Buckling of extensively welded steel cylindrical shells under axial compression

Pao Yue-kong Library Electronic Theses Database

Buckling of extensively welded steel cylindrical shells under axial compression


Author: Lin, Xiang
Title: Buckling of extensively welded steel cylindrical shells under axial compression
Degree: Ph.D.
Year: 2004
Subject: Hong Kong Polytechnic University -- Dissertations
Shells (Engineering) -- Testing
Cylinders -- Testing
Buckling (Mechanics)
Department: Dept. of Civil and Structural Engineering
Pages: xix, 393 leaves : ill. ; 30 cm
Language: English
InnoPac Record: http://library.polyu.edu.hk/record=b1781102
URI: http://theses.lib.polyu.edu.hk/handle/200/132
Abstract: Thin cylindrical shells in large civil engineering steel shell structures such as steel silos and tanks are subject to both axial compression and internal pressure. Buckling under axial compression often governs their design. These shells are commonly constructed by joining a large number of rolled panels using many short vertical welds and continuous circumferential welds. This fabrication method leads to unique geometric imperfections which have been argued to be severely detrimental to the shell axial buckling strength. In addition, welding-induced residual stresses in these shells may also have a significant effect on this buckling strength. A review of the existing literature reveals that few previous studies have examined the buckling behavior of these shells with a rigorous treatment of the geometric imperfections and residual stresses in these shells. This thesis presents both theoretical and experimental studies aimed at correcting this important deficiency in existing knowledge. The first part of the thesis, consisting of Chapters 3-5, presents the results of theoretical studies into the characteristics of geometric imperfections measured in three large steel silos at Port Kembla, Australia, and their effect on buckling strength. In order to accurately represent these imperfections measured at non-uniformly spaced sampling points, a new iterative Fourier decomposition method was proposed. The traditional Fourier decomposition method cannot produce accurate Fourier representations, and the reasons for this difficulty are clarified in the thesis. This new iterative method was then employed in deducing accurate Fourier series representing the measured imperfections. It is explained in the thesis that these real imperfections, due to their non-zero values at the boundaries, can only be accurately represented using half-wave cosine series in the meridional direction, an important phenomenon overlooked by all previous researchers. The characteristics of the imperfections are next examined by studying the spectra of the deduced Fourier series, including both ID and 2D Fourier series. It is shown that the dominant components of the imperfections in each shell are in the lower Fourier terms, with some of these components being closely associated with the weld pattern. Finite element analyses of the three silo shells (referred to as Shells A, B, and C) showed that these shells have buckling strengths which are much lower than those of shells with many known forms of imperfections. A systematic parametric study revealed that the buckling strength of Shell A is dominated by significant axisymmetric imperfections, while those of Shells B and C are governed by highly non-symmetric imperfections. Based on the specification of Eurocode 3, an automated computer procedure was developed to numerically 'measure' the local dimples and out-of-roundness of these shells. This was undertaken to determine the quality classes of the three shells so that their buckling strengths as predicted by Eurocode 3 could be found. A comparison between the predictions of the code and finite element results from nonlinear analyses reveals that the existing rule in Eurocode 3 is too optimistic about shells of Class A but too pessimistic about shells of Classes B and C. While detailed measurements of imperfections are available for the three large silos, the residual stresses in these shells due to welding are unknown. As residual stresses may have a significant effect on the buckling strength, an experimental study aimed at an indirect evaluation of this effect is presented in Chapters 6-8. In this experimental study, an innovative two-stage model fabrication technique was proposed and implemented with a sophisticated experimental set up for the fabrication of small models simulating large cylindrical shells with many welds. Three model shells were fabricated and tested under axial compression, each of which had a different weld pattern. The initial geometric imperfections in these model shells were precisely measured and decomposed into Fourier series. The imperfections were also found to be dominated by lower Fourier terms and closely associated with the weld patterns. The buckling loads of the model shells are very low compared to the value of a corresponding perfect shell, demonstrating the serious deleterious effect of the imperfections in these shells. Finite element buckling loads obtained from non-linear analyses with the measured imperfections accurately modeled are also presented. These buckling loads are close but slightly lower than the experimental buckling loads, indicating that the effect of residual stresses and other factors in these shells is small but beneficial, and may be safely ignored. Finally, a parametric study is presented in Chapter 9, in which the buckling behavior and strength of the three real silo shells under the combined action of axial compression and internal pressure were explored. Nonlinear finite element analyses revealed that the strengths of Shell A are considerably lower than those of Shells B and C, as the dominant imperfection components are significantly axisymmetric in Shell A, but are highly non-symmetric in Shells B and C. Comparisons between the finite element results and the predictions of Eurocode 3 indicate that the code overestimates the failure strengths of Shell A considerably. Modifications to the existing design rule are thus presented to achieve safe predictions.

Files in this item

Files Size Format
b17811028.pdf 16.49Mb PDF
Copyright Undertaking
As a bona fide Library user, I declare that:
  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.


Quick Search


More Information