Good morning,

I am a researcher. I am modeling a sandwich panel in Abaqus 2022 consisting of three layers: two very thin outer layers (skin) and a thick inner layer (core). The core is of the honeycomb type, where the elements constituting the core are wooden elements with a cylindrical geometry and a circular ring-shaped cross-section, positioned with their longitudinal axis perpendicular to the plane of the panel. These elements are unbound and not constrained. The skins are made of fiber-reinforced composite material. The thickness of the skins is 0.6 mm, while the thickness of the core is 13.1 mm. The skins are glued to the rings with an epoxy resin, so each wooden ring is individually glued to the skins forming the panel. The panel is subjected to three-point bending tests. The dimensions of the panel to be tested are: Total length 300 mm, support distance 220 mm, panel width 35 mm.

The supports of the beam and the load cell are modeled as cylindrical rigid shells with a diameter of 15 mm, to which boundary conditions are applied through reference points (RP). Constraints have been imposed on the supports, while the load cell is allowed to translate in the direction transverse to the longitudinal direction of the specimen.

The skins and the core rings have been modeled using 3D elements with the extrude function. For each element, a local coordinate system has been imposed. In the model, the skins have been constrained through cohesive contact interaction with the individual rings, specifying the stiffness coefficients K, equal to 150. Using the Damage option, I specified "damage evolution" with the "Quadratic traction" criterion and Maximum nominal stress equal to 36. In damage evolution, "Energy" and "Linear" types were chosen, defining Fracture energy equal to 12.5. Viscosity was specified as 0.0001.

Additionally, hard contact and tangential friction interaction with a penalty value of 0.1 was inserted on the contact surfaces between the cell-load and the skin, as well as between the supports and the skin. The same was done inside the core, i.e., in the contact areas between the rings.

Composite skins and core rings are modeled using an eight-node linear brick 3D element (C3D8R).

The problem I encountered is that once the model is run, the response in terms of normal stresses is not symmetric. In particular, along the base, the stress distribution is not symmetric, which should not happen. I cannot understand why. This problem does not occur when I impose a TIED interaction type between the skins and the rings constituting the core!