Search

This paper describes a finite element analysis (FEA) and design of cold-formed steel built-up closed sections with intermediate stiffeners in the flange and web under axial compression. A finite element model (FEM) was developed and validated using the available experimental results. The validated FEM can be used for further parametric study on strength of built-up closed columns. The results obtained from the parametric study are compared with the current direct strength method (DSM) in the North American specification of cold-formed steel structures for those specimens that failed by only local, distortional, and flexural buckling. It is shown that the current DSM is not quite suitable for the design of cold-formed steel built-up closed columns. Therefore, the improved design equations are proposed by modifying the current design equations used in the DSM for flexural, local, and distortional buckling. The reliability of current DSM and proposed DSM was assessed by reliability analysis.

The structural steel design specification in Canada, CSA S16-14, uses the same equations for the design of rolled and welded shape beams for lateral torsional buckling (LTB). A recent study has shown that the current design equations might overestimate the capacity of welded wide flange (WWF) beams. This paper evaluates the performance of the current design equations in providing LTB capacities of welded I-shape beams (WWF beams). An extensive finite element (FE) analysis is performed for simply supported WWF beams subjected to constant moment. In total, 256 FE models are analyzed and it is observed that both CSA and AISC overestimate the LTB capacity of welded I-shape beams by as much as 37%, particularly when residual stress measured at Lehigh University is used in the analysis. Also, the Eurocode is found to be conservative and the proposed equation by MacPhedran and Grondin in 2011 provides better predictions of LTB strengths of WWF shapes than the current CSA approach.

Recent developments in novel engineered mass timber products and connection systems have created the possibility to design and construct tall timber-based buildings. This research presents the experiments conducted on the steel-wood connection as main energy dissipating part of a novel steel–timber hybrid system labelled Finding the Forest Through the Trees (FFTT). The performance was investigated using quasi-static monotonic and reversed cyclic tests. The influence of different steel beam profiles (wide flange I-sections and hollow rectangular sections) and the embedment approaches (partial and full embedment) was investigated. The test results demonstrated that appropriate connection layouts can lead to the desired failure mechanism while avoiding excessive crushing of the mass timber panels. The research can serve as a precursor for developing design guidelines for the FFTT system as an option for tall wood-hybrid building systems in seismic regions.

It is well-known that the lateral–torsional buckling capacity of a wide-flange beam is lower when transverse loads between brace points are applied to the top flange instead of at the shear centre. Although top-flange loading is commonplace, the load height effect can often be neglected due to lateral restraint provided by the loading mechanism itself; however, there are several common cases for which it must be considered in design. This paper provides the research background to the current provisions in CSA Standard S16 to account for the load height effect and proposes an effective length method that expands on these provisions to reduce their conservatism, while maintaining a simplicity appropriate for use in design. Employing a large and diverse database of steel shapes, the proposed method is compared with a comprehensive variety of other design approaches and available numerical and experimental data.

This paper addresses the design of wide-flange steel members subjected to torsional forces as well as axial forces and moments about their strong and weak axes. The current Canadian steel design standard (CSA S16–01) gives no specific guidance on methodology with respect to torsional design. Codes of other countries (American, British, Australian) provide useful insight but are different in format from the Canadian standard and cannot be used directly in conjunction with it. Specialized second-order finite element programs have the capacity for torsional analysis, but are too complicated and costly to use in design practice. There is, therefore, a need for a practical design method that would allow engineers to account for the effects of torsion simply and accurately. This paper, written by a practicing design engineer, suggests a number of approaches that, subject to discussion and approval by experts in the field, could constitute the basis for design of steel members in torsion.

This study investigated the effectiveness of retrofit methods in improving the deformation capacity of existing moment connections in composite beams. Nonlinear finite element analysis was also performed to identify the cause of the connection failure prior to testing. Tests included one conventional specimen and four retrofit specimens. Three retrofit methods were introduced: a reduced beam section (RBS) only, an RBS with bottom flange reinforcement (RBR), and an RBS-shaped bottom flange reinforcement (RSR). A composite beam performs differently than a bare steel shape because the concrete compression flange greatly increases the tensile demands on an unreinforced bottom flange. Therefore, these retrofit methods were applied only to the bottom flanges of the beam. Quasi-static loading tests were performed. The main conclusions were as follows: (i) the deformation capacity of the existing moment connections in composite beams is not sufficiently improved by RBS alone; and (ii) both the RBR and RSR details move the plastic hinge away from the face of the column and reduce stress levels in the vicinity of the beam bottom flanges, sufficiently improving the deformation capacity.Key words: composite construction, seismic performance, reduced beam section (RBS), flange reinforcement, ductility, deformation capacity.

The strength of a slender section is governed by local buckling of its elements. The local buckling mode is mainly affected by the width to thickness ratios of elements forming the section and the member slenderness ratio. A nonlinear finite element model was developed in this paper to study the effect of the mentioned factors on the capacity of biaxially loaded slender I-section beam–columns. A group of I sections was selected with varying flange width to flange thickness and web width to web thickness ratios. Different member slenderness ratios were chosen to study the different modes of failure. The results show that the linear interaction equation is safe and provides a good fit to the ultimate capacity of members with intermediate slenderness ratios. It is conservative, however, for short member slenderness ratios. Lastly, compared with the newly developed finite element model, the 1996 edition of the American Iron and Steel Institute specification for cold-formed steel structural members and the European Committee for Standardization Eurocode 3 conservatively predict the capacity of members, especially for large ratios of minor axis to major axis bending.Key words: structural engineering, steel, stability, buckling, biaxial loads, ultimate strength.

Both the Canadian Institute of Steel Construction (CISC) Handbook of Steel Construction and the American Institute of Steel Construction (AISC) Manual of Steel Construction discuss prying action in tee-type and angle-type connections subjected to tensile force, but no guidance is provided for determining the prying force in bolted steel circular flange connections. However, such connections are very common in leg members of guyed lattice communication towers. To determine the magnitude of the prying forces in such connections, an experimental investigation was conducted on 10 bolted steel circular flange connections. On the basis of the test data, it is proposed that for use with equations in the CISC handbook and AISC manual, the length of the flange tributary to each bolt (bolt pitch) be taken as the circumference of the bolt circle divided by the number of bolts.Key words: bolt pitch, bolted connections, circular flange, prying action, tensile force.

This article presents an experimental study on bolted T-shaped end plate connectors subjected to a static load. Testing models may or may not include stiffeners on the flange of the T-shaped section. These connectors are tested for modelling the section in tension in a beam-column connector bolted with a projected end plate. The thickness of the end plate (flange of the T-shaped section) constitutes the main parameter in this study. Test results highlight the role of the end plate in different possible failure modes. A numerical calculation based on a model taken from the literature permitted comparison with test results. The analysis of results showed some discrepancies. Improvements must be made to the model used to reconcile the theory and the experience, and to adapt it to connectors that are under tension.Key words: bolted assembly, end plate, experimentation, stiffening, modelling.

This paper presents results of an experimental investigation into the effect of flange geometry on the shear strength of point-loaded, reinforced concrete T-beams. A procedure to normalise the ultimate strength and calculate a concrete contribution is implemented. This is used to discuss the effect of varying the ratio of flange width to web width and the ratio of flange depth to effective depth on the shear strength of a reinforced concrete T-beam. An increase in the ratio of flange width to web width is shown to produce an accompanying increase in the ultimate strength of a reinforced concrete T-beam, providing the ratio of flange depth to effective depth is above a certain minimum value. This increase in shear resistance with an increase in the ratio of flange width to web width continues until the flange is wide enough to allow formation of a failure mechanism whereby the load point punches through the flange.Key words: shear, T-beams, flange, reinforcing, strength, failure mechanisms.

Plate structures are often subjected to differential compression due to nonuniform bending during their service life. In practice, most beams are subjected to transverse loads that create moment gradient; the resulting loading system for such beams is a combination of longitudinal stress gradient with associated shearing and transverse stresses. This paper presents a theoretical view of elastic local instability of anisotropic composite beams which are treated as assemblies of symmetric angle-ply composite plates buckling under nonlinear varying, uniaxial compressive forces. Uniformly distributed load and central concentrated load are two common loading systems acting on simply supported, wide-flange I-section beams. The Galerkin method is applied to the resulting plate equilibrium equations. The present study shows that the buckling load obtained from the proposed approach for composite flanges is greater than that obtained by existing solutions. Results also show that for beams subjected to the uniformly distributed load loading system, local buckling may be initiated in the web, especially for short-span deep beams. However, under the central concentrated load loading system, the compressive flange limits buckling strength.Key words: local buckling, stress gradient, fibre composite, pultruded beam.

Lateral torsional instability of I-beams considers the relative displacement of the unstable compression flange to the stable tension flange. It is commonly assumed that little or no distortion takes place between the two flanges. In this paper, an approach is proposed whereby the section properties that control lateral torsional buckling are adjusted to allow for the influence of cross-section distortion, by the use of simple spring models representing the relative stiffness of the flanges and the web. The model is developed for elastic, inelastic, and plastic cases and compared with the results obtained from finite element models developed by other researchers. A method of quantifying the lateral distortional buckling resistance of I-beams is of particular importance in the hogging moment region of continuous composite beams; the ability of the proposed model to deal with this complex problem has been previously demonstrated by the authors. In this paper, the proposed model is used to illustrate the influence of distortional buckling on doubly symmetrical I-sections.Key words: lateral, distortional, I-beams, elastic, inelastic.

Perforations are often provided in the web and (or) flange plates of beams and columns of cold-formed steel (CFS) structural members in order to facilitate duct work, piping, and bridging. This paper is concerned with the establishment of effective design width equations for the determination of the ultimate strength of such perforated members in compression. A proven finite element model has been used to study the effects of perforation parameters on the ultimate strength of perforated members. The finite element model consists of short columns of lipped channel CFS sections, discretized using nonlinear "assumed strain" shell finite elements, and utilising experimental-based material properties models. The parametric study covers web slenderness values between 31 and 194, perforation width to web width ratios up to 0.6, and perforation height to perforation width ratios up to 3.0. Effective design width equations for plates having square perforations and elongated perforations were developed. The efficiency and accuracy of these two equations in predicting the ultimate strength of perforated CFS compression members have been verified through a comparison with the ultimate load results of several experimental studies from the literature.Key words: cold-formed steel, compressive loads, local buckling, perforations, finite element analysis, experimental, post-buckling strength, ultimate strength, effective width, design.

A full-scale laboratory test was conducted on a unique hybrid bridge girder with a reinforced concrete web and steel flanges. Half-width precast concrete deck panels were compositely attached to the top steel flange of a 17.1 m long hybrid girder to construct a "half-bridge" that was tested to study the service load behaviour and the behaviour under increasing load until failure. It was observed that the concrete web of the hybrid girder cracked because of the combination of dead weight and restrained shrinkage. Under the service loads, the concrete web had numerous closely spaced cracks that were reasonably well controlled - the maximum crack width was 0.20 mm. Although there was significant diagonal cracking in the web of the girder, the ultimate behaviour was dominated by flexure. This paper presents the methods used to construct and test the half-bridge, and a summary of the important results. A companion paper presents a detailed analysis and discussion of the test results. Key words: bridges, composite, cracking, girder, hybrid, reinforced concrete, structural design, tests.

Test results are presented and discussed for 18 beam-column specimens subjected to flange local buckling resulting from applied axial and flexural forces. Flanges classified as Class 2 by CAN3-S16.1-M84 “Steel structures for buildings (limit states design)” are investigated. Three different sizes of W shape with different height and web thickness ratios within the Class 2 range were used in this investigation. The flange width-to-thickness ratio was kept constant at the current Class 2 limit [Formula: see text]. Results of this investigation show that the conventional value of plastic moment reduced in the presence of axial load, as based on studies in the late 1950s, does not apply to Class 2 sections. This discrepancy had previously been predicted by an analytical study using an extended Rayleigh–Ritz technique and presented in 1980. Experimental results presented herein verify these predictions. Key words: axial, beam-column, buckling, compact, Class 2, interaction, slenderness.

The behaviour of cantilever-suspended span systems is influenced by the type of loading and the presence of flange restraint. An experimental program consisting of 33 full-scale tests was undertaken to examine the stability of beams in a single overhang configuration. The results indicate that torsional flange restraint significantly enhances the buckling resistance and is particularly effective when combined with a web stiffener. Lateral bracing of the bottom flange at column supports is of considerable importance in maintaining stability. Simulating the proper boundary conditions is essential, since test specimens are very sensitive to unwanted restraints and can buckle in a higher energy mode.A finite element model was developed that takes into account inelastic material behaviour, residual stresses and cross-sectional distortions. The predicted buckling capacities are in good agreement with experimental results. Key words: cantilever-suspended span, steel beams, distortional buckling, flange restraint, inelastic behaviour.

In this technical note, a spline finite strip for curved bottom flange of box-girder bridges is developed so that the spline finite strip method is extended to the analysis of continuous haunched box-girder bridges. This method is more capable in dealing with concentrated loads and is more flexible in treating discrete support conditions than the semi-analytical finite strip method. Key words: finite strip, box-girder bridges, spline function, structural analysis, composite.

The Canadian Standards Association Standard CAN-CSA-S16.1-M89 “Limit states design of steel structures” assigns welded wide flange (WWF) columns to the column curve for rolled H-shaped sections. This is conservative because of differences in the production of WWF and rolled sections. The WWFs are stipulated to have flame-cut edges. The residual stress pattern, with favourable tensile stresses, results in a delayed loss of stiffness as weak axis inelastic buckling occurs. Thus the weak axis and strong axis buckling curves lie closer together for WWF shapes than they do for rolled H-shapes. Automatic cutting and welding processes result in close tolerances on out-of-straightness. As well, the statistical variations in the geometric properties are favourable. A detailed analysis of data collected from mill records and on-site measurements was made to obtain statistical parameters of relevant geometric and material properties. A finite element program modelling inelastic behaviour, residual strain patterns, out-of-straightness, and material properties was used with the test results of others to establish test-to-predicted ratios of column strengths. Parametric studies provided an assessment of the effect of varying residual strain patterns and column out-of-straightness. This formed the basis for determining the factored compressive resistance of WWF sections for three different slenderness ratios. Key words: columns, inelastic analysis, out-of-straightness, residual stresses, welded wide flange.

Tests on full-scale connections between wide flange beams and rectangular hollow section (RHS) columns reveal various modes of failure as a function of connection geometry. Tension and compression moment plates with web clip angles or tension plates and seat angles welded to doubler plate reinforced RHS walls are investigated. Eight failure modes were identified from an experimental program consisting of ten full-scale specimens. Semi-empirical expressions based on assumed stress distributions predict failure modes and corresponding moment capacities reasonably well for various connection configurations. Further analytical parametric studies showed the importance of tension plate and doubler plate dimensions on joint strength and behaviour. Key words: RHS, W-shape, columns, beams, connections, experimental, analytical, design, steel.

The present work deals with formulation of theoretical and analytical methods leading to the development of column strength curves. The formulations were developed for both elastic and inelastic behaviour. Two types of reinforcement have been developed for strengthening the W-shape columns under load. Since the column strength curves are based in part on the magnitude and distribution of residual stresses, it is extremely important to consider the new pattern of residual stresses due to welding process. Also, the welding sequence will affect the magnitude and distribution of residual stresses. Theoretical formulations leading to a closed-form solution for the prediction of critical load were developed for two types of strengthening using the superposition of original residual, new welding, and initial loading stresses. A nonlinear finite element analysis based on the large deformation theory of stability was used to predict the strengthened column critical load. It takes into consideration the effect of cooling residual stresses and new welding residual stresses. The formulations were incorporated with gradual penetration of yielding, the spreading of inelastic zones along the member length, the presence of residual stresses, and strain hardening of the material. Experiments were carried out to determine the actual capacity of strengthened columns. Seven specimens were tested using two and four strengthening plates. The welding stresses were measured through a series of experiments, and it was found that the parabolic distribution is a very close approximation to the actual new welding stress distribution. Key words: reinforcement of steel columns, welding stresses, welding sequence, strengthening of existing structures, buckling, steel plating, finite element.

This paper presents the design and development of moment connections between wide flange beams and square tubular columns in rigid frames. The connection is effected by means of coped strap angles used to transfer stress from the beam flanges to the sides of the column. The testing program comprised 23 specimens, including two standard rigid joints introduced to give a comparison basis for the [Formula: see text] curves.The results indicate that after adequate precautions have been taken to prevent the occurrence of brittle fracture in the strap angles a connection is obtained which can develop MP and sustain appreciable inelastic deformation, making it suitable for plastically designed frames. The rigidity of the connection is less than that of a standard connection but high enough for the joint to be termed 'nearly rigid'. Preliminary theoretical studies indicate that the effect of the relative loss of rigidity on the sway behaviour of multistory frames may be significant and that P-Δ effects should be investigated.All joints considered here comprised beams of flange width equal to that of a column face.A design method is proposed in the Appendix.

End plates welded transversely to the web of the beam and then bolted to the supporting member should provide effective, simple connections to transmit shear. However, under loading, the end of the beam rotates and moments are developed in the connection. It is essential that the end plate connection be sufficiently flexible so that the factored moment can be attained in the beam without the development of excessive moments in the connection itself. Based on the behaviour of T-sections, where the flange of the T simulates the end plate and the stem simulates the web of the beam, both when the stem is loaded in tension and in compression, analytical procedures have been developed to predict the moment–rotation behaviour of end plate connections. Good agreement has been obtained between the predicted nonlinear moment–rotation curves and the experimental curves for eight connections covering a practical range of end plate thicknesses, gage distances between bolt holes, and connection depths. With the proposed analytical expressions the designer is able to predict the behaviour of a connection to ensure it has sufficient strength and flexibility. Key words: beams, connections, deformation, design, end plates, plates, rotation, shear strength, steel, stresses, tests.

The Tsing Yi South Bridge was constructed in the early 1970s to provide access between Tsing Yi Island and the mainland at Kwai Chung. It has a prestressed concrete box girder superstructure consisting of five independent units which are monolithic with their piers and have expansion joints at the mid-span shear hinges and the abutments. During routine maintenance it was found that the superstructure cantilevers were deflecting excessively, and consequently a thorough inspection and appraisal were commissioned. These included a complete visual inspection, concrete core extraction, radiographic inspection of prestressing tendons and a programme of load testing. The bridge was also reanalyzed to check its long-term performance. Subsequently, a rehabilitation scheme was designed to recover part of the deflection. This involved the introduction of additional prestress near the box girder's top flange. Key words: prestressed concrete, structural assessment, creep, shrinkage, radiographic inspection, rehabilitation, external prestressing.

The flanges of T- and I-beams are subjected, near their junctions with the web, to three in-plane forces: Nx, Ny, and Nxy, where x and y are horizontal axes parallel and normal to the beam axis. The finite element method is used to study the variation of these forces in T-beams under point or line load. It is concluded that forces Nx and Nxy can be determined for design purposes using conventional engineering beam theory. Force Ny, on the other hand, cannot be determined by available closed-form solutions. Based on a detailed parametric study, a simplified procedure is proposed for determining Ny. Key words: beams (supports), connections, finite element, flanges, loads (forces), T-beams, webs (supports).

The members of self-supporting latticed electrical transmission towers customarily consist of rolled angle sections. Nowadays, with transmission voltages in the extra-high-voltage and ultra-high-voltage ranges, the leg members are required to carry very heavy loads, especially in the case of heavy angle and anchor towers. Results of a study carried out on wide-flange shapes, to investigate their suitability as alternative sections to welded and built-up angles, are presented. The world's first transmission tower that has wide-flange shapes for legs and arms, Ontario Hydro's 500-kV double circuit heavy angle dead end tower type V6H, is described.

Concrete and steel–concrete composite beam-slab type and box-girder bridges are analyzed using the finite strip method, a variation of the powerful finite element method. The results obtained are used to determine the effective width of the compressive flange of the bridge girders. A study of a wide range of bridge models resulted in an empirical relationship that is in good agreement with the design criteria set down in the composite road bridge design code of the Federal Republic of Germany.The empirical relationship presented in this paper is intended to supplant the effective width specifications in current American and Canadian design codes, which in general lead to conservative designs.

A procedure to find the optimum proportioning of laterally unsupported homogeneous I-beams following the limit state design method of CSA S16.1-1974 specification is presented. Both bending moment and shear force is considered in the optimization procedure. The web is unstiffened and the proportioning is restricted to Class 1, 2, and 3 type sections. Some design charts covering a large range of values of unsupported lengths of compression flange and design moments are provided.

The objective was to determine experimentally and analytically two-dimensional steady-state temperature distributions produced in the cross-sectional planes of steel–concrete composite simple span bridges. The upper and lower surfaces were exposed to different temperatures.The research included the development of finite element solutions for steady-state temperature distributions from known boundary conditions and the calculation of strains and stresses. Temperature and stress distributions were generally nonlinear with linear strains through the finite elements. Temperatures were predicted to ±1 °F (±5/9 °C). The experimental strains are linear through the composite section, with the computed finite element strains giving generally slightly higher stresses. The conventional and finite element method computed stresses were compared.For positive curvature, the conventional method underestimated the compressive stress in the top flange by about 20% while the bottom flange tensile stresses were identical. For negative curvature, the conventional method overestimated the bottom flange compressive stresses between 15 to 27% and the top flange tensile stresses from 10 to 61%. The concrete slab stresses were overestimated for positive curvature and slightly underestimated for negative curvature. Slab stresses were relatively small when compared with the permissible concrete stress. Temperature stresses in the steel beam were shown to be significantly large, about 30% of the permissible steel stress, to warrant consideration in the design of these bridges. The stresses were calculated for short term steady-state temperatures. Transient conditions existing in the field produce greater thermal stresses.

A full-scale load test to failure of a 34-year-old three-span continuous reinforced concrete highway bridge located in southern Alberta is described. The opportunity to load test the structure to failure occurred as a result of the scheduled demolition of the bridge.Load was applied to the structure in two phases. In the first phase, precast concrete sections were placed at the centre of the middle span until a load of approximately 1900 kN was reached. At this load level a mid-span deflection of 28 mm was measured for the interior span. Additional load was applied by jacking the ends of the bridge at the abutment supports. Loading continued until a maximum deflection of 157 mm was reached at mid-span. Significant cracking as well as crushing of the compression flange at mid-span were evident. On removal of load at mid-span, a residual deflection of 118 mm was observed indicating that the bridge had been loaded well into the postyield range without collapsing.Results of the test are presented in the form of plots of load vs. deflection, reactions vs. deflection, mid-span moment vs. deflection, and deflection profiles at various loading stages. In addition the calculated flexural capacity is compared with measured values. Key words: highway bridge, load test, reinforced concrete.

A theoretical analysis has been undertaken for unequal width T-joints, haunch-reinforced, in square hollow structural sections (HSS). The unreinforced joint is merely a special case. The theory was based on a finite difference solution to the elastic chord flange plate equation and predicts the branch-to-chord joint stiffnesses under both bending [Formula: see text] and axial force (C = P/δ). The effects of varying the width ratio λ, overall haunch length-to-width ratio λ1, and chord width-to-thickness ratio bc/tc were studied for different plate edge boundary conditions. The latter parameter has the greatest effect on stiffness, but the other two, including haunch size, are also significant. There is reasonable agreement between the joint rotational stiffness factor J, obtained from earlier tests, and the predictions of theory for different edge boundary conditions. The unequal width, unreinforced connection is generally too weak to be viable in design applications. Haunch stiffeners improve strength and stiffness, and are particularly beneficial for large rather than small λ (i.e., 0.833 vs. 0.333).

This paper presents the results attained in an extensive research programme on Vierendeel joints in hollow structural sections (HSS). A total of 29 specimens comprised of five distinct connection types were tested and their strength and stiffness properties were recorded.As most of the joints tested behaved in a semi-rigid manner, a stiffness method for analysing Vierendeel trusses comprised of such joints is presented. Design curves for prediction of joint strength are proposed and design examples of Vierendeel trusses under panel point loading are analysed.Two joint types are recommended. These are the haunch type of connection and the chord flange stiffener, both of which have adequate strength and stiffness characteristics.

The carrying capacity of square hollow structural section T-joints stiffened by a rectangular flange plate is investigated for both branch bending moment and punching shear. The ultimate moment or load is determined from the simple yield line method of which one of three failure modes is applicable depending on the plate length. A large number of combinations of branch, chord, and plate sizes are analysed to provide a statistical basis for making recommendations of optimum plate lengths and thicknesses for stiffened joints in Vierendeel truss applications.

Present specifications to prevent premature elastic–plastic buckling of the column web in symmetric moment connections are based on limited semiempirical evidence because costs of a total experimental investigation are prohibitive and an accepted generalized theory for elastoplastic plate bifurcation is lacking. Finite element formulations, based on an incremental version of a deformation theory combined with appropriate iterative techniques, have been recently developed and displayed both accuracy and adaptability to elastoplastic plate buckling problems for general loading–unloading–reloading history.In the present paper the developed numerical technique has been used to conduct a limited parametric investigation studying the effects of a column axial load on the interaction between beam tension and compression zones as reflected in the stability of symmetrical moment connections. Both quantitative and qualitative measurements have been obtained but, more importantly, the behaviour and capacity of symmetrical moment connections as governed by elastic–plastic buckling of the column web are examined. The paper concludes by emphasizing the necessity to reconsider the present specifications controlling the buckling capacity of the column web in connections since parameters such as those considered in the present paper may have significant effects on the plastic buckling capacity. Key words: symmetrical moment connections, plastic buckling, column web, finite element, axial load, beam depth.

A flexible end plate connection consists of a plate fastened to both sides of the web of a beam by fillet welds. Bolted field connections are used. The connection should transmit vertical shear and allow beam end rotations to occur without the development of significant moments. It is common practice to evaluate the shear capacity of these connections neglecting the presence of secondary forces, which can develop when the connection rotates.Existing analytical models of the moment–rotation behaviour correlate well with test results. These models consider both the flexural and membrane response of the connection in the upper tensile zone and use an empirical relationship to predict the force developed in the lower compression zone. In some tests undesirable impairment of connection flexiblity occurred when the bottom flange of the beam contacted the support.A limit states design model has been developed, which better predicts the moment–rotation characteristics of the connection and which also predicts the reduced vertical shear capacity of the connection due to secondary forces. Consideration of the secondary forces results in connections being designed with more consistent safety levels.An interactive computer program has been developed to design flexible end plate connections using this model. The program is attached to a data file containing the geometric properties of all standard steel flexural shapes available in North America. By specifying any of these shapes along with the desired material properties, loading conditions, connection geometry, and associated connection component costs, a designer can use the program to rapidly prepare customized connection designs or design tables. Key words: beams, connection, end plate, flexible, interactive computer program, joints, limit states design method, plastic deformation, shear strength, steel structures, structural analysis, structural design.

The authors are commended for doing a useful study on the effective widths of flanges in beam and slab and box-girder bridges. The fact that the effective flange widths increase with the level of load in the zone of nonlinear behaviour is confirmed by several research workers. Dowling et al. (1977), after testing models of steel box girders, concluded that shear lag has little weakening effect on the flange strength. Maisel et al. (1973) , made similar remarks about concrete box girders. Consideration of the shear-lag effects can be important for working loads but at ultimate loads the usefulness of such consideration tends to diminish. In Fig. 1 the authors have shown BE to be measured from the edges of the beam flange for beam-slab bridges, from the centre line of webs for multispine box-girder bridges and from the face of webs for multicell box-girder bridges. Considering that the actual flange width of AASHTO and steel wide-flange girders was not directly considered in the idealization, why do these differences in determining 2B and BE exist? It might be appropriate to measure BE and 2B from the centre line of the webs in all cases. Such a simplification would make the use of Fig. 7 a and b even simpler. For serviceability limit states the Ontario highway bridge design code requires one-lane loaded conditions to be considered for all bridges. The empirical effective-width charts appear to have been developed for all design lane loaded conditions. Will the charts be still applicable if only one of the design lanes is loaded ? The authors clearly state that the purpose of the exercise was to determine effective widths of com