Sie sind auf Seite 1von 6

P .

P A R I S
1
Associate Professor of Mechanics and
Assistant Director of the Institute of Research,
Lehigh University, Bethlehem, Pa,
F . E R D O G A N
Professor of Mechanical
Engineering, Lehigh University,
Bethlehem, Pa. Mem. ASME
A C r i t i c a l A n a l y s i s o f C r a c k P r o p a g a t i o n L a w s
The practice of attempting validation of crack-propagation laws (i.e., the laws of Head,
Frost and Dugdale, McEvily and I llg, Liu, and Paris) with a small amount of data,
such as a few single specimen test results, is questioned. I t is shown that all the laws,
though they are mutually contradictory, can be in agreement with the same small sample
of data. I t is suggested that agreement with a wide selection of data from many speci-
mens and over many orders of magnitudes of crack-extension rates may be necessary to
validate crack-propagation laius. For such a wide comparison of data a new simple
empirical law is given which fits the broad trend of the data.
s *
I n t r o d u c t i o n
IEVEHAL crack-propagati on laws have been pre-
sented in the past few years, whi ch claim to be verified by the
experimental data analyzed in their respecti ve papers. They
are specifically the work of Head [l ],
a
Frost and Dugdal e [2],
McE vi l j ' and I llg [3 ], L i u [4, 5 ], and Paris, Gomez, and Anderson
[6,7,8].
Thi s paper will attempt to show that basing val i dati on of a
crack-propagati on theory on a limited amount of data is a poor
test of any theory. Moreover, a wide range of test data is now
available in a conveni ent form [9] whi ch may be empl oyed to
anal yze cri ti cal l y these several crack propagati on laws.
Therefore the paper will consist of three parts whi ch will i n-
cl ude:
1 A revi ew and compari son of existing crack-propagati on
laws.
2 The erroneous results obtai ned by compari ng each l aw wi th
a l i mi ted range of test data.
3 The results of compari ng crack-propagati on laws wi th a
wi de range of data.
A R e v i e w o f E x i s t i n g C r a c k - P r o p a g a t i o n L a w s
Crack-propagati on laws gi ven in the literature take many
forms. I n general they treat cracks in infinite sheets subjected
to a uni form stress perpendicular to the crack (or can be appl i ed
to that confi gurati on) and they relate the crack l ength, 2a, to the
number of cycl es of l oad applied, N, with the stress range cr, and
material constants C,-.' The single form in whi ch all crack-
propagati on laws may be written is
da
dN
= / ( f , a, Ci) ( 1 )
Therefore this format will be empl oyed in the discussion to
fol l ow.
Chronol ogi cal l y the first crack-propagati on l aw whi ch drew
wi de attenti on was that of Head [1] in 1953. He empl oyed a
mechanical model whi ch considered rigid-plastic work-hardeni ng
elements ahead of a crack ti p and elastic elements over the re-
mainder of the infinite sheet. The model required extensive
calculations and deducti ons to obtai n a law which may be
written as
da
dN
C^a'*
(C2 - 0>' / '
(Head's l aw) (2)
where Ci depends upon the strain-hardening modul us, the modu-
lus of elasticity, the yield stress and the fracture stress of the
material, and Ci is the yi el d strength of the material. Head de-
fined Wo as the size of the plastic zone near the crack ti p and pre-
sumed i t was constant during crack propagati on. However,
Frost [2] noti ced that the pl asti c-zone size increased in direct
proporti on to the crack length in his tests. Moreover, I rwi n [10]
has recentl y poi nted out from anal yti cal considerations that
Wo (3)
for the confi gurati on treated here whi ch is in agreement wi th
Frost's conclusions. Therefore, though Head adopted equati on
(2) with w0 considered constant as his crack-propagati on law, we
are forced here to i ntroduce equati on (3) i nto equati on (2) to
obtai n a modi fi ed or corrected form of Head's crack-propagati on
l aw; i.e.,
1
Also Consultant to the Boeing Company, Transport Division,
Renton, Wash.
2
Numbers in brackets designate References at end of paper.
3
I n this discussion a will imply the stress-fluctuation range and Ci
will vary slightly with mean stress in a general interpretation.
Contributed by the Metals Engineering Division and presented at
the Winter Annual Meeting, New York, N. Y., November 25-30,
1962, of T H E A M E RI CA N SOCI E T Y OF M E C H A N I C A L E N GI N E E RS.
Manuscript received at ASME Headquarters, April 11, 1962. Paper
No. 62WA-234.
da C3o~
2
a
dN
=
(C2 - o)
(Head's corrected l aw) (4)
Frost and Dugdal e [2] in 1958 presented a new approach to
crack-propagati on laws. They observed as was i ntroduced in
equati on (4), that Head's law should be corrected for the vari a-
tion of pl asti c-zone size with crack length. They deduced that
the corrected result, equati on (4), depends linearly on the crack
length a. However, they also argued by dimensional analysis
N o m e n c l a t u r e -
a = half crack length
Qo = initial half crack l ength
B = a functi on of stress range (and
mean stress)
C = a numerical constant
C; constants (whi ch vary slightly
with mean stress)
Ds = constants which depend on
stress level
/( ) = a functi on of
F { } = a functi on of
G { } = a functi on of
k = stress-intensity factor (range)
IClW = stress-concentration factor
m = a numerical exponent
M = a material constant
n = a numerical exponent
N = cycl e number
A
r
o, Nj , initial and final cycl e number,
respectively
da
= crack extension per cycl e of
d N
l oad
Ak = stress-intensity-factor range
p, pi = crack-ti p radius
cr = applied stress (range)
UQ = stress at a crack tip (wi th
finite radius p)
crnot = net section stress
w0 = pl asti c-zone size
5 2 8 / D E C E M B E R 1 9 6 3 T r a n sa c t i o n s o f t h e A S M E
Copyright 1963 by ASME
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 09/15/2013 Terms of Use: http://asme.org/terms
that the incremental increase in crack l ength da, for an i ncremen-
tal number of cycl es dN, should be di rectl y proporti onal to the
crack l ength a. Hence they concl uded that (i ndependent of
Head's model )
da
dN
Ba (5)
fol l owi ng arguments: I rwi n's [12] stress-intensity-factor k re-
flects the effect of external l oad and confi gurati on on the i n-
tensi ty of the whol e stress field around a crack ti p. Moreover,
for various confi gurati ons the crack-ti p stress field always has the
same form (i .e., di stri buti on). Therefore i t was reasoned that
the intensity of the crack-ti p stress field as represented by k
should control the rate of crack extension.
4
That is to say:
where B is a functi on of the applied stresses. Then they ob-
served that in order to fit their experimental data:
dN
= ( ? {f c } (15)
B =
C;
Combi ni ng equati ons (5) and (6) they obtai ned the l aw:
( 6)
da cr
3
a
= (Frost and Dugdal e's l aw) (7)
dN Ci
About the same ti me McE vi l y and I llg [3] modi fi ed a method
of analysis of stati c strength of plates wi th cracks used at NASA
to obtai n a theory of crack propagati on. Thei r arguments were
as fol l ows: Presuming that a crack ti p in a material has a charac-
teristic (fi cti ti ous) radius pi, whi ch allows computati on of the
stress (To, in the el ement at the crack ti p using elastic stress-
concentrati on factor concepts, the stress a0 is
(To = Kn(T net (8)
I CN = 1 + 2 (a/p,)'A
(9)
which is based on the elastic sol uti on for an elliptical hol e of
semi major axis a and end radius pi.
<Tnet = <J (10)
and substituting equati ons (9) and (10) i nto (8) gi ves
tr = <r[l + 2 (a/p,)'^] (11)
Based on considerations that under cycl i c l oadi ng work-hardeni ng
at the crack ti p will raise the l ocal stress to a fracture stress, they
concl uded that the crack-extensi on rate will be a functi on of <ra or
da
dN
- = F{tr} = F{KN<rnei\ (12)
(McE vi l y and I l l g's l aw)
Therefore for the special confi gurati on of interest here, i.e., i n-
troduci ng equati ons (8), (9), (10), and (11) i nto (12), we have
S - ' K - O T <
u
>
which is the desired form in this discussion upon considering pi to
be a material constant, in likeness to the Cs.
McE vi l y and I llg go on in an empirical manner to obtai n the
form of the functi on F { }, and suggest
logi 00509 ffiVcrnel - 5.472 -
34
(14)
KNane, - 34
(McE vi l y and I l l g's l aw empirically extended)
I t should be noted by the reader that McEvi l y and I l l g's laws,
equati ons (12) and (14), are not restricted to the special confi gura-
ti on here, as was the case for the laws of Head, and Frost and
Dugdal e. The appl i cabi l i ty to other configurations and an addi -
tional similarity to Paris' work [6, 7, 8] will warrant later com-
ments.
I ndependent of McE vi l y and I llg, Paris [11] proposed a crack-
propagati on theory at about the same time. I t is based on the
I n treating the special confi gurati on of interest in this di scussi on,
i t should first be observed that (11)
k = < r a ( 1 6 )
whereupon equati on (15) may be specialized to read:
da
dN =
"A!
(17)
Somewhat later, Li u [4] restated Frost and Dugdal e's [2] di -
mensional analysis in a much more elegant form and argued that
the crack-growth rate should depend linearly on the crack
l ength; i.e.,
da
= Ba (L i u's l aw)
dN
( 18)
where K N is the stress-concentration factor and <rnet is the net
area stress at the cracked secti on. For the confi gurati on used
here, i.e., an infinite pl ate wi th uni form stress a
which is the same result as equati on (5). Li u then presumed
that B was in general a functi on of stress range a (and mean
stress); i.e.,
B = B(o-) (19)
I n a subsequent work, Li u [5] notes that mean stress is of
secondary influence and, using a model of crack extension em-
pl oyi ng an idealized elastic-plastic stress-strain diagram and a
concept of total hysteresis energy absorpti on to failure, reasons
that
B{<T) = C6o-
2
whi ch combi ned with equati on (18) gi ves
da
= C5o"'a (L i u's modi fi ed l aw)
dN
( 20)
(2 1)
T h e E q u i v a l e n c e o f K
N
t o k
Hardrath [14] observed that KNanet for the special confi gura-
ti on empl oyed here, from equati ons (8) and (10),
KN<rnet = u[l + 2(a/p,)'/]
is similar to the stress-intensity factor, equati on (16),
k = <sa
h
(22)
(23)
if pi is small compared to a. For aluminum al l oys (2024T3 and
7075T6) his colleagues McEvi l y and I l l g [3] had already observed
that pi is less than 0.005 in. so that the condi ti on pi <K a is in fact
present for cracks of a readi l y observabl e l ength in crack-propaga-
tion tests. Thus Kj/O'net and k are known to concur for this
special confi gurati on and in addi ti on a proof that they are
equi val ent in general will be offered here.
I rwi n [12] observed that the equati ons for the stress field sur-
roundi ng the ti p of a sharp crack contai ns the factor fc/(2r)
1/,J
which implies a singularity of stress at the crack tip. Now, if a
hol e of radius p is drilled at the crack ti p, the maxi mum stress on
the peri phery of the hol e will be proporti onal to k/ (2p)
1
^ or
<r = CT/(2p)'A (24)
whi ch applies for any p whi ch is small compared to other planar
4
At about this same time Martin and Sinclair [13] attempted un-
successfully to correlate crack-extension rates using a similar parame-
ter but did not observe a correlation which does in fact occur.
J o u r n a l o f Ba si c En g i n e e r i n g D E C E M B E R 1 9 6 3 / 5 2 9
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 09/15/2013 Terms of Use: http://asme.org/terms
dimensions of any confi gurati on consi dered, and C is a constant
i ndependent of the confi gurati on. Fi nal l y, cr0 in equati on (24) may
be interpreted to be the same as KN<TNET, from equati on (8) and
its accompanyi ng discussion, if p is taken equal to pi. Then
K N a net = ffo
Ck
(2 p,)'
A
(25)
I n order to evaluate C, the results for the special confi gurati on
considered herein may be empl oyed, i.e., substituting equati ons
(22) and (23) i nto (25) and noting equi val ence for small pi gives
C = lim
= lim
pi>-0
a
[_
1 +2
(tO
a
]
( 2 p , ) I/ !
=
2
Putti ng this result i nto equati on (25) and rearranging,
Kn <rnet pi
1
/ '
k l i m
pi-t-0
(26)
(27)
which implies the general equi val ence of /fA.<J net and k.
S i m i l a r i t i e s B e t w e e n C r a c k - P r o p a g a t i o n L a w s
The result of the foregoi ng discussion, equati on (27), implies
the di rect similarity of McE vi l y and I l l g's law, equati on (12), and
Paris' result, equati on (15). A choi ce between the two is stri ctl y
dependent on a matter of conveni ence and cl ari ty of accompanyi ng
concepts in empl oyi ng one or the other.
6
The laws of Head, equati on (4), Frost and Dugdal e, equati on
(7), and Li u, equations (18) and (21), can all be approxi mated by
the form
da <T"a'
n
dN
=
( 28)
for the special confi gurati on whi ch is treated. Now, i t is evi dent
that Paris' result for this confi gurati on, equati on (17), i mpl i es:
n
~2
(29)
whi ch can also be deri ved from McE vi l y and I l l g's result, equa-
ti on (13), for pi small compared to a.
The laws of Head, and Frost and Dugdal e do not qui te
agree wi th appl yi ng equati on (29) to equati on (28), however,
L i u's l aw does concur wi th the specified form.
I t is perti nent to now show that determining m and n from a
limited quanti ty of data is a doubtful practi ce. That is to say
that pl otti ng data from single test specimens on a l ogari thmi c or
semilogarithmic graph on which laws such as Head's, Frost's, and
L i u's predi ct straight line relationships is not a reasonable test
of the val i di ty of a crack propagati on l aw.
E r r o n e o u s E v a l u a t i o n o f D a t a F r o m S i n g l e T e s t s
A typical crack-propagation test consists of a wi de plate with a
central crack of length 2a, subjected to uni form tension <r, re-
peatedl y applied. Duri ng a single test then tr is a constant and
data consisting of crack lengths and corresponding cycl e numbers
are obtained. L et the probl em of examining each of the previous
crack-propagation laws for a particular test be formul ated.
I n a constant maxi mum repeated stress-level test Head's law,
equati on (2), is reduced to
da ,,
^ =
a A
(30)
6
For very small cracks there is a difference between the two
theories which is left unresolved.
200000 ei o o o o 2 2 0 000 230 OOO
, N0.I5
575 000 600 000 625 0 0 0
NO. 10 ,
102 000 103000
F i g . 1 T yp i c a l d a t a f r o m Ma r t i n a n d S i n c l a i r o n He a d ' s s u g g e s t e d p l o t
;
TESTS BY MARTI N 8 SI NCLAI R [.13} !
SYMBOL SPECIMEN STRESS
MO. 10 27027 PSI
NO. 14 23 760
O NO. 15 21978
/ / '
NO 13 ,
575 000 600 000 6250 0 0
NO. 14
200 000 2 J O 000 220 000
, NO. 10
230000
[ 02000 105000 110 000 115000
N ( eyel o)
F i g . 2 T h e s a m e d a t a f r o m Ma r t i n a n d S i n c l a i r o n F r o s t 's o r L i u ' s s u g -
g e s t e d p l o t
TESTS BY MARTIN 8 SINCLAIR [|3J
SYMBOL SPECIMEN STRESS
F i g . 3 A g a i n t h e s a m e d a t a f r o m Ma r t i n a n d S i n c l a i r o n Mc E v i l y' s o r
P a r i s ' s u g g e s t e d p l o t
5 3 0 / D E C E M B E R 1 9 6 3
T r a n sa c t i o n s o f t h e A S M E
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 09/15/2013 Terms of Use: http://asme.org/terms
where D\ is a constant for a given stress range a. I ntegrating this
expression gives
rr = DiN + const
n '
2
(31)
Therefore Head suggested plotting l/ a^ versus N and, observing
equation (31), implied that obtaining a straight line for each
specimen whose data are plotted this way indicates verification of
his law.
The data of Martin and Sinclair [13] are employed here as an
unbiased source. From their data specimens nos. 10, 14, and 15
were chosen since those specimens were run at medium stress
levels and the greatest number of data points per specimen was
recorded for them. Fig. 1 shows the type of plot suggested by
Head and it is noticed that portions of the data do form straight
lines. Does this mean that Head's theory is verified? The
reader is warned that this might be a hasty conclusion.
First, consider the corrected law of Head, Frost and Dugdale's
law, and Liu's law, equations (4), (7), (18), and (19), respec-
tively. For a constant stress-range test all these laws reduce
to the form
da
dN
= D
*
a
I ntegrating this result gives:
log a = D2N + const
(32)
(33)
(34)
for a given material. T o test these theories, data from several
specimens should be plotted on a k versus da/ dN (or log da/ dN)
graph and these laws predict that the data of several specimens
will form a single curve. Again the same data of Martin and
Sinclair are plotted in Fig. 3. Noti ce that the points which do
not fall on the straight-line portions of Figs. 1 and 2 seem to be
perfectly acceptable here. Since the data were differentiated
to obtain da/ dN, an additional amount of scatter is introduced.
Therefore the data also imply verification of these two laws as
well.
Hence, all of the laws agree with the data and, since these laws
are not identical, the method of verification of crack-propagation
theories from a limited amount of test data is evidently in error.
An alternative approach employing a wider range of test data
must be employed.
C o m p a r i s o n o f C r a c k - P r o p a g a t i o n L a w s W i t h a W i d e
R a n g e o f T e s t D a t a
I n previous work [6, 8] i t was observed that data from several
sources [3, 9, 13] may be plotted in the form of Fig. 3 (semi-
logarithmic) to obtain a single curve on a range of stress-in-
tensity factor, Ak (corresponding to stress range) versus log
(da/ dN) diagram, where da/ dN covers as many as 6 log cycles.
However, replotting these data on a log Ak versus log (da/ dN)
graph reveals some pertinent results. The data are replotted i n
Fig. 4 and the three specimens from Martin and Sinclair's work
used in the foregoing for the purpose of illustration are shown to
indicate their concurrence with the general trend.
The authors are hesitant but cannot resist the temptation to
draw the straight line of slope 'A through the data in Fig. 4. The
equation of this line is observed to be
Therefore plotting log a versus N and obtaining straight lines has
been accepted as verifying these theories. Fig. 2 is this type of
plot employing the very same test data point by point of Martin
and Sinclair shown in Fig. 1. Again, the test data have straight-
line portions. Therefore it now seems to verify both Head's law
equation (30) and the other laws with the form of equation (32).
But the theories represented by equations (30) and (32) are
not the same!
Now let the laws of McEvi l y and I llg, and Paris, equations
(12) and (15), be examined in the light of these same data. Re-
calling equation (27) these two laws are equivalently expressed
by
da
dN
(Aky
M
(35)
Equation (35) fits the data almost as well as McEvi l y and I llg's
extended law, equation (14), and is considerably simpler in form.
The advantage of this form is clarified by considering its ap-
plication to the configuration employed earlier, in which case i t
becomes
da
dN M
(36)
The laws of Head, Frost and Dugdale, and Liu depend on a in a
manner other than a
2
. Clearly, their laws are at variance with
the data trend in Fi g. 4; i.e., Head predicts the slope of
l
/z and
Frost and Dugdale as well as Liu would predict '/* as indicated
in the figure.
Moreover, if one might be so bol d as to attempt to integrate
equation (36) i t becomes
J _
o
M
(N - N) (37)
10
AK
I0
4
'
O NASA
LIU
A MARTIN a SINCLAIR
X BOEING
LI U
- O - O - o j S ^ V
SLOP E 1/2
10
_L_
- 7
10
- 6
10
d(2o)
1 0 dN 1 0
F i g . 4 B r o a d t r e n d o f c r a c k - g r o w t h d a t a o n 2024- T 3 a l u m i n u m a l l o y
UL
e ye .
- 3
10 10
J o u r n a l o f Ba si c En g i n e e r i n g
D E C E M B E R 1 9 6 3 / 5 3 1
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 09/15/2013 Terms of Use: http://asme.org/terms
where a and No are the initial crack size and cycl e number, re-
specti vel y. For final failure of a speci men, the number of cycl es
may be equated to N r , as a approaches infinity whereupon equa-
tion (37)reduces to
N P =
M
cr'on
(38)
which does in fact resemble an S-N diagram. The authors do not
wish to i mpl y that equati on (38) is di rectl y useful, but deri ve i t
here to show that such considerations do not i mmedi atel y lead to
the conclusion that equati ons (35) or (36) are in error.
D a t a F r o m W e d g e - F o r c e T e s t s
The results of wedge-force tests of the confi gurati on shown in
Fi g. 5 are useful in cri ti cal l y exami ni ng the dimensional analyses
of Frost and Dugdal e, and Liu in deri vi ng their crack-propagati on
laws. I f the sheet is infinite, the onl y characteristic dimension
of the probl em is a. As the crack grows, stri ctl y dimensional
arguments i mpl y that for di fferent crack lengths oi and a? the
incremental rates of crack extension are
dN Oi dN a.2
which mi ght be thought val i d for all crack lengths or
5 =
S a
<
40
>
dN
Thi s result is the basis of equati ons (5) and (18).
I n wedge-force tests Donal dson and Anderson [9] observe that
the crack grows slower as i t gets larger. Thi s is contrary to equa-
F i g . 5 W e d g e - f o r c e l e s t c o n f i g u r a t i o n
ti on (40) and hence weakens any arguments based on dimensional
analyses which do not include further considerations. Therefore
the crack propagati on laws of Head, Frost and Dugdal e, and Li u
must be considered as unclarified, if not total l y in error.
Moreover, Fi g. 6 is a graph, similar in form to Fi g. 4, but for a
di fferent materi al ; i.e., 7075T6 aluminum al l oy. The data shown
are from two i ndependent sources [3 and 9] and in addition
wedge-force test data from tests empl oyi ng confi gurati on of Fi g. 5
are also pl otted. No further comment seems necessary.
C o n c l u s i o n
The results here indicate that the practi ce of using data from
single test specimens is not a sensitive eval uati on of a crack-
propagati on l aw's val i di ty. Randoml y chosen data from single
specimens analyzed here leads to an apparent agreement of
several contradi ctory laws to the same test data.
For that reason the authors suggest that laws which correlate
a wide range of test data from many specimens are perhaps the
"cor r ect" laws. The results at least i ndi cate that hasty concl u-
sions have been drawn in many earlier works whi ch should be re-
examined before any gi ven crack-propagati on l aw is accepted as
valid.
A c k n o w l e d g m e n t
Thi s work was supported by the Boei ng Company, Transport
Di vi si on, Renton, Wash. Thei r encouragement, as well as
financial aid, is grateful l y acknowl edged.
R e f e r e n c e s
1 A. K . Head, "The Growth of Fatigue Cracks," The Philo-
sophical Magazine, vol. 44, series 7, 1953, p. 925.
2 N. E. Frost and D. S. Dugdale, "The Propagation of Fatigue
Cracks in Sheet Specimens," J ournal of the Mechanics and Physics of
Solids, vol. 6, no. 2, 1958, p. 92.
3 A. J . McEvi l y and W. I llg, "The Rate of Crack Propagation in
Two Aluminum Al l oys," NACA Technical Note 4394, September,
1958.
4 H. W. Liu, "Crack Propagation in Thin Metal Sheets Under
Repeated Loading," T HE J OURNAL OF BASI C ENGI NEERI NG, TRANS.
ASME, Series D, vol. 83, 1961, p. 23.
5 H. W. Liu, "Fati gue Crack Propagation and Applied Stress
RangeAn Energy Approach," J OURNAL OF BASI C ENGI NEERI NG,
TRANS. A S M E , Ser i es D , vol . 85, 1963, p. 116.
6 P. C. Paris, M. P. Gomez, and W. E. Anderson, "A Rational
Analytic Theory of Fatigue," The Trend in Engineering, vol. 13, no. 1,
J anuary, 1961, p. 9.
7 W. E. Anderson and P. C. Paris, "Eval uati on of Aircraft Ma-
terial by Fracture," Metals Engineering Quarterly, vol. 1, no. 2, Mav,
1961, p. 33.
8 P. C. Paris, "Crack Propagation Caused by Fluctuating
Loads," ASME Paper No. 62Met-3.
1 0
& K
>
1 0
O N A S A
A B O E I N G
* B O E I N G , W E D G E F O R C E T E S T S
4 -
,-i SxrV
r r o ' f r *
OjQ
- j O - O
- 6
1 0
- 5 d ( 2 a ) - 4 i n , " 3
1 0 d N 1 0 e ye . 1 0
- 2
1 0
F i g . 6 B r o a d t r e n d o f c r a c k - g r o w t h d a t a o n 7075- T 6 a l u m i n u m a l l o y i n c l u d i n g w e d g e - f o r c e t e s t s
532 / DE C E M B E R 1963 T r a n s a c t i o n s o f t h e A S ME
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 09/15/2013 Terms of Use: http://asme.org/terms
9 D. R. Donaldson and W. E. Anderson. " Crack Propagation
Beha'ior of Some Airframe Materials." presented at the Crack Propa-
gation SympoRium. Cranfield. England. September, 1961: proceed-
ings to be published. Also available on request to The Boeing Com-
pany. Transport Division. Document No. D6-7888.
10 G. R. Irwin . "Fracture Mode Transition for a Crnck Travers-
ing 1\ Plate." JOURNAL OF BASIC ENGINEERINO, TRANS. ASME, Series
n. yol. 82. no. 2, June. 1960. p. 417.
11 P. C. Paris, "A Note 011 the Variables Effecting the Rate of
Crack Growth Due to C:vclic Loading," The Doeing Company,
Document No. 0-17867, Addendum N, September 12. 1957.
12 G. R. Irwin. II Analysis or Stresses and Strains Near the End of
fl Crark Traversing n Plate." Journal 0/ Apph'ed lIfechanin. vol. 24.
TRANS. ASME. Y01. 79,1957, p. 361.
13 D. E. Martin and G. M. Sinclair. "Crack Propagation Under
nepf'aled Loading," Proceedings of t he Third U.S. National Congress
of Applied Mechanics, June. 1958. p. 595.
14 H. F. Hardrath and A. J . McEvily. "Engineering Aspects of
Fatigue Cmck Propug-aUon," presented nt the Crnck Propagation
Symposi um, Cranfield, England, September, 1961.
Journal of Basic Engineering
DECEMBER 1963 I 533
Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 09/15/2013 Terms of Use: http://asme.org/terms