Narrow Results

The European Commission 6th Framework Project COCOMAT was a four-and-a-half-year project (2004 to mid-2008) aimed at exploiting the large reserve of strength in composite structures through more accurate prediction of collapse. In the experimental work packages, significant statistical variations in buckling behaviour and ultimate loading were encountered. During the experiments for the COCOMAT project, it was recognised that there was a gap in knowledge about the effect of initial defects and variations in the input variables of both the experimental and simulated panels. The effect of the defects and variations in the experimental panel resulted in some failure modes that were not predicted with the finite element modelling. This led to the development of stochastic algorithms to relate variations in boundary conditions, material properties and geometries to the variation in buckling modes and compression loads up to the first failure. This paper shows the development of a stochastic methodology to identify the impact of variation in input parameters on the response of stiffened composite panels and the development of a robust index to support the evaluation of panel designs. The stochastic analysis included the generation of metamodels that allow quantification of the impact that the inputs have on the response using two first order variables, influence and sensitivity. These variables were then used to derive the robust indices to quantify the response of two COCOMAT panels that were experimentally tested, including the response of the panels to simulated damage. The robust indices that are shown in this paper are functions of the robustness parameter which has been recommended in the final Design Guidelines for the COCOMAT project to measure the effects of scatter found in postbuckling loads.

A torsional buckling model of cylindrical shells with asymmetric local thickness defect is established based on the Hamiltonian system. The critical load and torsional buckling mode of cylindrical shells with defects are obtained by the symplectic eigensolution expansion method, which overcomes the difficulty of constructing the deflection function of the traditional semi-inverse method. Local buckling modes can be captured by this new analytical model with the superposition of symplectic eigensolutions. To ensure accuracy and validity of the symplectic method, the analytical solution with torsional buckling of a cylindrical shell is compared with the classical solution and the finite element method (FEM) solution. The results show that the most detrimental position of the defect is only related to the width of the defect, not to the depth. The local defect changes the circumferential buckling wave number of the cylindrical shell and concentrates the torsional corrugation on the side containing the defect. Torque symmetry is broken due to the asymmetric defect, and the most detrimental defect direction for buckling is the same as the direction of torsional buckling wavelet.