Pretest Prediction Composite Hat

Traditional analytical approach requires many different FEMs

The first model built was 3-D to capture the panel cross sectional shapes and was used to predict stress failure of the panel and ringframe. It was also used to predict facesheet local buckling but its mesh turned out to be too large. A second 3-D FEM was built that had five rows of hat stiffeners to predict instability interaction of the hats. A third model was built that was 2-D to predict panel buckling. This FEM used beam/rod elements for the hat stringers, and as consequence did not correctly handle torsional rigidity of the closed cells or the additional transverse bending stiffness due to the cells. Also the stiffness terms for the shell element material and property records did not include the effects of the unsymmetric nature of the membrane/bending coupling of the hats, nor their in-plane shear flows. Finally, the local buckling facesheet span widths of the distance between hats was greater than the actual design. For these reasons it was thought that another 3-D model would be necessary which discretely modeled all 43 hats of the test article to accurately capture panel buckling. This model was not made however, due to the required high grid count. A fourth 3-D FEM was built later in this process to verify HyperSizer's pre-test predictions. This model had just one hat stiffener row, as shown.

Return to the previous page...