Computational & Technology Resources
an online resource for computational,
engineering & technology publications
Civil-Comp Proceedings
ISSN 1759-3433
CCP: 79
PROCEEDINGS OF THE SEVENTH INTERNATIONAL CONFERENCE ON COMPUTATIONAL STRUCTURES TECHNOLOGY
Edited by: B.H.V. Topping and C.A. Mota Soares
Paper 26

Finite Element Initial Buckling and Postbuckling Initial Failure Analysis of Short Pultruded GRP Columns Subjected to Uniform End Compression

G.J. Turvey and Y. Zhang

Engineering Department, Lancaster University, United Kingdom

Full Bibliographic Reference for this paper
G.J. Turvey, Y. Zhang, "Finite Element Initial Buckling and Postbuckling Initial Failure Analysis of Short Pultruded GRP Columns Subjected to Uniform End Compression", in B.H.V. Topping, C.A. Mota Soares, (Editors), "Proceedings of the Seventh International Conference on Computational Structures Technology", Civil-Comp Press, Stirlingshire, UK, Paper 26, 2004. doi:10.4203/ccp.79.26
Keywords: pultruded GRP, short column, compression test, finite element analysis, postbuckling, failure.

Summary
Investigations of the initial local and overall buckling of pultruded GRP columns have been reported by Yoon et al. [1] etc. However, investigations of the postbuckling and failure response of pultruded GRP columns are scant and the collapse loads of these columns may only be determined by testing.

Recent work by Bank and Yin [2] on a pultruded GRP WF beam in four-point bending has shown that failure is by tearing of the compression flange from the web, caused by the growth of local buckling deformations.

Tests, carried out by the authors, on short pultruded GRP WF columns have also shown that failure may be due to tearing at the web-flange junction or longitudinal cracking in the web. In order to be able to predict the collapse loads of columns, numerical analysis tools must be able to model these failure modes. This paper reports on the authors' first steps towards the realisation of this goal.

A brief account is given of the material property tests carried out on web and flange coupons and web-flange junction specimens cut out of WF section columns. The values of elastic modulus and ultimate strength determined from the tests are also given. In addition, an outline of two series of axial compression tests on short columns with lengths ranging from 200 to 800mm is presented.

This is followed by details of the nonlinear FE model for the short column analyses. The ANSYS [3] FE software was used to construct the models from 240 eight-node shell elements with six degrees of freedom per node. The column end conditions used in the buckling and postbuckling analyses are explained. The principal methods available within ANSYS for computing the postbuckling response are also reviewed and the reasons for selecting the Arc Length method are given. It is also noted that the ANSYS software does not allow damage progression to be simulated. Hence, only postbuckling initial failure could be simulated and, moreover, only via post-processing of the nonlinear analysis results.

Two failure criteria are introduced for the web-flange junctions. They use the failure strengths derived from the tests on web-flange junction specimens. The Tsai-Wu failure criterion is used to predict the onset of failure in the web and flange and uses the strength data obtained from the coupon tests.

The procedure for identifying the postbuckling initial failure load and the failure site(s) is explained. In the first part of the procedure, the longitudinal distributions of surface stresses and failure indices are examined for a few values of load, which span the failure load. This ensures that the buckling mode is unchanged and identifies the likely failure sites over the length of the column. In the next part of the procedure, the failure indices are evaluated for all of the small load steps between the upper and lower load limits at the element nodes corresponding to the previously identified likely failure sites. The values of the maximum failure indices are then plotted against load and the lowest load for a unit value of the failure index is the postbuckling initial failure load.

In the penultimate section, the end-shortening displacements and surface strains recorded in the column tests are shown to accord well with the values predicted by the nonlinear analysis. The linear (eigenvalue) and nonlinear buckling loads are compared with the experimental buckling loads. In the latter two cases the buckling loads are derived from Southwell plots of the numerical and test data. It is shown that the nonlinear buckling loads are slightly lower than the linear buckling loads and are closer to the average experimental buckling loads. The predicted postbuckling initial failure loads are also compared with the experimental initial failure loads. Again, there is reasonable agreement between the results.

Finally, it is concluded that the nonlinear FE analysis, combined with the author's web-flange junction failure criterion and the Tsai-Wu failure criterion, is able to predict the local buckling modes of pultruded GRP short columns accurately, and their buckling loads and postbuckling initial failure loads reasonably accurately.

References
1
S.J. Yoon, D.W. Scott and A. Zureick, "An experimental investigation of the behaviour of concentrically loaded pultruded columns", in "Proceedings of the 1st International Conference on Advanced Composite Materials in Bridges and Structures", K.W. Neale and P. Labossiere (Editors), Canadian Society for Civil Engineering, Montreal, 309-317, 1992.
2
L.C. Bank and J. Yin, "Analysis of progressive failure of the web-flange junction in post-buckled I-beams", Journal of Composites in Construction, 3(4), 177-184, 1999. doi:10.1061/(ASCE)1090-0268(1999)3:4(177)
3
Anon., "ANSYS 8.0 Documentation", ANSYS Inc., Canonsburg, Pennsylvania, 2003.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £135 +P&P)