In industry the stress analysis is done using finite elements. For thin layered structures are used shell elements based upon the First Order Shear Deformation Theory (FSDT). But thick structures often requires volume elements. But using homogenouos anisotropic 3D-elements for the discretization of composite structures results in very large FE-models which can not be ecomical solved. This is due to the fact that three to five element layers are needed for every composite layer to get sufficiently precise results. The Composite Volume Elements which are implemented in some FE-Codes are not sufficiently developed. Tey are based for instance on quadratic approximations for the displacements u, v and w. But for the analysis of thickwalled structures these approximations are insufficient because the u- and v-displacents show a zig-zag distribution in thickness direction. Therefore the resulting out-of-plane stresses have a bad quality.
For the development of efficient volume elements different possibilities are practicable. At one side one can use higher polynomial orders for better approximation of the zig-zag distributions. At the other hand it is possible to use the 3D equilibrium equations for the computation of the transverse stresses. Another alternative is the use of improved stiffnesses.
Based upon this alternatives a composite volume element is developed, which allows the efficient computation of the full stress tensor for composite structures.