influences the ultimate compressive strength. Specimens that failed by delamination buckling tended to have a lower ultimate compressive strength compared with specimens that failed by either fiber failure or global instability modes of failure. This suggests that for CAI modeling, inclusion of the delamination damage mechanism may be required to achieve the necessary fidelity. 6.4 Numerical Modeling The other major task of this study was to develop a computational tool that will predict damage resistance and post-impact damage tolerance of sandwich composites to a low-energy impact event. The majority of past numerical studies of CAI have simulated a pre-defined state of damage in compression [6]. The goal of this study is to explicitly simulate the QSI damage event and then pass that directly into a CAI simulation to model both the impact event and the resulting compression test. Thus the damage resistance and damage tolerance can be modeled in one simulation. The numerical results will then be validated by the experimental results. A full-size model was developed using the ABAQUS/Explicit finite element software. The major challenge was to incorporate all the relevant damage mechanisms into the model. From the experimental work, the facesheets exhibited fiber/ matrix failure and delaminations, and the core exhibited localized core crushing beneath the indentation site. Facesheet-core debonding was not exhibited from the experimental testing. For this study, we aimed to model the fiber-matrix failure of the facesheets and the localized core crushing of the core. Future work will extend the model to incorporate delaminations introduced during QSI. The material properties for the constituent materials are shown in Table 6.2. These were obtained from the manufacturer of the prepreg and the honeycomb core [7] or estimated or measured experimentally. The method to experimentally measure these properties is discussed below. Each ply is 0.127 mm thick. The explicit geometry core was modeled using an elastic–plastic material model with E ¼70 GPa, ν ¼0.3, yield stress of 220 MPa, rising to 282 MPa at 1 % strain and constant thereafter. The outer facesheets are modeled with continuum-shell elements. A continuum model for the core would be far less computationally expensive than the explicit geometry model used here, but our prior work has shown that the continuum core model was unable to accurately simulate the localized core crushing during QSI. Therefore, the core was split into two regions where the top 30 % of the core adjacent to the top facesheet was modeled using an explicit geometry core. The remaining region that stays elastic during the simulation is modeled with anisotropic 3D continuum elements. The model showing the individual regions including the rigid indentor is shown in Fig. 6.3. 6.4.1 Facesheets The facesheets were modeled using 8-node continuum-shell (SC8R) elements with a Hashin degradation material model. The ply constituent material properties are based on [7] with modifications to (E1) to match the simulated and measured inplane and bending stiffnesses of the 16-ply facesheets. The ply strengths were also modified from the reference values in [8, 9] so that simulations of failure in EC and of four-point bend-to-failure tests match experimental results. The reason for the changes in properties relative to the reference values is that the co-curing process of the sandwich panel assembly, induced fiber waviness, particularly towards the core. The ply stiffnesses and strength are thus reduced relative to a facesheet with negligible fiber waviness. Table 6.2 Constituent material properties Material E1 E2 E3 G12 G13 G23 ν12 ν13 ν23 Density (GPa) (GPa) (GPa) (GPa) (GPa) (GPa) (kg/m3) IM7/8552 prepreg 139.5 ∗ 12 12 5.17 5.17 3.98 0.32 0.32 0.46 1,570 Continuum core 1.48E 04 ∗∗ 1.47E 04 ∗∗ 1.1 ∗ 89E 06 ∗∗ 0.17 ∗∗ 0.127 ∗∗ 1 ∗∗ 1E 05 ∗∗ 1E 05 ∗∗ 49.7 ∗ Measured experimentally ∗∗ Estimated 50 B. Hasseldine et al.
RkJQdWJsaXNoZXIy MTMzNzEzMQ==