Second-order analysis of steel structures has been studied extensively in the past decades by a number of researchers, who assume steel members are perfectly straight and very little work has been carried out on structural response by considering imperfections. The detrimental influence of imperfection on the buckling of axially loaded members is unknown and is most deleterious. Many design codes such as the BS5950 (2000), AS.4100-1990, Eurocode 3 (1993) and LRFD (1993) reckon the importance of imperfections in practical structures, and have stipulated as design requirements, however they fail to give clear guides for the fellow design engineers to tackle complex structural problems. Further, most research works on advanced analysis have been concentrated on framed structures and very little works are carried out on space structures and trusses subjected to high axial force and small moments. From structural point of view, trusses and shallow domes are more susceptible to buckling and the accurate assessment of effective length is much difficult.
This thesis is to investigate the design and analysis of practical imperfect steel structures using a performance-based approach. Both P-δ and P- ∆ effects are considered in order to model the response of a practical structure in the pre- and post-buckling stages by superimposing overall imperfections onto the initial structural geometry. A new idea of using equivalent geometrical imperfection to simulate the effect of imperfections in a second-order analysis is proposed in this thesis. The findings and recommendations will form the basis for fixture second-order analysis and design of steel structures allowing for global and local imperfections, which are unavoidable in real structures. It is hoped that the proposed method will be of use for practicing structural engineers in their routine design and for code drafters in drafting design codes.
The work has been devoted to the analysis and design of space structures that are susceptible to snap-through buckling. A one-element-per-member concept to include the initial member imperfection is proposed, which is almost a pre-requisite for practical second-order analysis, simply because of its consistency with a linear analysis model and also its simplicity to model member imperfections. An elastic eigen-buckling mode onto initial structural geometry for global imperfection has also been proposed, for which the conventional notional lateral load approach and effective length method are not applicable for determining the buckling load of complex structures with curved surfaces. The work will also be extended to the analysis and design of unbraced planar truss composed of symmetrical rectangular and circular hollow sections. A computer method for determining the buckling load of a structural system with allowance for the stiffening-tension member effect is proposed. Unlike the conventional effective length method that the inconvenience and unreliable assumption of effective length is eliminated. The advanced analysis allowing for formation of many plastic hinges is not investigated here because of the excessive ductility requirement for steel material, which may not be available for the type of structures studied in this thesis, which is slender and controlled by buckling failure.