Relative performance of metal and polymeric foam sandwich plates under low velocity impact

• Low-speed Impact tests on sandwich plates were performed using InstronR Dynatup apparatus.
• Numerical models for simulating impact were developed in LS-Dyna commercial finite element software.
• Relative performance of PVC and aluminum alloy foams is studied.
• Effect of core grading on energy and peak forces is evaluated.
• Peak loads from analysis are in good agreement with experiments and numerical models.

Relative performance of metal and polymeric foam cored sandwich plates is studied under low velocity impact loading. The metal and polymeric foam sandwich plates are constructed using a core of 40 mm thickness (with two layers of 20 mm each) and aluminum faceplates. Metal foam sandwich plates are constructed using aluminum alloy foam (ALPORAS) core while polymeric foam sandwich plates are constructed using polyvinyl chloride (Divinycell H80 and H250) foam core. Impact experiments are conducted with a hemispherical punch of mass 8.7 kg at a nominal velocity of 5.8 m/s. The effect of stepwise core grading on the maximum dynamic penetration force as well as energy absorption is studied. To maximize the energy absorption or to minimize the mass of the sandwich plate for a given penetration force, alternatives to Alporas foam are chosen based on either equivalent density (H250) or through-thickness compressive yield strength (H80). The increase in penetration force and energy absorption resulting from the choice of H250 in place of Alporas for the same density of the foam as well as the effect of decrease in mass of the sandwich panel by choosing H80 foam in place of Alporas for the same compressive strength of the foam is discussed. Numerical models were developed in LS-Dyna to predict the impact response (force-displacement history) and failure modes. Upperbound analysis is used to estimate the maximum penetration force. Peak force, energy absorption values and failure mode patterns obtained by analytical estimates, experimental measurements and numerical predictions all agree well.