Continuum Damage Progressive Failure Modeling for Crashworthiness and Static Analyses of Laminated Composites
Deleo, Francesco Roberto
MetadataShow full item record
A new energy-based composite continuum damage model is presented for laminated composite structures. The composite laminate is modelled at the meso-scale using the finite element method, specifically: shell elements. Each composite lamina is modelled with one or four integration points, depending on if reduced or full integration elements are used. Each lamina is modelled as a 2D plane stress orthotropic material and damage is included by degrading the stiffness matrix. Element deletion is used when all integration points have failed. The theory is developed to predict failure initiation using a modified version of the Hashin’s failure criteria, and failure growth is modelled using damage variables which are developed as part of this work. The theory has been written to be used for both static failure and dynamic crush analyses. Two sets of damage variables are derived, one for static analyses and a separate one for dynamic crush. When performing a dynamic crush analysis, the distributed damage model is utilized in order to ensure that the progressive failure crush mode correcty takes place. The damage model is energy-based using the fracture energies of four failure directions: the fiber and matrix directions, and in each of tension and compression. Mesh-size objectivity is assured via the adoption of the Crack Band Theory. The theory is implemented into an LS-DYNA user-defined material model (UMAT) in order to take advantage of the pre-built contact algorithms needed for crush analyses and for solving the governing equations of motion. The theory is applied to a static and crush problem, with excellent agreement between the experimental and numerical results for both cases. A tensile notched specimen is tested and simulated. Both the failure morphology and resulting load-displacement results at the constraint are found to match very well. A crush experiment carried out by the author is simulated. In addition to matching the experimental results with a high degree of accuracy, the theory discussed herein offers significant advantages over alternative damage models present in the literature. There is no need to filter the load – displacement results using a non-physical numerical scheme, and the high impulse loads are eliminated. In addition, the theory offers a level of robustness and stability which allows it to be used with any contact type, including ‘hard’ contact, and therefore eliminates the need to calibrate the penalty-based contact algorithm’s load penetration curve with non-physical stiffnesses.