7 Zelen, M., "Multi-Factor Experiments for Evaluating 16 Acheson, M. A.

, "The Whole Is Not the Sum of Its P a r t s , "

Reliability," Nov. 1957 (to be published in the open literature). Fourth National Symposium on Reliability and Quality Control
8 Epstein, B., and Sobel, M., "Life Testing," Journal of the in Electronics, Jan. 6-8, 1958.
American Statistical Association, 1953, pp. 486-502. 17 Geckler, R. D., "The Principles of Developing Solid-
9 Sobel, M., "Statistical Techniques for Reducing the Ex- Propellant Rockets," Aerojet-General T M 220, June 1953 (un-
periment Time in Reliability Studies," Bell System Technical classified excerpts from reference; main document classified).
Journal, vol. X X X V , no. 1, 1956, pp. 179-202. 18 Del Priore, F . R., and Day, B. B., "The Engineer and Sta-
10 Parsons, J. H., Wong, K. L., and Yeiser, A. S., "Statistics tistician Can Meet," NAVORD Report 4028, 1953.
of Electronic System Failures," I R E Convention Record, Part 19 Box, G. E. P., "Evolutionary Operation," Proceedings of
6,1955, p p . 69-73. the Symposium on Design of Industrial Experiments, Institute
11 Gunn, W. A., "The Reliability of Complex Systems," of Statistics, University of North Carolina, Nov. 5-9, 1956.
Western Regional Conference Proceedings ASQC, Aug. 20-21,
1956. Additional References
12 Lipow, M., "A Unified Model for Catastrophic Failure,"
Western Regional Conference Proceedings ASQC, Sept. 9-10, 20 Breakwell, J. V., "Economically Optimum Acceptance
1957. Tests," Journal of the American Statistical Association, vol. 51,
13 Molina, E. C , "Poisson's Exponential Binomial Limit," June 1956, p. 243.
Van Nostrand, New York, 1942. 21 Howard, W. J., "A Simple Failure Model for Complex
14 Weiss, G. H., "On the Theory of Replacement of Machin- Mechanisms," The Rand Corp., Report RM-1058 March 1953.
ery with Random Failure Time," Ballistic Research Laboratories, 22 Lawrence, H. R., and Amster, W. H., "Reliability Achieve-
Aberdeen Proving Ground, Report 982, March 1956. ment and Demonstration in a Development Program," Space
15 Meltzer, S. A., "Statistical Analysis of Equipment Reli- Technology Laboratories, Ramo-Wooldridge Corp. (to be pub-
ability," Radio Corporation of America E M 4194, June 1955. lished).
Downloaded by OLD DOMINION UNIVERSITY on July 10, 2014 | http://arc.aiaa.org | DOI: 10.2514/8.7324

Exhaust Nozzle Contour for Optimum Thrust

G. V. R. RAO1
M a r q u a r d t Aircraft C o . , V a n N u y s , Calif.

A m e t h o d for d e s i g n i n g t h e wall c o n t o u r of a n e x h a u s t Introduction

n o z z l e t o yield o p t i m u m t h r u s t is e s t a b l i s h e d . T h e n o z z l e
l e n g t h , a m b i e n t p r e s s u r e a n d flow c o n d i t i o n s i n t h e i m -
m e d i a t e vicinity of t h e t h r o a t a p p e a r as g o v e r n i n g c o n d i - T H E diverging portion of an exhaust nozzle is an i m p o r t a n t
feature for all engines which depend u p o n t h e t h r u s t
produced b y t h e exhaust gases. M a x i m u m possible t h r u s t
t i o n s u n d e r w h i c h t h e t h r u s t o n t h e n o z z l e is m a x i m i z e d .
I s e n t r o p i c flow is a s s u m e d a n d t h e v a r i a t i o n a l i n t e g r a l of on a nozzle can be obtained by complete expansion of t h e
t h i s m a x i m i z i n g p r o b l e m is f o r m u l a t e d b y c o n s i d e r i n g a exhaust gases t o t h e ambient pressure through a nozzle
s u i t a b l y c h o s e n c o n t r o l s u r f a c e . T h e s o l u t i o n of t h e designed t o give a parallel uniform jet a t t h e exit. One m a y
v a r i a t i o n a l p r o b l e m yields c e r t a i n flow p r o p e r t i e s o n t h e a p p l y t h e method suggested b y Foelsch ( l ) 2 for t h e design of
c o n t r o l s u r f a c e , a n d t h e n o z z l e c o n t o u r is c o n s t r u c t e d b y such nozzles. F o r jet engines operating a t high altitudes a n d
t h e m e t h o d of c h a r a c t e r i s t i c s t o give t h i s flow. A n e x a m - especially for rocket motors, one is required to design nozzles
ple is c a r r i e d o u t a n d t y p i c a l n o z z l e c o n t o u r s a r e given. for very low ambient pressures. E v e n t h e shortest nozzle d e -
signed b y t h e aforementioned method would be excessively
long a n d heavy. Logically, one would seek a nozzle of limited
Nomenclature
length, since length is a fair indication of nozzle weight. T h e
A = cross-sectional area of nozzle problem t h e n is t h e choice of a nozzle having a specified length
Fi, fi, Gi = various functions defined in the text, with i = 1, 2, 3 a n d yielding m a x i m u m t h r u s t . Semi-empirical investigations
h = Lagrangian multiplier, which is a function of y of this problem were carried out b y Dillaway (2) a n d Fraser
L — length of the nozzle a n d Rowe (3). A mathematically rigorous formulation a n d
M = Mach number some numerical examples are due to Guderley a n d H a n t s c h
V = local pressure (4). Their principal idea is t h e introduction of a characteris-
pa = ambient pressure tic surface as control surface for t h e m o m e n t u m , t h e mass
W = flow velocity (scalar)
flow a n d t h e length of t h e nozzle. B y this choice, t h e partial
x = coordinate in the axial direction
y = coordinate in the radial direction differential equations governing t h e gas flow reduce to one
a. = Mach angle ordinary differential equation, a n d a one-dimensional varia-
7 = ratio of specific heats tional problem is obtained.
8 = variation T h e present p a p e r proceeds in a similar fashion b u t does n o t
9 = angle between flow direction and nozzle axis specify in advance t h a t t h e control surface is a characteristic
X2, X3 = Lagrangian multiplier constants surface. T h e form of t h e control surface a n d t h e velocity dis-
P = density tribution along it are determined in such a m a n n e r t h a t t h e
d> = angle between control surface and nozzle axis t h r u s t assumes a maximum, while t h e mass flow h a s a con-
Subscripts s t a n t value. Obviously, this formulation fails t o include a n
c — chamber conditions expression for t h e flow differential equations, a n d t h u s one
C, E, F = values taken at respective points might be in d o u b t if t h e velocity distribution along t h e control
M, 6 = denote partial differentiation surface t h u s obtained can occur within a n actual flow. H o w -
t — conditions at throat ever, one finds t h a t t h e control surface becomes automatically
Presented at the ARS Semi-Annual Meeting, San Francisco, a characteristic surface. F o r this reason, t h e characteristic
Calif., June 10-13, 1957. conditions need n o t be included in t h e present formulation.
1
Research Scientist. Now with Rocketdyne, Canoga Park,
2
Calif. Numbers in parentheses indicate References at end of paper.

JUNE 1958 377

 This results in a great reduction of t h e computational work. T h e origin of t h e coordinate system lies a t t h e t h r o a t sec-
Moreover, the approach shown here m a y possess a m a t h e - tion, t h e #-axis coincides with t h e nozzle axis, a n d y represents
matical interest of its own. t h e radial distance from t h e nozzle axis. T o construct t h e
flow field a n u m b e r of points between T a n d Bf are chosen
Initial Expansion in the Nozzle a n d t h e values of x, y and 6 for t h e given contour are deter-
mined a t these points. Using these initial conditions along
Let ATBE, as shown in Fig. 1(a), represent t h e intersection TT' and TBB', the method of characteristics (6) can be a p -
of the nozzle contour with t h e meridional plane. Contour AT plied to construct a characteristics n e t a n d evaluate flow p r o p -
is the contraction u p s t r e a m of t h e t h r o a t and TBE is t h e erties a t t h e n e t points. Such a n e t of characteristics is sche-
diverging portion of t h e nozzle. T h e initial expansion occurs matically shown in Fig. 1(a) a n d is denoted as t h e " k e r n e l "
along TB a n d t h e wall contour B to E t u r n s t h e flow back to a since t h e variations in t h e nozzle shape between B and E d o
direction nearer t h e axial. Guderley and H a n t s c h considered not alter t h e flow properties in the region u p s t r e a m of right
this initial expansion to occur t h r o u g h a sharp corner. Since characteristic t h r o u g h B. Location of point B on t h e p r e -
it is advisable to avoid sharp corners in exhaust nozzle con- scribed contour is implied in t h e determination of t h e last
tours, one can prescribe a suitable contour TBB' in t h e t h r o a t right characteristic u p to which t h e " k e r n e l " of Fig. 1(a) is t o
region. However, t h e location of point B along this pre- be utilized in t h e construction of t h e nozzle shape.
scribed curve is left open in considering various nozzle shapes.
T h e location of point B, in fact, is a p a r t of t h e solution of t h e F o r m u l a t i o n of t h e P r o b l e m
problem. After t h e point B is determined t h e contour BB'
does not effect t h e construction of t h e o p t i m u m nozzle con- F o r computing t h r u s t on t h e nozzle and mass flow t h r o u g h
tour. t h e nozzle, let us consider a control surface passing t h r o u g h
Downloaded by OLD DOMINION UNIVERSITY on July 10, 2014 | http://arc.aiaa.org | DOI: 10.2514/8.7324

Sauer (5) gives a m e t h o d of analyzing transonic flow in t h e t h e exit of t h e nozzle. I n Fig. 1(a), let CE describe the in-
t h r o a t region in terms of t h e radius of curvature of t h e nozzle tersection of t h e control surface with t h e meridional plane. L e t
wall a t the t h r o a t . Using this method, a line TT1 (Fig. 1(a)) </>, a function of y, denote t h e inclination of line CE to t h e
can be defined along which t h e M a c h n u m b e r is constant. nozzle axis. T h e location of t h e point C on t h e axis and t h e
T h e flow directions a t various locations along t h e line can be function <j>(y) would t h e n completely define t h e control sur-
computed. I n t h e few examples carried o u t b y t h e author, t h e face. Along CE consider a n elemental length ds (Fig. 1(b))
M a c h n u m b e r along TT' was larger t h a n u n i t y and no a t a distance y from the nozzle axis. T h e elemental a r e a
difficulty was encountered in applying the method of charac- generated b y rotation a b o u t the axis is dA = 2iryds. Also,
teristics to determine t h e flow downstream of t h e line T T ' . ds = dy/sin <j>.
Let p, W and 6 denote respectively t h e density, velocity and
flow direction considered uniform over t h e element ds. T h e
mass flow crossing t h e elemental area is given b y
sin (cf> — 6)
pW 2irydy
sin <j>

and t h e m o m e n t u m flux in t h e x direction

p y 2 sin ( , , - < ? ) cos fl27^

s m cf>

B y integrating along CE one obtains t h e mass flow crossing

t h e control surface
sin (<£
mass flow • 2wydy. [1]
sin <f>
T' G
Similarly, t h r u s t on t h e nozzle can be obtained b y integrating
Fig. 1 (a) Characteristics net and control surface pressure differential and m o m e n t u m flux across CE

sin (<j> — 6) cos 6

LINE CDE
thrust
/;[*- Pa) + PW'<
sin <f>
2ivydy

•[2]

I n t h e present problem t h e conditions a t t h e inlet to t h e nozzle

are assumed to be given and hence maximizing the above ex-
pression is sufficient.
T h e axial distance between C and E is given b y

cot <>
/ dy
XE XQ
-So
Hence t h e length of t h e diverging portion of t h e nozzle is
fE
length = xc + I cot <f> [3]

Varying t h e nozzle contour would involve corresponding

variations in t h e control surface. One can leave point C fixed
and v a r y </> to obtain t h e variations in t h e control surface.
T h e location of the point C depends u p o n t h e length chosen
for t h e nozzle. P o i n t C can be treated as fixed in the present
Fig. 1 (b) Flow across an element of control surface problem, since t h e variations of nozzle contour are subject

378 JET PROPULSION

 to constant length. Hence t h e following condition m u s t be along DE, a n d
satisfied
/1 + X 2 / 2 + X3/3 = 0 &tE [10J

i: cot cf> dy = const. •[4] Since fSM and/30 are zero, one obtains from E q u a t i o n s [7, 8]

jfi.af/20 = /10/2M
Continuity of mass flow requires t h a t t h e mass flow as given
by E q u a t i o n [1 ] m u s t be equal to mass flow t h r o u g h t h e t h r o a t I t should be noted t h a t y drops out of t h e above equation,
section, which is invariant with changes in t h e nozzle con- leading to
tour. Hence it is required to maximize t h r u s t on t h e nozzle
subjected to t h e restrictions given b y E q u a t i o n s [1, 4 ] . Using cj) = S + a along DE. [11]
t h e Lagrangian multiplier m e t h o d this problem can be re-
duced to maximizing t h e integral This above relation shows t h a t t h e control surface coincides
with t h e last left characteristic in t h e nozzle flow, and t h e
conditions along this line are obtained b y introducing this
/ = f* (fi + X2/2 + \*f*)dy • [5]
relation into E q u a t i o n s [8, 9 ] . Hence
where W cos (d - a)
= -X 2 [12]
sin ((f) — 0) cos 6
[(P - Pa) + pW*
s i n <j> and
sin ((f) — 6) ypW2 sin 2 0 tan a — — X3 • [13]
Downloaded by OLD DOMINION UNIVERSITY on July 10, 2014 | http://arc.aiaa.org | DOI: 10.2514/8.7324

f% = PW
s i n <f>
along DE are t h e necessary conditions for t h e integral [5] to
be a m a x i m u m . Substituting E q u a t i o n s [11, 12, 13] into
fz = c o t 4>
E q u a t i o n [10], t h e following condition results
a n d X2, X3 are Lagrangian multiplier constants.
V Pa
sin 2d = cot a &tE. [14]
S o l u t i o n of t h e P r o b l e m

T h e solution of t h e problem lies in setting t h e first variation This condition relating M a n d 6 a t t h e end point of t h e nozzle
of / (Equation [5]) equal to zero a n d t h e r e b y obtaining
is t h e same as given in (4).
t h e required control surface a n d flow conditions along it. Let
F r o m E q u a t i o n s [12, 13] one can obtain t h e following rela-
u s first enumerate all t h e permissible variations of t h e q u a n t i -
tion dM/dy and 1
ties appearing in t h e integral. I n t h e following discussion, 8
denotes variation of a function, a n d partial derivatives are in- dO VM2 dM sin a sin 6
dicated b y t h e respective subscripts. = 0..[15]
dy I+ P dy y sin (6 + a)
As explained in t h e introduction, t h e initial expansion im-
mediately behind t h e t h r o a t region is assumed to occur along a
M
(
( 1 +
V )
prescribed contour TBB' (Fig. 1(a)). Let B indicate t h e T h i s relation is t h e compatibility condition between t h e
point u p to which such an expansion takes place, and let t h e M a c h n u m b e r and t h e flow direction along a left characteris-
right characteristic from B intersect t h e control surface a t D. tic. I t is crucial to this approach t h a t such a condition is im-
A n y variation in nozzle contour downstream of point B would plicit in t h e solution of E q u a t i o n s [12, 13], since according
n o t affect t h e flow between C and D. to E q u a t i o n [11] t h e control surface has t h e direction of t h e
For convenience t h e control surface between C a n d D is left characteristic. If t h e condition of compatibility were n o t
assumed to coincide with a left characteristic in the " k e r n e l " fulfilled, t h e control surface would become a limiting line,
of t h e characteristics net. This leads to 8C, 8M a n d 89 all i.e., t h e flow p a t t e r n would be physically impossible. E q u a -
zero in this region. </> = (a + 0) is a known q u a n t i t y along tions [12, 13], in connection with [11], give t h e form of t h e
CD, yielding <$</> = 0. T h e location of point D, i.e., t h e extent control surface and t h e velocity distribution in a form which
to which t h e assumed initial expansion occurs, is not known. does not require t h e solution of partial differential equations.
Hence 8D is n o t zero. I n this regard, t h e present paper goes beyond Guderley's
Between D and E, we h a v e 8D, 8M, 86 and <5</> all nonzero. solution. I n retrospect, one recognizes from t h e present a p -
Since only t h e length of t h e nozzle is prescribed, 8yE is non- proach, t h a t t h e additional Lagrangian multiplier h intro-
zero. M and 6 are continuous in t h e interior of t h e flow, and duced in Guderley's paper will assume t h e value zero.
<j) is also required to be continuous along CDE. Hence the
integrand in E q u a t i o n [5] is continuous. T h e variation of
point D therefore does not enter into t h e first variation of t h e
M e t h o d of C o n s t r u c t i n g O p t i m u m N o z z l e
integral / , a n d one obtains Contour
T o illustrate t h e application of t h e solution given in t h e
Si = 0 = fv* {(f1M + \2f2M + Xs/ 8 *)*M previous section t o w a r d obtaining a nozzle contour, a numeri-
cal example is carried out in detail in this section. A constant
+ (fie + X2/2* + \zfze)Sd + (U + X 2 / 20 + \^)84>}dy value of 7 = 1.23 a n d zero ambient pressure are used in t h e
example. T h e m e t h o d is simple enough t o m a k e t h e a p p r o -
+ dyE(fi + X2/2 + A 8 /8)attf..[6] priate changes for other conditions.
T h e first step is to choose a suitable curve for t h e nozzle
Since t h e variations in M, 6, 4> a n d yE are arbitrary, t h e above wall contour in t h e t h r o a t region. A circular arc of radius
leads to l.5yt (yt is t h e radius of t h r o a t section) is chosen for t h e
nozzle contour u p s t r e a m of t h e t h r o a t section. T h e asssumed
fiM + X2/2M + X3/3M = 0 [7]
nozzle wall contour in t h e t h r o a t region is shown in Fig. 2.
fie + X2/2* + X3/30 = 0 [8] Calculations according t o (4) indicate a M a c h n u m b e r 1.103
on t h e wall a t t h e t h r o a t section. I n Fig. (2), TTf represents
/10 + X2/2<* + X s / 3 0 = 0 [9] t h e line along which M = 1.103. T h e initial expansion im-

JUNE 1958 379

 mediately behind the throat is assumed to occur along a circu- For the present numerical example ME = 3.5 is chosen and
lar arc of 0.45 yt radius. Since flow across TT' is sufficiently the above equation yields the necessary wall slope SE = 13.22
supersonic, it is assumed unaffected by downstream condi- deg. Equations [12, 13] govern M and 6 along the control
tions. A characteristic net is computed (see the section on surface, and the constants X2 and X3 can be evaluated by in-
initial expansion in the nozzle) for these initial conditions, a serting ME = 3.5 and dE = 13.22 deg at y = yE. Equations
portion of which is shown in Fig. 2. The five right charac- [12, 13] can be rewritten as
teristic lines shown in the figure start from initial points on
the nozzle wall in the throat region, where the wall slopes are .^^ cos (8 — a) , „ cos (6E OLE)
M* = ME* • ^ ~ [17]
28, 30, 32, 34 and 35 deg, respectively. COS OLE

Instead of choosing a particular nozzle length, ME, the

Mach number on the nozzle wall at the exit, will be prescribed. where
This Mach number forms a parameter which describes a pos-
teriori the length of the nozzle. By choosing different values If* =
of ME, optimum contours for different lengths can be ob- 7 - 1 + M2
tained. Optimum nozzle contour for any particular desired
length can then be obtained by interpolation. For zero ambi-
and
ent pressure, Equation [14] reduces to
y _ I \ -7/(7-1)
M2
sin 2$E = COt CtE- [16] VE
( 1 + J—W~ M* sin2 6 tan a =
yME-
7 ~ 1 -7/(7-1)
ME2 1 + ME2) sin 2 BE tan aE. . [18]
Downloaded by OLD DOMINION UNIVERSITY on July 10, 2014 | http://arc.aiaa.org | DOI: 10.2514/8.7324

The above two equations can easily be solved by first choosing

pairs of M, 6 values satisfying Equation [17] and then ob-
taining corresponding values of y/yE from Equation [18].
Fig. 3 shows the values of M and 6 thus obtained as functions
of y/ys. These relations can be computed even though one
does not yet know the position of the control surface DE (see
Fig. 1).
The next step is to find the point in the characteristics net
2 3 (shown in Fig. 2) which would define the end point on the
x/yt
control surface. Consider the flow conditions along the right
characteristic from any point A on the prescribed contour
Fig. 2 Selection of the extent of initial expansion TBf. Pick a point A on the right characteristic such that
the values of M and 6 at A satisfy Equation [17]. The
dashed line shown in the figure is the locus of all such points.
/ From the values of M and 0 at Dh the value of y/yE at A can
•^ ft / be found from Fig. 3. Conservation of mass requires the mass
o .o /
flow crossing the right characteristic A A to be equal to the
6° mass flow crossing the control surface from A to E, the end

t
Di point on the nozzle wall. That is

M 3.5 2Tryt2ptWt . pW sin a

Bi ptWt
Pt cos {6 — a) y t \yj

CD^ PpW sin a

2iryE2 PtWt [19]
J1 ptWt sin (6 + a) yE ( - ) •

^ A
It should be remembered that the integration on the left-
/ hand side is carried out along BJ)i in Fig. 2, whereas the inte-
/
/ gration on the right-hand side depends upon the control sur-
24 face, as described in Fig. 3, and the point A - Also the ratio
D, P° of VEIVI in the above depends upon the choice of the point A .
The above equation can be satisfied by a few trials and by
20 \ / noting the error for each choice of the point A . In the present
/ example the point D shown encircled in Figs. 2 and 3, satisfies
,/ the above equation, [19]. By interpolating between known
/ right characteristics shown in Fig. 2, the right characteristic
0° 16 / BD through the point D, with respective values of M and 6
on it is found. This characteristic line BD is shown in Fig. 4,
indicated as extent of "kernel" since the assumed initial ex-
pansion occurs up to this right line. The location of the point
12 D as represented in Figs. 3 and 4 yields the ratio ys/yu
Equation [11] indicates that the control surface DE is a left
characteristic and this property is used to find X/yt for respec-
tive values of M, 6, and y/yt along DE. Thus the information
Q
given in Fig. 3 can be translated to define the control surface
1.0 0.8 0.6 0.4 0.2 0 DE in terms of yt as shown in Fig. 4. The length of the
nozzle is given by the ^-coordinate of the point E and is found
Y/YE to be 8.19 y% for this example.
Fig. 3 Mach number and flow angle along the control surface Starting with the above derived flow conditions along lines

380 J E T PROPULSION
 NOZZLE A
WALL CONTOUR-

NOZZLE A - ^

^
s

\ — NOZZLE B

-EXTENT OF "KERNEL"

'0 1 2 3 4 5 6 7 8
x/yt
x/yt
Fig. 4 Construction of the nozzle contour Fig. 5 Optimum nozzle contours—Pa = 0; y = 1.23

BD and DE, the characteristics net is completed in the region The thrust coefficients of these nozzle configurations, com-
between the two lines as shown in Figure 4. With the flow puted from wall pressures, are shown in Table 3, and are com-
field in this region known, the streamline passing through pared with conical nozzles having the same lengths and area
B and E is drawn. This streamline shown in Fig. 4 then ratios. Thrust coefficient is defined as
Downloaded by OLD DOMINION UNIVERSITY on July 10, 2014 | http://arc.aiaa.org | DOI: 10.2514/8.7324

forms the required contour for nozzle length of 8.19 yt. As

mentioned before, optimum nozzle contours for different _ thrust
lengths can be designed by choosing different values for wall vAt
Mach number at the point E.
and the maximum attainable value depends only upon the
ambient pressure and y, the ratio of specific heats. For zero
Typical Nozzle Configurations ambient pressure

The nozzle configuration computed in the preceding section

is shown in Fig. 5 and represents the contour for optimum
thrust when zero ambient pressure and a length of 8.19 yt are
-,(^P-"V^T
prescribed. The coordinates of wall points, Mach number and and one should remember that this value can only be obtained
wall slopes at the points are listed in Table 1. By choosing with a nozzle of infinite length and infinite exit area. The
ME = 2.6 and zero ambient pressure a shorter nozzle of length thrust coefficients of the optimum nozzles are also shown in
2.94 yt is designed and is also shown in Fig. 5. The coordi- Table 3 as percentages of the above maximum attainable
nates of wall points of this nozzle are listed in Table 2. value.

Table 3 Comp arison of thrust coefficients

Table 1 Optimum thrust nozzle A
(Pa = 0 , 7 = 1.23 L = 8.19F,) Contour A
shortened to
Wall slope Nozzle A Nozzle B length of
X/Yt Y/Yt M 0, deg of Fig. 5 of Fig. 5 nozzle B
0.25 1.08 2.11 34.4 Length-throat radius 8.19 2.94 2.294
0.33 1.13 2.19 32.8 Exit area-throat area 19.36 4.973 6.838
0.94 1.52 2.42 32.0 Thrust coefficient 1.7676 1.5829 1.5688
1.03 1.58 2.45 31.7 One- dimensional
1.17 1.66 2.48 31.2 thrust for the area
1.47 1.84 2.57 30.4 ratio, % 98.58 96.93 93.5
1.88 2.07 2.67 29.0 Thrust of conical
2.31 2.30 2.77 27.5 nozzle of same
3.37 2.82 2.96 24.0 length and area
4.20 3.16 3.08 21.6 ratio, % 102.3 100.5 102.1
5.43 3.32 3.24 18.5 Maximum available
6.50 3.95 3.35 16.2 thrust, % 82.7 74.1 73.4
7.98 4.34 3.48 13.5
8.19 4.40 3.50 13.1 1

Results presented in Table 3 show that nozzle A yields 2.3

per cent more thrust than a conical nozzle of the same length
and area ratio. On the other hand, nozzle B, of much shorter
Table 2 Optimum thrust nozzle B length and smaller exit area, yields only 0.5 per cent more
(Pa = 07 7 = 1.23 L = 2.94F,) thrust than the equivalent conical nozzle. If nozzle contour
A were cut off at a length of 2.94 yt (i.e., the length of nozzle B)
Wall slope one obtains a thrust coefficient of 1.5688. As can be expected
X/Yt Y/Yt M 0, deg this value is lower than the thrust coefficient of the nozzle B
0.21 1.05 1.96 28.7 which was designed to yield maximum thrust for the length.
0.29 1.10 2.01 27.8 To estimate the effect of y on the optimum nozzle shape,
0.63 1.27 2.12 26.9
0.91 1.41 2.20 26.1
7 = 1.4 is used, and for zero ambient pressure a nozzle is de-
1.52 1.70 2.34 23.7 signed having a length of 9.19 yt. This nozzle contour is
2.30 2.01 2.49 20.4 shown in Fig. 6, and differs considerably from the contour
2.94 2.23 2.60 17.9 computed for y = 1.23. Increasing the value of y reduces the
exit area of optimum thrust nozzle.

JUNE 1958 381

 scribed conditions. For example, nozzle A shown in Fig. 5 will
also be the optimum contour if, in addition to length of L/yt =
8.19, an exit area of A/At = 19.36 is the condition prescribed
NOZZLE
y.1.23-
\ in place of zero ambient pressure.
The nozzle contours presented in Fig. 5 show the difference
r =l JO between the optimum nozzles computed for the two different
lengths. On the contrary, Guderley and Hantsch (4) con-
cluded from their computations that for a given ambient pres-
sure all optimum nozzles of different lengths can be repre-
sented by a single contour. This may be a coincidence due to
either the sharp-corner expansion he considered, or the com-
°0 I 2 3 4 5 6 7 8 9 10 plicated nature of his solution.
x/y, The ratio of specific heats, y, of the exhaust gases has con-
siderable effect on the optimum nozzle contour as can be seen
Fig. 6 Optimum nozzle contours—Pa = 0 from Fig. 6.
Comparison of thrust coefficients shown in Table 3 indi-
It should be remembered that the nozzle contours shown in cates that the advantage of contoured nozzles is greater at
Figs. 5 and 6 are computed for in viscid isentropic flow. Similar larger area ratios.
to the methods used in wind tunnel nozzle design, one may
compute the displacement thickness of the boundary layer
along the nozzle wall and apply the correction to the contours References
shown in Figs. 5 and 6. Increasing the radial coordinates of
Downloaded by OLD DOMINION UNIVERSITY on July 10, 2014 | http://arc.aiaa.org | DOI: 10.2514/8.7324

the wall contour by the amount of the boundary layer thick- 1 Foelsch, K., "The Analytical Design of an Axially Sym-
metric Laval Nozzle for a Parallel and Uniform Jet," Journal of
ness would yield the exit flow for which the nozzle is designed. the Aeronautical Sciences, March 1949.
2 Dillaway, R. B., "A Philosophy for Improved Rocket
Conclusions Nozzle Design," JET PROPULSION, vol. 27, Oct. 1957, p. 1088.
3 Fraser, R. P., and Rowe, P. N., "The Design of Supersonic
By applying the calculus of variations a method is de- Nozzles for Rockets," Imperial College of Science, South Kensing-
veloped for designing the wall contour of an optimum thrust ton, England, Report JRL No. 28, Oct. 1954.
nozzle. The ambient pressure, length of the nozzle and wall 4 Guderley, G., and Hantsch, E., "Beste Formen fur Achsen-
contour in the throat region appear as governing conditions in symmetrische Uberschallschubdusen," Zeitschrift fur Flugwissen-
the formulation and solution of the problem. Typical nozzle schaften, Brauschweig, Sept. 1955.
5 Sauer, R., "General Characteristics of Flow Through
contours are presented in Figs. 5 and 6. Nozzles at Near Critical Speeds," NACA TM 1147.
A nozzle contour obtained for a given length and ambient 6 Shapiro, A. H., "The Dynamics and Thermodynamics of
pressure will also be the contour yielding maximum thrust Compressible Fluid Flow," Ronald Press, New York, pp. 676-
when the length and the corresponding exit area are the pre- 680.

Prediction of the Explosive Behavior of

Mixtures Containing Hydrogen Peroxide
E. S. SHANLEY1 and J. R. PERRIN2
Becco Chemical Division, Food Machinery & Chemical Corporation, Buffalo, N. Y.

This paper concerns a relationship between thermal sion, in synthetic organic chemistry, and for other purposes
properties and explosive properties for mixtures containing Mixtures of this kind are explosive within certain concentra-
hydrogen peroxide, water and soluble organic compounds. tion limits. The range of explosive compositions has been
It has been known for some time t h a t certain mixtures of determined empirically in a few cases.3 This is a laborious
this kind are explosive. I n t h e present study it has been undertaking, so that a way was sought to predict the proper-
found t h a t sensitivity to initiation is about t h e same for ties of untested mixtures. The present communication shows
all mixtures having the same heat of reaction. This rela- the correlation found between explosive behavior and AH, the
tionship is demonstrated for five different organic con- calorimetric heat of reaction. This correlation can be used to
stituents and for three methods of initiation. The findings predict the range of explosive compositions for untested mix-
provide an easy basis for predicting the likely range of ex- tures.
plosive compositions of untested mixtures containing hy-
drogen peroxide.
Experimental Part

Introduction Mixtures containing hydrogen peroxide, water and several

different combustible materials were tested. Only soluble
T ERNARY mixtures containing hydrogen peroxide, water
and soluble organic compounds are used in rocket propul-
"fuels" were used, so as to avoid the complications of two-
phase systems. Tests for sensitivity were carried out with
blasting caps, drop weights and static sparks, as described
Received
1
Sept. 9, 1957.
Present address: Arthur D. Little, Inc., Cambridge, Mass. 3
Mem.
2
ARS. Shanley, E. S., and Greenspan, F. P., "Highly Concentrated
Present address; Marquardt Aircraft Company, Van Nuys, Hydrogen Peroxide," Ind. Eng. Chem., vol. 39, 1947, pp. 1536-
Calif. 1543,

382 J E T PROPULSION