Probabilistic imperfection sensitivity analysis of axially compressed composite cylinders

• Published in: Engineering structures Vol. 17; no. 6; pp. 398 - 406
• Main Authors: Chryssanthopoulos, Marios K., Poggi, Carlo
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier Ltd 01.07.1995; Elsevier
• Subjects: Applied sciences; axial compression; buckling; Buildings. Public works; composites; cylinders; Exact sciences and technology; imperfections; manufacturing effects; probabilistic modelling; Stresses. Safety; Structural analysis. Stresses; composites; buckling; imperfections; probabilistic modelling; axial compression; manufacturing effects; cylinders; Compression; Sensitivity analysis; Methodology; Metals; Tolerance; Statistical methods; Defects; Design; Cylinder; Composite materials; Probabilistic model; Manufacture; Buckling; Strength
• ISSN: 0141-0296; 1873-7323
• DOI: 10.1016/0141-0296(95)00048-C

Buckling analysis of cylinders under axial compression is sensitive to the assumptions made in the modelling of initial imperfections. Normally, imperfection modes are selected solely on the basis of buckling mode considerations and their amplitudes are determined using tolerance specifications in codes or experimentally recorded values. Whilst this approach may be used for metal cylinders with some confidence, due to the many test results available for validation purposes, it is not appropriate for the analysis and design of fibre-reinforced composite cylinders where the test results are limited and the effects of manufacturing on the imperfection characteristics have not yet been studied in detail. This paper presents a methodology for probabilistic buckling strength assessment based on the results of a statistical analysis on imperfections on two groups of composite cylinders manufactured by lay-up. The dominant features are quantified and the effect of fibre orientation on imperfections is examined. Simple models describing the random variability of imperfection modal amplitudes are presented. Using these probabilistic models, characteristic imperfection shapes are developed for fibre-reinforced cylinders and their use in buckling strength prediction and tolerance specification is demonstrated