Validation of FE models for buckling analysis of woven GFRP shells

Citation
Mk. Chryssanthopoulos et al., Validation of FE models for buckling analysis of woven GFRP shells, COMP STRUCT, 49(4), 2000, pp. 355-367
Citations number
15
Categorie Soggetti
Material Science & Engineering
Journal title
COMPOSITE STRUCTURES
ISSN journal
02638223 → ACNP
Volume
49
Issue
4
Year of publication
2000
Pages
355 - 367
Database
ISI
SICI code
0263-8223(200008)49:4<355:VOFMFB>2.0.ZU;2-#
Abstract
This paper gives the details of a numerical finite element validation study for laminated GFRP cylinders subjected to concentric and eccentric compres sion. The laminates are of type 'Rovimat 1200' consisting of woven glass fi bre roving, with a chopped mat on one side, within a polyester resin matrix . Two and three-ply cylinders with various orthogonal orientations are cons idered, for which the nominal radius-to-thickness ratio is about 108 and 72 , respectively. The numerical results are compared to findings from a previ ous experimental investigation in which detailed measurements were obtained . Following a brief description of the experimental work, details of the de velopment of suitable finite element models are presented and associated li mitations are highlighted. Careful attention is given to thickness idealisa tion as well as the introduction of geometric imperfections into the numeri cal models. Both linear eigenvalue analysis and geometrically nonlinear sim ulations are undertaken using a general purpose finite element program, The correlation between numerical and experimental results is discussed in ter ms of buckling strength, axial stiffness, buckling deformations and surface strains. The analysis is shown to give a good representation of the buckli ng behaviour of GFRP cylinders of the type examined. It is also concluded t hat whereas the cylinders appear to be less sensitive to the effects of ini tial geometric imperfections than their isotropic counterparts, including s uch imperfections in a geometrically nonlinear analysis does improve the co mparison between tests and finite element results, and is considered essent ial for the derivation of numerical 'knockdown' factors. (C) 2000 Elsevier Science Ltd. All rights reserved.