Reliability assessment of buckling strength for imperfect stiffened panels under axial compression

摘要

Designing stiffened panels requires evaluating reliability of these structures with regards to the collapse limit state as it could be affected by the presence of general localized defects. Considering the case of a small square depression located on the surface panel, the buckling strength under in-plane uniform axial compression was evaluated through nonlinear finite element modelling. Artificial neural networks were introduced for representing the collapse load as function of the key intervening design variables. A full factorial design of experiment table constructed on these variables provided samples for the training phase, and complementary samples were used to test and validate the obtained models. Separating the various sources contributing to variability of the buckling strength, Monte Carlo method was used to evaluate the probability of failure as function of the applied compression load acting on the stiffened panel system. It was found that localized defects have a drastic effect on the reliability probability. For the considered geometric parameters and boundary conditions, the localized defect present on the central segment of the stiffened panel was recognized to be the most severe one.