Collapse Analysis, Defect Sensitivity and Load Paths in Stiffened Shell Composite Structures

Kurzfassung

Composite skin-stiffened structures can withstand significant loads after initial buckling has occured. However the application of composite postbuckling structures in curved aircraft panels has been limited to date due to concerns related to the sensitivity of the structures to manufacturing defects and service induced damage. Unlike stiffened metallic fuselage panels, panels made from composite materials are not allowed to have degradation below the ultimate load due to issues relating to certification. In addition the analysis of these panels to model the progression of failure to collapse is non-trivial even using modern linite element solvers. For undamaged skin-stiffened structures in compression, collapse is typically an explosive event caused by the initiation of separation between the skin and stiffeners followed by libre failure and delamination. For pre-damaged structures, such as those taken from service or those used lor damage tolerance and certification studies, the pre-damaged areas can grow under compression and contribute to the collapse if the are located at critical locations. A range of thin stiffened shell panels has been tested under the European COCOMAT project [1,2]. Finite element analysis of these panels in the Cooperative Research Centre for Advanced Composite Structures has successfully captured the experimental results for a number of the panels tested [3]. The numerical algorithms predict the initiation and growth of interlaminar damage and predict libre failure leading to ultimate collapse. The numerical algorithms enable the capture of the displacement behaviour and the analysis provides detailed information of the development and interaction of the various damage mechanisms. The work has now been extended to defect sensitivity. In the current pool of experiments it is possible to observe the scatter in results caused by manulacturing delects and material variations. The delects and variations are strong enough to affect the buckling making it difficult to predict the buckling patterns and possible failure loads using finite element analysis. A stochastic approach was therefore developed to introduce variability and it has been successfully applied to explain results that had previously been regarded as outliers in the experimental program. These methods will be described in this paper and extended to help design structures that will be less sensitive to the variations. A qualitative measure of robustness in the form of a Robust Index has been derived so that effective design decisions can be made [4]. Finally the research program has attempted to improve the post-processing tools available in finite element analysis by plotting the load distributions in the panels during the test sequence. The finite element is a method of analysis based on displacements and strains. Stresses are quantities derived from the strain field. In modern finite element packages the load field is not plotted and the distribution of the load in the pre-and post-buckled structures is not clearly exposed. An attempt is therefore made to enhance the interpretation of the failure mechanisms by plotting load paths based on the theory presented in [5] prior and leading up to collapse. The structures are most sensitive to defects located on the primary paths that are active just prior to collapse.