Adaptive Model Reduction for Local Post-Buckling Analysis
Résumé
Large aeronautical stiffened structures are subjected to local static instabilities like buckling. On the first hand this phenomenon does not lead directly to failure, but it changes the stress distribution and may initiate damages like skin-stiffener debonding. On second hand, mass can be saved by allowing local post-buckling behavior in the working range of the structure. The best way to achieve local post-buckling oriented sizing of structures is to perform large scale non-linear analysis (sereval millions of degrees of freedom) in order to predict local-global interactions and damage initiation. As computational cost is limiting, an adaptive model reduction strategy is proposed and then combined with a domain decomposition method, rationalising expensive computation steps