A probability-based methodology for buckling investigation of sandwich composite cylindrical shells

Cylindrical shells are widely used in primary structures of space launch vehicles. Due to the high imperfection sensitivity, buckling behavior of shells under axial compression exhibits a great discrepancy between theory and experiments. In order to meet the demand of aerospace industry for improved design criteria, a unified framework, the probability-based methodology, for probabilistic buckling analysis of axially-compressed shells is developed, illustrating it with application to two structures of Ariane 5 launcher. They are sandwich composite shells made of same material, but with different stacking sequence and geometric dimensions. One of them is also studied with the presence of three circular cut-outs. The methodology combines the Stress-Strength Interference Method and the Latin Hypercube Method to determine a probabilistic buckling factor, once the probability that shell withstands axial load without undergoing buckling is specified. Such a factor accounts for influence of manufacturing and in-service imperfections, but depends on required reliability, on sample size and on considered imperfections. Different sources of imperfections, modeled by literature, are introduced into the analysis independently and jointly. The main advantage of the methodology is the versatility as it can be used for buckling investigation of laminated composite shells and sandwich composite shells including different types of imperfections. In particular, the methodology allows for incorporating directly experimental data representative of shells of interest. To show this, buckling of an experimental shell is investigated by the probabilistic methodology using measured imperfections. An alternative method, the chaos approach, for buckling analysis of axially-compressed shells is derived using concepts of the chaos, with application to one of the studied shells. The goal is to obtain an erosion profile as function of increasing axial load. It illustrates concisely the degradation due to imperfections of the load-carrying capability of shells. The approach could be adopted when test-originated database of imperfections is not available to achieve a first assessment of the imperfection sensitivity of shells.

N/A