NASA NTRS2016
Two advanced composite tow-steered shells, one with tow overlaps and another without overlaps, were previously designed, fabricated and tested in end compression, both without cutouts, and with small and large cutouts. In each case, good agreement was observed between experimental buckling loads and supporting linear bifurcation buckling analyses. However, previous buckling tests and analyses have shown historically poor correlation, perhaps due to the presence of geometric imperfections that serve as failure initiators. For the tow-steered shells, their circumferential variation in axial stiffness may have suppressed this sensitivity to imperfections, leading to the agreement noted between tests and analyses. To investigate this further, a numerical investigation was performed in this study using geometric imperfections measured from both shells. Finite element models of both shells were analyzed first without, and then, with measured imperfections that were then, superposed in different orientations around the shell longitudinal axis. Small variations in both the axial prebuckling stiffness and global buckling load were observed for the range of imperfections studied here, which suggests that the tow steering, and resulting circumferentially varying axial stiffness, may result in the test-analysis correlation observed for these shells.