\ is

NASA TECHNICAL NOTE 1 NASA TN D-2603

CO
NO
CM

DISTRIBUTION STATEMENT A
Approved for Public Release
Distribution Unlimited

-^TTT
STRESS-INTENSITY FACTORS FOR
SINGLE-EDGE-NOTCH SPECIMENS
IN BENDING OR COMBINED BENDING
AND TENSION BY BOUNDARY
COLLOCATION OF A STRESS FUNCTION

xaoo aiaviivAv isaa

by Bernard Gross and John E. Srawley nous aaonaoHaaa
Lewis Research Center
Cleveland, Ohio
20010920 122
NATIONAL AERONAUTICS AND SPACE ADMINISTRATION • WASHINGTON, D. C. • JANUARY 1965
STRESS-INTENSITY FACTORS FOR SINGLE-EDGE-NOTCH SPECIMENS

IN BENDING OR COMBINED BENDING AND TENSION BY BOUNDARY

COLLOCATION OF A STRESS FUNCTION

By Bernard Gross and John E. Srawley

Lewis Research Center

Cleveland, Ohio

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

For sale by the Office of Technical Services, Department of Commerce,

Washington, D.C. 20230 -- Price $1.00
 STRESS-INTENSITY FACTORS FOR SINGLE-EDGE-NOTCH SPECIMENS

IN BENDING OR COMBINED BENDING AND TENSION BY BOUNDARY

COLLOCATION OF A STRESS FUNCTION

by Bernard Gross and John E. Srawley

Lewis Research Center

SUMMARY

/ A "boundary-value-collocation procedure was used in conjunction with the

Williams stress function to determine values of the stress-intensity factor K
for single edge cracks of various depths in sp^ ^nsg subtjected to pure "bending.
The results are of use in connection with ^iTYtoughness tests, which
utilize rectangular-section crack-notch beam specimens loaded in four-point
bending, and are in good agreement with published results derived from experi-
mental compliance measurements. The results are expressed in convenient, com-
pact form in terms of the dimensionless quantity Y2 = K2B2W3/M2, which is a
function of relative crack depth a/w only, where B and W Eire the^specimen —
width and thickness and M is the applied bending moment. / TT^^^^f/^c<H'1(itfJ~\
r ■ —'.' £-—-^-^Y ' -J
/ On the assumption that the condition for a valid K-j-c test is that the
maximum nominal stress at the crack tip should not exceed the yield strengths of
the material, the Kjc measurement capacity of bend specimens was estimated as
a function of a/w. The measurement capacity is proportional to the yield
strength and to the square root of the specimen depth, and it is greatest for
a/w in the range 0.2 to 0. 3.

Values of K for single-edge-notch specimens subjected to combined bend-

ing and tension were obtained by superposition of the present results and those
of earlier work for specimens loaded in uniform tension. These values are of
interest in connection with the use of single-edge-notch specimens that are
off-center pin-loaded in tension. It is shown that the KT measurement ca-
i .

pacity of such specimens is not very sensitive to the eccentricity of loading. \~fo .,

INTRODUCTION

It was shown previously by Gross, Srawley, and Brown (ref. l) that the
value of the stress-intensity factor K for a single edge crack in a flat
plate specimen of finite width could be computed accurately by a boundary-
 value-collocation procedure applied to an
P/2 appropriate stress function. For uniform
tensile loading, the computed values of K
for various values of the relative crack
P/2 length a/w were in good agreement with the
M - PL/2 corresponding values derived from experi-
mental measurements of specimen compliance
by Srawley, Jones, and Gross (ref. 2), thus
providing confidence in the reliability of
ah- the mathematical analysis. The significance
of K and its role in the measurement of
plane strain crack toughness Kjc are dis-
cussed in references 1 and 2, and in greater
detail by Srawley and Brown (ref. 3). (All
symbols are defined in the appendix).
M = PL/2
P/2
In the present report, the application
of the boundary-value-collocation procedure
P/2 —L Jl to the case of single-edge-notch specimens
subjected to pure "bending is described, and
Figure 1. - Single-edge-notch specimen subjected to pure the results are presented. These results
bending. are compared with previous results obtained
by Bueckner who used a different analytical
method (ref. 4), and also with results derived from careful experimental com-
pliance measurements of four-point loaded notched beams by Lubahn (ref. 5).
While the results of references 4 and 5 are substantially in agreement, there
is sufficient discrepancy between them to warrant a third, independent treat-
ment of the problem in view of the practical importance of the accuracy of
Klc measurements that are conducted with four-point bend specimens.

The use of single-edge-notch specimens loaded in tension through off-

center pins has been discussed by Sullivan (ref. 6), and the results of experi-
mental compliance measurements for two positions of the loading pin holes are
given in this reference. In the present report, the general case of off-center
tension loading is treated by appropriate superposition of the present results
for pure bending and the results of reference 1 for uniform tension. This
method gives a good approximation to the value of K for any position of the
loading pin holes.

In deciding what design of single-edge-notch specimen is to be used for a

particular application, an important consideration is the extent to which the
KIc measurement capacity C^ is affected by the design. The K measure-
ment capacity is the largest value of Kjc that could be measured with accept-
able accuracy by using a specimen of given width and adequate thickness and a
material of given yield strength. For a given specimen design, Cjg- is propor-
tional to the yield strength and to the square root of the specimen width. In
the present report, estimates are obtained of Cjg for single-edge-notch
specimens loaded in pure bending and in combined bending and tension.
 METHOD

The method of analysis consists in finding a stress function X that

satisfies the biharmonic equation Vh. = 0 and also the boundary conditions at
a finite number of stations along the boundary of a single-edge-notched speci-
men, such as shown in figure 1. The biharmonic equation and the boundary con-
ditions along the crack are satisfied by the Williams stress function (ref. 7).
Because of symmetry (fig. l) the coefficient of the sine terms in the general
stress function must be zero, hence

z
CO

X(r,0) = (-1)n"la2n-lr'1+(l/2>[-COS(n-|) e + m-2

2n + 1
cos
K>
n=l,2...

+ (-i)nd
2n
rn+l
|_
_cos(n - 1)0 + cos(n + 1)0
■]} (1)

The stresses in terms of X obtained by partial differentiation are as follows:

Ö2X J
X cos20 ä2X sin 0 cos 0 ÖX sin 0
2
a
y = ÖX
...2 de ör r <3r r
ÖI

„ dx sin 0 cos 0 ä2X sin20

+ 2 g-
r be2 r2

=
ö2X = ö2X sin ? e + 2 ö2X sin 0 cos 0 ^+ ÖX cos20
°x ^2 57 3F37 ? S7 -T-
(2)
^x sin 0 cos 0 52X cos20
2
' Se r2 Ö02 r2

ö2x 2
Ö2X+ cos 20 S X sin 0 cos 0 ö2X
xy dx öy = sin 0 cos or 2 r ör d0 Ö0<

sin 0 cos 0 ÖX cos 20 o"X

r 3r " 2 cS?

The remaining "boundary conditions to he satisfied for the case of pure

bending (fig. l) are as follows:
Along

A-B X 0

ax' X" a<

B-C X + + ax + o
BW

C-D M ÖX
X = =r
B* 3£ = °
The collocation procedure consists of solving 2m simultaneous algebraic
equations corresponding to the values of X and either dx/ox or dx/dy at
m selected boundary stations, thus obtaining values for the first 2m coef-
ficients in the Williams stress function, the remaining terms being neglected.
Only the value of the first coefficient d-L is needed for the present purpose
because this is directly proportional to the stress-intensity factor K.
According to Irwin (ref. 8), the stress component a in the immediate vicinity
of the crack tip (as r approaches zero) is given by

K
l(1+ sin -p
a = cos sin
38^

while in terms of the Williams stress function, as r approaches zero

-ch
=
y? cos
2 y1 + sin
2 Sln
—j
and similarly for the other stress components, hence,

K = - -J2^ d1

For given values of M, B, W, a, and V (fig. 1), the value of d, com-

puted by the collocation procedure will vary somewhat with m but will ap-
proach a limit as m is in-
creased. This is illustrated by
the example shown in figure 2 for
specimens having a/w equal to
0.3 and various values of v/w.
For a given V/W, as the number
of boundary stations is increased
in steps of three, the computed
d-|_ oscillates about and converges
12 14 16 18. 20 22 24 toward a limit that is close to
Number of boundary stations, m
the average of the five computed
Figure 2. -Value of coefficient di against number of bound- values. Accordingly, the value of
ary stations. Specimen width, I inch; actual crack length,
0.30 inch, applied bending moment, 200 inch-pounds;
d-j_ that was used to obtain K in
specimen thickness, 1/16 inch. each case was the average of five
 i>!vin3
i i i i i
values computed for m equal to 11, 14,
Relative crack length, 17, 20, and 23. The computations were per-
a/W
| formed on a digital computer by using
14 double-precision arithmetic (16 significant
0.50
figures).
13
A value of K obtained by this method
45 corresponds to a particular set of values
■? 12
of M, B, W, a, and V (fig. l) selected
O
for convenience of computation. For ap-
c 11 plication, the results are better expressed
3

JO .4U more generally in terms of the dimension-

CD
1—

"K 10 less quantity Y2 = K2B2W3/M2 (or in terms

<T3
- of Y), which depends only on the dimen-
.35
3 9 sionless ratios a/W and V/W. As will be
shown, the effect of V/W on Y2 is neg-
<D . 30 ligible for values of V/W greater than
•a 8
8
unity, so that for a specimen of adequate
u
25 length, Y is a function of a/W only.
Fir

Consequently, a table or graph of Y2 or

.20 Y against a/W is all that is needed for
6 calculation of K for any given values of
.15 a, ¥, B, and M measured in a test.
5

.10
4 1 RESULTS AND DISCUSSION
.4 .6 .8 1.0 1.2 1.4 1.6
Height to width ratio, V/W
Pure Bending
Figure 3. - Single-edge-notch specimen subjected to
pure bending. Specimen width, 1 inch; applied
The results of the computations for
bending moment, 200 inch-pounds; specimen
thickness, 1/16 inch. pure bending are summarized in figure 3.
Each plotted point represents the average
value of dj_ of five values computed for
m equal to 11, 14, 17, 20, and 23. The plot shows the dependence of this
average value of d-|_ on v/W for each of nine values of a/w ranging from
0.1 to 0.5. The relative distance from the crack to the boundary at which a
stress distribution corresponding to pure bending was imposed is represented by
V/W (fig. l). In practical terms, it corresponds roughly to the ratio of one-
half the minor span to the depth of a four-point loaded beam. It is apparent
that the dependence of d-^ on V/W is negligible if V/tf exceeds unity.
Essentially, the same conclusion was reached for single-edge-notch specimens
loaded in uniform tension (ref. l). Consequently, the K values were calcu-
lated from the uniform values of d.-^ obtained when v/W was greater than
unity. Strictly these K values should be considered to apply to four-point
loaded bend specimens only when the minor span is 2W or greater.

The final results are given in table I in terms of Y2 (i.e., K2B2W3/M2)

as a function of a/W. The results obtained by a different analytical method
(ref. 4) and those derived from experimental compliance measurements (ref. 5)
are also tabulated for comparison. Experimental compliance measurements are
used to derive values of the strain-energy-release rate with crack extension
 TABLE I. - COMPARISON OF RESULTS OF PRESENT WORK WITH

CORRESPONDING RESULTS FROM REFERENCES 4 AND 5

IN TERMS OF DIMENSIONLESS QUANTITY

K2B2W3/M2 AS FUNCTION OF

RELATIVE CRACK LENGTH

Relative Results of -
crack
length, Collocation Reference 4 Reference 5
a/w "bounäary-
procedure

,r2 K2B2W3
M2

0.10 12.4 12.2 11.8

.15 18.5 17.4
.20 25.3 25.2 24.2
.25 33.2 32.15
.30 42.8 41.9
.35 55.2 53.9
.40 71.4 _____ 68.6
.45 92.7 88.9
.50 123.0 151.2 118.0

gf, rather than values of K directly. As discussed previously (ref. 2), the
gf values were converted to K values according to the generalized plane
stress equation K2 = E_f, where E is Young's modulus.

In reference 4, results are given for only three values of a/W; of these,
the two lower results are in excellent agreement with the present results,
While the value corresponding to a/W =0.5 is considerably higher than the
present result. The agreement between the present results and those of refer-
ence 5 is good over the whole range. For the practical purpose of Kjc mea-
surement, the range of a/w between 0.15 and 0.25 is of most importance. In
this range there is satisfactory agreement between all three sets of indepen-
dent results.

The following empirical equation is a compact expression of the present

results for values of a/W up to 0.35:
2 2 3
2 = K^ETW°
M2
139 221 + 783
= I" (I (1/
140

120

!!i
Ü:
100

;i!

I
g
60

40

20

Relative crack length, a/W

Figure 4. - Curve of least-squares best-fit cubic equation representing results for values of a/W from 0.1 to 0.35. Use of this equa-
tion outside this range is not recommended.
The equation was obtained by a least-square "best-fit computer program for a
cubic in a/W, incorporating the a priori condition that K should be zero
when a/W is zero. Only the results for a/W up to 0.35 were used in fitting
this equation "because it is considered undesirable to use bend specimens having
cracks deeper than about 0.35 W. Of the several quantities used to calculate
a value of Kyc, the one subject to the greatest uncertainty in measurement is
the crack depth. Since the sensitivity of the value of K to a small varia-
tion in a/W increases with a/W, it is undesirable to use specimens having
large a/w. On the other hand, the efficiency of bend specimens is low when
a/w is less than 0.15, as shown in the next section. The optimum range of
a/w for bend specimens appears to be between 0.15 and 0.25, and the usual
value is 0.2. The curve representing the fitting equation is shown in fig-
ure 4, together with the collocation results. The curve is shown dashed in
the ranges of a/W from 0 to 0.1 and from 0.35 upwards to emphasize that the
equation is not intended to represent the collocation results outside the
range of a/W between 0.1 and 0.35.

Kjc Measurement Capacity in Pure Bending

One of the necessary conditions for a meaningful Kjc test is that the
specimen width must be sufficient to ensure that the stress field in the
vicinity of the crack is sufficiently well represented by that of the assumed
linear elastic fracture mechanics model. For reasons that are discussed in
reference 3 it. will be assumed here that the useful limit of applicability of
the model to bend specimens will be reached if the nominal stress at the posi-
tion of the crack tip reaches the yield strength of the material
6M _
B(W - a)2 " ffYS

where a is usually taken to be the 0.2 percent offset tensile yield

strength. On this basis, a test result will be valid only if the abrupt crack
extension ("pop-in") occurs at a value of M not exceeding aY„B(¥ - a)2/6.
The K value corresponding to this value of M may be defined as the Kjc
measurement capacity of a specimen of given dimensions and yield strength, de-
noted CJK, providing also that the specimen thickness is sufficient as dis-
cussed in reference 3. Substituting in the equation

. _ K2B2W3
Y°2 =
M2
and transposing, we obtain
l
c2

a
2
W
Ll
i2'^
36
^ <

YS

This equation expresses the K measurement capacity of a bend specimen in

terms of the dimensionless quantity C2 /o| W, which is a function of a/w
.30 only since Y is a function of a/W
only. For any given value of a/W,
there is a unique value of CjK;/aYSW'
.25
and hence CW is proportional to the
yield strength of the material and to
.20 the square root of the specimen width.

Figure 5 is a plot of C^/olgW

.15"
1 1 ♦ against a/w, which shows that the
W
/ _ 1
.-*-
t » — range of a/w for greatest KIc mea-
T a
surement capacity is "between 0.2 and
.10 /
0.3. In this range, the C^ for a
specimen 3.5 inches deep and of adequate
.05
thickness is numerically about equal to
the yield strength of the material.
1
.1 .2 .3 .5 Combined Bending and Tension
Relative crack length, a/W

Figures. - Dependence of Kjc measurement capacity by Superposition

CIK on relative crack length for specimens in pure
bending.
In the case of a single-edge-
notch specimen loaded in tension
through pins (fig. 6), the Kjc mea-
P surement capacity might he expected to
depend on the loading eccentricity
4H parameter e/w as well as on a/W. It
is therefore necessary to study this
effect of e/W in connection with
standardization of design of single-
edge-notch tension specimens (ref. 3).

The load P acting through the

off-center pins on the single-edge-
i
a_h
notch specimen shown in figure 5 is
statically equivalent to the combina-
r^W
tion of an equal load acting along the
specimen centerline together with a
couple of moment Pe, where e is the
distance of the specimen centerline
from the line through the pin centers.
It should he noted that e is taken
to be positive when the pin centers
are on the same side of the centerline
as the cracked edge and taken to he
negative when they are on the other
side. By the principle of super-
Figure 6. - Single-edge-notch specimen subjected to com-
position, the value of each stress
bined bending and tensile loads. component, and therefore the value of
K, is equal to the sum of the values

9
 that would result individually from the
10 i
action of P along the centerline and
Relative c rack from the action of the couple Pe.
leiiyui,
^e
The component of K due to the
action of the couple Pe is readily ob-
0.5/ tained from the present results for pure
"bending. A good approximation to the
component of K due to the centerline
tensile load P can be obtained from
ffi > the results of reference 1. It should
| .3
—jw- be appreciated that the results of
.2 reference 1 relate to a specimen loaded
uniformly in tension normal to the ends
which is not exactly equivalent to a
^ specimen loaded through pins on the
'■£
/£■ centerline. A pin-loaded single-edge-
-.4 -.2 0 .2 .4 .6
Loading-eccentricity ratio, e/W
notch specimen bends slightly in propor-
tion to the load, the net effect being
Figure 7. - Dimensionless stress-intensity coefficient
as function of loading-eccentricity ratio for single-
that K/P is slightly less for a speci-
edge-notch specimens off-center pin-loaded in ten- men pin-loaded along the centerline than
sion. (Ratio is positive when pin centers are on same for the same specimen uniformly loaded
side of centerline as cracked edge.)
at the ends. The magnitude of this ef-
fect is discussed in reference 2 in
which experimental compliance-
measurement results for pin-loaded
specimens are compared with the results
.30 " of reference 1. For the present purpose
Loading **■
— . '(PL re ben ding) the magnitude of the effect is negligi-
^eccentricity / ble.
ratio * '&
?"5 e/W /
^ Figure 7 shows the dimensionless
//// \ \ -1/?
\ -1/4 5^~ quantity KBW1/2/? against e/w for
?n
i 4 -0
various values of a/V. By super-
position, this quantity is equal to the
f / sum of the component (KB¥1/2/p)t, due to
is e
1 the uniform tensile load P, and the
II component (KBW1/2/?)^ due to the couple
II — ah
in —U- Pe. The tensile component was obtained
1 from reference 1 (in that reference the
~U—
// w symbol P denotes load per unit thick-
OS ness, which is P/B in this report).
/ *y The component (KBW1/2/^ is equal to
Ye/W, because M = Pe, and therefore
1
.1 .2 .3 .4 .5
Relative crack length, a/W 2
^K2B2W3 ,1/ w /KBW1/2
Figure 8. - Dependence of KIc measurement capacity
CjK of single-edge-notch specimens on relative V M2 e I P
crack length and loading-eccentricity ratio. b

For a given value of a/w, the values of

10
 30
(KBwV^/p)^ and of Y are -unique, hence,

25
/
/
KBW-1/2 KBW"
1 "S
Supe rposi ion, /;
20 Ipre; ent w ork)V
is a linear function of e/w, as shown in
e = W/6
/ figure 7.
A -
\ /
© / '
15- When the same criterion is used for
/ this case as for pure tending, a test re-
TV / sult will he valid only if ahrupt crack
ts 10- w- ' /' extension occurs at a value of P not
© o
Compliance
measurerieni,
exceeding a„„B(W - a) /(¥ + 2a + 6e).
^ (ref. 6) This condition was applied in the same
■fS' manner as previously for pure "bending,
2 / 2
and values of C_f/oyqW were calculated
^
.4 for combined "bending and tension on the
.1 .2 .3
Relative crack length, a/W "basis of the values of KBW1/2/? shown
Figure 9. - Comparison of results of present work, for
in figure 7. Figure 8 shows curves of
single-edge-notch tension specimens having loading CfL/o|. W against a/W for values of
eccentricity ratio of 1/6, with results of experimental IK. XD
measurements. e/w equal to 0, 0.25, and 0.5 and also
shows the curve for pure "bending from
figure 5 for comparison, It is apparent that the measurement capacity is
greatest in the range of a/W between 0.2 and 0.3 in all cases. The measure-
ment capacity increases with e/w, hut the magnitude of the effect is small,
for e ¥ =
the difference between CTK-/0YS / °° (Pure "bending) and for
e/w = 0 (uniform tension) being less than 10 percent. Uncertainty about the
validity of the basis of comparison is at least of this order, so that for
practical purposes it can be assumed that the measurement capacity of single-
edge-notch specimens is independent of the manner of loading. For other rea-
sons it is recommended elsewhere (ref. 3) that e/w should be zero for single-
edge-notch specimens tested in tension.

A comparison of the results obtained by superposition for e/w = l/6 with

the experimental compliance-measurement results of reference 6 for a comparable
pin-loaded specimen is shown in figure 9. The results of reference 6 cannot be
considered very accurate for reasons discussed in reference 2; nevertheless,
the agreement is fairly good for values of a/w up to about 0.3. The increas-
ing discrepancy with increasing a/w beyond 0.3 is, in part, attributable to
the fact that the bending, which occurs in the pin-loading of the compliance-
measurement specimen, has the effect of slightly reducing the effective bending
moment. This effect is neglected in the superposition calculations, as dis-
cussed earlier.

CONCLUSIONS

Stress-intensity factors computed by the boundary-collocation procedure forj

{t^

11
t single-edge-notch specimens in pure "bending were in good agreement with results
!
derived from experimental compliance measurements. Because of this agreement
"between two entirely different methods, either result can be used with confi-
dence.
AH
The range of relative crack length a/W within which the KIc measure-
ment capacity of a "bend specimen is greatest "between 0.2 and O.T.—This esti-
mate results from the assumption that the nominal stress at the position of the
crack tip should not exceed the yield strength in a valid K^ test. .

. Stress-intensity factors.for single-edge-notched specimens loaded in com-

bined bending and tension can be calculated by appropriate superposition of the
available results for uniform tension and for pure bending. The KT measure-
ment capacity of single-edge-notch specimens that are loaded off-center in
tension is only marginally influenced by the eccentricity of loading, j

Lewis Research Center,

National Aeronautics and Space Administration,
Cleveland, Ohio, October 29, 1964.

12
 APPENDIX - SYMBOLS

a crack length or depth

B specimen thickness
C
IK estimate of maximum value of Kjc that can he measured with
specimen of given dimensions and yield strength
d
2n'd2n-l coefficients of Williams stress function

E Young's modulus

e distance of centerline of single-edge-notched tension specimen

from line through loading pin centers

# strain energy release rate with crack extension per unit length
of crack border, or crack extension force

K stress-intensity factor of elastic stress field in vicinity of

crack tip
K
Ic critical value of K at point of instability of crack extension
in first or open mode, a measure of plane strain crack tough-
ness of material

L length of bending moment arm

M applied bending moment, PL/2

m number of selected boundary stations used in collocation computa-

tion

P total load applied to specimen

r polar coordinate referred to crack tip

V distance from crack to boundary at which stress distribution

corresponding to pure bending was imposed

W specimen width

x,y Cartesian coordinates referred to crack tip

3 2
Y dimensionless stress-intensity coefficient KBW / /M that is
function of a/w only

0 polar coordinate referred to crack tip

ü
v>ö;T r stress components

13
a
YS ^'2 Percen"t offset tensile yield strength
X stress function

Subscripts:

~b "bending moment component of Y

t tensile component of Y

14
 REFERENCES

1. Gross, Bernard, Srawley, John E., and Brown, William F., Jr.: Stress-
Intensity Factors for a Single-Edge-Notch Tension Specimen Toy Boundary
Collocation of a Stress Function. NASA TN D-2395, 1964.

2. Srawley, John E., Jones, Melvin H., and Gross, Bernard: Experimental Deter-
mination of the Dependence of Crack Extension Force on Crack Length for a
Single-Edge-Notch Tension Specimen. NASA TN D-2396, 1964.

3. Srawley, John E., and Brown, William F., Jr.: Fracture Toughness Testing.
Paper Presented at ASTM Meeting, Chicago (111.), June 22-26, 1964. (See
also NASA TM X-52030, 1964.)

4. Bueckner, H. F.: Some Stress Singularities and Their Computation by Means

of Integral Equations. Boundary Problems in Differential Equations,
Langer, R. E., ed., Univ. of Wisconsin Press, 1960, pp. 215-230.

5. Lubahn, J. D.: Experimental Determination of Energy Release Rate for

Notched Bending and Notched Tension. Proc. ASTM, vol. 59, 1959,
pp. 885-913.

6. Sullivan, A. M.: New Specimen Design for Plane-Strain Fracture Toughness

Tests. Materials Res. and Standards, vol. 4, no. 1, Jan. 1964, pp. 20-24.

7. Williams, M. L. : On the Stress Distribution at the Base of a Stationary

Crack. Jour. Appl. Mech., vol. 24, no. 1, Mar. 1957, pp. 109-114.

8. Irwin, G. R.: Analysis of Stresses and Strains Near the End of a Crack
Traversing a Plate. Jour. Appl. Mech., vol. 24, no. 3, Sept. 1957,
pp. 361-364.

NASA-Langley, 1965 E-2801 15

 ■a W O
T) H O U ~ CO
S-c
n
o
Q 0 ^3 ^
1-3
m >>„ Z <z ffl -z <
z
co" .0
co <
o
< Ora Z
Ü CQ Z
. w B
. w B

03 0
H cd
Ü "
bO "
0 CD
^ »
CO 43 4«!
Q 0
S jogg
CM
.SFl..
OS« « m H g -o 0
> CO 0 0
a ö
O CQ
c
>> CO ' s a
isCO ° o >H Ü 2 w "" 0 b a
CO H m
z 2 43 .5
« S-S5- a^ O a
a m o 3 B ■ 6'?, o 00 0
'■O H
ES S
r"i1ZIJ
X! O o a-0 -^ >—1
cd
O 43 CQ

0
< h
h 03 ^ H ti m >> 0 43
S C ^ 43 0 o S 0 .« J
0 ai Sgp»8 ■" - -o ^
Ml 0)5 1
as WfflRBU
ft K Z 3 B «Sä*
B ® > CO

PSgM s »ft
«Zh^g <D 0 ö
H
S fe 5 OM C CO Ü Ö 5-0
CQ S pq >>OQ "^ a
CO o
0
o 2
bo 43
zg«ü< a 0 ™
CO
S B L 2, t>D
O a ^ co co -—
CO w
H > *g co .b
o
^, ^
o ' ' CO ^
CD Cft a o > 73 cd T3
CM O W CJ S < CQ . .S C 0 B
0 0 Ü a
'S 5 a
I SH f-. T3 —
B ü 3 g
CO c -a ■ ■§
is 532^ 0 0 CO CD
W ! T3 G
BS^| 0 0 fi c.
< o „
_ H «D 42 to O i3 43
cojui H Z H _. u i - S "
CH 0 -a
a
CD
bo u
< cU E-i 0 SH a o
Z Z CQ Z ffl U co a i T3 3 co Ü §g
■M
C

• CO
T3 W O
■oWo
*H ^ CD ^ - to
d S CM
cd H CM
B -£• I
S 2 Q CD ^

(0 ~z < PQ ^^ <z
L
z - (D
- 0

2 gS
Ü CQ Z
Om Z
. -; B . w B

•a
CO ai W
w CD -*-"
4Ü cj O - u 01 CO
to
Ü
Q Q O 2 £«
bo « bo 0 0 co O +J cj 0 .
Q *" ü St H 3 ft E-> o
.5F 5 <u B«
""
■o c
• CQ Cd TJ
0 Z
0

2 J -a N CQ 3
m
cd £<
> to Ü
to
"°CO So "S
CD
o
ä W Son
' Ö J T3 CM
A
CQ ^ CO
a ^
« o J o i cO
Sw n
? -IS a ^ ;Ä
cd iJ s°!U: O
«-I
CO
c
•—i « : .3 B S ä ü a
! cd '^

2 'S W b0£
CO J
£0
O
""
CQ
E
0 cd B
>i CQ '
CH G
S | § '3 M a
a
tS
■a*
0
"O .2 43 So a CD CO Z M a || •° o »a 0
C
3 33 +3 G a G OT
;
a to cj B 'SS a T3
CD O CD in 43
2
c, S °5
u a> o £ co a ■a FH a« 3
ag rh (=; z ei rj
o a. CJ o B
CO 3 CD 73 O
o a-*
m
^
?> "*s
S ■ u >> 0 0 CQ
< fa >> <" < fa u0
H n H <
01 CO •O S
CJ
-* C• .w
S CD
B
0 CQ
Sa m2t,u
c- &K
a E -Q 5 §2^d
/§43 0^-a g a a
>, CO
CD ^
ä a
zixy gS 'S -o ?
o Q.O
CQ H ™ a a '
^ "r1 t_. 3 T-(
.2 i .. ™
co H 2G o. i> CO b0 CD ™ CD §« CQ bD dJ
w
*• u 0
a
qj
» T3
> .2 'S . •°< go» H -a
C en o -I ! o T3 -u «« *■ '
Ö "O " "
0 <Ü (H

§ fa m
B to
'V 3
CO h
co m • <C i oM
a 5 " T3 Ä T3
co • < i -S O CO a bo 0
CO k, CO 2 o CO g
bo cp
_
CO SH co z H tfco O, 0 cd c T3
Zgtüü< . C B a c z g pä ü < rt a 3 1 u c
B^
£ 3 SHE- gz £ ^ bO 'ro «
■ 3 55 H S H S Z ^ B B 5 3 «>
o aJ i u) D 3
O
c* S ^ —<
3 <B
CO *" S3| So md bo o CO

CM o H
be a
5 a h °< o >^ S CO
CM o H
E ft
< CO . a - O > T3 Ü t, CO X!
0 a
a
i u FH XI CO •a ■- 73 £3 SH
i u H « §.22 a
•M y
B C B rt
%° ■'S ^Sfa° Ills v «"a
CO CD
z^* O rt CO CD 0j2
H gg
Cj
CO o CD t« §
T3 C
S.S«a S ö 2 g .2' S a .s CD B
H u o 8 % 0-
0 0
< 5 w O Q 8 3 oT CD
3 CD 3 0*i > Ö ^ J3
bo u
< o
tU H Z H _, cj
a0 o0 -sa S» °0 .§43
bo cj
H Z H_ cj
B M s g 0 ., CO -43 ri bo 0"B c In a O 43 In
< cd F_ O H < ^ t
< rt H O H < ^ "C
Z « O cd a > CD
o ° 3 a°£ §g(3
+J Z Z CQ Z P3 U cd a
^ T3 o ° 3 3 co o H cO ■M §1 C
 "The aeronautical and space activities of the United States shall be
conducted so as to contribute . . . to the expansion of human knowl-
edge of phenomena in the atmosphere and space. The Administration
shall provide for the widest practicable and appropriate dissemination
of information concerning its activities and the results thereof."
—NATIONAL AERONAUTICS AND SPACE ACT OF 1958

NASA SCIENTIFIC AND TECHNICAL PUBLICATIONS

TECHNICAL REPORTS: Scientific and technical information considered

important, complete, and a lasting contribution to existing knowledge.

TECHNICAL NOTES: Information less broad in scope but nevertheless

of importance as a contribution to existing knowledge.

TECHNICAL MEMORANDUMS: Information receiving limited distri-

bution because of preliminary data, security classification, or other reasons.

CONTRACTOR REPORTS: Technical information generated in con-

nection with a NASA contract or grant and released under NASA auspices.

TECHNICAL TRANSLATIONS: Information published in a foreign

language considered to merit NASA distribution in English.

TECHNICAL REPRINTS: Information derived from NASA activities

and initially published in the form of journal articles.

SPECIAL PUBLICATIONS: Information derived from or of value to

NASA activities but not necessarily reporting the results of individual
NASA-programmed scientific efforts. Publications include conference
proceedings, monographs, data compilations, handbooks, sourcebooks,
and special bibliographies.

Details on the availability of these publications may be obtained from:

SCIENTIFIC AND TECHNICAL INFORMATION DIVISION

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

Washington, D.C. 20546