N A S A TECHNICAL

MEMORANDUM

[NA S A -? :*' -X-7 2 825) I3P& 0V E3 S T R BSS-I IT E ISIT Y N77-27427

FACTCr COii SEflI-ELLIPTICAL SOBFACE CRACKS
IN fi4f.E-THICKNESS PLATES ( H A S A )3 1 p HC
'03/?!F A01 CSCL 20K Unclas
63/39 29151

;APROVED STRESS-ZNTENSITY FACTORS FOR SEMI-ELLIPTICAL

SURFACE CRACKS IN FINITE-THICKNESS PLATES

by
I . S . Raju and J.C. Newman, Jr.
Langley Research Center
Hampton, VA 23665

TECHNICAL PAPER to be presenteL at the 4th International Conference

on Structural Mechanics in Reactor Technology, August 13 - 19, 1977,
San Francisco, California, U.S.A.

This informal documentation medium i s used to provide accelerated or

special release of technical information to selected users. The contents
may not meet NASA formal editing and publication standards, may be r f i .
~ - -. -
'
.. '

vised, or may be incorporated in another publication. .- L - '. 4

is\
,

1%e> .
'%
1
'8 3
JUN 1377
RECEIVED '-.I
;

NATIONAL AERONAUTICS AND SPACE ADMlNlSTRAflON \..*NASA STI FACluu

INPUT BRANCH
. !,
;
LANGLEY RESEARCH CENTER, HAMPTON, VIRGINIA 23665 ,' :._.
y.c
 1 krab 3.km-1-m
1 NASA yx x-72825
4 Tmerrds~mnk E m o n ,

Improved S t r e s s - I n t e n s i t y F a c t o r s f c r Semi- ,

2530
7- r-Olruea-(Yo

I. S. lZaju* a n d J. C. Newman, Jr.

1owarLun((yo
~Morcrr*DI.*abor,wR.rdAdch 505-02-31-01
NASA-langley Research Center 11 conactorcmtuD

Hampcon, VA 23665
13. lrp d Roan rd hd
12 u A c r * b n e r r d -
T e c h n i c a l Memorandum
National Aeronautics and Space Administratio.il4. -
,
A c ~ h
W a s h i n g t o n , D.C. 20546 1

I n t h i s p a p e r s t r e s s - i n t e n s i t y f a c t o r s for s h z l l o w a n d d e e p s e m i -
e l l i p t i c a l s u r f a c e c r a c k s i n p l a t s s u b j e c t e d t o t e n s i o n are
presented. To v e r i f y t h e a c c u s c y o f t h e t h r e e - d i m e n s i o n a l
f i n i t e - e l e m z n t models e m p l o y e d , c o n v e r g e n c e w a s s t u d i e d b y v a r y i n g :
t h e number of d e g r e e s of f.resdom i n t h e models i i a n 1500 t o 6 9 0 0 . '
T h e 6900 d e g r e e s o f f r e e d o m < s e d h e r e were more t h a n t w i c e t h e
nwnber u s e d i n p r e v i o u s l y r e p o r t e d s o l u t i o n s .
I

17. K.r *h l s u l p c r d bv Aumorls)) (STAR o t q p r v underhndl Spmnm

18 D~rntb~ttm I
Fracture I
S t r e s s I n t e n s l t y Factor
i
Three-nimensional F i n i t e Element Unclassified - Unlimited
S u r f ace C r a c k s I
I

19. SIcuih o l i t . lot c h n m ) 20. S I M I t V clad. (of Ihn W) 21 NO. cf p.pl n.R s e ' I

Unclassified Unclassified 28 $4.00

 IWRCVED STRESS-INTE5ITY FACTORS F W SHI-ELLIPTIC3.L
SRFACE CRACKS I N FINITE-THICKNESS PLATES

I. S. Raju and J. C. Ne-, Jr.

NASA-Langley Resesrct! Center
Hampton, V i r g ‘ s i a

Surface cracks a r e among t h e nore common ‘.‘laws i n a i r c r a f t and

pressure vessel components. Accurate stress analyses of surface-cracked

components are needed for r e l i a b l e predictton of the* crack-growth

rates and A-acture strengths. Several crdculations cf stress-intensity

factors f o r semi-elliptical. &-face cr&cks subjected t o tension have

appeared i n t h e literature. However, some of t h e s e solutions are in

disagreenent by 50 t o 100 percent.

In t h i s paper stress-intensity f a c t o r s f e r s h a l l c w and deep P.emi-

e l l i p t i c a l surface cracks i n p l a t e s subjected t o t e r s i o n a r e presented.

To verify t h e accuracy of t h e three-dbensional finite-elemeot m o d e l s

employed, convergence was studied by varying the number of d:ggrees of

freedom i n t h e models from 1500 t o 6900. The 6900 degrees of freedan

used here were more thatl twice t h e number used i n previorsly reported

solutions .
 -2-

L'urface cracks are among t h e more cannon flaws i n a i r c r a f t and

pressure vessel components. Accurate stress analyses of t h e s e surface-

cracked components are needed f o r reiiable prediction of t h e i r crack-

p o w t t rates and l'racture strengths. Exact solutions t o these d f f f i c u t

problears eze not avajlable; therefore, approximate methods must 3e used.

For a semi-elliptical swface crack i n a finite-thickness p l a t e subjected

t o tension (Figure l), Brovning and Smith [I], Kobayashi [2], and Smith

and Sorensen [ 3 ] used t h e alternatiw method and Kathiresan [k] used

t h e finite-element method t o obtain the s t r e s s - i n t e n s i t y factor varia-

t i o n s along t h e crack front f o r various crack shapes. For a deep semi-

e l i i p t i c a l surface crack (with a / t = 0.8 and a/c = 0.2) subjected t o

tension, t h e stress-intensity f a c t o r s obtained by Smith and Sorensen [31,

Kobayashi [2], and Kathiresan [ 4 ] disagreed by 50 t o 100 percent. The

reasons f o r these discrepancies are not well understood.

This paper presents stress-intensity factcrs f o r shallow and deep

semi-elliptical surface cracks i n p l a t e s subjected t o miform tension.

To test t h e v a l i d i t y of tne present analysis, t w o crezk configurations,

both embedded i n a large body subjected t o unifom. tension, were analyzed:

(1) a c i r c u l a r (penny-shaped) &rack and (2) an e l l i p t i c a l crack. These

results a r e canpared with exact solutions from t h e i i t e r s t u r e [51. To

verify t h e accuracy of t h e soiutions f o r surface cracks i n f i n i t e -

thickness p l a t e s , convergence was studied by v a g i n g , from 1500 t o 6900,

 -3-

t h e number of degrees of freedom i n t h e finite-element models. The

6900 degrees of freedom used bere vere more than t v i e e t h e l a r g e s t

number used previously [ 41. These models vere composed of singular

elements around t h e crack Pront and isoparametric ( l i n e a r s t r a i n )

elements elsewhere. Mode I e l a s t i c stress-intensity f a c t o r s vere

calculat 3 by using a nodal force method.

The computations reported herein vere conducted on a unique

computer called t h e STAR-100 at t h e HASA Langley Research Center.

 -4-

depth of surface crack

half-width of cracked p l a t e

half-length of surface crack

Young's modulus of e l a s t i c i t y

stress-intensity boundary-correct ion factor

half-length o f cracked p l a t e

stress-intensity factor (Mode I )

shape factor f o r an elliptical crack

applied uniform stress

p l a t e thickness

Cartesian coordinates

Poisson's r a t i o

parametric angle of t h e e l l i p s e
 -5-

Several aLtempts [6-81 have been m a d e t o aevelop s p e c i a l three-

dimensional finite-elements t h a t account f o r t h e stress and s t r a i n

s i n g u l a r i t i e s caused by a crack. These elements have had assumed

displacement cr stress d i s t r i h t i o n s that siaulate t h e square-root

s i n g u l a r i t y of t h e stresses and s t r a i n s at t h e crack f r o n t . Such

three-dimensional s i n g u l a r i t y elements [g] were a l s o used herein f o r

t h e analysis of finite-thickness plates containing embedded e l l i p t i c a l

or s e m i - e l l i p t i c a l surface cracks (see, f o r a m p l e , Figure 1).

Finite-Element I d e a l i z a t i o n

Two ty-pes of elements (isoparametric and singular) [gl were used

i n combination t o model elastic bodies w i t h embedded e l l i p t i c a l cracks

o r s e m i - e l l i p t i c a l surface cranks. Figure 2 shows a t y p i c a l f i n i t e -

element model f o r an embedded c i r c u l a r crack i n a l a r g e body

(h/a = b/a = 5). T h i s model i d e a l i z e s one eighth of t h e body. Various

numbers of wedges were used t o form t h e desired configuration. Fig-

ure 2 ( a ) shows R t y p i c a l model w i t h e i g h t wedges. Each wedge i s

composed of elements t h a t a r e i d e n t i c a l i n p a t t e r n t o that shown in t h e

4 = constant plane. The arrangement of the elements around t h e crack

front i s shown i n Figure 2 ( b ) . The isoparametric ( l i n e a r s t r a i n )

elements (denoted as I ) were used everywhere except near t h e crack

front. Around t h e crack f r o n t each wedge contained e i g h t " s j n g u l a r i t y "

elements (S) i n t h e shape of pentahedrons. The "singularity" elements

 -0-

had square-root terms i n t h e i r assumed displacement d i s t r i b u t i o n and,

therefore, produced a s i n g u l a r stress f i e l d at t h e crack f r o n t . Details

of t h e formulation of these t p e s of elements are given i n Reference 9.

The finite-element model f o r t h e enbedded e l l i p t i c a l o r semi-

e l l i 2 t i c a l sdrface crack was obtained from t h e finite-element model f o r

t h e c i r c u l a r crack by using an e l l i p t i c trmsformation. This t r a n s f o r -

uation was needed because t h e stress-iritensity f a c t o r s must be evaluated

from e i t h e r crack-opening 5isplacements o r nodal forces along t h e normals

t o t h e crack front. I f (x,y,a) are t h e Certesian coordinates of a node

i n t h e circular-crack model and (x',y',z') are t h e coordinates o f t h a t

s m e node i n t h e e l l i p t i c a l - c r a c k model, then t h e transformation i s

given by

2 2
c - a

y' = v

and

z' = z

f o r x and z not, a t tP.e o r i g i n . Figure 3 shows how c i r c u l a r a r c s and

r a d i a l l i n e s in t h e x , z plane of t h e circular-crack model are t r a n s -

formed i n t o e l l i p s e s and hyperbolas, respectively, i n t h e x' ,z' p l a n e

0:' tkle e l l i r t i c a l - c r a c k model using equations (1). Because equations (11

a r e not v a l i d a t t h e o r i g i n , a c i r c l e of very small r a d i u s , a/1000, was

used near t h e o r i g i n in t h e c i r c u l a r crack model. The small c i r c l e

maps onto an extremely IiCFT\)W +ipse i n t h e x ' , z ' plane. The use of
 -7-

t h e s m a l l c i r c l e avoids ill-shaped elements near t h e o r i g i n i n t h e

e l l i p t i c a l - c r a c k model. Figure 4 shows a t y p i c a l finite-element model

of a f i n i t e p l a t e conteining ap e l l i ? t i c a l crack. The transformation

reduced t h e b/c r a t i o ; therefore, i n order t o maintain b/c -

> 4,

a d d i t i o n a l elements were added along t h e x'-axis t o eliminate t h e

influence of p l a t e width.

S t r e s s - I n t e n s i t y Factor

The s t r e s s - i n t e n s i t y f a c t o r i s a measure of t h e magnitude of t h e

stresses near t h e crack f r o n t . h d e r general loading, t h e stress-

i n t e n s i t y f a c t o r depends on three basic modes of deformation (ter-sion

and in- and out-of-plane shear). But here only t e n s i o n loading was

considered and, t h e r e f o r e , only Mode I deformations occurred. The

Mode I s t r e s s - i n t e n s i t y f a c t o r , K, a t any point along an e l l i p t i c a l

or semi-elliptical crack i n a f i n i t e p l a t e i s taken t o be

where S i s t h e applied stress, a is t h e crack depth, Q i s t h e shape

f a c t o r f o r an e l l i p s e and i s given by t h e square of t h e coniplete

elliptic integral 01' t h e second kind [ 51. The boundary-correction

f a c t o r , F ( a / t , s/c, $I), i s a function o f crack depth, crack length,

p l a t e thickness and t h e parametric angle of t h e e l l i p s e . The present

paper gives values f o r F as a function of a / t and 4 f o r a / c = 0.2 and

1.0. The a / t values ranged from 0.2 t o 0.8.

 -8-

The s t r e s s - i n t e n s i t y f a c t o r s from t h e finite-element models of

embedded e l l i p t i c a l and s e m i - e l l i p t i c a l surface cracks were obtained

by using a nodal f o r c e method, d e t a i l s o f which are given i n Reference 9.

In t h i s method, t h e nodal f o r c e s normal t o t h e crack plane and ahead of

t h e crack frmt are used t o evaluate t h e s t r e s s - i n t e n s i t y f a c t o r . In

contrast t o t h e crack-opening displacement method [61 , t h i s method

requires no p r i o r assumption o f e i t h e r plane stress o r plane s t r a i n .

For a surface crack i n a f i n i t e p l a t e , t h e state of stress v a r i e s from

plane strain i n t h e i n t e r i o r o f t h e p l a t e t o plane stress a t t h e

surface. Thus, t h e crack-opening displacement method could y i e l d

ermneuus s t r e s s - i n t e n s i t y f a c t o r s .

RESULTS AND DISCUSSION

In t h e following s e c t i o n s a c i r c u l a r (penny-shaped) crack and an

e l l i p t i c a l crack completely embedded i n a l a r g e body su>.fected t o

uniform tension were analyzed using t h e finite-element method. The

calculated s t r e s s - i n t e n s i t y f a c t o r s f o r t h e s e crack configurations are

compared with t h e exact s o l u t i o n s to v e r i f y t!ie v a l i d i t y of t h e present

finite-element method.

For a semi-ci: .ular and s e m i - e l l i p t i c a l surface crack i n a f i n i t e -

thickness p l a t e , convergence of t h e s t r e s s - i n t e n s i t y f a c t o r s was studied

while t h e number of degrees of freedom i n t h e finite-element models

ranged from about 1500 t o 6900. The s t r e s s - i n t e n s i t y f a c t o r v a r i a t i o n s

along t h e crack f r o n t f o r semi-circular ( a / c = 1) and s e m i - e l l i p t i c a l

 -9-

(a/c = 0.2) surface cracks were obtained as functions of a / t v i t h

h/c = b/c -
> 4. Whenever possible, these s t r e s s - i n t e n s i t y f s 2 t c r s are

compared with r e s u l t s from t h e l i t e r a t u r e .

Exact Solutions

I n t h i s s e c t i o n a comparison i s made between t h e s t r e s s - i n t e n s i t y

f a c t o r s calculated from t h e finite-element a n a l y s i s and from t h e exact

solution [ 5 ] f o r an embedded c i r c d a r (penny-shaped) crack (a/c = 1)

and 8x1 embedded e l l i p t i c a l crack (a/c = 0.2) i n an i n f i n i t e body. In

t h e finite-elenent mQdel, h and b were taken t o be l a r g e enough that

t h e f r e e boundary would have a n e g l i g i b l e e f f e c t on stress i n t e n s i t y .

The boundary correction on stress i n t e n s i t y f o r a c i r c u l a r crack i n a

cylinder of radius b w i t h b/a = 5 i s about one percent [lo]. Therefore,

t o simulate a l a r g e body t h e finite-element model w a s assigned t h e

dimensions h/a = b/a = 5 , along with 3078 degrees of freedam. The

calculated s t r e s s - i n t e n s i t y f a c t o r s along t h e crack f r o n t from t h e

c i r c u l a r crack model were about 0.4 percent below t h e exact soli .: ion.

The embedded e l l i p t i c a l crack ( a / c = 0.2) nodel, with h/c = h/c = 4

and t / a = 5 , had 3348 degrees of freedom. The influence of f i n i t e

boundaries on stress i n t e n s i t y was estimated t o be about one percent.

The finite-element a n a l y s i s gave s t r e s s - i n t e n s i t y f a c t o r s along t h e

crack f r o n t generally within one percent of t h e exact solution, except

i n t h e region of sharpest curvature of t h e e l l i p s e , where t h e c a l c u l a t e d

values were abo-it 3 percent higher than t h e exact solution. Further

refinement i n t h e mesh s i z e i n this a r e a gave more accurate stress-

intensity factors.
 -10-

hecause t h e present. method yielded s t r e s s - i n t e n s i t y f a c t o r s f o r

completely embedded c i r c u l a r and e l l i p t i c a l cracks within 0.4 t o

3 percent of t h e exact solutions, t h e method w a s considered s u i t a b l e

f o r analyses of more complex configurations, provided t h a t enough

degrees of freedom were used t o obtain good convergence.

Approximat e Solutions

To v e r i f y t h a t t h e finite-element meshes used f o r t h e c i r c u l a r

and e l l i p t i c a l crack models were s u f f i c i e n t t o analyze cracked p l a t e s

w i t h free boundaries, through-the-thickness cracks with crack length-

to-width r a t i o s ranging from 0.2 t o 0.8 were analyzed under plane

s t r a i n assumptions. The meshes used here were exactly t h e same as

those which occur on t h e Cp = ~ / plane

2 of t h e c i r c u l a r and e l l i p t i c a l

crack models. These meshes were t h e n used t o model t h e cezter-crack

tension specimen. For crack length-to-width r a t i o s ( c / b ) ranging from

0.2 t o 0.6, t h e finite-element r e s u l t s were within 1.3 percent of t h e

approximate solutions given i n Reference 10. For c/b = 0.8, t h e f i n i t e -

element r e s u l t was 2 percent below t h e s o l u t i o n given i n Reference 10.

These results i n d i c a t e t h a t t h e mesh p a t t e r n along azy plane has

enough degrees of freedom t o account f o r t h e influence of free

boundaries on t h e s t r e s s - i n t e n s i t y f a c t o r .

Convergence

In t h e previous section t h e mesh p a t t e r n c.long any Cp = constarit

plane was found t o be s u f f i c i e n t t o acccunt f o r t h e influence of f r e e

 -11-

boundaries even f o r very deep cracks. I n t h i s s e c t i o n , t h e mesh

p a t t e r n i n t h e angular d i r e c t i o n , 4, i s studied. Fig- 5 arid 6 show

t h e r e s u l t s of a convergence study on t h e s t r e s s - i n t e n s i t y f a c t o r s f o r

a semi-circular surface crack and a s c m i - e l l i p t i c a l surface crack,

respectively, i n a f i n i t e - t h i c k n e s s p l a t e . The weer numbers of degrees

of freedom are associated with smaller wedges i n t h e 6-direqtion and

a more accurate representat ion of t h e crack shape.

For t \ e semi-circular crack (Figure 5 ) , h/a = b / a = 5 and

a / t = 0.8. This configuration was chosen because t h e c l o s e proximity

of t h e back surface t o t h e crack f r o n t was expected t o cause d i f f i c u l t y

in achieving zonvergence. The number of degrees of freer?mi ranged from

1500 t o 6195. The two f i n e s t models (4317 and 6195 degrees of freed.om)

gave s t r e s s - i n t e n c i t y f a c t o r s within about one perceEt of each other.

Therefore, t h e moc'A with 4317 degree: of fretlom WELS used subsequently

t o obtain s t r e s a - i n t e n s i t y f a c t o r s as a funci.ion c f .~/t.

Figure 6 shows t h e convergence study f o r a s e m i - e l l i p t i c a l

surface c r m k (< : c = 0.2). The h/c and b/c r a t i o s were equal t o four

and, again, a / t was chosen as 0.8. The number of degrees of freedom

ranged from io92 t o 6867. The model with 4797 degrees of freedom gave

r e s u l t s within about one gercent of those from t h e f i n e s t model.

Therefore, t h e model with 4797 degrees of freedom was used subsequently

t o obtain s t r e s s - i n t e n s i t y f a c t o r s and crack-opening d i s p l a c m e n t s as

a func+,icn of a / t .
 - 12-

S e r i i - C i r c u l a r Surface Crack i n a Finite-Thickness Plett

Figure 7 shovs and "able I p r t s c n t s t h e s t r e s s - i n t e n s i t y f a c t o r s

f o r a semi-circular surfiice crack i n a f i n i t e - t h i c k ? e s s p l a t e as 9

function of t h e parmetric angle:, 4, and t h e crack depth-to-plate

t h i c h t s s r a t i o , a/t. Bear t.he i n t e r s e c t i o n of t h e cracic and t b e free

surfaLe, tht s t r e s s - i n t e n s i t y factor increases mre r a p i d l y with a/t than

at t h e d e e w s t point (e = n.'2). For each value of a/t, the .-tress-intensity

factor calculate3 from t h e mdel with 4 3 7 de ;rees of freedon is largest

at t h e free surface ($ = 0 ) .

For a semi-c5rcular sawface crack uith a / t = 0 . 5 5 , the stress-

i n t e n s i t y f a c t o r s talc-deted by Srowning and S a i t c Ill, vho used the

alternating method, ve;e about 1 i o 3 p r c e n t b e k x the present results

for various - d u e s of 9 -

Semi-Elliptical S-ufsce Crack it J finite-Tnickness P l a t e

?igure 8 shovs aqd Table I p - s e n t s t h e s t r e s s - i n t e q s i t y f a c t o r s

f o r a semi-elliptical sm-face crack (a/c = 0.2) i n a finite-thickriess

p l a t e as a function of t h e p a r m e t r i c anqle, I$,and *e crack Ieptk-to-

p l a t e thickriess r a t i a , s/t. For each valtle o f a / t , t h e maximii stress-

interlsity f a c t o r occurs at t h e deepcst point ( @ = n,*2). Also, t h e mximim

s t r e s s - i n i e n s i t y f a c t o r is l a r g e r f o r l a r g e r values of a / t .

Figures and 10 show s t r e s s - i n t e z i t y f a c t o r s obtained by s e v e r a l

i n w s t i g a t o r s f o r a semi-elliptical surface crack ir, 3 finite-thickness

plate. Figure 9 shows t h e r e s u l t s fcr a s - n f a c e crack w i t h a/c = 0.2

and a / t = 0.8. Smith and Sorensen [3] an? Koba,yashi f 2 ] used t h e

 -13-

alternating method and k t h i r e s a n [k] used t h e finite-element method

ta obtain s t r e s s - i n t e n s i t y f a c t o r v a r i a t i o n s along t h e crack f r o n t .

These three solutions disagree by as much as 50 t o 100 percent. The

ressons f o r these discrepancies are not vell ucderstood. The present

results, shovn as s o l i d symbols, are considera53J higher than t h e

previous solktions [2-4]. The results from 3 a i t h and Sorensen [3: are

generally c l o s e r t o the present results, though 10 t o 25 percent lover.

Figure 10 shows a comparison of t h e maximum s t r e s s - i n t e n s i t y

f a c t o r r 7btained by several i n v e s t i g a t o r s f o r a semi-elliaical surface

crack as a ?unction o f a/t. "he maximum s t r e s s - i n t e n s i t y factors

occurred at 41 = r/2. The present reasilts are shown as s o l i d symbols.

The open symbols show t h e results fram Smith and Sorensen [3],

Kobayashi [2,ll], and Rice and Levy [12]. The results f:wn Rice and

kvy, obtained from a line-spring model, are about 3.5 percent below

t h e present results c,-ccr an a / t range f r o m 0.2 t o 0.6. For a/t > 0.6,

the Rice-Levy s o l u t i o n show a reduction i n stress i n t e n s i t y . The

dash-dot curve shows the r e s l t s of an approximate equation proposed

Sf Beman [I31 fcr a vide range of a / c and a i t r a t i o s . iiewraan's

equation i. w i t h i n 25 percent of t h e present r e s u l t s mer an a/t r.nge

from 0.2 t o 0.8. Sewman's equation gives a good engineering estima;e

for t h e maximum s t r e s b - i n t e n s i t y f a c t o r .
 -11-

A three-dinensional finite-element e l a s t i c stress rnalysis vas

used t o c a l c u l a t e s t r e s s - i n t e n s i t y f a c t o r v a r i a t i o n s along the crack

f r o n t for completely embedded e l l i p t i c a l cracks in large bodies and

f o r s e m i - e l l i p t i c a l surface cracks (crack depth-to-crack length r a t i o s

uere 0.2 and 1.0) in finite-thickness p l a t e s . Three-dimensional

s i n g u l a r i t y e l s e n t s were used at t h e crack f r o n i . A nodal force

method vhich requires no p r i o r assumptit;:: Jf either plane stress or

plane s t r a i n was used to evaluate t h e s t r e s s - i n t e n s i t y f a c t o r s along

t h e crack f r o n t .

Campletely embedded c i r c u l a r and e l l i p t i c a l cracks vere analyzed

to v e r i f y t h e accuracy o f -he finite-element analysis. The stress-

i n t e n s i t y f a c t o r s f o r t h e s e cracks w e r e generally about 0.4 t o 1 percent

belowtle exact s o l u t i m s . However, f o r the e l l i p t i c a l crack t h e

c a l c u l a t e d s t r e s s - i n t e n s i t y f a c t o r s in t h e region of s h a r p e s t curvature

of t h e e l l i p s e were about 3 percent higher than t h e exact solxtion.

The numbers of degrees of freedan i n t h e embedded crack models were

about 3000. A convergence study on s t r e s s - i n t e n s i t y f a c t o r s for semi-

e l l i p t i c a l s u r f a c e cracks i n f i n i t e t h i c k n e s s p l a t e s showed that

convergence was achieved for both t h e semi-circular and t h e s&-

e i l i p t i c a l surface c n c k w i t h about 4500 degrees of freed,%.

For t h e semi-circular s u r f a c e crack t h e maximm s t r e s s - i n t e n s i t y

f a c t o r occurred a t t h e i n t e r s e c t i o n Df t h e crack w i t h t h e free surface.

On t h e other hand, f o r t h e s e m i - e l l i p t i c a l s u r f a c e crack (crack depth-

to-crack h a l f l e n g t h r a t i o of 0.21, $he maximu?J1 s t r e s s - i n t e n s i t y frsctor

 -15-

occurred at t h e deepest point. For both t h e semi-circular and semi-

e l l i p t i c a l surface crack t h e stress-intensity f a c t o r s were l a r g e r f o r

larger values of crack depth-to-plate thickness r a t i o .

For t h e semi-circular surface crack, t h e stress-intensity f a c t o r s

calculated by Browning and Smith 111 using t h e a l t e r n a t i n g method,

agreed generalkv v i t h i n about 3 percent with t h e present results.

Hoverer, f o r semi-elliptical surface cracks (crack depvh-to-plate

thickness r a t i o of 0.2) Sbaith and Sorensen (31 using t h e a l t e r n a h g

method gave stress-intensity f a c t o r s i n considerable disagreement

(10 t o 25 percent) w i t h the present results. For semi-elliptical

surface cracks t h e results from Rice and Levy (121 f o r crack depth-

to-plate thickness r a t i o s less than or equal t o 0.6 sad an appraximate

equation proposed by Neuman (131 were i n good agreement w i t h the

present results.

The stress-intensity factors obtailed herein should be useful i n

correlating fatigue crack-growth rates as w e l l as fracture toughness

calculations f o r t h e surface-crack configmatior.& considered.

 -16-

ill B/cl;-ning,W.M. and Smith, F.W.: "An Analysis for Complex Three-
T i ensional Crack Problems," Developments i n Theoretical and
.lprlied Mechanics, V o l . 8, Proceedings of the 8th SECTAM
Cmference, April 1976.

121 Kobr yasid, A.S. : "Surface Flavs i n P l a t e s i n Bending," Proceedingt

of .'.he W t h Annual Meeting of t h e Society of m i n e e r i n g Science,
Austin, T
-, October 1975.

[31 SPith, F.U. and Sorensen, D.R.: "Mixed Mode S t r e s s I n t e n s i t y Factors

for S e d - E l l i p t i c a l Surface Cracks," U S A (33-134684, June 1974.

[!I1 I nthiresan, K. : 'Three-Dimensional Linear E l a s t i c Fracture

Mechanics Analysis by a Displacement Hybrid F i n i t e Element Model."
P!i.D. Thesis, Georgia Institute of Technology, September 1976.

I51 Green, A.E. and Jireddcn, I. Y. : "The Distribution of S t r e s s i n

t h e U?ighborhood of a Flat ELliptical Crack i n an E l a s t i c Solid,"
Pmc. Cambridge Phil. Tc~c., 46, 1959.

161 Tracey, D.M. : "Finite Element f o r Three-Dimensional E l a s t i c Crack

Analysis," Nucl. Enqi.. and Design, Vol. 26, 1974.

[7] Barsoum, R.S.. "On t h e Use of Isoparametric F i n i t e Elements i n

Linear F h c t u r e Mechanics," I n t . J. Nun. Meth. Engr., V o l . 10,
no. L, 'anuary 1976.

[81 A u r i , S. and Katkiresan, A.: "An Assumed Displacement Hybrid

F i n i t e Element Xodel f o r Three-Dimensional Linear Fracture Mechanics
Analysis," Proc. 12th Annual Meeting of t h e SOC. Ehgr. Science,
UnivereiLy of Texas, Austin, Texas, October 1975.
 -17-

191 Raju, I.S. and Neuman, J.C., Jr.: "Three-Dimensional Finite-

Elenent Analysis of Finite-Thickness Fracture Specimens ,"
NASA TN 0-8414, 1977.

1101 Tada, H.; Paris, P.C.; and Irvin, C.R. : "TGe S t r e s s Analysis of
Cracks Handbook," D e l . Research Corp., C. 1973, pg. 28.1.

(111 Kobayashi, A.S.: "Crack Opening Displacement in a Surface Flawed

P l a t e Subjected t o Tension or P l a t e Bending," presented at t h e
Second I n t e r n a t i o i a l Conference on Mechanical Behavior of Materials,
Boston, Mass., August 1976.

[121 Rice, J.R. and Levy, N.: "The Part- Through Surface Crack i n an
Elastic Plate," Trans. ASME. J. Appl. Mech. Paper Ho. 71-APM-20,
1970.

[13] Neuman, J.C., Jr.: "Fracture Analysis of Surface- and ,%rough-

Cracked Sheets and Plates," Engineering Fracture Mechanics,
Val* 5 , 1973-
TABLE I: Boundary Correction F a c t o r s , F , for S e m i - E l l i p t i c a l
Surface Cracks Subjected to tension

(u = 0 . 3 ; F = K/S dT6)
,

0.375 0.863 1.009 1.237 1.504

0.5 0.968 1.122 1.384 1.657

0.625 1.046 1.222 1.501 1.759

0.75 1.099 1.297 1.581 1.824

0.875 1.130 1.344 1.627 1.846

1. G 1.140 1.359 1.642 1.851

a b
h/c = b/c = 5 . 0 ; h/c = b/c = 4 . 0
Fig. 1.- Surface crack in 3 plate subjected to t e n s i o n .
 x
0
a
k
u

a)
a
z

N
-2 -2'

(a) Circular crack. (b) Elliptic crack.

Fig. 3 . - Circle to ellipse transformation.

 t

-
X J .
 1.6

I
4

1.2,

Degrees of freedom
A 1500
0 2439
.4
a 4317
0 6195 t 2a
u

0 .25 .5 .75 1
-
26
7r

Fig. 5 . - C o n v e r g e n c e of s t r e s s - i n t e n s i t y f a c t o r s f o r a d e e p semi-
2
c i r c u l a r s u r f a c e c r a c k (Q = IT / 4 ) .
 2.0

1.6

1.2'

Degrees of freecorn
.8
A 1692
0 2727
Q 4797
.4 0 6867

0 .25 .5 .75 1

-
26
n
F i g . 6 . - Convergence of stress-intensity factors for a deep s e m i -
eLliptica1 surface c r a c k (Q = 1 . 1 0 4 ; a / t = 0 . 8 ; a/c = 0 . 2 ) .
 1.6

1.2

K
s @ *a

.4 t i - 2a4

0 .25 .5 .75 1
@ 29
R

Ficy. 7 . - Distribution of stress-intensity factors along crack f r o n t

f o r a semi-circular surface crack (Q = 7 # / 4 ) .
 2

1- 6

1.2
c -
K
S

-8

0 1 .25 .5 .75 1

F i g . 8.- Distribution of stress-intensity factors along crack front

for a semi-elliptical surface crack (Q = 1.104; a/c = 0.2).
 1.6

1.2
K

.8

Present r e s u l t s
.4 0 Smith and Sorensen 131
0 Kobayashi ( 2 1
0 Kathiresan [41
I 1 1 1
0 .
I 25 .5 .75 1
-
24
71

Fig. 9.- Comparison of stress-intensity factors for a deep semi-

elliptical surface crack ( 0 = 1.104; a/t = 0.8; a/c = 0.2).
 2.2r
Present r e s u l t s
0 Rice and Levy [121
/
2.0. 0 Smith and Sorensen [ 3 ]
A Kobayashi I 2 I
1.8 B Kobayashi [ l l ]
---
b

Newman [131

- K
1.6 -
1.4.

0 .2 .4 .6 -8 1
-at
Fig. 1 0 . - Cornparisoil of t h e maximum s t r s s s - i n t e n s i t y factor for a
s e v i - e l l i p t i c a l s u r f a c e c r a c k as a f u n c t i o n of a / t
( q i = n/2; Q = 1 . 1 0 4 ; a/c = 0 . 2 ) .