SAMPE Journal - January/February 2017 - 49

Feature Article
Table 3. Process parameters for producing overlap shear specimens.

increased. No voids have been
created by the matrix squeezeout (Figure 13). Only at 450 mm/
min process speed, small voids are
found in the joining area. This is
in agreement with the simulation
results (Figure 11). The small areas
of surface melting lead to a local
and minor delamination of the
surface (Figure 13). This result
was not predicted by the process
simulation. A possible reason is the
simplification in the thermal FEmodel. All bodies in the thermal
simulation
are
homogeneous.
Because the laminates are locally
inhomogeneous
(e.g.
crossing
points of fibers) local temperature
maximums cannot be calculated.
As a reference for the reached
bonding strength, overlap specimens
have been produced in an autoclave
process. The lap shear strength
of autoclave specimens was used
to transform the ´state of joining´
(calculated percentages of bonded
width) into a lap shear strength
equivalent value (Equation 3).

Figure 12. Overlap area of overlap specimens after welding.

Figure 13. Cross sections of the overlap shear specimens and areas of delamination
and voids.

[3]
τsimulation is the simulated lap shear
strength. τautoclave is the measured
lap shear strength of specimens
produced
in
the
autoclave.
bonded, degraded and voids are
the percentages of the joining
width from the simulation that are
bonded, degraded or having voids.
The reference lap shear strength
of the autoclave specimen and the
calculated as well as measured lap
shear strength regarding Table 3 are
shown in Figure 14.
The overlap shear specimens
have an overlap of 12.5 mm and a
laminate width of 100 mm. After
welding the plates were cut into
four test specimens of 25 mm width

and tested. The specimens are
numbered according there position.
The welding takes place from the left
to the right, means from specimen
number 1 to number 4.
The lap shear strength of specimens
2 and 3 are higher than the lap
shear strength of specimen 1 and 4.
Specimens 1 and 4 were taken close
to the edge of laminates. Therefore,
some of the heat input of the coil
was lost due to the edge effect by
induction heating. Therefore, the
heating pattern changes and heating
efficiency in the joining area is
lower than calculation. Thereby, the
temperature in the joining area was
below the calculated temperature.

SAMPE Journal, Volume 53, No. 1, January/February 2017

The starting and stopping procedure
of the welding process must be
optimized to consider speed changes
and other edge effects to increase the
bonding strength of the welded joint.
For 200 mm/min and 350 mm/
min specimens 2 and 3 are at the
calculated lap shear strength level
and the simulation gives a realistic
view on the expected bonding
strength.
The lap shear strength with 450
mm/min is higher than the calculated
one. A possible reason is that areas
with voids are assumed to have
zero lap shear strength. This is a too
conservative approach.
By the used method, the lower
49
Table of Contents for the Digital Edition of SAMPE Journal - January/February 2017
