7.7 Validation based on different joint configurations
151
The simulation results are displayed in Figure 122 a) in terms of force-displacement
curve and visual damage pattern. The simulation matches the test results with regards
to initial stiffness accurately. The strength is generally overestimated while still being in
the range of the test scatter. There is significant scatter in the tests, which is assumed to
be largely originating from the specimen placement in the fixture. This is due to the lack
of stoppers which handicap repeatable specimen placing. Since the simulation indicates
a strength at the upper end of the scatter range, it is assumed that the specimens with
lower strength represent unfavorably placed specimens. Regarding the further curve
progression, the simulation indicates a less distinct load drop and plateau in stage ③.
Instead the simulation goes directly into stage ④ with increasing load after the load
drop. The visual damage pattern of the simulation matches the observations during the
tests well. In sum, the virtual test agrees well with the physical test. The deviation in the
post failure regime of the later stages is tolerable since the primary objective is to forecast the strength. In addition, the simulation indicates all relevant mechanical effects in
correct sequence. Therefore, the given example validates the developed approach.
Bending test
Analogous to the shear tests, the bending tests are analyzed by characterizing the forcedisplacement curve and the visual damage patterns during the test. This is illustrated in
Figure 120 b). The curve progression generally resembles the force-displacement relationship of the shear tests. Therefore, it is split in the same four stages. However, the
governing mechanical effects during the test are different. Stage ① represents elastic
deformation, while stage ② describes the in-plane core crushing of the 26 mm panel
due to penetration of the 10 mm panel. This damage mechanism is generally undesirable, since it leads to premature catastrophic failure, which is governed by the low inplane material properties of the core. In the tests, this is caused by unfavorable bonding
of the panels such that there is a gap between the outer faces of the bonded panels. This
enables unconstraint penetration of the core. Stage ③ represents debonding of core
and face of the penetrated panel leading to a force plateau. Stage ④ describes the final
curve section with increasing load, due to densification of the penetrated core as well as
change in angle of the penetrating panel. Figure 120 b) illustrates the core penetration
and core to face debonding. Therefore, the test results indicate that the failure progression is entirely dictated by core mechanics. As a result of this a detailed model with explicit time integration is defined as model framework. However, in order to reduce computation time, the core is only modelled in detail in the contact area of both panels in
order to ensure accurate mechanics in the penetration zone. The remaining core is modeled using 3D-continuum elements leading to a hybrid model. The implemented simulation model for the bending test is illustrated in appendix Figure 121. It largely coincides
151
The simulation results are displayed in Figure 122 a) in terms of force-displacement
curve and visual damage pattern. The simulation matches the test results with regards
to initial stiffness accurately. The strength is generally overestimated while still being in
the range of the test scatter. There is significant scatter in the tests, which is assumed to
be largely originating from the specimen placement in the fixture. This is due to the lack
of stoppers which handicap repeatable specimen placing. Since the simulation indicates
a strength at the upper end of the scatter range, it is assumed that the specimens with
lower strength represent unfavorably placed specimens. Regarding the further curve
progression, the simulation indicates a less distinct load drop and plateau in stage ③.
Instead the simulation goes directly into stage ④ with increasing load after the load
drop. The visual damage pattern of the simulation matches the observations during the
tests well. In sum, the virtual test agrees well with the physical test. The deviation in the
post failure regime of the later stages is tolerable since the primary objective is to forecast the strength. In addition, the simulation indicates all relevant mechanical effects in
correct sequence. Therefore, the given example validates the developed approach.
Bending test
Analogous to the shear tests, the bending tests are analyzed by characterizing the forcedisplacement curve and the visual damage patterns during the test. This is illustrated in
Figure 120 b). The curve progression generally resembles the force-displacement relationship of the shear tests. Therefore, it is split in the same four stages. However, the
governing mechanical effects during the test are different. Stage ① represents elastic
deformation, while stage ② describes the in-plane core crushing of the 26 mm panel
due to penetration of the 10 mm panel. This damage mechanism is generally undesirable, since it leads to premature catastrophic failure, which is governed by the low inplane material properties of the core. In the tests, this is caused by unfavorable bonding
of the panels such that there is a gap between the outer faces of the bonded panels. This
enables unconstraint penetration of the core. Stage ③ represents debonding of core
and face of the penetrated panel leading to a force plateau. Stage ④ describes the final
curve section with increasing load, due to densification of the penetrated core as well as
change in angle of the penetrating panel. Figure 120 b) illustrates the core penetration
and core to face debonding. Therefore, the test results indicate that the failure progression is entirely dictated by core mechanics. As a result of this a detailed model with explicit time integration is defined as model framework. However, in order to reduce computation time, the core is only modelled in detail in the contact area of both panels in
order to ensure accurate mechanics in the penetration zone. The remaining core is modeled using 3D-continuum elements leading to a hybrid model. The implemented simulation model for the bending test is illustrated in appendix Figure 121. It largely coincides