Tab. 1: Comparison between finite element analysis and experiments Shear Compression Hexweb CR III 5/325052-.0007 L direction W direction Bare Modulus Strength Modulus Strength Modulus Manufacturer data 220 psi 1.51MPa 55 ksi 379.3 MPa 165 psi 1.13 MPa 37 ksi 255.2 MPa 100 psi 0.69 MPa 19 ksi 131.0 MPa Simulation results 1.46 MPa 385.5 MPa 1.08 MPa 249.9 MPa 0.68 MPa 129.0 MPa Error -3.4 % +1.7 % -4.4% -2.0 % -1.4 % -1.7 % This made it possible to characterize the thickness of the node bond line and the phosphoric acid anodizing (PAA) treatment (Figure 2). Virtual optimization A virtual aluminium honeycomb model was created through FEM simulation and calibrated. The results obtained with the finite element model, using quasi-static conditions as described in ASTM C365 and ASTM C273, showed good correlation with the experimental results (Table 1). (specimen sizes, rate speed, bonding...) on Alcore PAA 5052 3.0-3/8-.0020 honeycomb and EC3D modified honeycomb 5052 (specimens made from Alcore PAA 5052 3.0-3/8-.0020). To avoid errors, the results were first compared with the theoretical values of the original samples. All the tests were carried out at the LARMAUR (Laboratoire en Mécanique Appliquée de l'Université de Rennes A modified honeycomb model was also created to predict its mechanical behaviour in both compression and shear (Figures 3 & 4). Specimen manufacturing Specimens were made from a plate of Alcore PAA 50523.0-3/8-.0020 hexagonal aluminium honeycomb. The original aluminium honeycomb was naturally slightly under-expanded. This explains why the density was the same before and after the modification (Figures 5). Fig. 4: Simulated comparison of shear behaviour: 3/8-5052-.003 hexagonal cell configuration vs. modified one Testing conditions and results Comparative tests were conducted in the same conditions Fig. 3: Virtual honeycomb model Fig. 5: Influence of expansion on density No106 July 2016 / jec composites magazine 51