Journal of Structural Engineering

The feasibility of connecting steel‐to‐steel with adhesives is examined. Specifically, the mechanical properties of structural adhesives, fabrication procedures, and properties of bonds are discussed. The mechanical properties of structural adhesives depend on temperature. In general, adhesives behave viscously at room temperature. Above certain stress levels, tertiary creep and creep rupture occur. An important adhesive material property is the diffusion coefficient, which controls the rate at which water enters an adhesive. The static strength of a bond depends on the cohesive strength of the adhesive and on the adhesion between the steel and the adhesive. If adhesion is the weaker link, then rupture is characterized by interfacial debonding. If adhesion is the stronger link, cohesive failure occurs within the adhesive, often due to tertiary creep. Creep rupture data for bonds made with two adhesives are given. Durability data on bonds stressed in 1984 are also presented. Studies on the fatigue and fracture toughness of bonds are briefly discussed. S‐N curves for bonds made with a modified acrylic adhesive are presented.

The long‐term durability of steel‐to‐steel bonds is affected by the ingress of water in a bondline. Sorption of water by an adhesive changes its physical properties and degrades the adhesion between the adhesive and the steel. Diffusion coefficients and equilibrium water concentrations are reported for a specific poly‐methyl methacrylate (PMMA) based adhesive. Changes in modulus with water concentration are reported. The effects of water immersion on the rapid tensile strength of circular butt bonds are determined; the ingress of water produces annuli of adhesion failure. The scale of the bond affects the fraction of the initial strength that is retained after a period of immersion. Fick's diffusion equation is used to compute evolutionary water concentration profiles in bondlines subject to cyclical boundary concentrations. The diffusion coefficient of the particular adhesive is such that daily boundary concentration fluctuations are filtered from all the bond area except an annular region extending about 0.07 in. in from the boundary; yearly boundary concentration fluctuations are also filtered except for an annular region extending about 0.6 in. in from the boundary.

This paper clarifies the definitions of eccentricity used in two of the approaches to calculate story torsional moments in the design of torsionally unbalanced regular multistory structures. One approach uses the floor eccentricities, in conjunction of the static equivalent loading, to calculate the floor torques. The story torsional moment is then obtained by summing the floor torques above the story. This approach was suggested in the Recommended Lateral Force Requirements and Commentary by the Structural Engineers Association of California. In this approach, the floor eccentricity should be defined as the distance between the generalized center of rigidity and the load resultant at that floor. In the second approach, the story torsional moment is obtained directly as the product of story shear and the story eccentricity, as suggested by Uniform Building Code. In this approach, the story eccentricity should be defined as the horizontal distance between the shear center at the story and the resultant of all lateral forces above the story being considered. It is shown that if the proper definitions of eccentricity are used, both approaches result in the same story torsional moments. A simple example is worked out in detail to show the steps involved in each of the approaches.

The bending strength of lumber is a complex problem. Bending strength depends on three main factors, the ratio of tension to compression strength of the material, nonlinear ductile behavior in the compression zone, and size‐dependent brittle fracture in the tension zone. In this paper, the relationships between bending strength and axial tension and compression strengths are explained in detail. The results of large‐scale, in‐grade lumber tests are used to predict ultimate bending strengths throughout the distribution of lumber strength properties.

This paper presents a detailed description of an experimental study of the flexural buckling strength of cold‐formed steel columns. A total of 68 columns and 25 stub columns are tested. It is observed that some types of columns show lower strength than the value predicted by the American Iron and Steel Institute column design formulas. In some cases, the differences between the test results and the AISI predictions are found to be larger than 15%. From the experimental findings, the cross‐sectional dimensions and the magnitude of the residual stresses are found to have a consistent correlation with the weakening of the column strength. Other parameters, such as the stress‐strain curve (sharp or gradual yielding) obtained from tensile coupon tests and the method used to form the sections (press‐braked or roll‐formed), are not found to have a definite influence on the strength of the columns.

A two‐dimensional algorithm and a computer program are developed for the incremental analysis of cast‐in‐place segmental prestressed cantilever bridges with variable depth. The model takes into consideration effects caused by differences in age of bridge segments, shrinkage and creep of concrete, relaxation and curvature of prestressing tendons, and vertical and horizonal movements of the supports as well as their rotations. It also evaluates redistribution of internal forces with time after continuity. A three‐span highway bridge composed of two box girders with variable depth, each carrying two lanes of traffic and a sidewalk, was analyzed as a case study. The results obtained using the developed algorithm show appreciable differences from current design practice and clearly indicate the necessity of including the difference in concrete characteristics of the bridge segments caused by their different ages when evaluating: (1) The camber needed during construction; (2) the redistribution of internal forces after continuity; and (3) the internal forces caused by the time‐dependent support displacements.

The designers of bridges must assure themselves that bridge bearings can transmit the high horizontal loads of current designs. The proportion of horizontal to vertical load is increasing due to the current trend toward continuous, joint‐free, redundant structures. From the 10‐20% level of some years ago, horizontal forces are now found at and above the corresponding vertical loads. These horizontal forces require that bearing stability be investigated. Two levels of stability criteria, for normal conditions and for extraordinary temporary conditions, are presented. An overturning factor of safety of 3.0 for rectangular surfaces (5.0 for circular surfaces) is proposed under normal (i.e., dead, wind, and live) loads, while a factor of 1.5 for rectangular surfaces (2.0 for circular surfaces) is proposed under extraordinary (i.e. seismic or construction) temporary loads. Examples of pot, disc, and spherical bearing stability checks are presented.

A power series method with the use of a coordinate transformation has been developed to solve analytically the buckling problem of uniaxially compressed rectangular plates with linearly tapered thickness. The compressed edges are simply supported and the unloaded edges may be arbitrarily restrained, e.g., simply supported, clamped, or free. Accurate buckling loads with rapid convergence are obtained in comparison with other approximate solution methods. The influences of thickness variation, plate aspect ratios, and boundary conditions on the buckling load are shown graphically. The buckling load is highly dependent on the thickness variation. The present exact method is a general one applicable to the buckling and free vibration of the rectangular plates with various thickness variations, although limited to rectangular plates simply supported on two opposite edges.

Based on some rational assumptions, a harmonic shear‐lag analysis using plane stress elasticity for the stresses in flanges of simple or continuous beams with I‐, T‐ and box cross sections is developed, and it is accurate and convenient enough for a wide range of problems. Since harmonic analysis is used, convergence of the solution and the peak stresses in the flanges is studied in detail. An effective technique is presented to improve the convergence. The shear‐lag effect upon the longitudinal distribution of the stress resultants of a continuous beam with wide flanges is studied, and a general procedure to account for this effect is developed. To show the accuracy and applicability of the solution, typical numerical examples are presented, and the results compared with other solutions.

Based on an empirical matching of analytical results for stresses in wide flanges obtained using the plane stress elasticity given by Song (1984a), simplified empirical formulas and some diagrams for the determination of the shear‐lag effect in simple beams under various loading with I and box cross sections are presented. For T‐beams, the calculation can be performed using an equivalent I‐beam, according to the formulas in this paper. For continuous beams, the results can be obtained from the analysis of a corresponding basic simple beam formed by removing all interior supports using superposition. Some simple and essential formulas are presented, by which the most important features of the shear‐lag effect for a simple or continuous beam can be estimated quickly in engineering design. To demonstrate the use of the empirical formulas and to verify their accuracy, comparisons of results for typical numerical examples are given.

This paper presents a three‐dimensional finite element algorithm for computing internal forces in intermediate and end cross frames of multibeam composite steel bridges. The developed algorithm was applied to four medium‐span [137 ft (41.75 m)] simply supported multigirder tangent four‐lane bridges, with X‐type intermediate cross frames and end cross frames of the Warren truss with verticals type. One was a right bridge and the others had 20°, 40°, and 60° skew angles. It was found that: (1) Most members of the intermediate cross frames may develop compressive and tensile forces; (2) maximum compressive forces occur in the cross‐frame diagonals attached to the ends of exterior stringers, close to the obtuse angles of the bridge; (3) maximum tensile forces occur in the cross‐frame chords, at midspan; (4) the effect of skew may be neglected for skew angles of 20° or less; (5) the higher the skew angle, the higher the maximum forces induced in the cross‐frame members; (6) the larger the cross‐frame members' sections, the higher the internal forces induced in them, and vice versa; and (7) vertical deflections are not affected by the change in cross‐frame members' sections. However, the horizontal displacements of the girders' bottom flanges are reduced with the increase in the cross‐frame members' sections, and the maximum differential vertical displacements between adjacent girders occur near the obtuse angles of the bridge between an exterior girder and the first interior one.

A simplified seismic design procedure is proposed for structures equipped with an innovative friction‐damping system. The system has been shown experimentally to perform very well and could represent a major new development in earthquake‐resistant design. This paper first presents an efficient modeling approach for the seismic analysis and design of friction‐damped structures. The hysteretic properties of the friction dampers are derived theoretically and are included in a Friction Damped Braced Frame Analysis Program (FDBFAP), which is adaptable to a microcomputer environment. FDBFAP is then used to perform a parametric study of the optimum slip‐load distribution for the friction dampers. The results of the study lead to the construction of a design slip‐load spectrum for the rapid evaluation of the optimum slip‐load distribution. The spectrum takes into account the properties of the structure and of the ground motion anticipated at the construction site. The availability of this design slip‐load spectrum should lead to a greater acceptance by the engineering profession of this innovative design concept.

Fatigue evaluation and design methods were developed in a recent project for steel highway bridges. The evaluation method enables an engineer to estimate the safe remaining life of an existing bridge. The estimate of safe remaining life is needed to make cost‐effective decisions regarding inspection, repair, rehabilitation, and replacement. The design method is consistent with the evaluation method. The reliability‐based concepts and data base behind the development of these methods are discussed and illustrated through some examples. The evaluation and design methods were calibrated for both redundant and single‐load‐path bridge components using safety (reliability) indices computed with existing fatigue specifications. The proposed procedures lead, however, to more uniform and consistent safety indices. The evaluation and design methods also follow to a large degree present procedures familiar to highway agencies. These methods, however, contain options to use site‐specific traffic and inspection data and improved structural analysis. Both the evaluation and fatigue design methods have been incorporated in two recent guide specifications published by AASHTO.

Experience with field testing of highway bridges in Ontario, Canada, during recent years shows that nearly every bridge has some aspect of behavior that can escape the attention of even experienced bridge designers and analysts. This paper lists some of the various surprises encountered in bridge testing that may have a significant influence on the load‐carrying capacities of bridges. In particular, reference is made to the behavior of bridges with steel girders and concrete deck slabs, and of steel truss bridges. Lessons drawn from tests on these bridges are summarized, so they can provide an advantage in the load‐carrying capacity evaluation. It is also shown that in some cases the appearance of a bridge can be misleading with regard to its true load‐carrying capacity. In such cases, field testing is shown to be the most effective means of evaluating the bridge.

Test results of 24 reinforced concrete wall panels in two‐way action (i.e., supported on all the four sides) and subjected to in‐plane vertical load are presented. The load is applied at an eccentricity to represent possible accidental eccentricity that occurs in practice due to constructional imperfections. Influences of aspect ratio, thinness ratio, slendemess ratio, vertical steel, and horizontal steel on the ultimate load are studied. Two equations are proposed to predict the ultimate load carried by the panels. The first equation is empirical and is arrived at from trial and error fitting with test data. The second equation is semi‐empirical and is developed from a modification of the buckling strength of thin rectangular plates. Both the equations are formulated so as to give a safe prediction of a large portion of ultimate strength test results. Also, ultimate load cracking load and lateral deflections of identical panels in two‐way action (all four sides supported) and oneway action (top and bottom sides only supported) are compared.

Experimental results from a series of earthquake simulations on a one‐quarter scale model of a moment‐resisting reinforced concrete frame are presented. Using the test results, some of the U.S. seismic design provisions are examined. It is observed that even though drift capacity was amply provided, the magnitude of drift was not adequately anticipated by the design process. Moreover, the structure was several times stronger than required by the code design forces. The contributions of floor slab, actual material strengths, and detailing requirements are identified as the significant sources of overstrength. The effect of these parameters on the overall performance is evaluated.

Effects of setbacks on the earthquake response of multistory building structures are evaluated. As part of the study, measured responses to simulated earthquakes of a ductile moment‐resisting reinforced concrete test structure are examined. The test structure is a one‐quarter scale model of a six‐story, two‐bay by two‐bay building having a 50% setback at midheight. Experimental observations are supplemented with a series of analytical studies. Guidelines are established by which to detect configurations for which concentrations of damage in tower members are likely. A static lateral‐load design method to improve performance is proposed.

Global damage indices based on equivalent modal parameters are defined using the vibrational parameters of an equivalent linear structure. In continuum mechanics‐based damage models, effects of the growth and coalescence of microcracks are accounted for through the definition of an internal or local damage variable. Damage to engineering materials essentially results in a decrease of the free energy stored in the body with consequent degradation of the material stiffness. It is shown that parameter‐based global damage indices can be related to local damage variables through operations of averaging over the body volume. The relationship obtained is applied to the case of numerical simulations of damage events as well as to the case of actual structures tested on shaking tables. Some results are also presented, relating parameter‐based global damage indices to averages in space and time of local plastic strain.