A Space-Time Multiscale Coupling Strategy for Predicting Delamination Due to Impacts.
Résumé
Composite laminates are increasingly employed in aeronautics but can be prone to extensive delamination when submitted to impact loads such as from bird strikes or tire failures. Current analysis tools allow to determine whether a given structure can sustain given impact loads, while appropriate safety margins are considered. In order to enhance the industrial design, more advanced numerical techniques are necessary to gain accuracy for complex materials. In the case of laminates, meso-scale schemes, in which plies and cohesive interfaces are modeled in detail, are necessary to provide robust analysis tools. However, the computational costs associated with such fine modeling would be excessively high for the engineering practice. So, macro-scale schemes, e.g. shell models, are usually employed but are not able to represent complex mechanisms like the delamination propagation. Our aim is to develop a bridging coupling between a macro-scale, applied to large structures, and a meso-scale, applied only to the front of delamination for predicting its propagation. The transient nature of the impact problems leads to explicit schemes and, because of the stability condition, the spatial multiscale coupling involves a temporal multiscale coupling too. The classical application of a Domain Decomposition approach would require to re-mesh the global model in order to follow the delamination propagation, which would be costly and complex to implement. In this paper, we present the basis for a less intrusive approach, called the Substitution method. This latter is developed in order to leave unchanged the structural macro-scale analysis, coupling it with an iterative adaptive finer meso-scale analysis. Two versions of the Substitution method have been developed, based on different bridging formulations. The first one, inspired by the multi-timestep method developed for Domain Decomposition, and the second one, inspired by the enhancements, based to numerical energy considerations. Two-dimensional applications of this method will be presented.