Quick and clean technology for composite cylinder stability analysis

• About
• Latest Posts

Harri Katajisto

Harri holds a Master’s Degree in Mechanical Engineering from the Helsinki University of Technology with a focus on lightweight structures. He has been the member of the ESAComp team since 2000 and joined Altair in 2017. He considers himself as a middleman working closely with industry, research and software development aiming to provide practical software solutions for the Composites Industry together with the Altair Composite Technology Team.

Latest posts by Harri Katajisto (see all)

• ESAComp introduces “through-the-width” debonding/delamination analysis - September 12, 2017
• Structures are not ideal – Challenges of buckling analyses - October 15, 2015
• Partner Spotlight: HyperWorks 14.0 brings further integration with partner products - March 22, 2016

Thin-walled cylindrical shells (radius-to-thickness ratio of 160-500) under different loading conditions have been studied for structural stability (Figure 1). Based on this study, guidelines for the selection of the element mesh density have been set.

Figure 1. Cylinder load cases

A set of problems presented in References [1-5] (Figure 2) has been solved with ESAComp using different mesh densities. The load types considered were axial compression, torsion, and bending. The number of elements in the hoop direction was set to vary from 40 to 240 while the element aspect ratio was set to one. Convergence studies were made and the results are presented for the axial load in Figure 3a. Respectively, results for the torsion and bending cases are shown in Figure 3b. The reserve factor against buckling (indicated on the vertical axis) has been normalized against the value obtained with 240 elements in the hoop direction. A short summary for the problems are presented in the table shown in Figure 2. The deviation (%) indicates the difference between the ESAComp results with respect to the reference results at the converged state.

Figure 2. Cylinder configurations

As an outcome of this study it can be concluded that in axial load dominated buckling 160 elements are needed in the hoop direction to achieve reasonable accuracy, i.e. the deviation from the converged result is less than 10%. For torsion and bending dominated buckling, 120 elements is a sufficient number to guarantee the same accuracy since in these problems buckling wave lengths are longer. It should be noted that in this study cylinder diameter-to-length ratio was equal or close to one.

Figure 3. Convergence as a function of elements in the hoop direction

The study demonstrates that ESAComp is reliable tool for the design-analysis of cylindrical shells and a valuable benchmark tool to be used together with other software tools.