This research studies the effects of a damage arrestment device embedded between a carbon fiber facesheet and foam core to find whether there is an increase in the structural integrity of the sandwich composites. Experimental and theoretical finite element analyses are implemented for two different composite sandwich geometries; plates and beams. Each structure consisted of the same loading criteria and was restricted to the same vibration fixture during the experiment. An accelerometer was placed on the composite plate to record the amplitude and the natural frequencies of the composite structure. Each composite specimen is then fixed to the surface of the Cal Poly Shake Table by two aluminum block fixtures. The mechanical properties of LTM45/CF1803 pre-impregnated carbon fiber and Last-A-foam FR 6710 polyvinylchloride foam were experimentally analyzed using ASTM D3039 and ASTM D1621 standards respectively to determine the material’s mechanical properties. By using the finite element program COSMOS with the pre-software GeoStar, accuracy representation were created to compare numerical, analytical, and theoretical results.

A study was conducted to investigate simple, cost-effective manufacturing techniques to delay skin-core delamination, micro-buckling and bearing stress failures resulting from fastener-hole interactions. Composite sandwich panels, with and without damage arrestment devices (DADs), were subjected to monotonic compression at a rate of 5mm per second, and compression-compression fatigue at 50% yield at an amplitude of 65%, under temperatures of 75, 100, 125, 150, 175, and 200 °F. The sandwiches tested were composed of two-layer cross-weave carbon fiber facesheets, a polyurethane foam core, and an epoxy film adhesive to join the two materials. The most successful method to delay the aforementioned failures involved milling rectangular slots in the foam core perpendicular to the holes and adding three additional layers of carbon fiber cross-weave. For the monotonic cases, the ultimate load increases were 97, 87, 100, 131, 96, and 119% for each of the respective temperatures listed above with a negligible weight increase. For the fatigue cases, the number of cycles for each test case was nearly identical. This still represents a large improvement because the yield used in the loading condition for the specimens with DADs was 97% greater than the specimens without DADs. The experimental results were compared with a finite element model (FEM) built in Abaqus/CAE. The numeric and experimental results showed a strong correlation. All test specimens were manufactured and tested in the California Polytechnic State University Aerospace/Composites Laboratory.