Some concepts and results dealing with the design and safety checking of members integrated in plane steel frames are presented and discussed. Initially, attention is paid to the identification and clarification of a number of ambiguities related to the application of the buckling length concept to frame members. Then, the safety checking of columns integrated in frames is addressed and it is shown that, if their buckling lengths are "correctly determined", only one particular column, designated as "critical column", needs to be checked − a finding which leads to the proposal of a "frame optimisation procedure". Next, the safety checking of beam-columns integrated in frames is dealt with: the application of the interaction formulae appearing in the upcoming EN version of Eurocode 3 is addressed and particular attention is paid to the appropriate choice of the buckling length and "equivalent moment factor" values, both in terms of safety and accuracy. In addition, one proposes an alternative approach to use the beam-column interaction formulae, which is based on the results of genuine second-order elastic analyses. In order to illustrate the application and assess the validity and advantages of the concepts and procedures presented throughout the paper, one presents numerical results concerning simple (two-bar) structural systems and these results are compared with "exact" frame ultimate (collapse) load values, yielded by second-order plastic zone analyses that incorporate member initial imperfections. On the basis of the above comparative study, it is possible to draw several conclusions and, in particular, it is shown that the proposed approaches consistently yield accurate and conservative frame strength estimates.Key words: Steel frames, columns, beam-columns, buckling length, equivalent moment factors, non-dimensional slenderness, frame slenderness, Eurocode 3.
INTRODUCTIONIdeally, the design or safety checking of steel frames should be carried out on the basis of rigorous geometrically non-linear elastic-plastic structural analyses, which incorporate member and frame imperfections and provide "exact" frame load-carrying capacities. However, in spite of the fast growing popularity of the so-called "advanced methods of structural analysis"[1,2] − some of them are already allowed by various existing steel design codes (e.g., the current and upcoming versions of Eurocode 3 or simply EC3 [3,4]) -, their use remains prohibitive for routine applications. This stems mostly from the fact that (i) the computational effort required is still quite high and (ii) the vast majority of designers lack the appropriate theoretical background. Thus, the most "traditional" approach, based on first-order internal forces and moments and individual member checks through beam-column interaction formulae, continues to be widely adopted by practitioners. Nevertheless, since the interaction formulae must incorporate all relevant second-order and plasticity effects, an intense research activity is still going on concerning th...