Launch vehicle structures, such as payload adapters and interstages, are increasingly designed and constructed using composite materials due to their high stiffness- and strength-to-weight ratios. Therefore, it is important to develop a validated finite element modeling methodology for designing and analyzing composite launch-vehicle shell structures. This can be achieved, in part, by correlating high-fidelity numerical models with test data. Buckling is often an important failure mode for cylindrical shells, and the buckling response of such structures is also often quite sensitive to imperfections in geometry and loading. Hence, it is crucial to understand the model parameters and details required to accurately predict the buckling load and behavior of composite cylindrical shells, especially if the shell is buckling critical. The inclusion of as-built features, such as radial imperfections, thickness variations, and loading imperfections can help improve the correlation between test and analysis. To demonstrate such an approach, a validated modeling methodology that was used to predict the buckling behavior of a scaled component for a launch-vehicle-like structure is presented, and results from the model are compared with experimental results. The modeling approach presented herein was used to successfully predict the buckling behavior.

@en

A subscale solid laminate composite cylinder with a 31.5-in. diameter, a 48.2-in. length, and a [23/0/-23]4S layup was designated NDL-1 and was designed to fail in buckling after a series of subcritical tests. NDL-1 was experimentally loaded in axial compression until buckling at 466.3 kips. Visual inspection of the test article after the test revealed that a shallow area of delamination occurred near the buckling initiation site. After a successful first buckling test, there was an opportunity to test NDL-1 in axial compression to failure a second time despite that the test article was not designed or analyzed for multiple buckling tests. During the second test to failure, NDL-1 reached a peak load of 390.4 kips, 16.3% lower than the first test. Buckling did not initiate in the same location during the second test to failure as the first test to failure. The second buckling event caused a large V-shaped crack that penetrated through the thickness of the barrel. Though buckling initiated at different locations, a similar radial deformation pattern was present just prior to buckling during the both tests. In the end, NDL-1 maintained a significant amount of its structural integrity and a similar prebuckling radial displacement pattern, even after the first buckling test. @en