JEC COMPOSITES MAGAZINE - Issue #106 - July 2016 - 27

Experimental characterisation
of initial ply waviness effects in
L-angles
All the investigated laminated L-angles
were manufactured from T700GC/
M21 unidirectional plies with a prepreg
areal weight of 268 g/m2. Two disoriented stacking sequences were considered: a [452°/90°/ -452°/0°/45°/90°/
-45°/0°]s layup noted 20/60/20 and a
[90°/45°/902°/ -45°/0°/45°/902°/-45°]s
layup noted 10/40/50.
L-angles were designed with a 5.24 mm
total thickness, a 6mm internal radius
and a 30 mm width. The angle of the
corner was set at 90°. In this project, the
first set of specimens manufactured by
the subcontractor showed unexpectedly
large initial ply waviness, as reported in
figure 1a. These defects were due to the
low number of vacuum debulks on subsequent plies. It can also be noted that
initial waviness is a function of the stacking sequences considered. Indeed, the
waviness in the 20/60/20 laminates was
systematically higher than that observed
in the 10/40/50 laminates.

Two different types of mechanical tests
were considered in order to validate the
proposed methodology for a wide range
of applied loadings. An electro-mechanical Schenck machine with a 150 kN
maximum capacity was used in combination with a 10 kN load cell. The tests
were performed in the machine-controlled displacement mode (0.1 mm/min
constant displacement rate). Each test
configuration was repeated three times
in order to estimate scattering. Moreover,
all the tests were multi-instrumented
with stereo-digital image correlation,

acoustic emission and micrography
under loading to better understand the
influence of such a defect on delamination.
The first type of mechanical tests corresponded to the four-point bending
(FPB) tests [1] on 20/60/20 L-angles
performed at Onera, as reported in figure
2.
The test principle is also reported in
figure 2. The two cylindrical support bars
were fixed while the loading was applied
through the two cylindrical loading bars.
The characteristic dimensions were the
distances between the support cylindrical bars (X1) and between the loading
cylindrical bars (X2).
Two different test configurations were
considered: Configuration 1 corresponding to X1=33 mm and X2=67 mm,
Configuration 2 to X1=48 mm and X2=82
mm. The cylindrical bars were machined
in steel and could rotate on ball bearings
in order to minimise friction and thus
simplify the finite element simulation
analysis. For pristine specimens, this test
configuration made it possible to apply a
quasi-pure out-of-plane tensile loading,
inducing a large delamination crack
located at mid-thickness of the specimen.

For non-conformal specimens (i.e. specimens with defects), the failure pattern
was much more complex, as evidenced in
figure 3, and governed by the non-conformal microstructure of the material.
The failure load of the non-conformal
specimens (with large initial waviness)
was reduced by 29% as compared to the
pristine specimens for Configuration 1,
and by 35% for Configuration 2.
The influence of initial ply waviness on

Fig. 3: Failure patterns observed for 20/60/20
L-angles subjected to FPB tests

No106 July 2016 / jec

Feature Aeronautics

After modifying the manufacturing
process, the second set of specimens was
geometrically conform, as reported in
figure 1b. Therefore, because the lay-up
and the geometry (except the waviness in
the radius) were identical, it was possible
to estimate experimentally the influence
of real ply waviness in L-angles on the
apparent onset of delamination.

Fig. 2: Presentation of the four-point bending testing device on pristine L-angles

composites magazine 27

Table of Contents for the Digital Edition of JEC COMPOSITES MAGAZINE - Issue #106 - July 2016