
r LL
> as m ic A A C c 0- -o 4.1 a
C$ 31 W V U 6 4.) M4 v D 0 zo w c -o 'cL rZ C" C 3 VI M I > Cl U a' di L. 'A Ij
mv Cx
W a7 Al a a, N rj o;: r. z o >. x
c Qu a ffl Is m u if L) W W co 1- 06 41 C.-M 4J M M0 L)
 C >,Cc z
.7 u 4 a-
41 Im r, W
0 ti- -C 4, ,p
x z Lu'n u J3 z 0
:1 Ln
a, 0 c . z ck 4 0 u n M 3 O.J
- - - .0 m _ 3 a, . %. _j z .
0 do 'r, r (n .0 q Z-ZW 0 0 M L,
-0 0, 7m
"i fo r. a v > r, 0 a, a. a 3. U r , L u
aC U m 02., r i
't
Ox a > a, a, C, :j z
U V 0 1 0 Z m p m 0 LAS 3 ul Ul z
m 0)
Q. Im 41 V, U, L. z M 4
I a w W x a ol a
'a 7 S o a E w r, 'm
u a, i w 0 0
0 d, O4 E u c ka Lj ui

u 0
" u '.. -, a m C 2
fo m w . CO
Oco . . Z u W
, 1 0- 0 4 0. 0 =wz ! , >, .1.
Z.u -1 u ..ta 0 w 0
aj M r w w a =_ Qj_ <1 ED
Go . = r: 0 0. 1.
w U. t"
> 3 (x z
0 L) U V CL a M En li
.0 W 0 Li c I QJ 4A
fq! to u I EnV En cu a w c 0) 4,
4 0 0 M uj
3 C u 3 a 0 Zx
c 0 _j (In Ln M2 . 0 11 10 a, a, U-1_J W L)
CL -cu m-jol %J m 0 u
cu .0 Ln _j W. :1 41 10 D Lu W.
to w 0 OX z ffl M > >
cn 4 - k. _0 Li
0 r 19 u u
w :3 Cn Cox La V 0 1 a,
=: - * w a, C U Wr E M 41 a- L
cm 0 a . U. : 9 m - to
a 0 v o cu 0 0
U, r 0 Ul
nu 0 a U. E 0 Ir uj
m 1 0 L m -P
16M 6 0
0 - 0 z
0 4 IE w cn 0 0. 0 (n C_
OD
al rd M 4, 0- p 2, m k.:
m (D Li (d 0 U,4, S.
K v a M _j (V I
ru m c .0 z, Z. 41 IV m u C u
Z 6 V A 0.0 L. -M Ln Z C6 w y w
d o t o c a UOJ
& a O_ . 40.0 > :1 n uj 0
u > 0
U.C C v 4, CLCu 0- C, 4 U. U-
0 IF 1 0) > 0 D 0 41 U L,
41 v cm U: o o 4, S. 0 u V. M 0 C L c a W
r
u w lu > 4, a 0 U u .0 at 4. ww r U) (a
- lb
4 0a
.4 4,
10 4 M 0 .0 9: w 0 N 0 Z 44. w
Ix -C Lo x
MUvWoo< A 0 rj D U
I., Ea 01 lb > T 03 4,
0 N tM 41 U U UJ O4' Z S Z (n cc
0a w di w C) mx 4, Q 0 m Ul 0 4, 4 tft U7 -1
0 CK = 4110 0 01 (4 u I X.. > m0 M -0 Z 0 W Lu
L'i t, 0 m0 Ic MQ v P 0) 41 - = En
q W _j 4 , 0
Cn 0 U 0, fa . 61-P 3 C WZ U
CZ41 :1 61 U .C 0 r r 0
a w 3 w 0 tm:> rj W to a 0
, m 41 W (A u W r U u 0 ej
C L " 0 a z m w CM I . D a
cl M 0 1 w Uw0 v
0 a 0 IL - 0 .0= 0 u 0 al I.- x N
Ir v ri 9 4, > r OL in C, j.
c a to 0 r _j > 0 -P ul co
al 'o m u C,
C; m Pi MU 4. 0 J X La
0... P_ W
_j 1- 4 cn LI) 4 33 41 z CL T C
C z z
CO W u C Vj M U 0 E
0 0, 0 tj a w 4 a 40 cx u 0 "a 'D a,
0 j r- L
z z - c iu 14 (n 0 1& 0 41 u > I C z u G17
0 0 to a LO 0 Q a, u 4., 'D D 0
4 M ; - 0
o -0 2 9L u > 4 E
(x w 0 s. co u a A 11 a w X W a (D
0 m Q > 0 L" z
mz a . n S. 0 0: Cu D 1, ej z m oz w w
a, >. 6 a 0 0 1 - x 0 a d, = . 41 ! 0 A -1
u Q z C v v > 0 La O -
0 c N V_ 44 a zo 0 0 c 4, 41 - UJ
,I to i C = c - 0' 0 'm w c a, cl u 0 to 0 0
7, 40 0 Z: L,) M cr) m : >
c 'I t-i
E (M v cx 0 _j . 0
vcw x < LL 4 r z z -M u 0
G, T (m c S 0 c c 03 0u M 0., 0v C)
-1 0
03 Ux 1 0> M.0 do m 5
71 re 0 E L, r 04
ft a CL a CL a) u < 4 a Q,

wm to o P Ci o cn c
a lu
ap 0 a > m ai a > w En 0
m ed o > a W, u 4) a, 4 .0 a- CL
m rx (n j
r a, -C kft
_j ui
I -(n _j
L" D 0 C,
a r Q v c 1 N V) 0.
> 4. r= to x 1: C. to N W 0 c
'A. al a, 0) 0 3
D k" = c .. i
< c u (D Ul el X (x
o-
m ro ul
41 0 r0 .0 U. z m
OL
C Lft40 M C 0 > W kd -N 0. lp
41 q) Cj Vt Z In a m aim a m w
C, m CU 40 L. > m a M.. w (n
j = 0 0 w a
0 C'j 04, .a 0 -A m z . a)
CL co cu A, ci 0 0 ;; 0 0 a
0 u I S En .0
0 m , 2 o IT-
4. >
CL 0 a v m In T Iz,
PI 4, 0 ej a a0 .43 an 41 .0 va V, m w w .. 4
CK %. 0 m m C2 C I C;
- Cc 4. . (M_ 0
N CL. 0 w al m 1 3 w
jo L., 0, U0 U0 (n 0 w cn x
",a 0 OK rd n .0 m -i Li C: ol
a) re (U 3 (D I _j z I In
R'a 40 11 a I a w v N -4 w Cj
maz v T
0 C2 u
4ZI a -1 3 0) c fa6 E 'a C, z d, I
w 0 14 1 IV 0 .0
w w .0 Z Uv > L W 'o u w z z 10 co
V a m do W .0 Q, M (1) - Ia. . >. . 11 0 11 c
z N <1 _j r
m
a 1 ..mr ,m 2c w w ej .M..I -m a, um- 3 ,I
0 tr w D * , ED07 -1 _ . a,
U) 0) 4A a, .1 u 6
3 C w th N go Ln c uj a =1 0 0 a w .1 z L-1LU W W c
a c > 0 0 w c =I m 0 0 L > m m T
U)
4; r i' m 'n "at 0 cex
d. 91 'n w Ln vlu (D..ko V a c 0 w 0. 0:
0)
c m 0 _j 0 c 0 0 lb x Lj (n a_j
> 0, 0
- v
a ul a, - 0 v - I
- In C 0 w 0 . 'D f
Er Lu
w :3 41 U
0 al 14 M.C a A Im 0) 0
1. 1 0 1- L, w LU N a,
41 0_'o tmo 0> 15 .. > (L 'T _j oc a,
ID
-ov 0 al w a- a r 01a, (n 1. Ij z m OLUO
oz 0 cc _j m N 0 4 U Q tu I
0, U w 0 . u w 0 Z 'm 'o u co 4.
S. V, 4,- _1 41.0 11 3
CL 91 CD CL r. E C : o- 6, 0 u m li a, CMCL C ca 'a uj
C, N N 0 'm CK En 17-
0 41 u
C, v 4, :3 uj m
-0 0 ul 0 w m Id U ul c at M OL LA9: 0 M.9 C u L. m U) cu
-0 v >. w P EDca x m -P 131W
-0
V) M w
re li 0 wz 10 w 01 *:
w %. x m m . 2 o
0 1, In I
10 1, m .4, ""P 0
-C T 1)
OIW 'L f 0 Im. 1, Itl.45 4U
OD
G,- > 0 .03 a. i a 4 >
1. 2 4 4 0.. 0, 0 4. w 3 (4 a z
P.. 0 PWI a 0 .1 W U 'A 'A U. 0 41 cilo
4,a cn
x <C 0 M Z) m Ic 0 L" _j Er,
ow a c a _
at En c 0 cg 41 .0 m W w ir U) LI)or w 40
U U c 30 W
X.z u * C
z 0 CL u V > I# t, w I a, w IL z > 0 U . 0 4) cu m
0 -2 a. a.. 0 _ ;, 0 uj CL
< u
, 11 .. a ;; . p ft u 00 > 11 z 0 w
11 La L
0 '0 0 = w CK O 0, U
rU 7- 0u '1 9m CR0 4
0 C
to 0 0
D. v (U 0 1
0 1a (n 0 u w 9L
m u w 14 E 4C x X
CLC c u C.C Lq C, 5 v a) r M 4 U C L.
l No aj .0 -1 _ 0 m U, 0 0)
d? P 6 -W ti ? .. a 0 -7
tr 0 2 L
0 M , c Z (L d a G. IL x z u _lj U) c r
a,- .7 0 gw u k.) x 0>. z0 0_ u 000 0
0 , A, c 3 0 0 0 0 r.=- a,=
U) co L.)C 4, al CL0 .0 tg CL w Ix Yy u
E a La ki co ."x .0 0 IA u U
 m~ 230 t1c '0 ..- too l.50 MWLW.W w V 0 mO6z0Q0 v0 m*aO4>0.
me oo - of
t1t4fl44
Wv 14 a c 44 0.0 4. 41,v04444 0 m )U.-4r
u
06- - 0 m~L4s 0 w 06.4IA4 Z
1' >a0 -Jo m al 44SnO .41
C2 IVo-
mn' 9, Ou u0 aCA 'i's
1 0U u-nl. C
(.1 00 -t 44 wy '
44 N u- -- LA 'a ..-.L. 4i50 4, W6-
.0
-4.-.6 lE 0l442a,
U MW m 000Q.uO6t
m4-. m *0 a '-o to r
-C4, 0 4
.-t 44
x a0.-. 00 . * -. > z 444. 341 a06- C
0-P
oil 2ONc 0, 0o 0
fl0 a6L.660 X6 x4 .0
ZNN fl:). 3 4
00 0C3Z44
m0 -L U a 6.-'
tO. to00
. - $.4 m0 t44
W, . .444L0a044
t1 1"i- un %U . Mn.4. a),u. I m1 I C U '- UWOn 4-4.6 44)-aZ .43
m 0 =02 X 0 * 0 0 WE 20 WOW
0 w0 0. 4V mt.al44 W0, Ofl63
0 E W W2 U s'4 -, P- m
M 345..
a, Z -f.u
44 o o L
j .0
040. C- w 4,W 441
6- t064Q4V - W 0.C V U1'- x 6. Ws a,1 ,I.b..04.4.C. W044
0 04 j N 1 a
C O000.S 3 (-OU uxa I3 MO
44 a, Wfi S.4.aO
3 6--Ls j W a0 .0 r m4"4 > W4 >4, s.W- '2 zI 40OW
to owo V1 * - 44V4 0 <Ca, t N2 Va On. 6.0J4.44. Q-li
0r 640a4IVC c z-~- W
2 k.%m 21' 16-rU OE.N*= W u a- 4 u
flO - 02 6 01 > ru ) 0 5.
x.n.~ 0 cms
040n, '600
(ioS 4d 0. o11U44 V0 a.0 02
*. 1"- 00 .. I a,0on 0- 0
03.. 0 NO, 4.N A0 C Vta
>440 6U 1-4 V46
,fn 04)41 546.>44Mm4
mo 000) m a, r OZC-u
144f -c cl* ,Uu644
4446 .0
ls>o .- ,- 440( -Ca 0. -'u=026f m a2- c 5'6 >6004 m
'-'4444~' Vm 44z 44z04 0 0.W- . 0 I60 . Q-O W4- 40,
V 4
3-0:a L UCN (6 0n 0 a. 3.06 N 04 0 u6 t"4 r c00so,046U
.P- 4 .4 C -. 4 n InU'.V 40 nV on *- cm -L . t. 0>1.. 4-04 V 0(6
6-f > CE 0.-00 04 fl4 GO..0 00,4 4, 0a,0 064444 z4 3.(tU W0
0u4W4 04 CL
u 4.4 9.-60 40)cc 4 44 CD V2tA0
04 00 U..4W 44n(J6W
U10 aW.
Ut - 0. 440 6. 00 04
cc 00 uS W0 U> 42 r 2 1- 0

E3s
fi 4
O 96 ac
L44
0>4
xL6-a
6
0
(6
.40
6.0 A
0
m
0-VV4a
4u. CtoV 30C
a0 941U40.
..
4t-0
6
S
U
a
CNU
~ .
z
6.
Go
..

(n
U
l
Z C;
a
66
ao c 00.a) 4
.Ot.W'M'>VC
0
U36
C.4
a34..
0
zU a.. 0mvc0>-6a.
Sn.
d# N0
460-
0D L 10 U1 0 A.440 4404 U-1In 1 Z.4. Uc 0, 10 f3l atii 0444 CL2n.L"
z400 N(t cc tis 0A to4 4 .0. C4 w- O l- u1 x fa a,
... c ) r 0 *-'Ovaf'A6-
30 2AxL, L 0 oa .- I ; '" 0c -.-
a0 W.Ot 's ct 66
a L..( 66>46 ua c 0,6

-- .. 444 2 2 U4 uC) 0 n 6.
z4OU 00 4 Is 0a 0 W~ C0 0,4of 3C 4W Sn
0 0c 20 440 .0L *Io440nraI oW20 - 640 m
V4 a.t00 n.C 0 0 a-
Tma EZO x N is 0 u6N >4.0 a0. 0 N<WU7)> "A. u uU 0
N 66.4 >6. 66 6. 004 004
.0- NWT N -0 6o .0. 4,3 V a-V 4~ W26 66.. 3 .42 o Iw "O' 4 W
36.0
0-~4 - u
m-L 6 0.0 ". 0 U-' 440 Sn 4,0 . U 4 0 = 0
44c.4000
04 5 .
n- wN M42u 100 atm 10
ru nr MnOS&LAtoc a.000S 44.o vt 4* 34440Oo U
010' C
10 020
- ~ Wx
0. 06A0.
.- Goa6..
0Km
. '
mw u
V0.4
344.z
amo=~t 6.0
0L
0.-oZ ~ 41-
W
M l4S26
0
. V0
.IVD
0.U) c
t
ii. .44S
4(Qd rmO0
aS -0.
- 0644
o
=l4. or6. W 0.-44, .0 u urnS' 1 .L ' OW -643 S a 0..0 0 W . 0> CL 6 >0. m.46
1- 0.1
r6f QSN U4 11 6. C, 40 .. 6..ft l6. -. ( -w(- 07n >P 4Sn , C 44 l14 c. 0,SI&0V6.6.00a
sl3
6- x1 w- an j' U Ins a,06(6- c V NCL U . : 1 W CL" a. 046.U' to L'. C

'40 U
22 C.4 0m as M-
- 0~S.
ms 1o a
1
. -' 0. 0 L 4 C 1 4 0 0 00 0 do0
V do-4 ~
4.0 a . C~.0
MtZo 02 0. - =U u.0,o46o,,4n
C44 x4VU.. ;0r444 444,r 0
mu to ) - 0.w1 0 00 >CL .( 4 a,0-U 0 0-WOS 4=a 0.M0;0-
.46W
Z0 W0'4.
11 U4 0 c0, 1 4, Wn4-6-4-
w-404.4. uxf3 z4#a6a Uo ofCdo
44(> 1-4-b
40 .u 0 m .S (4- >6 kA.C 44 C
q4 -uo' v
Co fNw 404 0L a6 T
041 0 m06 0 2 .4420 0 S t 4 030.u06
44 4 4 41 n j a c -0 t n (n 0 4 0 4 Wor40 w40 . > 6 .-.4E al . 4 4tctm n t tt4
au z
anmt0016. 0- . V.4 VI 4 = ts.0 .4240.-4.( Wt2-U 4,
""0UO 440 444 400.fn
V a U0 0 446.02044416.
<,,0.6 0.0al4.)6. 4440 40
CO -t M0 0W w .l C... W4 I UOc(
Vw000xcm0P0.b0020023 0 Sn 000 =
-f.l :N .4 U ' 4-0 3 moumn>4a,4anmis'
6 aD C 4 -Z 4.4 02 6 a 0 0 451

A. 0 MDWc ( 6- 6-6 00m 44a

0 m ouO Ov4 OW '0
(a3 u 3 0.0.
.4~0=64 0 4 4 LJIJW tfC 66 4 0- 6 46 0 104 u C O m
N (6. 14-4Sn"60 M 11 c-' C.000 c0.4 00
a'. a nOCa -4.4. a
3.4
w
44a4 mD -
0446 W .. .USU 44 L
Ou4.4-446 u A 34Q EW 0- 004, E0. U44
.4 0. 0.C 2ODv
m.um 0 vW 2nVD'4-4) m 1 1 C 0.4, 0, 006z42 a
0,0 U C 0
03a,4 an.44
K4006
m 432c
0 4 01.0.a W 0 0V04W0.40
u ' -X4-0
u4 w4 -i Q64ZUE 0.4 . $- 00 0fl-43 Sn 0 A 6. 0.
.00S44
WLU 0r4lf4 'Cu 0i b-(f 0 min m 00 s'
036-1fl c0 00.0 - 10.-64S
OCVO, A) 0) 60 000
am 44u0.V -z u CA 6.
fta 10 a)0.440 WE_ , 0 0 0 nn. 3 '.404

006.0200- 44N 6 0 u09 400= 006.01W 6-2

(A 'a -. 4 4t
- o
tin U4
4444 140004-
22 .S0S U 040 00004n
0K o 441- 1.4 2 06 SnUm m 4 00 066.4.1 0 N 00S-..l1mn
00 6.O n f n NOW Sn a4.C 62 00 >, 36. L0040 4 . 16 ta 4
t6. 64 44 4 V2 . .00
mutU M2
0t 146
(V a,44).0
0641 006..-
0t
m m 0w 0 0 .
6-a t mm 6-0 0 On fit 4 06.
6,0 CA 2'4-.00W L. V 6.6.60010" W V0, 6- k' uw4t.
00 '0044 W 0 0 .0 44. 0 V W
0 la . WU 3 f0 CL
'00.6 V
Ma 044.4 CC34 05U o g5Q. a4 ccMt
44p0 0 k" .02 .04 u4 d. 0
CL -00 1. U 40 WW C "044. 06t6 T4 Zl I06.0wxE
66 s c
6 *-410.0m4
0 om..0-
. c.0602 a Go44 '-r to
mx 000 6u 0 016. 4 W6- 02
Co 445 . 0 V 06'
44.0044
a4 0 U ) 6.0-
LOU ,44.044
.0 a 05 4--O 00 .14W
m(55
CD0 m 5.-WI . 0 (fl>0'- V j
0..4140 n4011a VmWO0 L.t(6 ' 6- Sn 44 0. . 0
Via fl13 O 01 L
400 n 44 3 W a .46-Sn a,0 6 Cie 0p a,4 0 43M
0 0 4N0 2 LL
4 2 0l 0
a06 0W ol
6..40 66m4220 4. 004 4 21 2 W 2
4a0 0. ftu .. O 444a.44a1 0O~6 Sn 0- .5 06.5 $.
506CL U 0.044- W n uN(n 3L. 404S. .0>'m di a' a.
040 >,0 a 0
00
>0 '4 S
a, 2
n01-44
0 .
0.
N
'n I6 I
06W-1i 0
CC
u
a.
a~
~
0
a46
024
m
V
4444W
(0-v
46
80.466L.
2 6
1
xn4
11
=..
44
V
41
0 6.04
0
0.. 6
0 .4
03 %.0
0)4'4
0,LL
On m 242>.~ Vt. -r ON
!; Sn Es' 11mIa,V 21 SNt SnLo 4n .qm
0al
-.- 02 44 -'0?2 0 4,40 tO64,VUI 0 N 6-44 4,0n00
.4C 644 0 4,. m 40 0 6zU.2 x TV 6. 6. .40440 26x6 j11,3
0 6- 0 40.42
Na 4,41 =- 040 a, 00 04 0- c'S 0, X M .0 4 0
m- 20d 3 01,1
ai 0 4

04
0>0 . 6 C.0 0 02 040 (440> az u 1%u.- VW .0 00a, 0 5 ua 0
0$: WU (40 = z U c,4 03 C 36. 3 L W
Co a) r rd
.41.5s 2 0
In444a 6.4.Sn 4 0 2i 4U 0. 004 ,O. U *4 0, 'a W
4
V 4. .0 WW
>0666 a WOO 'EU 0 aL,4o4 20 V
>fM... ' z U~.0WCA0S 44
, a40.'s
6.A
0z "01a~'6.0
000t4
p =2 U
Cis -
In0 44m 'a-.
ol"- 04 I
at3
-1
44mS
0 0 V
-VP)
6 0
NO
N, 44.
.
a0 ~0-.06 .
W
44
004
Z 044
i
o . 41'uoa
.0. 22 N 4 z.ti64T
cm 'C 0 V40 > 0 N 0 c Vmm
o'
V 0
1146 0 W aC UO 0-Wa6.I.(-(0C,0-.fl 0.1x,4 ... I 0-.9a .4- 6-IV a0

'o u
04 z 11.4.4
4 w.(n0
W LU0( -uO 4m z
W0.. .
,- .U WI? 00 a 04 o , 44
> .44
6-~~.
46
Vi4.4V
4C
a4E4It-U 44
0M
ca
0
0.0
NI
Wfl
00 ~ 6-
m 14
06V. In
40C.'0
0j > a
-
Sn
4421
U
04
U Sn 44
'a C4
4*4,
W4 1.'30
4 .
s4 0
2.;4Q.-a
-rn
6
4
4
4 44 06
=
.3
NO

4404
044~~
02,n4.4.N
' 4- 0-V41 V x :
4 0
44 0
~ .0 0-0
0.0 CO4~o .4
404I~
m2- u.rs'u
6.
0.'-s In- I0 V
a,
0000
-
0.4 t
0
Sn 6.
> 30 >0 .4T 6- V 0 .964,-m 0 WC 4- IT'0 0.
'J.S4
4 ' :
C1 6 6. 40U
Sn 0 -2 2, 00 U 4 0. U 6.V 4 - -* 4 . 4 S V .4l
-06000 Sn C(L 0L'443 04 ) 0 44 V44 V
W0 >t 02 V M 0 . ,0 0
0 M40
2554)D.
n W V.WQ-S Cn
0.44 .0
~~.-n 44 0> 4640
C0
4,M4
fl.04-4
4 a,0 0 6.0o
4.-
LO
. .
Sn
0.
Um,0 I.44
.40 v. 44
&30n l-'44 C'm
a1. ,W WMm
5), -4-U0 (d M Q406.004, .. 0 04f'.
Wo L 0 0.
0 .fl.4M....
4 0 . 0 0lfl.-.l '' 0 .11 OlC 0 ..
2~~I
Er(.03n 4 0.20 0 USnC 0.E 44 aa 0 .". (U P
.40

Lj 4 s z u m zp mc 14 -1 10P c p 3 En L.'
 0.
17 41. 0, LftLl to ..0 a to a)
" . 4 0 w
CV
0: J
'A IS .4 a -0 0
41 0 6 0 ua w 5: N
0 a 2 CL ILU u
al
r
u q. L 41 U m<N
0 4 a Ul 4J - 3 k:44 X 9L 0
3 C, CA u 0 a 41 1 v < " z
-T 0 4, p cc cX Li om -P -rA,01 c"m-P is M m
CY a - In 0. U) c
w a
0
do w cn w 3 a "S
.0 c 41CL4"". 0
: a
C 0 D.,
L;
0 w
0 N 0) a
T 0 z W;g
ce 41 14 0 $.- 07- 0 a' C CL u
,4
L U", 41 vi ft app En E do- 0 m OT
0 OL0 w U) to . m-Pr
w Z W
2 >-:to aw - 1>1
41 cuso 1 3 , , s.. In
L,
C,
jC, m 0 41 (13
r
a,
0)
m
IL >
z 1 41 c 0 N (n
0 en <1 w cm
c: E
cr w
4, ci u u u v0 ; . =1
r a, Z. 0 m -j u 7. 0 p en (r
11 3 -j CK 'n 0 go.1 v
In w 4 > cl) to 0 0
a M <XX 61
u
a a 0 " a m
< T- CL
a 0
- .0 'd 6, 0ka 41 '1 LO0
cj m m 3 N- "o.
do ru At 41 CU Z C '1 0 w z
96 U) , . N 'm . (-I U!
r 0 (n i (n -j 4, 1
0,
a u u
m ; .1
r W 0 W
0 a, W 14 0 (u w X: Iq a, - u -d - 0 'd c cx -0 0
A rd do D Lu 0
C6 z (n (A
> > U4 >
IK In I W 41 0
Lo w . 3 a W C q. fj
f w w uj 1 0 O.U m -CAto 0 v 0 z -T
L, C3
at r 4 z C, 4 4, Uj L"
ce J
- 41 1 00 OC i
In E 0 do
. -T
U) - ..
a
: M 0 LL m W 0 3..
La w ,
n m a.. 1 0 0 V.?0
cn M r -W
a 0 : - = 0 3 .v X. ;l. a, 0) fn
CL r CD N C'u E 61 NOD PI .Q 20 MEW
, = -
IL 11 IX C, 'a .15 0 -Z
cu :.a ru t, a, 0..
cm.$- Q.U .0
CL a, r (n 1,
a x w W d" 0 -' 'a 04 0 3 A U. ca G) ge
U- 17
41 m >- . v V a, ; to S. r.
3 v a cu K 6, 44 a a c w 17
Z6 x Z of CL m m cz M 51 U 11, 16,ml, Ilu I I I
OD > 0 -j m .1 3.0 u (36 01
Lo IT 0 -1 -J -P .6- u x ZW
a ;,n A Q 1 ID 4 1" 0. (2 Z
cV w > uj U,
0 m w UJ
it CL" as E

a I X Q 41 V) CL 4, a .01 3 0 In 'WL
L! il z > r (Y
w IZ 0 & V E L
0 >
>, < N
Ln a 4 0 Im 4 0 u - z c-,-ru
CC cu " u 0 4" 1 0
41 0- 06.0 r
0, w U) 0 1, w 0 Lo14
- 11
4 u
Ic w 0 uj < > w ..ri
OU z a, 4. M 0 z c >- 0 a 0
0. > m - 'c
cl, 1 Q CK 0 0 CU 0 E 4 J - . ,

z Id - m N -P 0 F,0 * ul
:1 - - - z
LA
'Z" 41 Z3 u 0 L, . a La z -0
.0
j
a
> 0,
A CL C u
w 1
0)
0 2 Ld 0
00
LL,0) 0) 0 u u cc
In I- z -1 :1 11 010 -1 a, Uu
0) 110 w W
V, > u c w 0.0.
0 La
1 0 E (v 0 ,E, , .0
0- ci 0 Q OX" ru
z 41 < 000 kft mp0 0 cr w 0-
L.)tn > Q 0 uuz 0, cn tj x P, Z. 1. U 0 0 3 0 0 >
Lxu m
CL 4" a 4-0

_j
-a u al tn Q 4; 0 v L V, w Cc cc fe 0 c U, <C 0
1 2: a, (n -i 40 0
z 0 a =I 4 IJ 0 W a, 0
C'
0 0 Er zt cw 2 w a , *
z 14 m ki w cc
>,, . 0 U > C f A ffl U U
w V) -i -
Z 1 0 31 W :7 0
CA 0 L"
z w Lfl 4, n z
(n li a C Lin 0 X: z 0
0 0 w z c
> m 41 Ll m
0 Li :m <1
a, w 'c 't, a,
LL v lb w 3 N z rt u a 06
U) W
1:
E cl, 11 11 W,
441,
0 0 1 trl ou
v co 41 -a w 0
rd E tr 0 UN- En W 4- W 0 (D W kA
cr -P (11 z -1
U , Li a 4, rd a3 LL
ex
0 um a t- tn a u
41 - t- 0 r-aj o M
c x z E >
co I In 0 > X: w T
I.L - - 'a . C -0 411, IV, > m .4 r. ul .0 0 cn kA
:3 cc w 0 (0 u ..u V) "
lo j W
C 0
En W) a cio a 0:
0 w
.0 LA 0 a 4, :1 Z 0 U,
m 0,
0) c: u
0 0) 0 m
41 z > uj
u
=1 -) ;i
0 (n 0 'v 0 c , a) 1 -c . '40 0.
wo'c') 3 > Z, Lj
> Z u 0, u a
< C 0
, 4, 0 cu 'c -0 U. M
z Xc 40 LmJ Irl 01 rl . v w 0 if,
k) Lj -j 0 0'
0 0. 5
cmn W c -2 - -x
a CIO 1 0 a) m- 0 A
-
11 0 0 En D cn
cli
'a

u cr,
>
0e ej > m 0.

Ln

Iz z41
00 0
K 0
NO 0 a

41 -
a

a, 23
> 0 '0
ul 4, a

cc 0r,
z La
a 'r, 0 Z

0 0 m 'I e 1, 1, 1, .. 41 Q
.0 L, r z M -j
:3 2.
w .0 0- Ln
In 11 .. L" 0 ki > 0
Iz .1 m w L", .11 illa 10, 0 cn 2 w
W LL 0 j
-P z ri Q, > w 0 z
0 u a a ;c
a 11 L'i
uj Xc Q r n c LU
3 w 0 E
cl, w 4.1
:, $' L -
0 > (n 21 0 wu z w 6 0
V 0 40 U .0
01 1., U, I C, rj j > C3
N - 0 = rd 41, C -0
ZD > w C, u .0 3 0 IT 0 a 0 0 w 0
. rl C 3 'o 3 1 C2 IX 0 2
0.2 1"
-P a 41 kn
U.(11
fn 0 Ld 'I -
4 3 LU
= - 'D -
0 ft
-
m a; (n W -c L t A 0, z
Z 01 w a .0 01 re 1,
c- t 1 11 "1',
u ce V 0 0'0 x
li ..L. a IV do-P 0, M Im
u w
W) a,
En m - -
u m
-o n 0 -P ul m
. 41 u 0 W i m , 1= 00 0 .1 1
0 m 1
0 2,, 1
#0 0 0 fi u z - I
v , I ;a , 'n
0 x -j
CU4WU, C w .,10C, C 1 0 cli
0 0 z 0.= m 0. 0 Li 'v.) 7Z
z N Na Q 0 41 -x 2 :0
Li (1) zIT N41 a I& a 5, -
CLfn 1. _j'"U W
W w 0
.C M 0 0 0
4 CL 'p A 0.. 0 0 n V

4
 0 0 0 a~r 0 Oi
(A -1 0 '-.O40 CJ in-

0 OEOc 2iD
a~ uaom0 z ua4 a U) a.
aC-au
I~ '-
z~~.
ZL. 0 CE
a ) a 03 0 .40 c~ft-4 c X 0 04- a
-0 k g no, a 4u 0, m Wa
m ma-P -
w. w 2 a- a- tu a- z-a
-0 cl -LI
IW-.4L.
. m -C C.N 0. >40 LE
CD .c V'.04 C-'f U~ -j
'ae-oa oro
X,.o .# 044 044.0 wZ' e -N z W U
-~~~~~C i-'-.i-- -- n4i=ri "',N 0 I.. 0-
c Q.0 a.
E 0' z 4 1 .0..-al . = 04 a-a.c . c
044>Vv tn0 .3 W 0a 03E (04- a--n-c :1 11 U

.10Sx 22I4 80, C4'-0 .' 11 - 11. 1.~ 100 .4

01-.4V z..i-
<4nc-~O4*.. N4 u4 M00 M01._1'.t -. 5 0>-'4-

u44
to InJ. UN 10c 2 4- 0. W
0.i cu - o a-n

u 0 44 a -W . a a. a- -.
= a4., .a- 0O a-j - 000 0 'C
I-,>w a=-, %. mz L, 1i 0 a,4x o-
-P- CU)Z D mr~ n -j--
> 0'.5~ A.i 1 m0 .W '04 a 4V00 X.- in
D m .- 'I W 0W
a,7- P.i.4v (m
) .0 0 01-0..- *W0 O..0 *.- 1. vi.-.- ,fa
W..a 1Z.a--.

A - CA>4C0 10
M'o u -L' .- aI
:10 ao-
m 4 a
Coe N00 U.20 g a-El V 1 u 0 32 0 .WCL a , NzwU
r0 S cos a 0 .05 a,
0,i~ o w.-- N a i 1 4 a. a cQ cu
.0 0i4 0B>~ a-6 0 1
-'4 a~-P' a0.0 -> 4 4--40,, 2
.1-C0 x lb Z 0 aC 004. 4L
C.0. (.0 Nn 0
0-a-~l04 03
OWI-f Ola, I WWC4 00 Caz4
ua0- 0, aa a ,u Cl'4 044 ZZ5-
0VI: a Z 0 0 a S 0.'-ao U) i 0) 0* 45. i. L-j
m L t

-- rcrI. 'd4-W 'n I-5O OE(v-55-

co5W-C
m- OOao S.a , 0 M 0 IV

IWc4 1~I
_W N 54- .0
* a tc.- kn-io
01 .- N ' C 0 aS
.. uc N... .- a-u v a. a-44 40'MO4
m--a~-
u-a r. 1024. 4

>.W w0 a U1 u - .aI> 05 4 . N - 6 >'C 0a- '4-.

-- C> N

0 11 -a 0 0- 1E Mn-'. *1 a1.6W-'. 0. 5 10 - l 00 a.w

V In .4U
.-. 0a. C-5'44.
U3 a-iU~in 0- -E n. a-a 0W. 0
w- - 0.3
1 4.auaa0-- -
a4 a .0. W. (ni P.4 .~-a-0 0 -w.4 I --wS- 0- r 4 CE 4.0 V O iauU
a- E-Aa. 5-.4 >.'4
4- 0, r 1- -W . 0 * I-> , N 0. a C a,24(

-) Qa . c 00.5 - 0 0.40c a-.

*( w0. 0*u
ea-
'a o-U2 .i3
m- za- .-
1c
.--- ~ Ui
4 1 ,a- z, a' t
,WN'oa-
al.- ' -- .'.'-
0 4-
>,
. U
VO
a-c
0-c1-'45
UI0-.00
-I-440
.LJW
a-o-o~~~~~~~~~~
4Z..
'00
C-~0
C004N 1
.I4
W -'
T-- t:0---m1(Iic
>
0 -a
I a-~ ~. :
-i'' 7
00a-0l
af~-a2
C
iC
.
,
a-s.S-WC
-O
. 0C
---
0 02
a-nxi
4212
i
0
zl~'-
S-43
.- - C (0.- Q- ' 0" -5-s 4 Wa,-' aL-o -a - r0 tw, -a ua 0 - I
a --P ' 0 U oP0- C,2 UE
-. . d,4o s o a . 44 .wL- 5 ' l0a'5 .0 a ! -4 - 4 . T44 -
a-aC
Ca--in--'- aECU C-o,0a-aoo 3- 4C CCL l-a- oo~-ia. U CO
.~-CIna- 0- ZN- O 0 a -0-.s..
)a ui0W .1wi- ac '-
t501.iC
C I-4 :.l a I --' NZ C,0M Us-s V 0- 0 W .'. - UX
C.

C0 E a-L' a-04'- 00- z Z w0-2 !E0 , 0.

u -C0
I 15- 1
0-411
0.4-' C15 0 In :3U-liai 00m M4ul N I 04~ ia, - Sn0 4 -a
-a- a-u' - W -2o . 41f aa~ 00 <C M.~ 44. Q-ac~a--
a 4-G- 3 r

-'Ma ian Cl -j Da 4- >. Ea

M0 v~- 0 o- Jr M2-
- s 0 0.
Oa 00-'a 17 1. C0 '-'C N -540 UO - CIp V -Lg 2 UL U
Eu~~IZ
-1W--
0 144a-s-5
4-'
a-C~~~~~~~~~~ -- 04--- ~ ~ ~ ~ IZw.4
000
0--U-
k-,
o
--
a
-. 00 a-
4 N~-a.-.i
10--
,14
-4'-'-
,'-a
- I~-
4C
W-
cE44
!o coi
20 o
a, z a-1 .C It- I0 -C a 14 -- -5.)- -Ma-a
4 0-
Oia'''4-I--S .4 C
04 0-a *.~40 4) C.E m- 0 .

~ C 0a ' 0 )C
a 0 ,W " 4 al0 00W CaNi, C 0aIn
&'a C' 0 -. 0
0 4a--*0 , Q 30 L4 -- M.0i-a0-,4 CL D- -i0 41
L C CA'Wm-0-.- UT. 0 It440-0a..'~ ID q 0 4 LL 11 xa
-
44 X- U-44Q'-'wE0a-5I5p4c a-Ia a-
i-Wa NO 0, a ,U a UUTU .- .
C = >I 44 0 -i-.a N U)I ca ua- a-00WaE4a,
D2 2ia- . en 4n
m- 0 C1 .. s
d4.a- 6-, aa- a- C I, 0 aV- C 2- a
04. -. --
~ ~ ~~
a
~ 4-.~"IC--OEf.
0 W.V.~ 41a. ~CL..aLM
M-N a,
UO ME E~
0E a-
0
IiU
a,--0
~ 4 i
04 21
~ 0
~
1 -P r-

~ ~~~ ~~~~~~~~~~~l
s-B~
II~~~ 07 i N U 0
. a
401E
-VaO)
0 0Z-
')
W- u
-
0CO 1--i
0
U-
2a
-
0
* aa
.-rVa5.aW
a

Ca 'aB ~
a-- - U) ft 00 -- i-~-, 'l mo0-oW
4 -2C

N.-; W. 0- a.-a
.=4-Nu-a . va
w40 W 5 a 1 0 41' a 4 4 W w- zul a. u
a-WV~ a, 0.aaso m a- .. a-a
"1 O a. 0-U -o0 u4 10 a- o-

010V a, P $O~...u0 3a 2 '5 44 0 N 0444 a--'P5--. w- E) Ea

-~~~ 'a N)C -.- 5--4E 0 020
fu a
0. 1 04,5 01 -aia-uc n -2 N - I aa
D.-' a-a. U'M a-W u5-i '.0L. .0 w. a 4CE r 5 Via a-ON, NJ a-ia-v
S -o CL.- '-A u aId DM a.0 IaIao0 C: 5a. a,>S- Q'a0 u 0 7C1 CO-C
01 0 -- 0U oa--- El CD W-0 >
,-~a 4 mW0 -0 D -Paa0-- 5 0W 0 Ca.

.- ia-ia-l m- z U) 0 W N
w-.40-.- no . 4'a VL VMO m 00 N>~20
44-a0CI. - , BI)- .. 00 . . --
u050 acaa '4 fO4 a-i

a-C0- q ora-~C.4 a, i--w w R - ao .1 kfs W, Z V . uC i4 2W 44 a-1

w .a..i l-- =O 000 . E 2t , 044 0.'I I41- " w-a . - - C
0- ' .C -- 1 r1 i M 44-6C
0 a4 0,-'-' 0 0 -1 00 4 4 - 02 w 1 1 0 a -
N' 44440 ai--.5 0 CUS -4 COO. a cam.-.z- -- ' V C )-j a,.4 a--
o 0r-
.. 24.-025Z ~ ~ ~ ~ Ca-W--.-s I-. U' a - I 'a,
0 U(4 3 2 z.
t l-ain
01-.N-u
-- 41
W z-
30 0
U a -
M C
W
ia- --
a E
W
0
U
4)
'4
5 0 0
4 N U
0.

N ~ a- ~ ~ ~ 0- II 005
00-'-00 U0 0- a-C -- 0-0 UO4. E 0- rd0 a-
WaE4U.4L-a 1 15 4'j, N- Iuul3 4U) au S 5 N='0

a,v -)--
Zci-- - -- - - - -.--.-- -ED-- - - -
 u U) 0 z V a Ma v c
0, a wo - 4 w M I En V) 41 . YNM 10 U 45
go w N w
"o q 0, Lu 3 0 u w (n 'a
x 0 1 a 0
z 'A L"10 0 Z c in Q
a .10 0 3 N c Wx w -C
CL 3 r I M&
a
>,
= -
0
I
al a, - 0) , a, C, to 1. go c - ! 'd ci u
31

ch o Z a o u ME c 3 41 > , a 0 1 1 .01_u
LA': 1 0
ou r 'aa u 0 a o Ln CLu a 'o, L
0: aA a a, 3
L. 0 p .. a, w w (n
c 16. -P U. a al In tft (a w C, IM, 41 11 5
u cn .0 0 z M 0 ft . L.
14 'A . 41 C, 1
. In
11 1
wi o %" 4 d, 0 11 0 N. 1: L')
41
0 )c 11 0 C, M u . " .1 11 ; *
CL In I r ri a cx 9 Q a u
a 0. U w
L U) x
a 07 x M a, 41A4 a 0. n
3 ci : 0,
N a
.0 11 M '0
r, W N '0 11 1- -X w (n LU
M 0 0 z CL
z c ac (n M M w z ou 4 o
uj (1 0- S. u u 0 :1
L,
>t I IL
OZ IP U S: 1, a 0 0 1
M'O CUL I a
z z
0 c
-u 0 u _j
C U) UJ 0 0 U 0 .1 .1
a 0 ME . .
9 0
$. 0 C c p Ij
a z
> :I w rn -4 19 a 11
Ln v 9 f
41 a 0 a 6 u 4
. 0 c 0 u
.. Q ed 1 0 lb ol >
4, a, fn
v 0 4, 0 zw I CL a
0 .0 0M 0- c CL 0 C 0 0 M a 0 0 A u 'A CL 0 v
a, a 0. 4. o M W 0.0 C > w x a 0,
rM & 0 L m ff 0 U) _j cu It . .*
tM (1) aj 'a 0 1, 14 x z 0 U-P v M
I ol
01 1 Ir 3 V W 0, C d, d.
0, .L, U - W N 3.
z , -W
w 11 1, a -P a 0 a W 0 c 41
0 a P 3 uj 0. 1. > C
0 x Z c 00
M..
M CL
a 0 0 3 z w (n 0 I1 0
.P IM 4 0 w j w a Q -a a 0 a P r u N tj In
0 -Az CL
t, > w 14. . - A 9 = M3 4 0 4.
a C 4. & -i U. tj I I L'j .. I - 3 .. a cm C. 0 a', a J 0 LL 0 4- S. V E U- 3- 61
40 M C W X CUU U . CM to . 0 0 1 1- 4, 0 'n C 0, 0 a, 0
C C, C g e V " C CD 0 - a = - X - S.. u 14. V. x M tL cn co 4, 91
- > u
6 u 4 5, ki x ru Lu M r . u 'a 0 wm
C rn at v IL i 'o
X.0 .0 40 a Ix r Z2 o ; ull 4- 41 S. 10.13
0.0 u a 31 0 w 0 K A Al 4 M 4 0U r V L.
.1 cla 0 kq 0 4.. 0 w 0
41 L& (P.w OiN (I DW AM Z.0 w d a. al M a = W
r 0 11 0 4J Z
w
A- 0T 4 a to 0 E
,V 6 = M
'L "cn P N 1 111 E1 0 1 11 41 ,_ 9"-,- w V) 0 N 0 .1 11
a U L. C A4 -
CL &Z > 0, u 0 PN
C 3 ca I V Z u a In 1, La n X > u Cl 0 2 M u a 3 0 9 a
a 41.! 4 C, 0 ra P.. 4 z td 0. -W(a M
I Iul M 0 z u U. 11zm
w U) Iv al 0 3 W CL 0
ft 1.
.0 41 W w Q-A >Z w al 4 w 3 1 L u (M x
,x 0. :, :3 a 4A4- 0) M u > z ul
4 Z 04 00, 1 CL0 0 1 0, L, -
In 0 41 't,
4, 0 ra 1 . c 41
aE .,lp U - N
0 -6 ;,co -doV)
z V 0 Oi%T3 > Cd 1
41W
CL0 0
Nc E
1. M0all,
w 11 ro a. . u 06 N W . ry A 1 4 41 11 M
a 'A 0 =
z co to
as W S. a x a .,acu
0 4. 0
'A $.. '.10
a; 0, 4 ..0 0 CLQ c L .- . . , .
C.- > 0- N m . 0 .. = z - :. 0, 3'. Aj :2 a 4) .0 Ln u 0 0OL .. 04,
Gi 9: 41 am.. A 41 W X V V (D 4, La 7 0 _;,M w u M
10 .
OL3 w Ld w I
4. 41 41 U
4.41 to D
a IM11 lp do I >
-M nocz 0 zzu " u (A. A w w IS
. uowzon cu =Cc- ocN c- Z L 7, E -a" r0)11
C 0 az
Q-Ato -M
0 c
V. wx c WM 0 0 0) (M
c fa
a M C
Z) a
am u < M
M C 0 0 . 0 M > r
< , all M 0E
Ol c 0 11 0 u _ . 0)
Q w u v a 0 C6 1-1 co 4. $. u a u > Ix L, 0. N w CL v 0 (Mvc v

04.'a a a lb 6 'a cn 2 P Tu w a V ww
r 40 0 . 40 1- j w a -0
=N 11
z 0 A.M
c a v a c
IV U) x -1 4) c al tv pi u N u 3 u :) L4 La w
16 11 5 '.16
1, M 2 CU
41 0,
M 9
4, z
41 0 q 14A IN ;Go
co u 0-0 vt M 6 ol
It 0 w I a a u A .0 at MI u
Ma Is, U 10 10, In 0 a 0) u _j 0 c 0 u 2 Z a
3' a M w 0 11w a - W x N 'n 0 w 41 3 m re - li rM 3 W N 11
S. C 4, M 1 .0 3 U. w u 3 u P 0 cx 0 w 0 a A z vi am v
M CL -. o 4. .- J) w .. 0 3 a 1 a > a w
- , u 0 u 4..
.P cl .. u N 0 ca a ol a ,
w - ;
ul V.
0 "
o 1- 0 M o >, 0
a 0 0 11 10 C 9 .. M 2u 0 W=f u fd .1 a
0 -0 0 - 4.
%. P I w 1. L -'a vi -3 r 0
1' wz 41 a, U S. c M a CL
-0 w 0 91 0. " w 6 Q,t " a , " 2t Z c X
a a, 'a to >. 0: W E 5 cn r, a > 0. 3 a G. - P _j u 0 0 1 . . .. .. .
v z 4. 44Ift .0 0 U 4
;p 0 cc x C6
0 o, 1 u 0 w
41 to fd W
0 W U.U.Z
a . 4 w 41 a
u v 4,
3 0 ul CL 11 0 14 -1 w 4.Ul 11
v v w c A, u V, a (A 0.0 k" D 0) v 0 W rjl of cn z 0 E 0
a x > r. . 4) w a E a -i z - I W..= MM 0 0
CK 9 :
op % '. 1:1 W. 2 W
=I 0
,z CL co n ;V p 0.0 0 > :) a%- 11 C,
11-4 n. 0 fe LL E
rv 0 u 0 V w V a N , N , - 3 ji (D R 3 w u 0- M E a' L)
W U) x u a 0 In U 4
.0 :E
I ag M L -P
g 4 W 4- 11 P > t
(L Z 1 .. a (n 4 op 0, do 4 0 " r :)u
0 M a -
IED I.' 6
= 06Vj
a M C, 5 3 z (n cm(n M-P V) N
a 0. W 7 41 9 'o
ot IV ww 41 a 0
0 1 40 31 .. In
uj. - 0 CL a _
> z
a 0
w w -1 4 M . of 0 Z IV M so I
lb 0 -a 40 #A a Gou z '02- lE IOJ1111go
> c 0 CL M Lj Mv cm w w 0 T CL
C6 4 0 0 c 0 5. 0 z w w w -7 u c 4.. 0 011
'b a u 0 >. WW, al
r 4 COU_ I ol r14 *1 Z 1- 0) >. ds 'a gxU. Z q. E 0
9- 0 4 c 4.1 V 0 0 U (') S. A 41 p U 6 9 w 4 fn 0 0-
0 4 v 11 0 zzwo o 3. T a C Z w
.0 do- r a Ll -1 (A IV 4 Cj v 0 CL w -
L. a Qj 'we M 41 j W. nw
( , 4 42 'rr w ul. r. . >, 11 L .0 ,
, 'o = W.. V-..= w 4'
z ; >:
0 > A I I I Ux-z
c 0 0- L. as 0,v p4 u p IZ
x n Cft z x w 2 0 41 Mo- 0
w
u 0 L .. < En " u
o -4 r'w i. (n u 4
w 0 0.. u 6-0 z 41 1 . CIJ0 0 .,a .0 li 11 1* O 11 M r w 11 5 In
41 00 w C, a C do > w 1.
'L 0 4 M u , p 'L :1 (n 0M 13 Z 11 > CL0, a,
u 0> a 1 13 r cn >
O-Owo Q
a u
t r
0 a, a 41 a a 0 m ju w OLM10 .. .9 M = (n w w
M S.. 0 0 lu Z c 3 0, 0 1 Q0. 0) V 12 > . .. m
Z 'n w _j w 4., cu N c u 41 L, . to 0 x ol ; __ 0 0 u
>: o 8_4 0 u L6 z OL W 0 CD N M 1 0 0 v rn w
.. zLj z 04 v u z
r- ,J'7 a
00 - , 16 .c ..0 , c 0
a 4. @j a 4 (M Wp 0 z 2 a v :3 ko 0 = 0
0 M
c a- C,
- _j
a M is a. k" ul (0 -Xj0 .- -0 0 0 00 do to w 0,C 6 -01- - E
D 9 WL, W 'a V) U 01 W ;Moo- ww N 6 C
u me a-- is z -1 0 rl 41 0 u ol 41 1 IV 'L u
K C 1 0 C W0 M VI CA W Ul
Z 41 (L U
q L) I aq. w V M g
4 a$ P w U) 1 0 c 0 0 l 0 a 1 0 M a q. -1
L6 W r 0 Q u .0 < . K t" w rl M 41
0. 1 u a z M I u ao >. -C 11 0 uj T at CM N 0
a u 4s 'o a. - 2 ".
N VIN _ C.. - . 0 , 00 >..&.0 (D
CL M t1n, 0 a 1,0 al > (A Z
V 0 r .0 ej) cu .1 M z a
I 0 0 v a .1 Lu _j I..Io .o 1 3- r I I w r 0 M4,
jm- . I a 0 4 U) v _j 4 Ic ru r w 3 0 ro L -P .. I c 0 .1. TOz -Wi LL; of
a ip a 0 0) go M w al w C4, 1 0 1.) L. 'A 40 40U D 0
U, 4 a C CL cu lz 0) 1 C D0 0 0 -0 ru I loo1=1 1. 0 0 10
CIO a-, 1-
0 I. U U. x a C I- w Z W r, 4 x Ct aw 4 u 6 t7 > 1. 3 Q 41 C N a
4, D '0 a t I 41
> 0 4 0 a. (D fq C W r 0, w
.j L u r 0 M w .1 0 Z
 4-0 0 0. 0 V4 co >.n Oc Ic r a L 10 a a
. 4 - 0 v IMO -z 0 6 z C a is 71 W a 0 CL CL 0 C
0 M M cn w u if 2--- VV a q a (n -1 A to
c wall Z 1 0.02. I'll. W1: 00
0 C - Z E) -Zi
Of
0 .1 r u
a 4. 0& CL
CL
Ima )- a 3 x M a zw 0-- 't .12
-1 a Ix '0a0 0 v
3 cL &uo"O"
u 6 u -P a Qv u -j a 224 Mao V 0 4m _j W.C 0 to v
v 6 cm 0 a L" 'n a co 0 I cl OoI so 4 '- r .4 1,
w a a a u
I 61Q1 0,
a
o WW 0 r
a o r u q, x . It.
111.at Al $-x ok a 6 a Oj Ln It-, .
cc a w. 1- 4 0 1 -C '* 1- 1* a o.- IK 'n -M 0 . .. A cm
N, La. a ale ID cL 6 u z o
LU
o M M 8 a 1.
0 z Lq 4 4 0 at r
o - Ic - a . a 4. c o o
an zt- ra.u o m a a, 1, .4.1 1. L" In al
0 a I * I I I a 41 3 o 4 u) .1 IL 'A
6 ID L. a w c ni 4) (D 6) m 1. a Q %.
3 a at Mz v a La
` 6 6 4 c 0 I IL 4 114., 1 .a 6a, : . a
0 , I o I.- c ccn Oc o %.v Z c o -j 3 c cv Is
ca 3 1 otIc e
co- c (nL Iw 5 -P
4 o - 6 a a Q. ' :)
o ;4 4. W %..a vo o >" >. -c . 6
344*1c -n o . I . I I x- ol p a 10 o M 9 Q 21v a c N
o V C -P A a 0 M0 j3 r Io q. 6 D >o R u1 6 u
11 c) .. C 14 a u a o 4 (6 C
L z Mu a 0 o a x IL 013 v lo 41 o a A v 6 o
Id cn r a lo w * 4>1
s u
r-1, Is-
D r. a, a- > x a 'c o o o v ej
21 11 z C, ,
* * 82 c oC'u
Zz 4 x * o - . 'o o
cL a a 4 s. u 1. c a Zno to q roo u 3 a L
c, v r a w . 1 o . 6 A rj 14 0 3 4- U 4
XF46 'it A o 3 lo > ww 6 %. o & to
z 4t u 3 IL (L 31 u o I a (x, o (n o a a, M
w 'o -, 0 3 z6 V Qj
to C 'n :)
'n 6
>. * c
L c to ro
G a, u 6 -0
o oa. o $.- am a to 4W
6 z2c cc jo 4 -z z a
.1 3 o o a ff Entj o tx 6 -4 a v a . a a 3 WZ
fe Z
6 o , Q ,j o v u
ck D, cq a - 4 0 41 0 1
c 2 m a tft M al'o 't Ge CK o o o CC11
ck & C 4 4o Z w a a to Q A W M C
Al .c M u) Lo . 6S $.o CwKw
z oQj 41
(A x K
o 4
ol.
I11 o C 0 c 14 o o. u -j w
ej a a (mc C- $-
w D 3 a Z rCD
(n c w rj o a-
o w a, 6 4,
MQ3
w M
o'
CDW 10
D.-_j cu tft c, . m
w.-1" 11
u '2 u En j
u u, o 1 .1 C I-A C v= oM .. - z -x LL w In u o
I I-
o o am
w m a a . 0 w G o (d u
D M 14 co ta o S, o a M u
's = z o v .1 C -P to -j m I
u a a b (r tj u a In .o L, o a
2 z in :j (v o
Goo5 3 2 cj
Rl ou u 4 .1 3 c (4 o 3 D -4 q- A = :5 . , a 4 ul =1 11 m
6
.c 3 w 4u a
cv 2 w 11 o 41o r.
Q).. z .. 1") 2 C
4 -1 oc I m Xc w do 0 o " (A u v f :5
a 'A
> a tn K oa o r o Q a, >- .. ot x - q aw
u L z cc u a 0 C 4, a L-j j w P.- C It " a o
m Q,- 4 - - -; :3 v - I x - M a 'A K 11 v . W 4, 'o '& 'o - x
2 uj
t ,, c : . :: 11
m o.4 QI
, ! 31 ImD a c3 c fj v U CL o a t, o 3 4 - W Rl w c u
cal
-;,z N.- w,jaIT o
In a o 4 vvow zw u,3 I -o 1A a o j o xm _j Ln 4M U
rj c 41 z rN 74 9 tft
o o'. " $. o L, * cL 6, ffl cL c. m
oc -4-.1 =' Ecn cc o
Im,
'a o. 'A
lo o
"j
c
a, CK=41 "1 3 L
M o La .u al X
CL
a
m0
31 A
v.
A
a a 6 >U lo -ool
z
tv w
o
z , ac 4-
o C I' In - L, N
ow 5 c Z a Z a w ! :1 o .2 . . I : o :v 0 3 tn .1 Q N
o l- tn
In
3 n, Ll W t- t 03 ib - o 0 .0 u 0 tr ov w
cc ej
0, 3 o
z o o o o a, u o.Q o L r 3: MI a'- g 32 1 U, z a, 0 c x v 7
E Z 'X
X Zm a<
.3 c az
o m
u < -
D.- ce z ck: a D a 91 ul u Q >- al Ij c 4A j <lj)m<o o o
M .uv p x 61u m, :) 214 - o = T 13 o cx z co 0 o o) L o
Q6 4 N t, x In o Al .0 +1 > u u W N c U or V W

ks N _j r.
c
AS N Al A kAu Go M w a Al41
o 4w - A,
cu lp -
.c a o a u v 4 m w u u v3 Q o jj a o d, 'I a v:
ov u u al -1 v
Lj rj o a u &C m wzc 13, u 1. a 1 !621
owl
I -- a a ej 4 ch o o 4 Qc vi . ,u j,
z &
x, cn I o' cL A u 2 cL x: - cl vi Z o :)
0 o al r cl _j
Mz a %j'a L o C (n
x u) w
11 u n o 40 z a, u f 12
Ij
c. 91 W tj rm -V %.Vla4 > o o a Ln
L: z w co 2, >, o ej = z W 4 11 o , u . = oo. I
. w * K M 10 E 4- W gL 4 4 2 V a, z
Q a 1.- -1 z 30 1 4 o *wD cm o 0 J LJ W
m (1) c - z 'a,
uj 3m u 4. x
ED 11w
o a
"m (a 'n c .5 'o vo Bo tn X 4 o 3 4 o
1: a x m.. C M M o 4. a ul (n o to u m x ch
al x M o C -a C e W - X G. _j oo M dl o n (D U
-j o .. - u , 9 0 , U CK-J U M r 4 <E 11 Ir cu v
a o o 41 e a m M
a o
o
imoc lo a -$. cn r m a a
a, cc c A
L
'a Lo= 3 a, 4lk m
11 o o a c
'A Mo D W :7 c)
04, M0 U L w z C a M d'
0 Ul 61
0 0 t-
C
U 2 0 Z c u) 4E o w c;
46, Uo 41 4 M I
4o 6 11
C 0Q 41
& C* m
6 0 7-
E,w w
6 (D tn 6 11 o N a o 0
a Lo al u W q 0 CLI In 41 W X c .4 D, a. 0 0 41
cv- u A, 0 do a 3 w W0 CL o >- ka .. C . 4,
u ff a, 4.1o w (n L (n c) a E o o-- 61
o > o w "m . 6 w
. !u or c U, W
L.
z w
z) LL
In o c .1 As3. u
.4 o >oc Go tL Ix D= Mo a 04, a a u) N ffl -x GoC o-I z 'a 'o
o .1 w w, LL o
9 a) a u) u o 4"1r. o u
rl Z 6: o j- 1(3 w w C to 4 '1 4, :
v, ol 3 w 4"o a L ej .
tM t" C m 0 x z x a: Zo 3 n,D Z - 4 . -4CL V1 oI cc Lu
IAI& L o' v 01 . at o 'T o
r a o u o .. 3 x W = u) V o m on L w u u
(I vi lu o ki n 40
In u- n w co s. w N >. cc o.. a. o
cx 14 a , w > o u a, o 'n
6 .4 >- M u 41 a o m cLcn z 3 K a W L. 41 o o a, = W
a I.. tn alD o CL o 40 a w o M o w = u w ff 1. Im cLo A 3 41
a, I o 11 Cl-I 41 co w M > S o c 11
-j > z o 41 11fon1-1 o3 'IT
10,QIO 01 -0 tl
cj CLal W too *Oo u
u Go lz 11:
wc v n o Mxz ol w u > w -4
La 21, c W- a) ID W m m
Ok .1 . u) o c a x w . I - - w r x: cn- a 4) .. Z
r. a o Q. u w 11 ru
C z w _j v w W q) a .1 o u u 0 c z -7 cu o i u-z &" o1cc ov o 20 c P- v
- v a = '- . I c IP 11- - - u w o LAacc W 4 , oom
u CL rp.w o I ip I w ul o w
"o :3 ul 6 4, > -4. o 41, If -1 uz !' = 6 -a & -T z Al m 11o 1c Ix 6, I
c- . 2. (K tn 0 a u C. o: z L. I !, o > m (-n
2 w=ro co o w 6 o o o u 0 -. c A _j c In
(n s. o w 8 w
u IL 1 114 N In u. >
.4 a I z 1 13 n > t. o rl o Mv Ic W > m
L, -1 u at cn o r
4-- . q. aI o
u 0o cl
o Im ci M
C v I- zw (n
w 4r u 0 a
at 41
a
Is 4.
Ll
<c
u)
68 ItIj11 1. 1 cm
61 lo M ac En w
i. aa z 2E
44
4 CL 4, m c oV w Nw a GIN a. v Lu
U L3 Oil
u " C . n fn cL z v o ac
t. - 1. 1; . . o 0 4 > V o o. u. 08
M U w a. -w v C a 1
4. Oco o, :: u) m a* -o .1 %. a Omoo IoA$3
. .21
z 4.11, -a w w- " o R ca -T-- c to ., z > a rj o -c a)
o A N I- c w o.co u & t" - a - m 6 ct v) 1 if I c t o
4 K rfn ca o al to o .. o A Q,
N v w a.- a, 4 o 4, .1 c w
4A :) pe w
e-71 'Tu v'c, w
4
z
-j -A
kj 'n

w4lzo -o" a.
cl o 4 wc
TXIDK u
a a
o'll
m cx ka
cL o o
vo Q Z
to L 0
o
u M _5
a
Z o U cL -
cu -zz
4 c L. 0 C
a W
LmE
W-
4 ru
a 41
c o
In
In
X U U 0
u D

o
0 oc

Ll Z W
0 v o xww4 ID z
N C "ZKoOe22
6 o Woor L Q L a. w rl a cD
F, 6-0 W V) 0 o 6 t>n*. .2 .6
 B-19

0 C ps N o m 3
a z 0 m 0) CAu 0 m 0) 0 a "0
a 0 C z 0 3
r c a
z 0 u 4 C, a S. 3 D Az > t,
w c V 0 C , is u O.C cc 0 C
0,41 0,0 r Z
e IC z cf > ol >
M= ON X
0 u L .0 . j 0 w ri .0 a, I F. T a Lu
W.-.0 u a W, U.
Go- In
41.
> j w :,
$.
E .4c 0.0 9 - al w ci -P
> a, 0 x 9 > Co 0 0 10 a, . -3 - 1 4, , 3 3 W
U 0 it s 010- a,
o t
0
0
>
M
.. 0 : - 'I. a ro >
1- 01, w - 0. z L, w
4. -M N M 1 0 _j E 01
0 1-0 0 W le u 'o C, r, I I V 'J" z
0 3
- A 0 Z Q CL CL 'f!: En :r, 0
n
> w 0.
>
U)
>
U) al
0
> V c 14 d, C 0 V7

u a: ro >
fa U40 U)
u - , .6 0 , o " CL m 4 _j
IL, 11 11 1,
a: C, E a, w - 11 1 V c I I w Z C
> le OX 0- r- OE
0 C. 0 CAm
W w 04
0,13 a > o a -P Z
11 0 0 -j -j 0
a ol r 0.,.-.. a, a 11 m .0 "1 1 m
IV lb x
>. CL 4
4 >
a, 'o :1
p 'V
u uo -. :." 4
0 a Z Ll. 0 0
0.0 0 a I v 0 0 CL 0 6 m a z
=X0 0 u P 11 m .0 3, a g
> 00.
0 m M.P a a o I m li w
a 0 a n 0 0 a. L.
u
cc 0 40 CL. P
m
-4
0
03 a
a, 41 r u
0 %ftr a E I >,
3 4 'd 11 1
01 0Cc 0a
ol cvu c v r, few T -1 0 9-0 C,
4 lu z 0) a 0) a, > CZWL I (n Qj
.1
u =
ri n
n 1 4, z
Q 0
a, at -,)
a > .:., 2, - , 0 f 12 (1 x C13
.1 CL- r IVQ x Iv4. 4 0, L'i
k, 1. ko 314, 06 E 10 r -
t" Q cx E 41 m 0 -
0 0 0 r F ;JW=
w ui X cm E
u IV,1410> a, 0 v 0 a 0
04, L 1> I- 1" 11 Z - a w (n 4,
0 v m c a w 0 d, N w u Q 0,
=I 11 a, rl .1 WC. w
w I r N 3 0
t- 4 1& 0 0 D 4.4. a, 0 x a
60 > u r_ 0
CL-P m 0 1 tj CL 99 0%. c 'o
'L 'a-i m c r to 41 c Z D - t-
w Im 0 0 14. .1 Ir. Li
11 *1 fl (4
ol (z
'0 M-
.1
(n
cc 11 CL
L.
m a, 0
- -
.0 0
- 5 En 0 L'i w
-C CL 0D .0r 4.3 0) a a >. ri
> C v Z) 6,
W 00 ri M,- . 'i 11W, Uz
Co LJ U) 4 - E < v
w u m 3 a. z
.
a
z
z Ix a, 0 1- 11 3
Ul 0
lu M Z a-
al ed
-
IM; CL Z Q..C W V 41 010 u u W-P v 0 v co
t . 4 . u 0
$0 v f0 -j 3 rd at
c 0, c m 11 1
0 t r In L.
m 0-- 0 N' C T u A
3 4 z 0 j 94, t _5 L
9) Ir . . I MI . t 0 1 M - -6 . r, = -
0.0 -IIJ c W0 -
0, I m w ur Q
w LL 0.5 .0 Z a c a o C) 0
ri
C CKn 4 0 =.a -0
0 v CL E 1,
:) L
1-1 01, 3 a.,.
(D 41 0. 6
> 0 41 a X

C 'a C a
V cc -0 w m (i 6 N "a r w
a r. Ic,'0.
0, 10 <r Z 0 0 11 C li0
a -P 13 m 3
0 En z -p a 21 Im c IS
C, t 0 -j 0 3 . c
m to (117m -.3 1 2.1 - ct. U
J I c 0 41
't 1 11
0 W 2 11 U. im C a, V
=I" .1 1 1 E 11 4) a) o-
0 M;c - .1; MU 'n a
u M- a, In m a z 0 z m c = 0 a
0 0 n
41 3 z 3 0 M 0 rJ U
0 3 0
'a. c , = 0
U -1 0
5 M
D a v 'n
vi I CL M 0 E 0 C- 0
u 0 0 4J -P W .10
w 0 Z v 0
a ri a, Lu
c cv , QF1 13 .0
44 W-
Q LnW LD CL 4 -P V.C %1
a, a, w (1) N a E C V -x
C. w p -C ff 0 a,
:1 43 w 41 0 CD -)
z X L) a U
r - ou W.P a, 5
c 4. - * .1 " :5 ,ru
<
3 r
D CD U. >
M V
w ZO re a, 1 0 Z
lu 0 0 ri T
w
Cc 10 Li; 0 0 v
-0 0 > 10 0 0, 0
4- 1 (n U) w
1 GI r. W
0 U 0 C, 4J
t. C 4J do tj Co Z c U u En
Z E m
41 C, 0; ri 0 r 0. a
w 4. N 3 a
a i to
C, 0 m 41
u = - r vi
u al w M go K 0 0
3 c 3 C ft-
fo 0 E' - c 4. m
> m (3 a CII
m cn
or
0 m 0 -A re m
ift E E 01
0 ko W 4) 4. (n 0 41- E 1 3-. 41
:1 WII- 04- = 4,
41 r K C7 C 41 11 41 w N < Ul
C
z 0 Ll N rl a qj 0 V UlId
1 o 10-
13,
mW - D . 0 - I a m14
In 2 -C 4 w C4. r C; 0 ;;
lu .1 00
C, w E am.1 Ir C, lp u > m 0 (r
.c 1 3-0 (D a L to CL W
m c m :1 cm a uo a r ol'o z D I
a LA 0 m W C) LLI
D 14 >
'10 1 01 40 41 0
f Z-0 0 C 111 J, w
V C > u C 0
z li 0 ', 4, '" li M 0 Vk m
(W 0 U
w ;: 4 u u 11 %
u 'I
P u Im m 0 m 0.
I 1 0 ID
> W 0 w a U :1 (ou41 0 -P . r U
= 0 -0
I U. , M t, 06 p = a, 0 0 :1 C Ir. Sr Ix CL- tA cn
.1 M4 w rd c
0 - a z :5 1 0
13 ull > >. U > 2z
o 00 W o a. 41 al m w ri
L. 4ro o x li m 0 .0 a.- a
0 > & c . (a o 0..
0 6 0 q. z to c u m .. 0 0 u P 11 1 *1 41 0
0 0 0 0 . 4x
I I -a w -j ri,
0 w 11 L a, 121 .1. - 41 1=1'o = 2 0 -j 0 4, 13 0 11 .011
-P -C - 0 3 E p c .
2 .0 :3
ozw
U z
m w 0
C3> w
.c
0 C
0 -t W1 .0 a - 0, 0 U 5 N -P IM > 4 M > .0 3
S. m C, d, M >. 1
2 go c z U 11, U I I CN)-V 10,1
0 z 41 41 u
L. w 0 a
all
0 =
op 0
:I
tr 9
A 0
u
w wo di 0) a 0 ro al -P a 10
fa 0.0 o
wli
w C 4
0a m D0 F > u
a ,L a, 11 0 w 43
CL a 0 0 '7 C, 0 U
e)a.vy
00 41 >.
0 m w ci 1. E -
40 :,j

110
gr
-
01W.0 O O'o a,
0
I
M0 L. a) - Go 0
0 10 U
a, I
CA
m a, z E 3 j 0 3 3..'. z
4.0 Q Q M 0 1
w CO .0 41 a, 0
tn u cr (V Q.- 41 4 0 %. > 6 W U) 0 m Im m
C KI-
t" a moz a
W(no"Nv 0. N -j .4 z d, a -i
0 0, al W 2 U) CL. -0 4u, 10"0 U. U3-I, u
41 -J >
.0 m .'ft 4)
.0 114, 11 4J C
41 0 t- w c c
I.- > C, LA E hft-1 a n
0.3 9.0 0 a. cn
X E .1 r. 3.p 14 0 w U.
Ln Go ej 0 a 0 L -1 W .0
W cc LO ci 11 U. z a, cc w
th = 34, E .4. 0> 11
to
0$ 0 pi co 4 W
z -Ox< IL Ij 4 4 c m u i. S. - I 7. o
u 11 (; 5 go IR u CL A
-j .. 'L.
-01, r. :) .. 0
1 P0 11. w u u a
I . "j CL w 0 M
L. a a - zx Ic 0 Z 0 M -0
06 -0 L. > n r 0 .0 al 9) , - a mi... 5. C, n
0.- 0
01 0 W A 41 U I -C 4.1 3
0 1. 01 4- W w 0 X QP"
o < z 4 OA . to -Z
-1 0 0 W cl (d 4 <
LL U cm 0 c 0 V.! m -W a 11 w
w 4 ul w 4.1z -P 4. a. 0 , -P oc
u .5 00" 3.4 =Aq-
> -z' c a "W 0 w A 0 c 00 0 S. x w 0
4 (f oj 0 0)
CL j N A u CL w li CL4
CUE C 4 3 -V U I.. CL p 1.. 4.
,OV4..Pw 6
B-23

!
0 OZ V Z r
.4 u 41 1; 1
0: 1. so to 00 S -0 1. .0 u nI 40
> L3 U. 0
.0 0 D > a wa A, m 0 0 m 0
ro W a.Gj:9 0L)r0
p a 0C, a> mm 3 C
M
CL 0. 0 a In
*6140, LZO
w Q .0 .. W -j .0 a N I ou 0 0
on C, t 0
0 K O.C w 0 - 0 '1 0 , 3 w m
u a- Q, v 3 N mc - < p
< 0 r w u
- CL m a,.o .0 " A Cu
11 01 :! 11x" . -0 L I IF j 4. u m
1 1. to a
L. 4. U P a
ov, - 0 0 71t1
L,0. 0- LaM
r r w S.
t3a 0 0 u
61 0 0-0 u 0 x -T 1 4 11
1, 4
IN 0 0 a 0, 0 U. LL 3 0 4
0 u o
Gr z X 11 f-
14 vi 0 .9 A. r
w 0 Ir :2 0 r. m v
IV CK 0 A a CL
9 0. 4 0 v 4 C Z. a CLW W IV mo11
11L.qa z 1: W w .0 s a 3 Go u 0-i 0
u a .0 I C w -1 -j z .10 vt IV -
N
41 W ! Q! w w 11..11 "1
W- 11
U > a ev, L, OM I, uj -P
W .20 w 2' 1>1
0 S.
a 'W-0 =
cpif x w > D z 0 0 a a S. a) 0 w LA .. a 3 z > Qj
0 9 u a o to c w Lu .0 0 x 01
z cc 0 N 4, t, 0
3 e C' .0 0
a 5- 32 41 u IC 1 '21 .0 0
Ch .0 a .0 w w z U. 0 O-Ma, . L 0 z CD
a a C. 0 M- E oz.. zoo Q! X 4 .0 0C, w 0 A,
ox CX>. 6 Ld q. 10 KC 0 1 0
E 0 u -P u 0,
u > 0- 0 u 0 = cl w > .1 10 v - a 0 -0
Z 0 ; ; z 4 'a 6
0 x -j > CO- I
0
11> a 003 a .0 1 11 00
CL C 4.1 a * CL r 0 U0 0 a0160 00 "U -000 1.. v 0 1 u to 0
a -P -00 w 6 z Ixi a 0 1 3t " cCL 0 6 30 0 d. 0
1. In 6 r 0 0 1 0 .0 0 r
.1
COK
A- 9 40 a0 L
0 0 3 6 m a u %.
u 3 CL 6 C
0 M
1 0CU 0a. 00 -P o
K 0- - X
a
0- 0 .0 w 0 wa u .1
k", - M '. 0 1. LL 3 C3 W
t r, .. . , 6 0 IC to C, = I o- -0 U 21
IL 0 - 0 4 9 0v to -1 - 0
L. 41
0 1.. 0 a a a
3 to
6.. r
0 3 x
in 4 C ONa 0 v
0
uc
3
-j
a Lu 0, ON0 0 CK 0
0.0 04 0 C oz
to
1 K 4' q. v 0 f9 z
.10
p rl P D a 0 4*1' 1 a 0
11 1 C 0 0 x - If is M r " 0 Cu
P, ui
> 4' v 0 'C 14 r- o
W -4 11 0- . 0 '1 u - u 'L .. L. 0 No
a I- a iq 2 V.1!
0 'A CL M 0 00 3 v3 V C o.
z
ki 6 0 aa 0 J:
u C x
-j j z", &L, 4, u v
.0 & 'L 4_0 2 a'1 j11 0 0" w 0 U u 0 .9 CL
.0 0%..14,
0 U 0
4- 0 -0 tooV - 0 0 D 0
S-41 -- C, 0 r -P j k! 0 =3 m -1 u CL=
a. m 0 z - En 4. 4, E 0 u a W tn W 0 a 0 -A 4
oll W -W 0 -104- at, 1
C X: 31 to 117.11, - "I
- 41 41 .. I u Of a r 0 10
tull I- lb Us Im
41 .8 4 . M M. =.C W = 3 u C Ix
CL Cv D 41 CL 4
CL 0 X rn > 91 0 6 -j 0 0 9 U.0
0 6 0 a V a io 9 :3 SQ. > w > 6 6 0 E
- 3 m , 0, - -V m .0 a I-l" ,, a = 'I, u 0 4, C
a 0 0 C j C 0 - -- , 0 a r
0 or (A v 4 00 0- 0 1 . 6 w 14. 8
w 0 '1 & 0 cx .0 z r 0 CL.:; 0 U. 0 u
0.. 0 >. . - . C 01 q.
0 c .. w 4 00 1 .c wwu o al ". 0
r- 0 V,a - a0 00 or) u 40. w v 40 a, (n IV %.41 Cv 41 of GoCY 64, (94-2.0
-10 (a 0 .p ow go & z om 40 >I 4 0
v -P (D a
do.'20u uC . -1 a m
a w IC wL'ICk.0m
a- C V. au 0 a r P, I fG
z W,-
C Qa.,..
4 -P .00 C v - '0 ffl 4- 'A
a a0 0 CL-0
.1 U0
0
1. M 0 3 u 0.4,r. 0C , 0
0 .. .. U, 0 01 41 w 0 1 . u 41-0 0 11 3 "0
11 11 r m PCLm"W" 0 4 x I- 0 -0 4?
31 > tr 0
w v a.-,, ut - .0 w
U a z I I.. L0 3ww3
w co
z Im Imr ,a I It 2 11
C u 4 0
I ORa, Ul 0 0 & w 0 =I 0 C, 0
Nr 4, C a 40 z K u
x w C u w > 3 il K CL -1 0 gr C If, z u a
N 4. tj al W 4. M-0 W a

S: -P N Cl- r U
.0 o 1; a W x (n 0 9 4>w.
CI of 3.0 8C of
&
0) 4. Ww W Lu a 6 0 4 = 0 a
a M..1 6, 3 $.*. 0--a- - w - 'T
S. .0 *: .. $. - x !i 0 x m0
.0 1 0 0 a ttt4l m- $. 2 v
a a, 0 > W
0 C,
Z 0 -4 0 0 0 .0 a D C 1C
61 m u 0 U7
> 0 C W C a a
11 0 101, .1 W w
a u 0 > 0 0 W z 6 0 1 1 cm W
a U Id a r -P
r. = 0
60 &
1. >. 3 -P 01 0C CL 0 do 0-
0 ..1.0 u w li ..u Cl U >`0 > m 0,
T 04,
-
z C '-p &I at -CP0 1 'IV
a m u 3 IV u 4A C alula 3 U)
m 40 M- 0, -P a U, -j 2 w u cm -j LLI(0
at 0 N. 4 1 1, 3- if 10 3 4. Z Si 4;
0 a a . v a 41 to v 0 ri 0 u m
-j U7 11 :3
-j
.06 1 -j1 a0 0 0 1 %. 6
ko 0 %..5 U 1. a 11 >1 0 0 Wl
N m 0 -P -A 0 wo,0 36
14.1Iw- - zcx
r 35 In 4 X CL W
X C4 '2 1- w
0 a C = fm a C 11 3-- _j
- Ip .-j
Cc 16 *L . , * 0 0 ua CD
q. .0 3
m a 14
m . a .. m - 11 . IN o
V 1- 0 oo,-C C 61. a w 1- 3 40.02 u u K a
C,if 600 v 4 a uj a a :3
m Cl GoIV.0 w do w beM (n
41 0 . . - >. a 6 U i to.9% m 0 .10 ae w 0
0 z 0 W A u .1 0 -P a,
CL 0 S. - 0 > 0- > a >.al-m In z Ix -T x
3 01 CL. 0 - 3 41
C S cx 04., S. 'A >. C C. 34 v o 0
CL UC 6 w 0 a r. C w j
6No" a 4. to $. 0 a .1 go u 0.
0'0 0 4 60 IL 0IL C (L.r. a.. 0: -j
W LLI -0 do
in C - u o- a 0 .0 CL CL D r m w a >
x T uj
n > m CD 'm w a 0 4) v CY 40 a N 0 -P 'a a m 0
N- 6 A
0 N Aj Cw- 0 r 3 0 0 t. '0 00. 0
00 0 0 a -10, 1"
0- u r id 2 v 0 Wr C. 41 10 C',
41
a
U .1 m
r
a a, .. M 0 1 !.A P N C
u 4 IV
L. a 0 a to m 01 N 0 L C 0 -= - !: a w w j 0
u 1,.41 1 40 : 00 6 war
0 CL OZ. w 0C U 4,
to a m 0
W a CL
0 u S. au a m 0 & > 'a
Is i ww
N za
L 11 Im U. 'K N
U. C2 $- W P C6 C 6 CA
.0 a 41
1 0 C,16 3 0 0 1110
MO- m . L. IC : I-
0, , . 00 >
4.
4 0 C u 04,0 a . 03 00 >.
bl M > U 0 CL
.
- w A3 r _j 0 $. Go4, 09 u
D -I u
L 0 u
mmvz,.z
0 41 m - J
a C Go 0 134
zCU2
z U fD x
C; 3 ow CL 1 0 a 0- - 0) cp -P o- .0 L
C, - 0 '1 - " ft - T I wzm
a - o 'a a r- N if (A ON GAN & u 0 .. " -
v 5 o.. N.. , . .. . C z n rd V a k? A 0 I'M
C m 3,0 al 0 W 4
(A m go .40 a
60 4- 0- 041 6-0 .4 a99 1. 4
> 'C -to>-
x w L, C &
N u Cc u u n 4, M CL C, W. 2 -T
0 -a z w M r 0 4 0 3 w W - w
v 10 6.. .0 N L, a 0aim w z 4 a u 0
6 a ,CZ CDW w W 0 0 CL w 61m if .1..v 11
4 .. -0 1 CL- 0 W w I 1.0,
a 0 11 6 .7A
- -P a& C
IRa 0 a 6 m A w C
0Iil.OOWCIWI IV
C
- x V.C
-
L .. 0 a u 0 u w <C:1 . 11
" .. . 1
a 0. C
. 3r 4-
-j x En
U)
64 . ;.. a - M w m 0) r- , 1, - z C & u 0 0 0 0 u 0 0 u
Cv W t, .0 2 4. agw w o - 01ki C. mCuN 0 0 - 3 4- u 9 > 0 of a$ a "Ji W z o 0
c 0 0 o m cL >... a . -P to m ip a Q. " & p - 0 m - 2 W
10 1 4 2 1,, p " 7- =1 J Z a a 67. 0 a z 7
0 'L p 0 a _ 'L 0 3 0 0 4 . mm< w P,z 'Az u CLCL - > 'I '1 3,
A 0 0 53 0 0 a C3w -i U 04 0 0 .1
6 -0 C
r 0 44'.. 6 u 4 C u cr z 0 -1 O W W D CL 4 u 1- La 41 CL v ", C M
W'"' z 0n
0 41 Cx ..
AJ al CD-0 0-0
0 m 'm cc w
-r =) C; x11 CID00'n x
a
p, 'd ON 0to W.4.a a a LLIOw
oz
0. 0. 0 b, w
C2, n um 3 m N w
6 m v
-P 0 a 0 tn >,:I 0 M 10 6 d, 3 a - - V 0 L. L.
40 f = 0 1 a = . . 0 tr 0C , v 'a C I U 11 a 0) m
0 0 0 Z. (mt)=1 ae w -L
>
3 j 0 4X -0 1100 a v W
a C a 40 C 11 > 4. , a "4, 4*
41 4.1a ov
41
a 0 w -P 4. 1 ul 4 0 C U 1. r
u P CL:: u v A C .0 > of 0 W 4 0 -
v a, Z L'i m w . 0; 0 " oz,,! 6 W4 D > m 4
9 0 & u of or cc
C, 0 w
a- ON02U z W '44 u 0, u I 40's 1.11 r.
v Cv
on
u
0-
CL -
0 a
I 1 0
j u CL 2 3 -49L- 6 lb a u in z
3 6 0 L u a 0 A 0 Z0 o rcm
'r 41 " 41 .51 3 0, '.10
4CRdo
6
.0
11 6 a- 0 CLa I. w -j 0 ri 0 N a C p
C. U .%rCj z C - 0 p &.p 0 7 1- L,! X
.0 40 C C --
L aN 06 V u 41 9 0 4 i'l Ift 4- -P CK U)
VIX rw r -P NO- D. 0 L CL >. 4 ; -J D 'I M- a
0uc'Z" 0 w -0 0 0 14 11 a rd a
0 4r, 0* -0 0 0 :1
tn C. a 6 2 C.A. u a6---
v 3 v a. u En a z w 3
U. x N <U=Z 0CW
A i aL
0 0 1- 94- U 0 9 t W 0 a
C 3 M Q 0 0 a M
C 'A a 64 (j) r
4 M x u j 4w -C u 0 Ix _j 40 CL U
 B-21

0w
U~ V-
-- 010 z uI- 91 'a-9:4C4 ma ONO . 4.
1, 0 a0 -i-cZ-.-
1 o,'o.
3, E %-.'-
0.404 *Ct.0 0 0.Jvi alE5t-.~ Xat -1 0..0 0 4 -.1 0,3t
a0 cuE 5. e 05.49.0
S. a =.4
2 2G u .cO .s4pvc, m -'
u--atw m s a-. - -P4.a ta wtc ON 0c0 u-5
'm3. v a E
,a IV 04000 a 4a0w- 5. 0.
aaC .4.Ojt 1O O1 >
- -W a 50.0 tf 2 (n S .N 044.'o C c305 40 5.4. 0s-.
k a An.N-c 1t u
0.045 ,-t 0'. ut x a Ct m 3 Ct-0 'fl -al Vt
* , - -4 0, u C
.ucce-0.C- LC -E to- 5. . C>.MP as *U z
CL -- 05In a-'
0.~~~~~
0,( .
~~ a-14.04.
4. 0(.CatO
UZ
ce C.-. 0S5C44 0 V -r. I.%.C
OMalOt .0 0 =at 0 1104 tx.c

a, ON - Q0 0I 2 -tww . , o n- 5-n (0
z C E .u- t C . 4

In a >s c- c Nt- Os u- ac 0 a, 0 a j

05~~~~~~~~
O tt4 455.fEI 5 I0 - 0 w 10 0.4. 0 CV s-C
4.1~
w5
E 5.e
.a Cc- Cu E a3 t-c M 0.- 0,-1 -' SC0 ri 31. 0 Coloa0 0
n cua S ,a 1w3 0a. u! C6 ;rd t- q- - -. J a Go a Ce 5= .
b-I 0

* v-04a .343 t
4.IU~
004.S-0
~~~~~~
a
.
st
V
04-N.
..0C20> 0t4.-E
s
0
-C
. -a
.Es-a
00' 0>O
0. c
me s
0 c0.4c 2
0,
a~~~~~~~~~~ 0 - u as t.0-5-4. L..0 '-.00C C. tl O o ses
tO MC M.4-450U. 4N(0.0 as z- 5..1 z005c0 5. 1n
0 tot
5C04..-$
Nc N wI&
- aI E u.C.4t4 24 -P 4I CEQ I-.P5 4. (5A

4-1u q.-0~ 0
0 f2a C C..
0(0 u 3 A .0t U -:a 0..u.C. 3 a t c>
- u 2 ov
L 14.0e .04 e
1.0 q.v m
vs 0 V0ta1 a 0lIC 1AC-.4-m47)
CC -. 0 - rn '04 - Cs aS- 105wC
0 5c U. C CM -0 A M .5.M 04. 154.2 5-
* Co05 Lw l EC
4,ua4.~ 4.5 >0 g z--~w. C c~l'
00 4.0 L.5 0--s u 5u
ty. -CU,
W5 45.
ac I Os 5.4 t ' as 0t4c.s s ( -. .. "
C 04$.t.-S
01U0 VrLaItas-.a3 041 c -a 14s ) 0 244. C CCC C E.ll a -Z >-. n e 01 :. .
0 0-.Cc ut f4 t
Vi - g>- .0 C -. 0
*.4 44 03 -as 16. utwt.Cv Wa.-.-.ca--Oq
01 0
4.su-C t
C0O *ac
5.0
to
t3 S.- ~'4oc6t.
u
oa
- -'
.a Vc.-
alU4-3-.
a$
" -
*
t
J00
20 '0 0 N0a, -. 0t-44, I
534.10000.1.;4.-.>
-4r CC ~ C I
0
do-,I d.0c. :c-0 4. 0, c .c as
10l&n 4. aZ
2.'Ut '.
- -0t
to
ass.
Go > .0t
U Cu) -
2. 0N
$~-
1.
-
4.
u t0.as3 -4C x =
zidc -. 0t (D
~ '4 NE .. 1
-w5 u (- :=
-4cc
as 4,ra 015
sCt
-- ~u .0 0 .l 0.4.4,c .0 01a GovUM .c4- . ca. 5C 0.C..5t. r D
0j 5 CU Swu x0 00,4 -0
S. flyIM 4 X . 14 L ffvn- (0,0 ;; N. - C c-t4D. t--4-
1.- .C 44 uG
C;
>~~~~~~~~~~4 CW 4. c u C-~. aset > .t tCO -4c~s. '.~C5 ac as
WC . asZ u I-M-i I 4.-. tJ4.C
C-05 to- IC 5. .C 04 1. 4
a. C
c. -- .M
> r um0~s C04.
wf CwZ
0~(u0 - C ---OC so as 4d .J *. uECcS-.4c!
'S
4.-O ri- 0-L Wt 0 U 3,- W. 5. 3 k10 . 0, C asv-P
4. tci a0 0 UPL.)t 4 E04 031 4) .,V0> v1u CA a&O as .
OE C=- vo
C 0 . W vut*14oc 2.4. v u E-qW.4CO tdtt. = = *~jN
-cN w4 a C.4 as 0-33s C
- C)-- t. to I at E Q5 o.tc. - m
0..4'. 0 00 aaCCr) 0 E
Ct-ot c 3a Cl - 0. U
-. 0 -t ' as-- - owu - a a- m.-a WC-c C - c -
co-
o .. tiV C '4 ' M -00 -t
ru. c4Jo CU>(
V Its -1 l--a.5.a'-cS.t--v C
o L w..5. Iott ozw0 05 0 wvl'tt-- 41 5.st 4 (N c0. n 074 "3 4.. !a# -C C. t
c--z4v~ L1
M 4-.I n-.20 4 =--v ct4-n 04a, 4.4-0> t L.-0 C -C f, -4 aC0 i3ncJ-..4a 4.d
o c .0 G -. 44 Cola
I-54.-,40 -- uEC atJC~4-4-P
r.1IL4.c 0-04.1 5 e~

4.a2u .4ct0..tt a: Vv*.w0 aat '- c-4'.>.t

> 00. COO N 40
r OWUU.DE 0WIS C44

20, Ct a -- .1 4a a, C.-c a cu4

I0> w5-5An-w 4.ci $.
0(0 CC' .U K, w 0.T
U. 0(10 40 cP-w Ctl t 1 M71.
-. MO"- 0,
. >4 '.3 0 w -tu ca-MS1 -t a(04XW4 05.a
o" n< ut.. ras Dvo t w
L74ii z0t o.U 4. w-C." 5 .51 0,0 nI*0 0 cu
2 x K Nt. ol.- "I5 u--atc
K>-st.- 4) 04
0C 0 mI
Os- 4.C C 5 uc4L
w4 5* a 0.
3U-- >..-4CUL 4. g 4a U'-

CW VQ.;Nw 050 .0
ro 0, V34 C (#J
4. I" .. 4- S N.

Z-tLo 0 L05
.11'-~ '-
ecc
1
~~ ~~~~
0 4.... 0 u, ',.
us-
ctc~
-ao - -
oud1ro
C - 0 0 -- tc
0 11 I
40- 0 1 04 015s.
C 0 C- .c .0,CCSO5>4 '1 (,12.4 0 04 0.ft 40-4
1(0.-In
N >I
45-0 n1 m. wtt.
5 -P1at$- tr Ci neat w ':d- U 00
; La

*20a$ 5 01.44000 (0 W! &..

4. 5 0 ss r c. 4.
cs c- -NO a>-c & .S ow' - a.. 1. . -010 t
I-O Oe tta0Cva.54. 0 - Os o-n00c : '-(su
0-tO Cs .- .
.4Csz 0-pV40 4 c
02( -'0 Is . .. 4C (0
x
0.Mct CL
Li40-
( -.
e
t..Ls.I-c~~~~~~
tLsCt
55.
a.U 2
'
s a> It 0
Ct
x
5.
04.
Lz4-
w
xt
0,
0 .4 -
C-.
rC- ~t0E
-
k w
.tacs4 -'45000 I* V-
0 C-45S L
0-
'44,4,1

Is C v. C
a s-c4. m .. cc -- a7 cc-o Cu 0 Un
ew~~~~~~i~
-0 t SU S'
CUSt-bt l-i.E
ac MW
0sa~----4~2
104-(0 XZ - -..v1 & 4 5
Z.C IW 0 00 1 E.4M A C s0ta .-c-.
w00 0 CuU .0u A4(4 n~ 0 a5 C'nc m Z0 t 4.ad
-Z0 uc3tm . 0. a Grt
C045- Wo *.4 C 0-"as)> 4.3 (0' 0
Cs0
'4~~~ . 000-n -- NO 0. a.an.vo 5
a57 0I r--u( n -u- ye rC3a 01- 10'0'.40 (0
-- acs- M .C
a,00 Dc-s M 00* 40-t c1 4. -- 0 t 4. m
00~~ 4. LPc
rl-O uac40 c0s1 c >t 4.1 0 a ct c 4
I--e.2w CC C u OE
m0 -4 as S
A,4 C v 4~u. S40
wc.

s-s 41J cas-cm

e-s-i -a cot cz 4.-t.4 a. m40. *c - .4i tdq.
-rx - t.c4>cC 0-o'4 .-n. .4 C.00 as-c~ w.c -'- w-l .
IT0S, I,.Ns Sac ci vu CIfm A x. K,5t
-o -. cMu u0 0(a.z4 . , 01
I. 0.- 4 -i 50 5) 4 55'0 0It v v 041CC v.t .
S 2LC 0.3
4.4 452 Mt1 5 S .40 0.5 .s., C.4 C nUl-.-5O i 4,I'0 400..-.- -
(14 0,4 040- IC-t W0sv
acS- E (0Cu. Li vta 5.C .t5-44 a,4( . -, (05
00, 4 .40C U-.44-5.19c- .
IS Z=5.0E.0 (nC 0-.tt4
N5..0 'I-AI- .a. r
.i.i
30.0 20.4 0 ac 0C
4.1 I-C CC,,~~~~~~~- 1 C4 as > 0 5 5 4.C
.ft0 .a 4c . . 0 . L o-*
-- NO
-v cc 0 . 00 ea.m.. * m v' -u0 0s
ED V40a CC 0 0. t6Iu. 0 > m w - aet 4 tS.
m0 a 4 -- e-. C0
tUx Cc1..
Z. . IV 0 04 uE0 0sd-4..Q44.. t.4.L *5 ( K 0j 4. -4.4.10 5SuL (e0r
T.-0
540 N-w a 6,-c-
e>.c 4. 40( 0 40C a
>.o r .45 C C e

t No
u 50a, a-4 C- uassa. 0 0- Nd 5. 012 C,IO" ks-- 04 N
C n~a N5 0 LORi 05 Nl 1 a c-
>-cfl ua.t4- -Vts 40(0 N)

4 r-;(
 0 1*-- C(M 'U) 0* -1 ~ -4.
jU)

UaU,'. aC 3U4i Gi >- 1UC~ ' 21.~

* U~ 4' ) C N'4~~.l
L'U..-
WzT. ZL. 1. "

0, C4 41a ~>,7 0' v))* OC1

'L 1..
1~o 0. .- 'e0 0 .
4, at > 0 =0 !M00
v 01LU A C4'C0404 .-w ,>0 .( ,ria

* 4o . .4 t' z aC , 0bo0
1. a. 4.4
0 l . 1.> C O'o I wwz0r -
a. a -w tx04 .,41 a~ -1-
. q .0 a-;1 6 D.4 a v'4 .)
w c -P, C)W

O 4>'41. 0 Z40 0C "4' '0 'KU' )11(1m)1zL3 -' IC d 4' W.

-1. t07 0> u Wi Lo')4 0,4 CCC. z

* 'v. O O C
4 4, . LS'b 01 4a - C fl0- 0) U.J.
oz' a-N.
P,
C 0,4, 0' . 31" U 0 U-
>EU 0).- 0 U a,
0- c- U
,4
Q413 1WCa u 0.P-w L) . :94C
c x. Z"4.4
,.' -0 C4 :4K
OS-C).0 4 -' .40 0- 0.4 .. *.4 u.'. -T 04 u01.2 0 LS.CU4
03.- M3 'C'. I- In-* .. 0 al* 'C .

* fl'x~c *0 x a..J140.4 *-UML 0.P 0n11 Mo.-,

*~~
04'. U)o' 00 '
a04 4' W a) U Zm
) U4-1 1 . *.
>. 2 4b 01 En-'4 0~-O
v. w4 z, 0 1.. j C1. 0. 43

4- 0 x 17 C I.' Ma 1 v-COa2 -..1

104 a *.2- - ma D047

0 OV. .'1W r .. 4 t 02 > .- 50 .0,40

1 VU11 0 W 0o - ( n) I.'. I. '
L. 0C CW0 w, 1 s.o
0 1 0. 0'. >.U)..
, , o- 0 0-. oCL--
> 1- 4 6 a'( C..-l I'2: "C0aa0 *1 u.1 LN 0' 14
104 0-a, 44 >1 w) .4. N')4a
C-lOU
a or
.0..
CCU0'..-
>'a 1 00. ' ~L w-UaC .( .4.>

.01 ,44 UD '0 - ) ,0 , Z 0

4.T 4, C1)
z v. (n I. 4)
114'0' 11 t, 04
o.4 UEU) 1
-14 0.4 '06 74 N U 4

t0 N
W 4,,4 . ) 6 (.1
f0 0 . 0 0V .00U- 0 ' U) ~ 400.4

4 ,1 UQ w: '0 41. 00 W C, a 0.4o 0. 04 4 , 0 N > 1 > 4'

0 - ml1 OIP'1 fl1 r .' 0I.I. 4) 0a, 0 1 0 T0 mlla

u0 .1 u , *'-.c'3 e 31 U) > u(.c U. x010) U - X lb NE.~ v z W - )44' D
0. 4.4u 0 *. 4 44(~. 4 1.' a, 0 -.C( 41 .(.a44. 090 .1 CLW
WICL 4WM 4 4.0". 4 )II,..fl.'a..1 4,
xNW W A. ' W.

.4,-'~~~~~~ : -0 S.' 01 ,. x' >t,

ff~)

c D Iu 0.Lo0.-.

x' ;e 0 lp 1140 01.0t. 1 I.. -. 1 ~ *l E11*'OOIJ

I)c3m' a
uj 4
v.I) 1 1 ,c, VC,, 4 q., (4q Z) 0 1 40 ' - 7
n. -
0-'.040 1 .4~ ,a.0 .4-. I>U(1 a 0.'4 -. 0 7 L,. 0w 4'
1. 4,0 13,
ft1 .0 04 014 1.I.'4 ECO c C 1. U 1)
K- NU 4Q
31 a,4 2 c m.''0 0-7 r5 a.. 414) NN r 1.3 ')1.0m

~~~~~~
~~~
* ' U4 I - ) 4 4 . . J W w -: 1 E . > OW U 4 '

1.71 . . 0
4WEo 01
a72 . 0 m z 'C a, .4' : )
CWa0 0414>
w441U 0 - * 14 1. *1
(1-C .. ,) 5 4 S . Oc 4'. 2 U 1 > l , C 3.
0- 2 u 3 0
14
U4K W....U 0') 10
w -....... 0 NE
C 'j 00.4 4-0 U a
v14 -
4WU)( 4 '
3.L..M04 P 014. a %. 0 u p. 0 1. w0 w1.0 E1%.-0I &44 w 72'.).) -PIC
j 0 ' 0 o v0.. 43-jU
ar w4 lb n-04 . 4
10 2 4m60 4'U)c d
~ * ~ 0 ~
1

~~~~~~
~ ~~~~~~ 14E U 4.41 ~ 1C'. G acC 0 o 0'. 1
00 >0U Q04*V
w 4 1)
E UC M0"
4 ~
~~~ *
. U.-'
-, 0-44
1.
)i r07 1 .
0
0- 3(1(
03 4 . 4 0
004
0o
U4'..
V'U.
4 CU44
I-
.0 J wow-1
Z.
I0 -L) J
- '.
04 4
a-~7 Lm a,- 31 mC M 'TKa.

472a370- Wfaa '0C 0.e

41072*.
G1,1C
LA
000
5.5'.
1042 (b*
U U.J
.3
00 014
mu
CA
010.1.
C4
I 1. >0 U)
=
00 VpP-)
)
4U0
C) a
72.19-~.
002. 4' 1..3 ( -- 7C 0.. C54C, . " 1 .W -" y4.w4
11.0
mco 0 0 w 6074 a * 041* r40. 0- .. a 00
a0 0 .. >, ,~ (114 0 o4.N
UK'.k 1- 0 p. P-3-O 0- 1.4
ai 0..1040
a-4. 0 p. ... >-U
11 0 2 K7 004-7 S2 U)- L.0 n 0 a' 0 r. CU rC. Z

w 0O
W.4 o .0 .a a) of .4 10 U'.. S N 0 U a. I.41 40 0 X"WO.'r
> 41

-'1440 'a a 0l cIa 0 41 mA

V.4. - vCU)m. a 40l m20.' 19-.
0)c 0-0.

0, 0 004440
01.' .. ,CN wC .. ') 0-.)o4. #A3 I0 t.CL
.. &.1N N KUE
0 -,c- U.,'0 701 -j hi N N La a :1 'a 00.0W N Q 72,0-4a.
In 0 10 0v4.1 -61 op 0A 1
 > t
0 WE CL W
-C (n -j z ru L.
-7 -j - 0 0 C4 Ol
c L, z 0 (r, V 4n CU:1 < W 0
z 01 0. -j r
v W
cv A a v -j U)
W .0.0 r a %. W (n
CL (n -w m 0 .. 0CL La
-j W a%Q 0 C.1 C-
CL 6) a w ".C IT
U
>
CL o to to P 0
WD Q. .0 Vt W -i
ZWE 9 ; 0
0 T w CY U)
0 - * K CL 0j I : Z) W .0 1 cx 0 >
z 0 L m CL r 41, e . z I
.2 -w U M x 4 . w a W U N 'A
c L z .1 .7 u L - c 0 CD c CL W 4,
W U- 0 X: a m 0 k.! - - , - I
. . - -X u 1, at j v 41
n 0 , a v m0 0 toQ Lj U CLO) Z
0, w 0 0, M
x w c a 4 0 n wU
NI Ix 4, 0
0 -c . WzW .-cm
cq U,I- >
0 -0 W & En
t, ; W ;q 0 WD W
cn V to u0
> a W U & W U
(0 0 M < a) 0 U j 0) Ul w 4 E U. , - U. 0
fa - > - W 4 U v Z (n a zz
c .0 x W 0- cli 0) '00 two U N 0 W 0 3
C3 a = Q x -= >. a c Z t, 0 00 La 0 (13 40 > -V
n rd 0 -4 tm C q. m 4 CL*,Vp
a C, cu c
>. a, cr
a = CD
>, MN (n M Z 4 :N CL a ni zx 3
z z 41
L :3 a 0 z f z 0 r 'n W a U -WW
0 n M U, -0 -j z (4r,w W cm.. X 0 0 0 0 0 w 0 z 0 a .1 w
fu tn-
04 E cm cu U) c wwr
IV Ln W 0> U K
41 N IX W
w p:0 0) P, CK0 C,
CL il D
u a 9 U x mv C3cnIn fm 4 CL 310
z a 0 CK (1 41 0 c
K L; 4- , 0 X M a 0 0 w C I In OW I u
-1 110- -c U 41
U W
a n 0 r Otw q m U.0 .4 w C 4 w 10
C, %: 0
x rjCLO
>. 1w . -7 cu I-- 4 3 0 -T p U
.7 1 1W 0 11 .1 -j m-P (n
D.W x a. z m 0
tn , 0.
.
00,q CL U*, 0W -j r- -i W
-j O'm by i Lj C NI cn '0 m tn :3
U SWMI I (Inn I In W .,(.u our a,.v $. 4 0 m (4 D 'a
Lj U) kAu a qp - x a
U. U. W N a W a W c(nn1 0 4 9 OA4. a 00C
0, mr "J W 'wo 4 a
tn Q. D t 6 0 0 0. IV _.p _j 5 4 0 cn 0
9- > r U > $. cc -i > Ift
xx 0 4. iv C P- (D a, U C)
0. 0 . -TW 0 . = w- a z
0 U) U 0 -a U -M r 4 0 CL U -j I W tn 5 c
, * z L .0 Z L. c al W >.3
z , - - . . " 4, . 11 rv 0 z r W CL 0
6 'n .
0 Z
2 Ooz -T N ac
z Cm ZU
"i W - - i- - x IV 0 U) cv 4, .1 W 0 0 u a 0 z T .0 a w C3 L)
-j P, 40< K P L W U) 0
0 W -i N 3 4 14 W 0 1 x t K r IQ 4 x ix m C', > p
W w a ex 00 W.00
0 I'M U W 0 (Y. W
9L 0 & x in N 3 1 (1) 0 Q. 0 Z (L M M 3CK041
u -i > 4 V04, .0 3 cr. 4ZZOO LA
1 0 cl, CL 0
0M (Z zv U7 CLM
41 -Mc Q C, 0
N -i vZU) 6Urm-M rC, 1w C to C En a
in cn M
CL Y) -i j

CO C; N
Ln w W ,z ri W
0 L'i > g _j 0 cv -j -, 1a
cmm 0 z z cntn
1 0. .2
0 w W
'L rv -i
CL CL 0. 1 Cd
'a %
Z wt cm U.
LLJ
Z L, t 0 M
tn a a, m
-P X CA C.Q. z :>
CD
n 3 cLm m M., - W w IL
W z bl a U) 0 -j
611 0 W U) Gj
IL w a a 09 v tn
in fn CL
01
ae -P a, -XII
0. im -n W 0 0
> > 0
tL z U 3.p
. w 2 _j
N W
N 4. . Cc (n
W a U, z . 4 0 0 V) U,
V) 0 0
CL
T uW 10
0
I. X
rl,m0 -
a . z 0 c W
0 , u
cc 3 ac CD 0 1.- 0 OA4.1 Id W W W LI:
u
W a cn W z
U)Z;
ka
Ic - "
CL M - - 4 3
C) 4Z
t - zuw0: CA:W
o Q 01 z 0 0 N W x D. 1 -0 D,
0 c cn '1 4
JC
0 -. 10 4 li 4J (M 0 E,) 4 = U)
0 a 0 u
U.
- I " E
V7 WW W
W (A $-- z . Z
Im in
U.
> -0 -j
CL 10 C'm .,n V 0 N z 11 T
S.. lp
L 0 *1 U) -j
= cn
.0 LLJ Q.
w a-j >
_j 0 a a - a 4. w L4
U) m rv w cog 0 0 a0 w
a P Cd u u 0 CL P- z
to, 0 0 '0 'n 0 tL O'T ri) t, a 0) (n .
0- tj 0, 0 to
>. -P w U. -w 06 -P
9 0im co
C z 'D a
00 su - 4
ff . ou-
- 5 7 04.
K T t. w0
ry M %,
U. W v w 0 q. - t a a, cr 'm 0 (n -i cc C3-T
-j
1. K W tn S. 0 0. >. 3 In $. 11 W
m CL W a S..m P 0. -j W P- en
gx a," 0 Ulc I- w n T 0 u CL C z Rr 0 .1 cx x 2 -
GO 6 r) W
I> 0 p CK CIO rl 0 VO) onal No U U 3 (AW tn > tn a W
0 0 4 C, Ot, u x co u
6 Z A, 0 13
14,.0. 11 0 Z! 0 0 u 0 41 N rl cvu Ut 0 W00 . Ex W
,GC4 rW L, z Iml 00 tn go 0 C, m t 0 -j Wc
-M 0 W 01.1 '0 cc w W 11w
Cq ri 110 'L m C, IL 0 ae " (n C6 cc= :Cn w x mv mr
m _0
W
:3
Cc I OU) 0 4aA. 0 W '7 IV
> .1 Owlu VO) Ir C', mw_ U a
Wm zw W tn m Wz 44, W
tn w
U) >
CK
x -,n a# U,W 0 a 0 L" ,n
U)a WIL10 C, 1 0 0. .. 40 En pw -a c tn U)
z W n
% 6.10 1-31
(Ln2. I'll" M b' I- W! 0 *t W 4 Oa
tn ts L
cm
w , L,; > 0 !01 > 0-
cn U. le m I.- z v c m , -,x *-0, c J
< '0 0; w,, 0 a, 9K W 0 0 3 W u W 0 0
z
-w cl a c = 0.
C U.
U m -0
a z 'a
z C 0 ,-jQ
om cv _j 5- Z
-0 'Z '- 'a 5 . 0 m tn L CLZ-O -Ell 3- 0 CxD 0 p 116
0 Ow MK;;,, 0 41 M 1 Ir) M ,A P,
>
Iw 3V)w cm (n 11
tn . co Q 0 OW-wK 4.109 w -P -j "U. IT 41.,ncc
Oon6 m oc :-x WO I Q ;mm A, W 0 _j 4 0 U.W K a U.0 m cww
ac > 0 m a ZOLCI a, 00 cu m < T v Ow
z W0 or fv 4 u wu Z CL3 c ev 4. U.0 m
> Ncc r a- c ;t n 0 n 0 Xn
t% CL K at tn GeW P, cc N tr

loam
 ON _j
w a 0 w 6
2 P- .1
I t ! - In N w ff < 0 :4. . C,
0 1. 11)
gr w-j W U) M
t* 0 1, 10 1.- If x 41 0 1 w a- .0 1. 0 (D a V)
u 6 a ZI-I- a 0 .2 cq 0 z 0
Lft 0 r- C, 1 w u 1 1, cwz
4 r a In I I L,: 0 CL x a;w w0 0
c. 0 0 c Ix c I- -j In -P S. w ul am N 1%1
I u .. w 11 Go 04 w 01 0 a x 0 MQ w a W
vw I -i > v w 41 9. I u 00.03 u w -1 0
0 0 a - r M cf) 41 z If K ri CL
.C U..0 ri w 93 w I. - w a z N 11
a, x x 3 5 ! = Ll 0, u ru IX 0 M
a, a C 0 0 M 4 w z 0 0 N a row L,
to U. w
31 -LJL 1 40 S. z0 1-
go go 0 Go a CU)
a J 0 In 4 w
%. L aw I. "I a c S.
a IL c 13 0 0a CL .0. M -4 I,, C, z
L 1-

0 p a 0 cr, 0 N Wz 4,
W t
41 3 v C -P w
-0 In 0 CL W,g:
:1, 41 w c OM w ui al V Z lu 0 0 (4 li 11 1* i 3 W
GI 0 c w 0 0
w I. U.Mv ul w Ift
OC CL .0 0 -1
z CA 0 z w 0 0 c m < 0 -0 0 0 M g c LO Ix I I.
u r. C. LU CL , w cc M " I I - 0 N W W
W,; CL - m z 4 1. a, t, In w CL 0 U)
K > 7 LUr 0'0 a, >-0
w z w -i
&D... 4 0 M
I. , C', L, 0 11 . M. C r do -30 % E (3 W
E 0 - z w Va 1 0 1 a z c -T -1
1'. 41 cm 0 z m 4 . . cog 03 A 0 C Lu Z
a 6 a0 r,.. .. zwo z M v r c u :3 cu V, 0. 111 > CL
.3 04 I I C, i- CL
.: 0 , S. xN 01-1 w W >1 A.
Ir ol 0- U., D-9
4, 0 al 4, Z v u
3 Go cU. z2 0 2 4, c . I w
M $.
0 a q 0 = Iw W
" E
m w
z >
S. I's
0 0
III a>
I.. m
-
u
I
x
. u 0
W LLI
0
-r
LL Cl) 4 r w 0 0 (d I a x ty, L. u un w
IMz M a) a 1 41 .1 Q a z r 41 d,
z <
z a) 0 To a _j M mx at Ip =- = w
'1 , =0 m
4,0 w > M 01-0
0 0 a a 0 U (3 T 0 x 4j cr L - > z 0
M 0 Z W -W U V W 0 0 En bl U In M III z M Cl
cv u R 0 > w 'r Cd 41 cn 0 Z ru aw .. p u M: j .0 CE <
-W En > M (n III M a0 N m r Cj ru 04
w M - a N w UJ -0 uj V) U Lj - V N 0
C-Wo xc. vNn 0, ui N 0L'i 0 90 a, 2'c 0 a. z 0 M .1 M C W w
0 a% -I w ..4 u III w q
00 u > w > wu a cn N E 1. Q w 0
w z LL
a 00 1111. Z,
Z:
Cll
(2 ZI, I = 1-1 1 0w 0 1. "C
M p z
1- w U)4. M
4, .. 0Iz " w
-Z 0) A
U, I woo c -0 3 0w
W .0 Ili cn v c -i a: N uj 2 < 0, " 9 x w x 0 C w L,; 0 M A 'A ki w
4 K TO 1110 -C L3 w 6 1. -X 2. 0 u C 100 &W
c x u p
> & zw r0 N >
0 > cm I-
r, w 0 0 t U) 111 0 c 'a' M C, - 'M-0 w In 0
W . 3 W A, Q 0 Z c .0 10 r G, 01 < Ir. rm ... Ms ILI= 0 A c 0 M 1- Z
c 14 111 CLIf,=I a m M w n 13 z 111
0 r -am M x I. ,,W 1. 7 (r) 0 z
Q0N (D 1- G, 0 Z W .r 0 C)W 0)
-j & W, 0 , s 1. 0: - N :3 m L. I - 1- a III - u
u a j 0 w ac w L.1 M Lj L III U4, 'j, z CL- CNC - c 2c c
L'i A u w a (,)Cn r In 0 M C) IV In V) u 0 a, CUILO 0 M C o CU a 0 0 ww In rd
ON ri I M ca w z 03 z w z 1, a
L" LU u > - 0 0 C- M -P 4. w t 0 -P M c W .7 1 En 0 1 CIC2
0 En w V) 41 0 0 'A 0- u 6: 1 c x , L..
0 'r C.0 c u 1- ZZ 2 1
UOZX> zcw:O,- UZ4
v ru I ILI0 1 uj -a Z M C' 0 0 N
N -Z I, C Z x ("i M w K r S. u a M co M

UJ C; W ;.V u cc W 0
0 N N
w w U) ell -j
In = wwM a 0 T
(11z 0
0 w =
0 1-
'a 0 140 1-j f 'A >
X: 1- C CL a, 0
ej -1 0 ,
w CD 0 N p M -a
z 3 0 ul
LL M 61 w L'i v zww N W I.-
zc, E C, . Z 3 CL X M u
-P 0 6, 0 w w rM C V,
N in (n u u UJ 0
11 In , z
0 LL,W
ci 6 0 >
0 w
0 n- IA L" a L, In xL'i
W Li w to 0 2 co WI I.-
0; I " C . 40 u L,. X
CD 11, m w
oxx
IV OWC tr)
U 9 cc L'i 9L W
(K u L'i Wz LA III= C, 1 (1) 0 w
u U) d, :3 -j 41 1 41 z
LL C, M
al ul
z =1 w 2 M u
0 0 . a I; . 1 1- -1 "0
0 V 0 a r U. U) - I0 I. rcw >
Z LU u a T, . W M W 0
:3 0 C=,<
z 0 ul z -j
0 L N W cn Li 0
'Z uj t, r ow Ll
U. V, ozi,
rl
LU
6 Vc a " 0L,. - Q, u
r LU w w o p M0 z r, c w IL1-
C V. 14 5, 9. > 0 'i,
Q Mw I r- r 0
at Ir. C) go 0 0 IM 0 -T w Q"'
,x
0 c M 0 w W a. N
z 4. a- 11 Z LU ru
M u Z 0
W-i ru. 'MM I I
> 9 E
IZ - 0 a r
U, I 'iy oc In w ID .1 0 z 0 , 0 z
CZ0 0 a, . w 0 0 1 1 U 0
V, E s 41 fl).. a a U-Xw
a, . c) I XC cl 0to F
& a 0 0 a z w
In In ki 0 > 0 4- M 0, , Q, WO, 41 Ll: C
-jw Q 'M - 3 -9rlmo i-ix W 0 M In ON 11 -1 n E Cd =) J w M a E X:
- I.) a, a . C'U I 1- 1- 2 '.. .j rd (1) M M 'It 'In I
T.0c N r) 4* 2Lo 1 U
0. 111 > a z
0, v
M M41 0 M
=, c
1. 4) IV 3 z
uw w 00 i 0 w A -T z U,
c w M
w N W-V u -j W, .1, 1 .0 u Z) x 0 c 0 wo, III a: 1- x 0 .11
GA>- L'i
41
, w M L'j z In CV IL0 le W-1
z QW-0A,3uc Z 01,W In!-, U I-) m Z
pi 0 r 2 u' 0 > 0
:3 A
X. zw 0 w
11 z I01,W
= 0 1" MI
En m:" 4 N ' U,
=, N
. vC,
c
tn -a M 4, Ov< -J- > 0 6 w w In(17r C
12
ff w UJU) 0 IM Lft a M w lrl "C km a ww M
, w =? a- a ,: a 1. 0 w :> 0. - 1 5. 1- - I D- I .. lu
> C, al w w c M awww > w . - - C z N > Q M Li
'r C) w M x 64 w -1 3 Qj t: I > :9
4, 0, 1, 0 a 0-0 W W.- CL u w (n w 0) A 11
> In
'I Z C,'.I -P IfL, a W
M Z
0 z 5 a" 0- 1. 6C toC z z
0 0 1, 2
U- a W M - t.,
0 4,
03,
a Zuz 'I 5,
a$ -C
3 w zcn a D 0 tr
=0 04 1 41 D IV
'A . -i -j cu 2 -j
xwc 0" = N
z wu C) a
cli I. c N
0=0 11
a M 1
u 44 111 4,6- In
1 vv 111 41 -1 x -j CL 0 ix M ln 4
0 U; CIP,0 -1 0 1. . ... a a C> =x w 'MU- & . w W w -P Z -1 . . ;
I, C7 r- 0 0 0 tr .f 0 c -1 0 - w - c < 0 X C CL a I w c 0 "W C W 0 a 0 = V
c C- moN ccw awro- u 0 C v x 0 C. u C-MOC S. U 0
Le a V : wcox= I. > C, w 0 v 1 0 ui
0 Ln K ZC * x 1- rl go'. K 4 01 N u l'- r -n
 Zoo C<Z .- U, L (A
z zmcw zzm W
0. -a
r z w <C Q
'.%
C.o 4 0 %J
0 : .1
C,
Lu c-, C

w
01 C-C0 Z L) 6, cz 0
U, '0 . O
w ,
i
Wx
X W u
w < -j L:
cr _j
w 0. In x
,,Mw _j cui I m i,
Zn U, L'i x a: u
-j CL W z w Ix
Ul z v 4 z N
, !
En uj 2
ox=o
or w 0 L'i N
Wm Z
0 a 6 0 U.0:3 -i * cc
c w . m <C 0 0 CL f
z w w U. -j pl- Z
U;;.j -i I., w >
C, W-j -C *w D 0 u m
ti) C3 0 z 1
z 0
U) z
Li
U. w tl L6 0 >- Q w N 2 ww c
w C) (n
-2 0 w I WL,0 uz s 2 w
cz W 00- -
c 11 3 , , - - 0 uj K w E
> :3 Ld -j
X N mzw- W uj
CL w z C) 0 w > C3
cl) r to w C, .0 >
m 0
<01 -Xm
cr 00<
<c r) I
Lij
I t Uw
w z =1
Orx
tW> 0. Q
w WE 0.
W
u
W 0. <- U., W
0, =xw 2 c ; c
- a 0 != L Z LU m
0 Ez - 0 10 Z >
L) ago 2
W-- 0 -1 _j Lu
K L) W a w 0
0 w o w
'D zcz uj W,: x> -.0 Q 0 wz
w C 0. w
0 u z X C, r w
0 z 0 V>U
"V - 0 1 -T
Z LL w . r
t4z W
W LL U. m CS 0 W
W
0 w - w cn
>. w -rl Ln m 0 o aZ Z La U X 0
w w (9 ce ul
C 8 MWOW a 0 U, W w
0 -j - uI
0 -Jci
0
0 -ZO Z 0
0,
u 00 L 0
N x Z c L, 00 -j W
0. z 0m V) 0 En
C3 w
C, Z (j) 'T w ww
CPO
as
Aj LU z 0 Z; WO f u Ul
w Mw
uP- cli -P <
MC2::, W.; C -W,
w
'2 0
> -M u 0
- -
w ul"U, o z 00- 0 'r q
'm 4,XjLj C; 0 Z X tn
ix L
z -i
zo or , - . w w Li
0 r Ouj
z V. w L, O.T U W
w V-020 w* 0I w
=. u
W
0
U < . , w =. w 0
:) z 'wcz . C, w w I.- wow
LL 0, Ul U Z 01
P, 7 O
=Wz
aM La w U UJ
M 0 -MWO- W Lj -
W a <rr o LjX c U, Iwo wqzwi- U. It =z LLi0 _j V)
0(D: -j x W w 0 w0 Ix =)z w 0
0 z x I , 0. -j L. 0 w W z
0 c a. 4- D" x m
-= Z- d, c- ow, LL
In w
-j U) > Na z
C LU Z r
>
- f > I
to w x
.,,.z U, wo.,Z,-00-
0 V)
> z -j
c
Q
'00
ul
ZOXOM wwww
_j 0L; 0: W C i GO0 =Cwzxu= C
2 2 0(3LL.
, ,1 41 .
. a -X
w z , a.
N Z
XC"
<r z =C Z--j
wouwm-j
0 W01m
'm0 02 W
QZ< C) z-TW 0 OXWOM.00 0r
w .w 0.La0 N 4-1
Cv CLZ 0
_j '40 C
w 0 z
-MmEnn L'i 10 ! C W r=,-Xoo
w OC (17 4=040 -,j W 0
< (z w <woo c 60cr 00 ko -j
0 C) w w OLLIM 2 w zo urwrz Z z -0.o<<N oo OtL- <Woc^,Oj
C, W, (no
z x M 0 m 0 ON L, U
a tW N 'm U) a In <
w

3 z

u z w
m
>
61 w r
L a 0 a w w
u
Im-j .61 P"
C'm w Z -j z
cc j z w cc
N w x v n - .
-= cn U.
w z
c ,w =Z 0 02. w
D 0- w L, Z 0 0 z
41 :3 -L,
w--j cq 10 w
U-
u z 41 - . -
ul > 0 LL
w M N -j I
In tv x w
z 0 LIA C, cv CD x 0
CL w<
to m Z. cr w a z-
A C9 C9-j
-M 'a
u w
! Cw wZ >0 CL 4, w
4 41 ui uj
4. 2"
zC ci " w Z -
C, N Lmc cn 41 lll: 10 wx
<C U-
w
; w - w
I a, w U! Z; W 0
.IF In a L,
w v >x La 3e z>
%.,
w w uj I > UJ
z
w 0
41 a,
v c) w 9K M 0 -d 7:-
-j 0" z R-0 u
c 0 w -Z
r a, (n -x ..u C,W,Wz -3 2,0 zCS Ll
a0,4 to cx
Lu (n
w v cl W x 0'
0 > L) r 4- r
0 a, fd
X W,:; -i
-; z 0
A o do . - j CA
td 4 m cc w
n 0 d w Ic U
wom 41 w
w x CDzs. 4 1 0
(n CK
-j E
00
ru U,L) -j li 0 X 0w wz 0 (n
w r- v
cv- L v x IK
D
j CL z u) N I Ix ki
0,0
1 Im L)0
4, z oczcKr w z 0 r w
ri =>a -V-.. " - w
C.PGo fj
'0 4
CA do 41 .- J Zww o-
z W cn o- v.1 a w CKCC M>z
.r L" 0 aa-, w w -O z -VODW- =e
In m
Ou
Cu -rl
w Ul_j co z U u - -j CL
m0 w u u
cc w C, 2 W'a x
woz
j
C; <
(a
2 01
a
MOE
-
K
g-
o 40 U
C a
6 31 > 0 .W-j" z zc
U 1 N 0 0 v 0 Z v vczu1 0 *f Coo' >
2
0, r
0
W fa S.
CPI
u
U,
v
0, 0W a > 0 0 w0 NWW
w
z
z
0, al
mt 4111 . 0 W
0D -P - 0
.14 0, m a. c W. <:;,
DO wx
a " ..- u 0
bi W L ul
w w 0 W2,=- z c w2u
'i
fri & 'n W,-.-w (Acv a MC z
on 0 CL. 1. m u r =.P 0 912 , 4, --
6,
M-
cl %. 0 > wo m
:<w > o- Q (n _j w 0 0
0 4, W (0 r CL cn w LL < LL w w
w Nw
-T c0 A, X Z
-0 U w W"I.- -Z
wj z .. c .. zo
w
3 0 it U, z 0 -
-ox - =0 m
ry a N M
.. -A wz wuu
w r bcc Cu
2 CU'f0i (n 06, lw 0- w w
,vw t c mw
W u v 'n 'm
9. ra -P 3 , a U)
41 m x cu U, UJ w Cj x w V) 1 4 4
.0 .. w - 0 M 0 z t 10 --, aowo
w "0 a m.cn
,c 0
L'i -T (n a x 0 a U,
czz
5. cx
CL
x f MC

0 CD p -j z z o'n Om X0 vur w 4 WX4

d w 0 w F, Q It ac K (D P, xr
 -0 In C, 11 > -Z
> ui w z C 0
z C)
C, U. z tr,
v can-
a
w En w0 w 0 w U. W M w
C X:.j Ix z L.
w z
Ww u 4 0t) 0 cr, U- >
uzw z z zw x 0 w to
L. 0 C u) Z
o W ; . !; 2 Co. 0
U. X U.W (L
LL' W Cror w 0 w uJ z > Z LL,C
w x v w Lu ,, C3
w w 42
.
L. r
X
w
MWO
_j to
r T. U,
CL
W UJI _j w x -
X: Ul 0: x z z
U. 0: < I LL, a I _ CK w ! z 0 11 xx
C: : w w !; 40 0 C)
0 U, 11 0 cc z L, w _j w tj
L'i
.- I *I u W Q
z
0
i
UJ 0 w 0 ww UJI
Z W
0-
z I
C, < 0 x z r, >
_j > w w . 0 (AIx
0: W =0j Z Z LJ = 1 0-X Lj
Uziw 2 W.
U.CL W a. 0 20 M U.z *w w W -
4" W 0 xx zzz
w z0 -Z u rx W.1 'm U < 0 0 L
0 Ldw
(f) 0, w W
w w Uj
ul a _0i
Z)0ul
_j _jLU w x Ui eraZn 0 - -
CO Q 9L D x z0 j Ul L, U L,
w k.)P-.ex 0 > U.C xwLJ 2 M r U. U,i z
0 En Q Wki C
- x wMa) U,x 0
LM Lj
cc" -0: WZ
> C w Ul
Z:)2 Cv -1M z 0- z z a' w O_
0 x .x w zC <
w w >
w u x x Ix 0 4T w u.
> 0: 4 w wo
2 U. 0 C Li M u r< r 0
11 w UJI
ujz
L'i z
w _j wzx
v
ww
Z 0
w Mto cc C N Q,
NN0 1 0. 00a T w C) Ce
r LjtJ _1 W
_j U,U) a 010 fu (n v)
> 2 ; W n L) CL x
Lj= Ul .1 _j z
W L)a >,m Z: to w W U. 3
wr- UC Z>X LLX w
Z _j u U'l
w La imU) 00) w 00 u 0 w w at U. a
z Gi, 0 0 (1) at w _j C4 q > w <
0, UUw 4x V_ coo u Lu 01 u; a tu x -1 w 40 _w C i
ft w 0 z
wzw 0C >5 -4 Z> q,v 0)
w U,
r
z
U,< a V
-Ko=w zMC 0Z :3 N _j_j w
La0W W CW
w u 0 z z w < -i
z
0 a 0 -1 Cw 0 Ci a Z W - =1
L, 0 ww =0 Mw1: 4. X
0C Li 2.- 0 = X 0", C, zz w 1)qT
Ck U. N
J T M o > U_ I w
0 0 r > C _j
0,Www M W P-- D LuL'iCK
Z r- LZI Iu L" a
M w- o =2
w Uj 5 w C or wm2wu 0 j 0 z 0 W W C I
CL '0 . !a M wu CKx (P so W-M z cc
x
r W-U,
s. x
0- Z 000a- Z 0 00 wMa
2 11 - wo w0 _ja - > _x
In .0 CL
T W w 0, =0 U. IT C
2' Za P_1wz P>- w
4 > = on .0 'L 1
0z 0 In I- _j w
_j > 0 W.- 1_ -NW
x
z w IL
L, Zn
Cj w
x Ow i
< 2 to ; LCl Mae "<r 40 .0 w jW C_(n
(n - .- o
z w w - Cacm .--M
Ii 0 > C z u z in
w(n cwujjmzz vI w (1) 0.
C z 0 M CKw Ol v LL U) z Z OZ
4, x a) w 3 u C 0
0-
-zz
ZZ
ui 0 jWUm. _= 2, r
4,r Q
zx U) 41 -C) 0w X.:tL rX -JZ M
(L_n & at
a. w w " < .cOWlw- 5- Z
322 .0 000= d 00 0 0OWI-2,WK: 0
1. w
0 z - <
wo :.I a
0 En.
WW
01. < w > N
VC00awl0wa
11-MU
U, Vic ="=W Qw P- ONO .1 - W 0 0 Qzzwj0Cw
z a T 4 -i 8=ULIW.ZC .rm Lp 0 C W:,.wu. M 0 uxw
0 CZ L. 0: m2o<rcro zLj =(nw =. w x w 0:
0, 1 r
'm 00 x C ruwa=0<8 Cw
. tj w 0W N w
ec-raw Q -1 Z 0 W 0 cowu
Mxx_)wCwccr z x 0 CL w C,3 0 -0
Ix
W x
COUJO CK
0Ul M
.0 Xw >UJ WXLJO
<C X0 M UIL,=
_j C w
U) u. X (n
0 _j L, U. L" 9L 0
Ix z

W z >
: 0 x C49 > x Z -am
x0zo eg Z
x 0 W
-i 0 w U,
(1) W w L'i W U. Z<U. C 4z<U, U)z 44
Cw z z W W0 CL <1 v x 0 CL w
-i _j .4 , _j w z v z.J 1.- 0 _j W- -
C CL =
CL 0 Ul
0 0
= ZWC%
W
0 uj 0
IL u -,
u: U.I w
0 _j U. z
w uww xf
_,> I.-W
Ww f2 Ul 4
I-
x U) .x
V) z 11
N
kx w
73
U. Ll W w IL L,
=V I =) . w ; w w Cul'i ;
0 0- 0 _j ONX0
5 W C
WwMw
w x .W ":
cc W CK
tj 0 'S
a, _j C - k.. .4 li 0 O:C 2 D, CK
LJIL Li LL0 0
0
z U. WUJI lid EnM z WW w
" a 1: .13 4
E ex ED CZ3 u v 0M .10
x 4 W L. V W r_
A. to ury I Lu
_j z ui U. 0 .0 w W
Mw ul 00M w n W 0 w a U; U. 0 cc 0 z U.
z -1 j -1 0 rl UL 0 U,Z2 OL A 01nm
TN
wx
u
W2 a
C, L. w0 N a CA W In M -1 W
N.410 w_!w W0 X 7 Cl)C') Z (1)
UJI
a: I w w
zofn . a Z to xw- w an jo > W L'i W 15 lo a 0ED z
W. 4 a v 0z 0 z z w 0 XN. rx a zwo c (n
M 0 (A 4 w Zx a V) Lu Uj U." x ED - " 0. 10 _ 11 z
rz* = V = w to z I x I C (a W w-
0 w 0 w Z
< 01 I
zv v W_jW _ aU.0W Za I w W au. to
_j w W Z
ww wZ , v -i Z VDX
X 0 r x u am u
w _j K 0 z < w wo 0
UZ > 1. C
v wIn 0
vz Z4 w4 1
L'I a C C LLw
00" W_ 0 0 =3 Cr
-Mx w x 3 0 0 tic or wa w
Q Ce CL
u ui Z w W ZW _j r X a', > < W > x
W j z z T 0
_j z jz w > tx k.A0 M w Co
CCK 0 w vomw Ul w
'i CDZ W>Clu w z
z Z. In 0 =I 0 w ULM 2! u
OL w CD_ 61 0 u 00, I .0 z
W >* 10 L, NEW W 4 C 2 In U.
wo >
1 6 w ; 41
zC, U)0- a w q
to : U,
z w C)
:.Q W>ww D, TUJz a rC kj x u 0- z>r Z r u 0
. 0 D w
Vw ul 02
U)Ij W I W 0 Ui v r) :;W=CN 0 0 wl
ov 0 tr =a- Z ,,u 0 0 w * V 03
w Ul w MXN 0 0 a z V) U) w O.Lz M wj
0 a, a
.4 A, w 0 Ll w MCuUj :! " _j wz>
.- z a N z w
0,>O NQW>.m w ZN> CL Wm w 0 t.) w
0 u Ce M 11 U. N W I x X'T 2vwx 8z to wx. MNW= 0
w 10 a x
L, P-
VW _j A IC w < z CC 0
CE UCJ U rN_ I . 0 wc d
X Z w =1 2 v Lu
1, Cj 0 4Z 0, C, ,uj'm
0 Z N CLCK 0 2 0. 0=0
W Lu w In * = T Z , I w" 0 - , a- Lj w w u0
x w W Lr w
0, w 'I LL)
Z Wt x _j IC XN =1 gx n _i -a
-M I -< , ; j _.; - W -i W wC 0
W U.
;020N 0 w 1-0 C6 01
> 4) wm - > 0
0 M 0 z- 0 a oui - v W.= z
Q. W-JA 0
M:.Om w 1 zw CL 40 w so
6 X, w
C4- L, CD
S. * u a to 0 :Cax
> z Gor = 0 > w <1 > 0 w 00.. 0 wwz
- 01 _ W L,
.0Z W j4o Ui 1 0.-"-' u
x 2 z _j , 0't 0 1 W
4 , P_
(n
Wo
WZO CK u L, C 1- E8 w I w wt0 C _
4Z Im ; Go w cc
ZZ> W,0; 0-jw C>M z ouli
v U.0 z;;Z:C;CLu 0 w *z. w W C6
U7 v
ezu Z = r. W 4 x
Z _j 0 L "I z
a Z > ao 7 Z0 (MM
0= ;a 0 L; V IL
z Ui w w a CCLz . 0 j CL
IL = 0W
w& 11014 0:11nu 1.0 0 N _JE;wcc"Wk 1 U7 C.0'L
_j
6C. M W0 W-4 In z , a q" (1) uj Q
0 > .0 W _j W 99 P, x Q u t, < 6 La U C', W" 0W4
0, W)2 C, C0:,CO UJIX C w 0= 0 x :1 CL UJ
Clic
Ora- _=)0OW,.,
0 0
061 IL 0
u0w <, W00UJ>.T u x -:3 <X0QLwCw-V. v COZWOC7vm vc < c z in a -c
C, 2w
rx 0 w . . . 0.
0-u CL C6
W >w
ug
XO cc
w Ew WOOM a. 61
 W 2 W2
z w w w ZNZ w wzu (n 0
0 Li W
w u in 0- N C) U) 2
z z w m U co
cc (0 w C; U) z -j cn 0
X 0 cc X W=W
-TLi W cn 0 L > z X 1 :3 LLI
ocrw-
0 w
11- w r Ix LU
11Q
moc 120 e Xto 0 CL. wk,W
001%
to 0 w -T
z 'n X
-a w >
a
z n n
0
=I
oww :iw01 z =
OD (D
gL :3
En mcCD
- 0. ac
X " ; cz - . -j U) C, U)
! 4 x U) to
-j w w CL
" C x cc w LU . Lj M w maz
2 Cc,w In X Ln (n CA W W ww or X 0 uj LOL, Lu Z
(n w 1 -3 z
01 -1 0 11 w= to - - z 'M 0 cr W W
(n L'i -j w 0 2 7 wZ CL Z LU
Z -jj U) Z; 34
I'- w M C 0 U z uj www 7 tj
0 r
x < . X LLJ C)
z u Ld > 0 W L. CL tL w
0 w C) (L x c z
k., 2 . I ZW" x 9L
U, ! 4 m
Mx 0 w CL
w LL X
0 z
r.a w w tr * z =Zm x
a z CL z LL -D 0 ad CK w -2 O=W= z
w u w w ru x
UJ u Er NN WU (n 11 x < 4z -
ki w -1 0 La 11LL
x in w > > U, CL C3 w CL u q w z
0 CoW V) (n 0, z z a L,)
z W 0; - w wW
w to c U) w r w CO
c 'm a z C k, W
v 0 w LZ c 1- 0 0 C'
0 v - Z." M M zww w w *:
". = uj W m W.-C
(n w -
U) L6 C
r4 X ki w ccl w
w ru 0;! rj 0 'T 0 Ld
w CK
w -i w
0
N w -j CCK z 0 z c
w
0, La C9
z 0 W 2: - D - li I.-
wz 0 Li
n rl -0" MO u w It X cmX- c!) cz u X -1 to coww (o
U W m
w 0 -1 m or w U) z < w W(L M zZ m C3
W gL U.
z rn W w N w c
to
41 m m > cc
Z<zw (I w X * w -J"W- (n
> x CD
IJU 2 X0 -4
c 0.
0
V
0, v W Ld CK -0 0 LL
j 0 L) 0 0 w 2: X 'n 0 to !4
(1) > - w8F0 1
-x 0- w
r Ow =Wwo
Xcp. >0110 - i w X IL 0 " w
4 (n r: z
D - Lq v w z 0 -j u
a m cN N 4 U. t 4 -,
z w 4 c z w >OW 1 0 11 w 0
0 x
a C) 41, :3 w = 4
w 0 w -COUI=: Z .. (n
n 0w
U),O
La - 0 x
0 U) u 2 w (L -0 InLU cn -i y w ft2 X L, in cn q w
a LL) 0 aw w - =3 m a Z r W< 8.
w 2 2 0 _j - L . $. Lr-lj c WIZ 4 1 u
0. _j -j 00 =3 z "<rw i Er
: vi gLxc
0 - > _).,o
> LIO- c Wc > -Irw lx > _irz 0
v w Ir a < z 0 0 7 C, En UJI4 cn
0 = w w w -j 0 U)m wm
ac C "-j m cr
< 0. W > 2 (D
OZUll 0
c
0 X L, L, < z u Cz,;
Go w Ow> lix z
IT :; i w =n :3 9L -x -Ow- x C)
-. z 0
In
gz 1 0
2WL-P) 0 0 z (n EW 00 -3 -0 -C CD . r z
OD 0
z
0w 0 D w C,
NL-) mmum-CO C uj e)- -
aco ICL . U) I N -4 m 0 z -j w
X: 10c Mw
a 0 ce r En0 a 0. w 0 w P- w w
QLC
w
z w m zo cm z 0- 0
r, U) gLc z 0 C (n Co
P 0
L, M M c 0 0 0 2. 0 z v
U2 .1 0 z w <1 k c-
0 zaoEoeo,
at a CL
u

U) Li 0 c w L
ul w 0
Li 0 ; w 0.
-C
L,; En U) P.-L, w
U w in & La In
c T m X < -1 z
> < w uj
(5 w Ll C r U)

x 0 c U"
c w L w Li
Q z .4 0 z w L,
X u ki
Lo U; 0 w a. z z
WO,
z Ml
IZ -x 0 0' (1 z m < Ix to
Z w r LL c 0 U0 w L,
w
U U) w f L, w
-j L 0:
v w X: x c 7- C',
z 0 m ul 0 Cl 'a Iz (n (n
'n , W k, - L "X -
0 Li
C4 X U, e 0 U:I , X -1 3
LLJ w m
7 3 z in m -j v
C, E')U 09 < r
w -ZLIU) > >
w -j X: (1: i L2 < z W U.,z 2; IV
z c N
17 -= 0
C, v En
to in .40
a -i , a) w x cc u U
z w to - z 4 w to 0 T
m w k, w * =:! :; a w -x 11 . r
uc, 11 7 .. z C (x
LL) r cn w
> rl -r 1, Z ' (L
x z 0. z w
X cz C 11 1 w N uj - C (11
-C :3 LL c ;z C)
Z.: X ;W'
w - a
Ix go L U, u Z ki
w V) w Ix w 'r La Z (f)
z w w 0 CL r It :3 Ci I a, I-
C, -1 "1 V- r
z m u Z, 0 w
0 w & -
,o : % rl
W (r.
w
or Ic x X G %.CDZ - L, I
-j in (m 0
x 0 co L" Ll . P- L" W C x w -0. CL
6j x Z w .nc r 9L w vi ZE to
x .3 w w in c r- z 1 (11) =, K Law
VI ; &C 0 LL:
U.- , 0. 41 -=<
0
AtLI,
w z 0
A. 01 In z 40
Ir
cn
'n Z
'N 4- (I 0
-7
,, r. 7. !a >
Q- Le) V, = z
rL! L
X Ll m Z - 0 09 w z ag
X
07
w 1!!: z
La'
=
L) L 0 L17 w :, - U
L; 0- . M
< W 0) M
L 0 D 17
7
Co U)
cr,I - w C, 61
om cl to w U"-lj 1-1cn 0 C,
U. -=EW W
> W L.1 w
7-
IL ac IL
aw N I I I- > -M
> -j CL
=C 6 -t Z CL0 M U7 CK W -Ul LIQL-J -CKON
> 'T z
0- U. -c ZN
-Z>r-=CxW=v'
C-0 m v w -r
ma LL n z w 4 z =;: z
OP. v c X to .,Can r. I:
Z Kz a Q M 01 :;z
z 1, *W= w C
r 'n z

v or, .0
wo 0,32",
%. w -C r c; En 0 .0-1
w cc - 3. :,ON 0 L
j 4 w :OT OX0 Wuuj
X-C n 0 - a Z : 0w
cc c -i EZ& rr
0. -Fwz 0 Q-o I i :,,
X<=fL33 wa;z 7 0 L, Z C, T r. z L, u < Z"'dZI c OX
N 0 z 7 = I - v , cc
z IL X I w cl z 0- X %.;u
 ZOC- W. '.4 -; o Z' a

W22,- L.0u10
LA oL.Jr .40 0
xzIJ 01
a of OX
W 1 2.7 N
ff .. 1,;
(n.v.
w~f
=.
-
0fU2
0. c2C
0
_CD
_
0.
0(2
_OUl ~ 2>'-,
a.202
wf x-a w4 >4 0>-

0'~~ b'' 2 00 .W 1-'

cl~ U) 0W4 W un U
0C 1, NC- 401. 02V) m
U (2 XUWI x-
19 a- j L.1M 0 Z
fL.f OW .LjWCA

xI' w-*.af 22<W41- fN v

In0 0N4 a- -
40.) r z.

WO.
LAWL NUW z W c-U 22
S1-0 01 .. 0 I .W 0 P'.0 2
z z rC zCS S . . W En V)'2

W1W.f W)-cC>4 m~ aN.I

wl~
Uf. ' . ;K Z - U 1.1>
4209Lf cc P. =111 *
V 0..> x W Z O~1
(U 0
I- M 40. 0.WC 4-mx >0
W.x2. i
_j 'a _j. 0 CN CL
47 0) S. I 4 InN W .4Z< It
-1..- F.U
3n%. c2 a En 1-0w '.A 210 Ow -

CK600zm 4 w uZmO.
0 L)~ I 0 m .'(CW 4.4
Z CZW
4 , z Go m cc
xfQ~ 4.a1x 11 0.
C 0a P U0 Z Z
1-4 1-4 -

.0 Wx'' WXCD,-, '.11'

W MI1 L' Cl C) 1UU
.0 (DU w. z OW.- -. U~f 1-1J1
O. I 0.
c.1- 0'W vW -(n x X1- .0 "- (lUO CC
D1- an w <n = 41.. .. (1- W I '.. .
03 -= ff2 W.. w11.. C n. 2 i W2

C7 VZ (A'-4 L , E -1 N In z! 2.0 22 01
rU =--- It W11 r... 0'~- ..o'(L4 WO 0(1 2
U> 0CC.
C uf > CV4 CL U %I-.vI- LC 0'I."
N 2-

02-Z :3
U ~44 ~ ff4-UU
.4 u
= _ >ZWO
0 0 ..
1)(In N~
aLJ1C
22.-C
W
W.
C0 .0
~ .
,U U WO
0. -0. f 111- UC A 2

-j
(A - CL W W 0z1 0 4 U- -4 * .4
N,-. X 0 4
w.4 ZUffCA
x 0 w CC)
X- mw -O
0 ~
O,0-. az, 4 0) 4
0103
',
4
0W
.. 0
U

U~(2 0 In 0a1 C-Cw 0 Ll '' ~m

W4a 424W W>- 0 02' >W 2 . w
1-40 0 II '.1' C.
- 2wCC I.0 W 3 c lz aoL C *0 U
.4-0 f.n.
0 (n0

0'.. 0n O 0 2 f 102 2 - - On
0 01

01- 0
L 2.4 01 P.~ .. -@2 zN 0 00 ,Wo =,U -W, 2 >W.
z0. 1- C-, 4. r l 0j OU Uz O '. - 014
x 42
01 WI U II'

4' n WW 2 0UC
of ) W.A 4' 0> O W4 - 01-
n 2 * . I
1 01-x- CL (40 W (') C-C In- 02 201 I, ILI
Cu. In 0 'U .

In UZW In .q4-C . -a ' 0''f 00 01 I -- M.12 co

.4z .. 0

1ol
'. -.
W04 > . 44 qe '-U cc -j w .00I.
. -(1 I..4 x0
0.4' < m
N (1) U1- '.W-U 21. 4N It 0 .4 2W'- 'W 0 1 .WnoI..
0.. WU- En' 0-I2 4 u
S'. 0 U) ZW 4 ZW 44) x1 -U . Z0 v - C

UI.
InU a-~l . N (D ~ I. a4W ('.,
I. U) WN-..
0 z' -a.
U
ZI1 U .0
W
CO
Wn-O
>'C
. W00U04 .O. -.fW1O
n0WW0
W LL- C0 -I 20 UC 0 --C .o.C Wnn 2 '0W 0)2 - x''n- W - WI
a22
a.''In IL. I- - . 0. -22 .) I4.- 0- Ujz U ES 0. 0 2 10z . 24 0

v2 Z 0-C u of M0 co - x XnA .; 0a -. In221 InI In. x). a

In - 0 ^-!:N co OO . 4 2 4C 0.'. 0. , .1O.

.'UW.~ffM 0 N2A -CC) 0 0.4 1)W W X00C;), 0.4' &::.U

N4220 q.ff41
N C0-. 4Z.4W -J;
.C4U01 4A-C1 0. q
-jI.-'f 4D IZnU . C30 2' C.
41-
f00 020 N
WI~~~t U.U
0002 m ~ f U)0..,2 C00U
4z =00 X.
w4. 2
'4 0A' 2
U 0. 0 I,'2
Q
00 a S. 50-I 2U) 'C 40 1~a 0)- a04t-

z N.1 rw
0-4Oc -j CL CI w U) n t w u: 0.
M-C~~~~~~~*
o 0 WI D a o xzz1 c(

a u >W a dXW 4 x .w
0 z N Z L
z coo 1Wc~lU, a2 a f) u z z -
 REPORT DOCUMENTATION PAGE
1. Recipient's Reference 2. Originator's Reference 3. Further Reference 4. Security Classification
of Document
AGARD-LS-136 ISBN 92-835-0360-0 UNCLASSIFIED

5. Originator Advisory Group for Aerospace Research and Development

North Atlantic Treaty Organization
7 rue Ancelle, 92200 Neuilly sur Seine, France
6. Title
RAMIET AND RAMROCKET PROPULSION SYSTEMS FOR MISSILES J
7. Presented ata Lecture Series under the sponsorship of the Propulsion and 'inergetics Panel and the Con-
sultant and Exchange Programme of AGARD on 5-6 September 1984 in Monterey, USA, on 10-11
September 1984 in London, UK, and on 13-14 September in Neubiberg (near Munich), Germany
8. Author(s)/Editor(s) 9. Date
Various October 1984

10. Author's/Editor's Address 11. Pages

Various 246

12. Distribution Statement This document is distributed in accordance with AGARD

policies and regulations, which are outlined on the
Outside Back Covers of all AGARD publications.
13. Keywords/Descriptors

Ramjet engines Missile propulsion

14. Abstract
This Lecture Series provided an introduction to modem ramjet technology; and applications to
missiles were especially emphasized. The survey and characterization of various types of ramjets
were followed by the discussion of ramjet components. Three of the lectures given on the second
day dealt with the main types of subsonic combusdon ramjets. There was one lecture devoted to
supersonic combustion ramjets. The experiernce gained from the research and developmer.t of
existing systems and components was covered in detail in all the lectures. '. '

The AGARD Lecture $&rieswas sponsored by the AGARD Propulsion and Energetics Panel and
implemented by the Consultant and Exchange Programme.

0,-

.-- ,/
.4..7-.
y . .
 0.0 !0

< <"

0 C6~
0 <0 0"
0~~~~O , '~u~~ -.0 .8

'p ,- -- > -- &

Cc C11'0 <~ 0 Cc CA

0 p00 45) 0 0 B

>0 0>

4) 0. (u~
99L Lv544r.

00

V~~~~< __ _ _ _ __ _ _ _ _ -

S 0.
U.~
E 0..ooo o

UJ I
-11- 16 P. A >
 I.

41

.k.4.

< <B
00.0.
S.

,-.. *.1
0.S
0. 0*
-.

,,....- "

c4)

U, e
-a I
0j
A

,. .... .O
,. a. - -_ s.oo -
4.
.. , : - *, .

.w.9
0 .

E-o