An equivalent imperfection-based FE simulation of the stability of dented cylindrical shells accounting for unintended imperfections

Metal cylindrical shells subjected to axial compressions are prone to buckling. The buckling strength is highly sensitive to geometric imperfections. Experimentally tested cylinders with intended imperfection have already, in most cases, inevitable (unintended) forms of imperfections, such as weld depressions at the weld joints, irregularity in the shape of the cylinder and imperfections due to loading/supporting conditions. The interaction between the effects of the intended and the unintended imperfections in the experimental investigations makes the determination of the sole effect of either one hard to be identified. To evaluate the buckling strength reduction due a certain intended form of imperfection such as a localized dent, the effect of the other hidden (unintended) imperfections should be accounted for, otherwise a misleading strength reduction value will be obtained. In this study, a simple approximate procedure accounting for the effect of the unintended imperfections is proposed and explained though a finite element (FE) analysis of experimentally tested specimens with intended dent imperfections. Geometric imperfections equivalent in their strength reduction to that of the unintended imperfections are introduced to the FE analysis and referred to as equivalent imperfection. The experimental value of the buckling strength for the intact specimen (without a dent but having unintended imperfections) along with the FE model are used to calibrate the equivalent imperfection that represents all possible forms of unintended imperfections. On the light of this study, general recommendations for conducting combined experimental-numerical parametric studies are proposed.

Automated summary

This study investigates the buckling behavior of metal cylindrical shells subjected to axial compressions. The research highlights the complex interaction between intended and unintended geometric imperfections in experimental investigations, making it challenging to identify the sole effect of either one. A simple approximate procedure is proposed to account for the effect of unintended imperfections in order to evaluate the buckling strength reduction due to certain intended forms of imperfection. The study also suggests general recommendations for conducting combined experimental-numerical parametric studies.