NOTCH STRENGTHOF COMPOSITES

James M. Whitney Air Force Wright Aeronautical Laboratories Wright Patterson Air Force Base, Ohio 45433

241

.-

__._ .. -

_. . .

II
a

a L R

(O/f

45/90)2s

T300/5208 cr,= 71.7 KSI

o.2b --Lq 1!t

I 0 I 0. I I 0.2 I 0.3 R, IN I 0.4 I, 0.5
242

0.4 -

I 0.6

STATISTICAL BEHAVIOR OF COMPOSITES

WEIBULL STATISTI CAL STRENGTH THEORY R (S)-EXP B(S) - RISK-TO -BREA.K

TWO PARAMETER BULL 01STRIBUTION WEI B(S)-/y

o(

dV

TENS ON: I

243

BENDING R(S) = EXP

(a + 2) vb s 0t 4@+ 112 0s,

b = 0

[- 1
4 of + 13 Yc=+ 2) Vb

3PT: -

BENDING(aPTI:

BENDING (4PT):

244

THICKNESS EFFECTS

TENSION:

BENDING:

(3

PT)

h2>hl

245

_ .~

(X-R),

IN

POINT

STRESS

CRITERION

Qy (R+do,O)

= co

AVERAGE STRESS

CRITERION

Rtdo L
00

Uy (X.0) R

dx=Uo

246

CIRCULAR HOLE APPROXIHATESTRESSES

X>R

CIRCULAR HOLE:

POINT STRESSCRITERION %
%

Q)
= C2+t,2
t

2 +3E,4-(KT -3)(5c,-7Ela)]

R 1 = R+do

AVERAGESTRESSCRITERION

6 2=

R R+ao
247

FINITE WIDTH CORRECTION CIRCULAR HOLE

Y3( 2R/W )=

2 +( k2R/W) 3&2R/W)

6.t

(O/90)

T300/5208,

K,=5.

II

FINITE B

ELEMENT (INFINITE

SOLUTION PLATE

FOR

2R/W 1

=I 13

5.r SOLUTION

4.t Y -F 3.f

2.C

I (

0.5

0.6

0.7

0.8

0.9 X,

I .o INCHES

248

2R O*bOp0.5.. I.0

(cm) 4.5 I 2.0 I 2.5 I 0 3.0 II.25

(0,901,, 0.4 Vf = 0.66 lq=S.I 0.3 I

T300/5208, - 1.00

c ;3

- 0.75 4 DATA POINT AVERAGE -0.50 0 3 DATA a POlNThERAGE

-z 2
IO0

lo0 0.2-

0.1 -

- 0.25

0.2

0.4 2R

0.6 (in)

0.8

1.0

I 1.2

2R (cm) 1.5 2.0 T30015208 Vf = 0.66 v. =638 MPa (92.4 ksi) 2.5

b \ ez b

2R

(in)

249

CENTER-CRACKED SPECDlEti

TAB

END -fW

3 E -

;-

2,_

.-

I
01

I
0.04 (X-C),

1
0.06 IN

I
0.08

I
0.1

I-

C-

0.02

250

CENTER CRACK

Cy(X.0) Jx2,cz = J&T) = ayX

POINT STRESi CRITERION
aD % cO

x,c

J.

I-ts2 s

Es=

c+d,

AVERAGESTRESSCRITERION
-0% (J-0

co
I-C4

I I+<,

cl=

C c+ao

FINITE

WIDTH CORRECTION - CENTER CRACK

= Yi(2C/W)ON

Yl(lc/w)

- 1 + O.l28(2c/w) + 1.52(2~/w)~ 2c/k 5 0.7,

- 0.288(i~/w)~

ERROR< 0.5%

251

2c (cm) 0.1 oo 0.5 I I.0 1 l.S I 2.0 I 2.5 co.25

0.07s 5 u

0 -0.1s F -0.10 O0

o.oso-

0.025 -

(0, f45,90),,

SCOTCHPLY Vf = 0.50

1002 - 0.03 I 0.0 1.: .

0 0

I 0.2

I 0.4 2c (in)

I 0.6

2c (cm) 0 0.5 1.0 1.5 2.0 2.5 SCOTCHPLY 1002 3

(0, f 45,90&,

Vf = 0.50 co ~320 MPa (46.4 krl) i,, = & /4 b

0

= 0.0909

cm (0.0358

In)

01 0

I 0.2

I . 0.4

I 0.6 2c (In)

I 0.8

I A.0

1.2

252

FRAC!CU~ TOUGHNESS

POINT STRESSCRITERION

c-Q),

K, -

CT0/2 ndo.

AVERAGESTRESSCRITERION

K,= J m(l-64) %
( I +<4)

2c (cm) 25 0 0.5 I --------------I---1.0

I

1.5
I

2.0
I

2.5
I

3.0
I

I-

-J

250

- 200 d,=3.63 &, = E,, /4 cm (0.143 in) in) -150 E c0 0 k - 100 (0.2 45, SOI,, V = 0.50 co=320 MPa (48.4 kril SCOTCHPLY 1002 2 = 0.0909 cm (0.0358

0.2

0.4 2c

0.6 (in1

0.8

1.0

1.2

253

250 20

Ti, ~ 3 . 6 3m (0.143 i n ) c

200

15

kg

= a 14 = 0.0909 crn (0.0358 i n ) ,

[0/-145/90],~
Vf = 0.50

S C O T C H P L Y 1002

2c

(in)

XODIFIED THEORY

= = = =

HOLE RADIUS NORMALIZING FACTOR

=

Ro
m
C

UNIT LENGTH

EXPONENTIAL PARAMETER NOTCH SENSITIVITY FACTOR

CASE 1: dO CASE 2: Rd

m - 0 - CONSTANT: ORIGINAL THEORY RECOVERED O<m<l

-,A-I,
m
n

*N --+A uO

'cr

CASE 3:

1 ON uO INDEPENDENT OF R

A - &
CASE 4: Rd m>l m,

A-

0,

*N a0

i.ao-

HTS

GraphtIm IOp/~451, Kp m2.943

-Epoxy

00 0.25-

loglo 0 -05 -2.0 -1.5 0 -1.0 log,,R a.5

R (mm) I.0 -0.5 (Inches) 0 1.5 1 0.5 20 I.0

256

INFLUENCE OF STACKING SEQUENCE NOTCHEDSTRENGTH ON

Srmb*l lso/o/:4sls

HTS

traphim-Eporv

0.25-

Lo~,~R 0 -2.0 -1.5 0.0 -1.0 Lqo R

(mm) . - 0.5 (mchrs) I;0 I.0 ~ , -0.0 , . . 2.0 2;o 0.5 . 1 1.0

SUPERIWOSING NOTCHEDSTRENGTH
MATERIAL 1 AND MATERIAL 2 @ @2 IF A2 =A, AND 't$'12 - 'Kf$

=
l

ARBITRARY RADIUS FOR MATERIAL 1 RADIUS IN MATERIAL 2 PRODUCING SAME PERCENT STRENGTH REDUCTION

Re

Re=
ll(m2-I) acm -'i?)

acmR m2-1 ) R ( ( ) % -

ml - m2

257

FOR CONSTANTm:

aC =
LOG aC

0
c1 5

1Al - m) ,
- LOG C21

FOR MOST COMPOSITES:

ON a0

INSENSITIVE FOR R 5

1.0

RADIUS-NOTCHSENSITIVITY FACTOR SUPERPOSITION

0.25-

l/KY I I

0 -I:6

0.0 I

i loglo R (mm) 1 I 1.0 ImJFl 2.0 log A 1/

3.0 1.6 2.4

4.0 3.2

-016

0.0
loQlo R

016
(inchrrl

258

DEFINE tiONENTIAL
Kp2

SHIFT PARAXETER :
c2

K&

Cl

q-

Re
-

R i$j

a m

LOG Re

am LOG R ml - 1

a=m

m2 - 1

EXPONENTIALSHIFT PARAMETER

log R*=a,logloR a&m-Mm*-I)

I 0.25 log R
0.0

. .
i I I (mm) I
1.0 ;

I I t I ,
i

log R*=a,log

01
-2.0

R
12

2.0
d4

-I:2 logloR

-0!4 I (inched

259

Mo8I,rr

Curvr

A 8

C 10.0 8.6 22.4 40.0 7.2 I I.6

m ii
03 0.36 0.15 0.42 0.40 360 S.60 2.24. 3.oa 3.oa

8 0.75.
0 0 A

0.50 C ml0 m =0.5

log,oR (mm)
0 -3.0 -1.0 -2.2 0 -1.4 1.0 -0.6 0.2 2.0 1.0 3.0
I.8

loa,oR (in)

COMPARISONBETWEENWEIBULL THEORY AND MODIFIED THEORY

-612

cxsSl2 a==ylL

-h/2

I E s h/2

w,

Y)

1 + 1/2pz + 3/2p' - (K; - 3)(5/2P where p = R/y

- 7/2pe)

u&,, = [l + 1/2[* + 3/25'

- (KF - 3)(5/2e6 - j/259]- where [ = R/(R + d,,) and rl, = Rm/K

Loading

geometry - volume of numerical integration.

In do = m In R - In K

260

Shape pwameter a 3 5 10 12 ii 30

Exponontirl prramotor m 0.31 0.39 0.52 0.57 0.61 0.56 0.57

Notch 8onritlvity K 3.9 5.5 9.7 10.6 16.6 14.0 27.5

Corrdrtion coofficiont r 0.985 0.987 0.990 0.966 0.996 0.983 0.977

a.

* Cw*l~Uon cwtlkl*nl lrom lho ImAll squAwA fll ml Iln~w owAlion, In d. - m In II - la I(; mlnlmum threa pdnla (radtl) wed for l ock glr*n

3s WEIBUL

1 IO

1 IS SHAFE PARAMETER

I 20 0

30 WEIBULL SHAPE PARAMETER 0

Exponential parameter m as a function of shape parameter CA.

Notch sensitivity K as a function of shape parameter ~1.

261

COMMENT An interesting observation has been made (ref. 1) concerning the Whitney "point stress" or "average stress" criterion. the center-notched unidirecIn particular, tional laminate with no notch tip damage is shown to have a square root type stress distribution with an equivalent notch length, which differs from the actual number of broken fibers by a small but constant amount, independent of the number of broken fibers. This is consistent with the assumptions leading to the "point stress" and "average stress" criterion. This is not the case if notch tip damage (in the form of matrix yielding or transverse broken fibers) is present, and it is shown that the damage produces stresses in the unbroken fibers which are less severe than a and also that an analogous equivalent notch length does not square root behavior, exist'. REFERENCE 1. Goree, J. G.; Dharani, L. R.; and Jones, W. F.: Mathematical Damage in Unidirectional Laminates. NASA CR-3453, 1981. Modeling of

James G. Goree Clemson University COMMENT The Weibull distribution function is based on a weakest link concept of failure, This concept underlies the size which in general is not applicable to composites. effect which would account for the difference between bending and uniform tensile strength in purely brittle materials. The difference between tensile and bending strengths in composites is more likely related to the "gradientn effect, which allows localized failure and stress redistribution, as in the case of elasticplastic materials. I. M. Daniel Illinois Institute of Technology

262

COMMENT The present point stress and average stress criteria relate the notched strength of a laminate to the average strength of a relatively long tensile coupon. Tests of notched specimens in which microstrain gages have been placed at or near the edges of the holes have measured strains much larger that those measured in an unnotched tensile coupon. Furthermore, orthotropic stress concentration analyses of failed notched laminates have also indicated that failure occurred at strasns much larger than those experienced on tensile coupons with normal gage lengths. Earlier, both Hahn (ref. 1) and Wu (ref. 2) presented data that related fiber strength to gage length. This suggests that the high strains at the edge of a hole can be related to the very short length of fiber subjected to these strains. Since the length of fiber that is highly strained is proportional to the hole size, this would explain the higher strains to failure measured at the edge of smaller holes. Lockheed has attempted to correlate a series of tests of several laminates with holes ranging from 0.19 to 0.50 in. Although the average stress criterion correlated well with test results for hole sizes equal to or greater than 0.50 in., it overestimated the laminate strength in the range of hole sizes from 0.19 to 0.38 in. It thus appears that we need a theory that is based on the mechanics of failure and is more generally applicable to the range of hole sizes and the varieties of laminates found in aircraft construction. REFERENCES 1. 2. Hahn, H. T.: Failure Mechanisms. Failure Analysis Fibrous Composite Structures, NASA CP-2278, August Strength Theories Wu, Edward M.: Analysis and Mechanisms of Failure August 1983, pp. 173-189. and Mechanisms of Failure 1983, pp. 153-171. of

of Composites: Status and Issues. Failure of Fibrous Composite Structures, NASA CP-2278,

Larry Fogg Lockheed-California

Company

263