Crack Paths 2009

the influence on fatigue life of both overlap length and comer geometry is greater than

that of stacking sequence and ply orientation at the interface. The results of the fatigue

testing program and their modelling are extensively discussed in ref. [6].

i. ,_

a ‘

I00 -

[cMP'a]m

. [lllhwr-ll'l m m

a [4541;]. w--l0 m m

A [451-U].w=4l] m m

4!} ‘

0 |U|,. w=2l1 m m

o [45th:],i Its-=20 m m

0- [-l5;»0|, w i l lm m

|||l

1

1 1 |i1|il 1 | | | 1 1 m

iES

I154

IE5

IE6

IE7‘

Cycles to failure

Figure 2. Influence of the overlap length on the fatigue strength of square edge joints.

(Data for [0]6joints are taken from [1]).

. '

O

200

. one o H 2 ‘ s e

1 .

_

o n C 0 0 . 0 c m

I

.

_ 1 0 n -

A A O - O ’ "

E

'00

E

‘ 3 6 % )

g

-

. [01h w=4lJ m m

5

g [45: 113]‘ w=4tl m m

m _

A Hig-fiLnF-ll} m m

0 [0]‘. w=2U m m

0 [4550;], w=2tl m m

A [45:50]‘ “ = 2 0l1l'l‘l

l l l l l l l

I n ]

| | | | | | | | l

IE3

IE4

IE5

IE6

IE?

Cycles lo failure

Figure 3. Influence of the overlap length on the fatigue strength of spew filletjoints.

(Data for [0]6joints are taken from [1]).

During the fatigue tests. joint stiffness trends were monitored by measuring the

instantaneous joint stiffness as the ratio between the imposed load range measured by

the load cell and the resulting range of the actuator stroke measured by the displacement

transducer of the testing machine. Figure 4 show examples of normalised stiffness

trends as a function of the normalised fatigue life: the global analysis of the results

indicated that stiffness trends are not influenced by lay-up. overlap length and only

slightly by comer geometry, with a tendency of the fillet corner to delay crack initiation

(identified by the beginning of the stiffness drop). This general insensitivity is quite

important in view of using measured stiffness data for the development of a simplified

971