Structural Significance of Ductility in

Aerospace Pressure Vessels GEORGE GERARD 2

Allied Research Associates Inc.
Boston, Mass.

The results reported in this paper include an exploratory experimental study of small gage length
fracture strains at stress concentrations and an analysis of the weakening and strengthening effects
of stress concentrations in terms of a ductility ratio. The structural significance of tests used to
evaluate ductility in the presence of stress concentrations are discussed, and a procedure to evaluate
the ductility ratio as a basic material property is suggested. Finally, the structural strength/
weight characteristics of various high strength sheet materials are discussed in terms of the struc-
tural design problems associated with aerospace pressure vessels.
Downloaded by UNIVERSITY OF OKLAHOMA on February 2, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.6248

I N BALLISTIC and space vehicles, a major proportion of

the airframe weight is designed from tension considera-
tions. This marks a distinct structural trend from manned
specimen of prescribed configuration was apparently satis-
factory. It is widely recognized, however, that such a meas-
ure of ductility is merely an acceptance test value that has
aircraft and other lifting type vehicles in which tension con- little significance for ductility values pertinent to various
siderations govern roughly only one third of the airframe structural applications.
weight and compression, and shear buckling considerations Stress concentrations act as weak links in a structure; so
govern the other two thirds. Consequently, the selection of it appears that fracture strains associated with the localized
materials for aerospace vehicles has generally focussed atten- region of the stress concentration may provide some mean-
tion on the maximum tensile strength/weight potential that ingful estimate of the structural significance of ductility.
can be realized primarily in pressure vessel applications. Accordingly, an exploratory program was initiated to meas-
It is important to note that the last statement implies that ure fracture strains in very small gage lengths of the order
there is an essential difference between the strength of tension of 0.01 in. on parallel-sided tensile specimens and on speci-
structures and the tensile strength of materials. This differ- mens containing a circular hole as a representative stress
ence is intimately associated with the presence of stress con- concentration.
centrations in the structure: those that arise at structural
discontinuities and are determinate in character and those of Test Techniques
an indeterminate nature which inevitably result from fabrica-
tion such as weld flaws and mismatch due to tolerances. A study of possible strain measurement techniques for the
The relative importance of stress concentrations on struc- investigation of small gage length strains on low ductility
tural tensile strength depends strongly on ductility. Quali- materials led to the selection of the photo deposition of an
tatively, it is well recognized that the greater the ductility, accurately ruled grid on the polished surface of the specimen.
the greater the reduction of the stress concentration from its During the course of the test, the displaced grid lines are
elastic value. Thus, methods of quantitatively evaluating photographed on glass plates, and the relative displacements
the structural significance of ductility become of paramount of the individual lines are measured with a special precision
importance in the attempt to exploit successfully the attrac- photomicrometer to determine local strains.
tive tensile strength/weight characteristics of high strength In Fig. 1, typical enlarged photographs of the gage lines
sheet materials. are shown after fracture of the tensile specimens of titanium
Some of the results presented in this paper represent the alloy Ti-7Al-4Mo. The local strains obtained from the rela-
preliminary output of a broad research program aimed at tive displacement of these lines were determined along the
obtaining an engineering appreciation of the structural sig- paths indicated on the photographs by solid and dashed lines.
nificance of ductility. Among the objectives of this program One of the significant results evident in Fig. 2 is that the
are 1) investigation of the basic aspects of ductility to assess strain does not increase up to the point of fracture but reaches
the significance for design purposes of available techniques a maximum within the fracture zone contained between the
and conditions of test for ductility, and 2) investigation of fracture locations on the top and bottom surfaces. This
relations between ductility and maximum strength/weight in observation is probably a result of the very small 0.01-in.
typical structural applications. The preliminary results pre- gage length grid used in these experiments. It is quite
sented herein generally deal with these twTo subjects. apparent that if a 0.05-in. gage length grid had been used, the
strains would appear to increase up to the point of fracture,
A second significant observation is that the maximum local
Basic Ductility Measurements strain at the edge of stress concentration shown in Fig. 3 is
The rather loose connotation of the word ductility is per- of the same order of magnitude as the maximum local strain
haps indicative of the fact that, in the past, structural engi- for the parallel side specimen observed in Fig. 2. Specimen
neers did not require a precise description of this particular geometry can exert an important influence on strain meas-
material property to design reliable tension structures. The urements, so further investigation was required before maxi-
commonly accepted measure of ductility as the 2-in. gage mum strains could be correlated.
length strain obtained after fracture of a structural tensile To investigate the influence of geometry on fracture strains
in a preliminary manner, the dimensions of the test specimens
were held constant except for the cross-sectional dimensions,
which were varied both in width and in thickness. Five
Received June 9, 1961; revised May 28, 1962. parallel-sided tensile specimens and two specimens with iden-
1
Prepared under BuAer Contract NOas 59-6257c with New
York University. tical stress concentrations in the form of concentric holes
2
Director of Engineering Sciences. were tested.

1216 ARS JOURNAL

 For the specimens without the stress concentration, the convenience, Eq. [2] is first rewritten in the following form:
local strains in the region of fracture were obtained for each
specimen along several lines parallel to the longitudinal axis kP = 1 + (ke - l)e [4]
of the specimens and evenly distributed over the specimen where
width. For the specimen with the stress concentration, the
strain distribution was measured on both sides of the hole. e = Esm/Esn [5]
The maximum strains of the two specimens with the stress Now, letting Esn = E for cases where the average net section
concentrations ( —lc, —2c) correlated with the results of the stress is elastic, and noting that in general
parallel side tension specimens if the distance from the hole
to the edge of the specimen was used as the effective width Esm — (Tfu/€f [6]
in computing the width-thickness ratio. The correlation is
presented in Fig. 4, in which the upper line represents the where ef is the appropriate very small gage length fracture
highest values of the maximum strain and the lower line the as observed in Fig. 4, Eq. [5] becomes
lowest values. e = atu/E ef = eb/ef [7]
where
Stress Concentrations and Ductility
Now that it has been established from the experimental
€b = (Ttu/E m
results discussed in the previous section that very small gage
length strains are involved in fracture of stress concentra-
tion specimens, it is pertinent to consider the behavior of
Downloaded by UNIVERSITY OF OKLAHOMA on February 2, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.6248

stress concentrations from an analytical viewpoint. In this

manner, it is possible to establish relations among the factors \
that influence notch strength: elastic stress concentration, ' V / )\ | l FRACTURE ON

geometry, and ductility. FRACTURE ON ^ [/

BACK SURFACEi /
1 \
1

Plastic Stress Concentrations \l /. |\ \

/
The reduction of an elastic stress concentration as the stress / ^
is increased into the plastic range has been analyzed by Stowell °y* V, J v V""*-
(l) 3 using plasticity theory. He solved the case of a small
circular hole in an infinite plate under uniaxial tension and
arrived at the following expression for the plastic stress con-
centration factor kp which is in good agreement with experi- LOCATION (inch;
mental results
Fig. 2 Local strain distribution for smooth tensile specimen
kp = 1 + 2 Esm/Es1 [1]
In Eq. [1], Esm is the secant modulus corresponding to the
maximum stress at the stress concentration, and Esn is the
secant modulus for the average stress in the plate.
Hardrath and Ohman (2) tested a variety of stress concen-
trations in the plastic range and found that the following
generalization of Eq. [1] was in good agreement with their
tests:
Kv 1 + (A;. - l)Esm/Esn [2]
Here, Esn is the secant modulus for the average net section
stress at the stress concentration.
In Refs. 3 and 4, Eq. [2] was used in an investigation of
the static strength of aluminum alloy specimens containing
fatigue cracks. The test data were correlated by using the
following relationship for the secant modulus: 8 12 20^ w
"24 28xlO" 2

LOCATION (inch)
Esn Gtu/€u [3]
Fig. 3 Local strain distribution for specimen containing a central
It is apparent from the results of the previous section that circular hole
Eq. [3] may provide an inaccurate estimate of Esm> since
the uniform strain eu was used instead of the very small gage uu
length fracture strains. nn
For cases in which the net section stress remains elastic,
an important simplification of Eq. [2] can be obtained. For
I ^"" ,-'J!/&
^r
1
Numbers in parentheses indicate References at end of paper. -•*^1^\ ^^v % 2222Zp
y j f e ^ -~ir~^

I0
20 30 40 50 60 80 '00

Fig. 4 Maximum strain at fracture as influenced by specimen

Fig. 1 Photographs of gridded specimens after fracture geometry
AUGUST 1962 1217
 which corresponds to the restriction that the net section
stress is elastic and on which the assumption of a constant
e is based, the trend of the test data is reasonably well pre-
dicted. For N values above approximately 0.9, e varies as a
result of the net section stress becoming plastic and tends to
unity as ke approaches unity.
Although a similar analysis can be made for the other test
data presented in Ref. 6, this correlation is reserved for the
next section in a somewhat different form.

S t r u c t u r a l D e s i g n S i g n i f i c a n c e of N o t c h T e s t s
Certain environmental and loading conditions that may
significantly affect design performance are associated with
the structural applications of materials. Stress field, tem-
perature, strain rate, stress concentrations, and size effects
may be included among these factors. To guard against
inappropriate use of materials under certain combinations
of these factors, metallurgists have used tests such as the
Fig. 5 Notch strength ratio as a function of elastic stress con-
centration factor according to Eqs. [10 and 11]. Test points of Charpy, notched tensile, slow bend, and others to screen
Ref. 6 for aged Ti-2.5Al-16V at various temperatures materials. In a structural design sense, these tests have
been useful in the rather restrictive sense that they establish
Downloaded by UNIVERSITY OF OKLAHOMA on February 2, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.6248

temperature conditions under which notch sensitivity is

likely to become important.
Thus, in the case where the average net section stress is For example, the features incorporated in a Charpy V-notch
elastic, which is equivalent to a notch-unnotch strength ratio test include a relatively high stress concentration, a rather
somewhat less than unity, e represents the ratio of the brittle severe stress gradient, and a relatively high strain rate at
strain eb to the ductile strain ef. As such, it represents a various test temperatures. The severity of the test has
ductility ratio ranging between unity for a truly brittle ma- proved effective in guarding against poor structural designs
terial and approaching zero for a very ductile material. As that ignore the presence of severe stress concentrations and
will be shown subsequently, the ductility ratio may provide fabrication flaws under low temperature conditions. This
a useful and meaningful index of ductility in structural ap- test, in a way, effectively isolates the metallurgical problem
plications. from the structural engineering problem since, in order for a
material to "pass" the Charpy test, it must have sufficient
Strengthening Effects of Stress Concentrations ductility to accommodate severe stress concentrations. Upon
passing this test, the structural designer is free to use the
Eq. [4] essentially accounts for the weakening effect of stress material with the hope that ductility will forgive all.
concentrations as a function of the ductility ratio e. There is Recent attempts to use high strength materials in aero-
also an opposing effect of stress concentrations which causes space pressure vessels have resulted in the recognition that
strengthening. This effect has been observed experimentally extreme care in design and fabrication is mandatory to
by Sachs et al. (5) for sharply notched bars and sheets of achieve anything near the full strength/weight potentials of
highly ductile materials. Apparently as a result of the tri- the sheet materials. Consequently, the notch tests used
axial stress field existing in the vicinity of a severe notch, in the past as well as those recently devised may now be too
the net section strength of a highly ductile material can be severe for those cases in which the requisite attention has been
considerably higher than the ultimate tensile strength as given to design and fabrication. It would seem, for ex-
measured on a smooth specimen of the same material. ample, that notch sensitivity could be incorporated directly
If ks represents the ratio of notch strength to simple tensile into the design process if this property were evaluated in
strength, then for a material in which e approaches zero. terms of the appropriate stress field, temperature, strain rate,
kp = 1/h [9] size, and particularly stress concentrations representative of
the structural application.
Data of Ref. 5 indicate that ks increases linearly from unity
with the cross-sectional area removed by the stress concentra- Evaluation of Ductility Ratio
tion and depends on the cross-sectional geometry. For a
material of limited ductility, the following modification of In order to indicate a possible approach to incorporating
Eq. [4] is tentatively suggested: the results of stress concentration tests into the structural
design process, it is pertinent to return now to Eq. [10] and
kv = l/ks + (ke - 1/h) e [10] the test data presented in Ref. 6. From a series of strength
It is to be noted that in terms of the more familiar notch- tests on specimens of the same geometry such as presented
unnotch strength ratio N in Ref. 6 in which only the radius of the stress concentration
has been varied to obtain different values of ke, values of kp
kP = 1/N [11] can be measured directly for each specimen.
In the region where the average net section stress is elastic,
Results of Analysis it is to be expected that the ductility ratio e may be constant.
Therefore, by rewriting Eq. [10] in the following form
In order to indicate the type of results which can be ob-
tained from the analysis presented, Fig. 5 has been prepared kP = (1A,)(1 - e) + he [12]
in terms of N as a function of the elastic stress concentration
for various values of the ductility ratio e. A notch strength- it can be observed that, by plotting kP vs ke, a straight line
ening factor of ks = 1.10 has been used in order to effect a of slope e offset from the origin by (1 - e)/ks should be
correlation with existing data of Ref. 6. obtained.
Also shown in Fig. 5 are test data of Ref. 6 on the aged A plot of this type for several titanium alloy and steel high
titanium alloy Ti-2.5 A1-16V for three different test tempera- strength sheet materials based on test data of Ref. 6 is shown
tures. It can be observed that, for values of N below 0.9, in Fig. 6. It can be observed that the data conform to the
ARS JOURNAL
1218
 straight lines corresponding to Eq. [12 J in a reasonably good 2.5
manner. For the titanium alloy Ti-13V-llCr-3Al, it appears
that kp reaches a maximum and then remains constant as ke
is increased. This trend appears to be in agreement with
Neuber's concept that crystalline aggregates tend to relieve
2.0 A
very high stress concentrations. /
It must be noted that the good agreement indicated in
Fig. 6 was not obtained for some of the other high strength 1.5 /
sheet materials presented in Ref. 6. For these other mate- / ^-~~^~^L

rials, there seemed to be discontinuities in the test data when

plotted in the form of Fig. 6. The discontinuities could pos- X --~^±
1.0
sibly be interpreted as ductile-to-brittle transitions. Unfor- -^^ ^
tunately, there were not a sufficient number of test points ^ A AMS 6434
e
.023
ks
l.ll

for any one material to show conclusively the presence of a O

•
4340
T i - 6 A l - 4V
.076
.049
1.33
1 16
.5
discontinuity. Consequently, judgment concerning the © Ti-5AI-3Cr-IFe 093 1.07
1.40
validity of Eq. [12] for these other materials must await fur- • Ti-l3V-IICr- 3AI .170

ther test data.

Returning to the materials shown in Fig. 6, it is possible to 0
, 7 9
1 13 15
evaluate e and ks from the slope of the straight line and inter-
cept for each of the materials. These data are shovm in Fig.
6. It is to be noted that the evaluations of ks and e are Fig. 6 Correlation of Eq. [12 ] with test data of Ref. 6
Downloaded by UNIVERSITY OF OKLAHOMA on February 2, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.6248

based on a limited number of test points and may be some-

what lacking in accuracy.
2x10°

Significance of D u c t i l i t y R a t i o
j ^
10°/ 0
o'
It should now be apparent that the ductility ratio e can be / —-—~" "°"S
\ /P = i.6xio ....
. • — • —

looked upon as a basic mechanical property that may provide B -°f

6
e1 b
a meaningful measure of ductility in a structural sense. In .7
— • " "

\ 10°/
°
combination with Eq. [12], it permits the determination '.6
. .5
of the plastic stress concentration factor for prescribed values c o Ti Alloys
of the elastic stress concentration factor. - .4
• Steels
It is notewortlw that evaluation of the ductility ratio in-
b=.3
volves a straightforward test procedure in which the strength
of specimens containing suitable stress concentrations is the
only measured quantity. The ductility ratio is evaluated
from these test data without requiring any direct measure of
strain. As such, it represents a simple test procedure from 1
.02 .03 .04 .05 .06 .08 .2 .3 .4 .5 .6
which the effects of stress concentrations can be determined
under the environmental conditions appropriate to the struc-
tural application. With a knowledge of the ductility ratio, Fig. 7 Material strength/weight ratio as a function of ductility
Eq. [12] can be used to incorporate plastic stress concentra- ratio
tion effects into the design process.

Structural Strength/Weight Considerations It is important to note that Eq. [13] reflects a convenient
strength/weight-ductility ratio "law" of materials which
Material Strength/Weight and Ductility hardly could have been anticipated. Because of the small
amount of test data on which it is based, however, it must
In order to focus now on strength/weight considerations be regarded as tentative, subject to further verification.
related to aerospace pressure vessel design, the data presented
in Fig. 6 will be used to provide some estimate of the rela-
Structural Strength/Weight
tion between the strength/weight and the ductility ratios.
For this purpose these data are summarized in Table 1. As discussed previously, tension structures are character-
These data are plotted in Fig. 7, and, although the data ized by stress concentrations of varying magnitudes. It thus
are few and somewhat scattered, the following relation be- appears useful to consider that the complex interaction among
tween material strength/weight ratio and ductility ratio fits strength and ductility of materials and the strength and
the data within ± 10% limits : ductility of structures can be idealized by assuming that the
essential difference between material strength and structural
o-tu/p = 1.6 X 106 e1 [13] strength is caused by stress concentrations. With this con-
Although the points for the steels seem to fall toward the lower cept, the tensile strength of a structure S containing stress
10% limit and the titanium alloy points toward the upper concentrations relative to the tensile strength of the material
limit, all the data are grouped together in view of the paucity used atu was given in Ref. 7 as
of data.
S = (Ttu/kp [14]
Table 1 By combining Eqs. [10 and 13] with Eq. [14] and letting
ks = 1
Material e <rtu <?tu/p
Vtu/P
AMS 6434 0.023 240 ksi 0.86 X 10 3 in. S/P = [15]
4340 0.076 250 0.90 1 + (ke - l)(oV1.6 X 106p)6
Ti-6A1-4V 0.049 170 1.06
Ti-5Al-3Cr-lFe 0.093 190 1.16 Results obtained from Eq. [15] for prescribed values of ke are
Ti-13V-llCr-3Al 0.170 195 1.11 shown in Fig. 8. It is most interesting to observe that, for
each value of kc, the structural strength reaches a maximum
AUGUST 1962 1219
 i.2xicfi reasonable accuracy, considering the few test points upon
ke=i • which Eq. [13] is based.
Also evident from Fig. 10 is the fact that, for very sharp
<~ notches (k0 = 17), high material strength/weight leads to
low structural strength/weight. If an elastic stress concen-
^^SNt tration factor of 17 represents a particular structural applica-
s^S
tion, then material strength/weight ratios in the range of
0.8 to 1.0 X 106 in. represent the optimum for the materials
S/p /X considered in Fig. 10.
y
\ o /max 6 p

Design of Pressure Vessels

The design problems associated with large liquid and solid
propellant powered vehicles generally require that pressure
vessels perform several distinct functions. The primary
'[/_ 1.4 1.6x10°
function concerns the load carrying ability of the structure
under various combinations of internal pressure, static, thrust,
crlu/p maneuver, and gust loads. In addition, the pressure vessel
generally seals and contours the propellants and provides
Fig. 8 Structural strength/weight as a function of material some measure of environmental protection. The latter
strength/weight for various elastic stress concentration factors
Downloaded by UNIVERSITY OF OKLAHOMA on February 2, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.6248

conditions, particularly the sealing function, generally impose

according to Eq. [15] important restrictions as to the method of fabrication when a
homogeneous material is used to satisfy simultaneously the
and then begins to decline with further increases in the load carrying and sealing functions.
strength/weight ratio of materials. The maximum or opti- In order to simplify this problem somewhat, the design of
mum strength/weight levels for a given ke can be obtained a cylindrical pressure vessel of diameter D and length L under
by performing the operation d(S/p)/d(atu/p) = 0 on Eq.
[15] with the following results:
( O V P ) O = 1.6 X 106 [5(&. - 1)] -1/6
[16]
(S/P)max = (5/6)(cr hl /p) [17]
Eq. [17] is indicated on Fig. 8. Results obtained from Eq.
[16] are shown in Fig. 9, which relates the structural condi-
tions (ke) and material properties (<n«/p).
The implications to be derived from the results of this
analysis are quite clear. Careful attention in design and S/p

fabrication to the minimization of stress concentrations is

essential if the high strength/weight potential of materials
is to be realized in tension structures. If a given level of
stress concentration must be accepted, then there is an opti-
mum strength/weight level for the material. Departures
to either sides of this level result in a decrease in structural
strength.
Ref. 8 contains rather extensive sharp notch test data on
titanium alloy and steel sheet materials of various strength/
weight levels. These data are replotted in Fig. 10 in a form Fig. 10 Correlation of Eq. [15 ] with test data of Ref. 8
suitable to test the validity of Eq. [15]. Also shown in Fig.
10 are the numerical results obtained from Eq. [15] for ke
= 17 and ks — 1.10 which correspond to the geometric con-
figuration of the test specimens used to obtain the data. It internal pressure p will be considered. These constitute the
would appear that the trend of the test data is predicted with prescribed design conditions. The weight of such a shell is
W = TTP LDt [18]
The structural strength required for a thin wall cylinder
S = p D/2t [19]
Using Eq. [19] to eliminate t in Eq. [18]
W/ir D*L = (P/2S)p [20]
Now using Eqs. [14 and 10] and assuming that ke is large,
the following approximation is obtained:
S = <rtu/kee [21]
Substituting Eq. [21] into [20]
W/wD'L = k.(Pe/atJ'p/2 [22]
Eq. [22] represents in a somewhat simplified form the struc-
atu/p, in. tures-materials-design interplay for pressurized cylinders.
Fig. 9 Optimum strength/weight ratio of materials for pre- It is interesting to observe that the overall weight efficiency
scribed values of elastic stress concentration factor according to (W/wD2L) is equal to the structural efficiency (ke), the
Eq. [16] material efficiency (pe/<rtu), and the design condition (p).
1220 ARS JOURNAL
 Of particular interest in material selection is the appearance W = weight, lb
of the strength-weight-ductility parameter (o-tu/pe) in Eq. p = density, lb/in. 3
[22]. crtu = ultimate tensile strength, psi
It is apparent from Eq. [22] that structural efficiency can €& = brittle strain, e0 = vtu/E
ef = local fracture strain
be achieved by reducing ke as much as possible. This re-
emax = maximum local fracture strain
quires careful attention to design in regions where geometric tu = uniform strain
discontinuities may occur and detailed attention to fabrica-
tion, particularly where welding is used. Since weld flaws
are characteristically regions of severe stress concentrations,
References
good design practice requires increasing the sheet thickness
in the region of the weld (sculpturing of sheet) and using an 1 Stowell, E. Z., "Stress and strain concentration at a circular hole in an
annealing weld zone. In this manner, indeterminate stress infinite plate," NACA TN 2073 (1950).
2 Hardrath, H. F. and Ohman, L., "A study of elastic and plastic stress
concentrations associated with fabrication can be reduced in concentration factors due to notches and fillets in fiat plates," NACA Rept.
importance. 1117(1953).
3 McEvily, A. J., Illg, W., and Hardrath, H. F., "Static strength of
aluminum-alloy specimens containing fatigue cracks," NACA TN 3816
N o m e n clature (October 1956).
4 Illg, W. and McEvily, A. J., "Static strength of cross-grain 7075-T6
D = cylinder diameter, in. aluminum-alloy extruded bar containing fatigue cracks," NACA TN 3994
e ductility ratio (April 1957).
= 5 Sachs, G., Sessler, J. F., Pray, R. F., and Yeh, T. H., "Relations be-
E = elastic modulus, psi tween the notch tensile strength of cylindrical and prismatic specimens of
Eam = secant modulus at maximum stress, psi titanium alloys and heat-treated steels," J. Basic Eng., Trans. Am. Soc.
Esn = secant modulus at net section stress, psi Mech. Engrs. 82D, 401-410 (June 1960).
Downloaded by UNIVERSITY OF OKLAHOMA on February 2, 2015 | http://arc.aiaa.org | DOI: 10.2514/8.6248

6 Sachs, G. and Sessler, J. G., "Effect of stress concentration on tensile

ke = elastic stress concentration factor strength of titanium and steel alloy sheet at various temperatures," Sym-
np = plastic stress concentration factor posium on Low-Temperature Properties of High-Strength Aircraft and Missile
ks = stress concentration strengthening factor Materials, ASTM STP 287 (Am. Soc. for Testing Materials, New York,
L length, in. 1960), pp. 122-137.
= 7 Gerard, G., "Structural interplay: design and materials," Aero/Space
N = notch-unnotch strength ratio Eng. 18, 37-42 (August 1959).
V = pressure, psi 8 Espey, G. B., Jones, M. H., and Brown, W. F., "Sharp-edge-notch
structural strength, psi tensile characteristics of several high-strength titanium sheet alloys at room
st = and cryogenic temperatures," Symposium on Low-Temperature Properties of
= sheet thickness, in. High-Strength Aircraft and Missile Materials, ASTM STP 287 (Am. Soc. for
w = width, in. Testing Materials, New York, 1960), pp. 74-95.

Magnetohydrodynamic Generators
and Nuclear Propulsion R. J. ROSA2
Avco-Everett Research Laboratory
Everett, Mass.

S o m e of t h e m o r e advanced possibilities for u s i n g t h e potential h i g h t e m p e r a t u r e and h i g h spe-

cific power capability of t h e m a g n e t o h y d r o d y n a m i c generator in nuclear propulsion are discussed.
Radiating and nonradiating closed cycles u s i n g liquid and gas core reactors are considered. Also
considered is a t h r u s t a u g m e n t i n g s c h e m e t h a t m i g h t substantially e n h a n c e t h e performance of
any type of h i g h t h r u s t propulsion s y s t e m .

D URING the past few years considerable interest has

developed in the possible use of magnetohydrodynamics
for electric power generation. At present the major effort
lowing. First, the device has no particular advantage as an
electric power source per se. Its virtues stem primarily from
its role as an expansion engine in which temperature, power
is directed toward its use in commercial powerplants, but of density, and power per pound can be much higher than those
course the question of worthwhile space application also has in a conventional expansion engine such as a turbine. Second,
arisen. The purpose here is to explore some of the more very strong scaling laws govern its design. Performance
unique possibilities which appear to take maximum advantage improves markedly in all respects when size (in terms of rated
of those characteristics that distinguish an MHD generator power output) increases. Third, the superiority of the de-
from other methods of energy conversion and transfer. Such vice depends on the fact that the energy conversion process
a discussion is necessarily quite speculative. However, the occurs in gas phase and hence is partially freed from the
obstacles to practical realization do not seem so great as to limitations imposed by solids. Unfortunately, it also is not
preclude its having a potential value. a "direct conversion" device; hence, its full potential can-
Some observations about the MHD generator to which not be realized unless certain other components of the cycle
most workers in the field probably would subscribe are the fol- in which it is incorporated also are not directly limited by
the properties of solids.
These factors motivate a consideration of the gas core or
Received November 22, 1961; revised April 17, 1962. cavity reactor as a heat source for the MHD generator and
1
This work was supported by Headquarters, Ballistic Systems lead to a search for a type of propulsion cycle which best
Division, Air Force Systems Command, U. S. Air Force, under
Contract No. AF 04(647)-278. takes advantage of the characteristics of both devices. This
2
Principal Research Scientist. Member ARS. does not imply that worthwhile MHD systems cannot be

AUGUST 1962 1221