Journal of Structural Engineering

The nonlinear behavior of a cable‐stayed‐bridge tower with slender columns is generally affected not only by material nonlinearity, but by geometric nonlinearity as well. This geometric nonlinear effect is accentuated by a sustained gravity load, an increase of lateral load and a stiffness reduction of the structure. A series of static analyses up to failure are conducted with various parameters that govern the nonlinear behavior of an A‐shaped bridge tower, under sustained gravity load combined with seismic lateral load perpendicular to the longitudinal axis of the bridge. The present study particularly focuses on the influence of creep deformation due to gravity load and of cyclic lateral load on the ultimate strength and deformation characteristics of the tower and the applicability of the current “moment magnifier method” in the American Concrete Institute (ACI) code to the analysis and design of this type of structure. Analytical results indicate that no significant effect of creep and cyclic load is observed on the ultimate strength behavior for towers in the range of slenderness ratio l/r less than 100 and having an axial stress due to gravity load ${\sigma }_0/f_c^{\prime }$ of less than 0.2. The current moment magnifier method gives results on the conservative side except for the low slenderness region in which considerable moment redistribution among legs takes place due to axial force variation.

The test results of 92 tubular steel, pin‐ended struts under different conditions are studied by statistical analysis. The significant influence of material effects, such as strain aging, strain hardening, and the Bauschinger effect on strut load capacity, is investigated by applying a two‐way analysis of variance test. These three material effects are found to exercise a significant influence on strut load capacity for both the rimming steel and the semikilled steel used. The Bauschinger effect plays a dominant role in reducing strut load capacity, while both strain hardening and strain aging have a significant influence on increasing load capacity.

Concrete‐to‐concrete collisions may occur between breakwater armor elements. In assessing the resistance of the individual armor elements, next to concrete technology aspects and hydraulic aspects, information should be obtained regarding the load‐time history of concrete‐to‐concrete impact. The latter aspect is studied in a series of dynamic experiments on different contact surface geometries. Variables in the experiments are the impact velocity, the concrete quality, the mass of the striker element, and the contact surface geometry, i.e., spherical/planar, conical/planar, truncated conical/planar, and spherical/corrugated geometries of the striker/target surface. A simple elastoplastic contact model is developed, based on four parameters that can be determined from static tests on the same contact surface geometry.

In order to estimate $T_r$‐year wind speed in typical sites where broadcasting antennae are located, the series of daily maximum speeds, recorded in several broadcasting stations, have been analyzed. The estimation from short‐term records was considered. A model of bivariate‐point process, for the peak speeds and relative event times, was used. This model consisted of assuming a homogeneous Poisson‐point process for event times and the independence and identical exponential distribution for the peak speeds exceeding a pre‐fixed threshold. The threshold was selected by event times when they were consistent with the Poisson model. Results were compared to those of the Cook‐storm and monthly maximum methods and to those of the largest yearly value analysis in which different statistical tools were used to determine Fisher‐Tippett type 1 (FT1) asymptotic distribution parameters.

Within the U.S.‐Japan Cooperative Research Program, a full‐scale model of an RC seven‐story building was built and subjected to pseudodynamic testing, with a large number of interesting results. In the present research, attention is devoted to a specific experimental result, showing an interaction between parallel frames. This phenomenon is related to the coupling action of transversal beams and has been designated the three‐dimensional effect. An interpretation of the experimental results is attempted through a finite element nonlinear static model, representing the first floor of the test structure. A concrete model including a smeared cracking capability was used, which allowed the sequence of crack opening and propagation through the shear wall to be followed. The large strains occurring in the tension side of this wall were highlighted as the mechanism generating the three‐dimensional effect. An estimation of the contribution of this effect toward resisting the global overturning moment is made. The computed response of the system is shown to have a dependence on the level of tension stiffening present in the concrete.

The ultimate load capacity of shear webs in a plate girder is modeled as the sum of the contributions of the web and the flanges. After initial elastic buckling due to a uniformly distributed shear stress, the web provides an increasing shear resistance by virtue of a changing shear stress distribution along the boundary, in which the shear stress remains at the critical stress in the compression corner, and increases toward the tension corner until shear yielding occurs. At this point, there is no stress normal to any boundary. As the web yields and distorts in the tension corners, the flanges are compelled to bend, and their resistance creates a normal force between the flange and web at the boundary, making an additional contribution to the shear capacity. Design expressions based on this model of the collapse mechanism give a direct prediction of the ultimate shear force, with no need for empirical adjustments.

The techniques of discrete field mechanics are used to perform an elastic stability analysis of planar X‐braced columns. Closed‐form analytical formulas are derived that are independent of the number of joints in the system; thus, the solution of large, simultaneous equations is avoided. The explicit formula derived is used to develop a series of curves showing the critical buckling load for various planar, X‐braced, truss configurations and geometries. A comparison of critical buckling loads for the X‐braced truss indicates that the modified Euler critical load may very up to 85% from the loads obtained by the discrete field analysis method. The study also shows that as the number of panels increases from four to 32, the buckling‐load ratios also increase. The X‐braced truss depth has a direct relationship to the shearing force, which in turn will affect the buckling load of the truss system. Variations in chord member areas create variations in buckling‐load ratios. The effects on the buckling ratios of the last two factors tend to diminish as the number of panels increases.

This paper describes the results of a research program to study the behavior of reinforced concrete beams exposed to severe fires. Data are presented from fire tests of six full‐scale beams continuous over one support that were fabricated using normal‐weight concrete. Four beams were exposed to the standard ASTM El19 fire, and two to a short‐duration high‐intensity (SDHI) fire developed using realistic fire‐load and compartment‐ventilation parameters. All six beams developed significant shear cracks near the continuous support rather early in the fire, but eventually failed from excessive flexural cracking and deformation. Mathematical models for predicting thermal and structural response of concrete beams exposed to natural fires, as well as the ASTM standard exposure, were also developed. These models predicted the experimental behavior with sufficient accuracy for purposes of limit‐states design.

In practice, light secondary systems attached to the massive primary system are treated independently in the seismic design. Evaluation of the secondary systems and their design forces are performed as equivalent static lateral forces applied to the approximate centers of gravity of the systems being analyzed. These design forces, since they are statics based, do not lend themselves to providing realistic design guidelines of secondary systems. In this paper, possible dynamic design considerations for primary‐secondary (P‐S) systems are studied. A simple yet efficient design procedure is developed. It is shown that, while an increase in damping ratio of the secondary system can always decrease the relative displacement between the primary system and the secondary system, there is an optimum damping ratio for minimizing the maximum acceleration of the secondary system. Compromises thus need to be made when there is a conflict in achieving the best global secondary‐system performance. For illustration purposes, a series of design surfaces for different system configurations under the case of white‐noise input are presented for obtaining the lowest maximum response of secondary systems.

This paper deals with the analysis of prismatic and nonprismatic members of any arbitrary variation in moment of inertia, and where the material is permitted to be stressed well beyond its elastic limit, thus causing the modulus of elasticity to vary along their length. The stress and deflection characteristics of such members are determined by the method of the equivalent systems, which permits to replace the original member of variable stiffness with one of uniform stiffness, whose elastic line is identical to the one of the original variable stiffness member. It is proven mathematically that the inelastic analysis of members with continuously varying moment of inertia and modulus of elasticity can be carried out by using equivalent linear systems of constant stiffness and applying known handbook formulas or methods of linear elementary mechanics. The member can be analyzed for both elastic and inelastic ranges, all the way to failure, thus permitting observation of progressive deterioration of the member's ability to resist stress and deformation, and establish useful, practical critical limits.

A finite strip method for the analysis of plate, beam, and thin‐walled sections with material nonlinearity is presented. In order to simplify the solution, the components of elastic and plastic strain increments are calculated separately for each loading step. The plastic stress increments were evaluated by following the Prandtl‐Reuss flow rule for isometric hardened materials and utilizing the elasto‐plastic matrix. A computer program has been developed for the finite strip procedures. Several materially nonlinear structural problems are analyzed and included in this paper to serve as numerical examples. A general theoretical approach for calculating deflections of beams composed of perfectly elasto‐plastic materials is also presented in the paper. Finite strip results are compared with theoretical solutions or other known solutions. Excellent agreements are observed in all cases.

The aim of this work is to provide further numerical and experimental verifications of the validity of the AISC load and resistance factor design (LRFD) interaction equations for the design of biaxially loaded beam‐columns without sidesway. Sixteen H‐columns are tested under biaxial loading, and a comprehensive numerical study of the behavior and strength of these columns is made using a computer model developed. Against the background of this information, the general validity of the present AISC LRFD beam‐column interaction equations is demonstrated. It is found that, in very limited cases only, the AISC LRFD equations can slightly overestimate the load‐carrying capacity of some beam‐columns. The present study concludes that the AISC LRFD equations are reasonable for design, give realistic representation of actual column behavior and strength, and meet the requirements for simplicity in use.

A large number of railing systems was tested—pipe, panel, and picket railing systems made of steel, aluminum, red brass, yellow brass, and stainless steel, and fabricated of solid, pipe, and tubular members of square, rectangular, channel, and molded shapes. The predicted design loads were compared with experimental data. The normal design procedure that assumes transfer of loads by the top rail to the posts was found sufficiently accurate for the railing systems under investigation. The deflection limitations recommended in ASTM Standard E 985 were found to be more stringent than the established strength requirements, as is indicated by the strength ratios of 1.7 and 2.4 for the allowable deflection and the observed maximum test load, respectively. Designers of railing systems and code authorities should give consideration to the serviceability of these building components in the light of the stringent deflection criteria based on standards of acceptance and overall safety.

The calculation of time‐step sizes to prevent interstep events in multilinear problems is addressed. The treatment is restricted to the Newmark‐Beta family of implicit integration methods. Closed‐form equations in the fractional steps are derived for single‐degree‐of‐freedom (SDOF) systems. Explicit solutions are, however, not feasible because the equations are cubic. Numerical results indicate that the location of events can lead to marked improvements in accuracy, when compared to solutions that handle the nonlinear error at the ends of the time steps. It is contended that the derivation of closed‐form expressions for the substep sizes, in general multidegree‐of‐freedom (MDOF) systems, is precluded by coupling. Identification of the steps that contain events is prerequisite to any event location technique. In this regard, it is shown that a customary test to detect unloading, based on incremental hinge rotations, sometimes fails. An approach that does not suffer from the identified shortcoming is presented. The numerical stability characteristics of an approximate event location strategy based on interpolation is examined using an energy approach. It is shown that the constant‐average‐acceleration method does not satisfy the requirements for unconditional stability when the events are interpolated.

The seismic response arising from strong ground motions of steel frames constructed with moment‐resisting connections causes significant inelastic deformations in the members and connections. The inelastic behavior is dependent upon several design parameters, such as proportioning and selection of members, magnitude of design base shear, and configuration of lateral force‐resisting system. The influence of these parameters was investigated by performing inelastic time‐history analyses of six variations of a structural frame for a typical five‐story building. The direct design procedure adopted in the 1988 edition of the Uniform Building Code was used in the seismic design of the steel frames that were providing all of the lateral stability. The computed inelastic behavior for each frame design is compared with the expected behavior as envisioned by the code. The results reveal the calculated story drifts are larger than the acceptable limits for the structural frames investigated in this study when subjected to the design earthquake. Design implications with regard to the provisions of the direct design procedure are suggested to improve the structural performance of low‐rise steel frames subjected to strong ground motions.

This study investigates strain distribution in regions immediately adjacent to the transverse post‐tensioning bands in a new class of flat slab reinforced concrete bridges employing bidirectional post‐tensioning. While longitudinal post‐tensioning is uniform, transverse strands are concentrated in the vicinity of column lines. The currently applied design assumes that the transverse post‐tensioning effect spreads along straight lines at a specified angle. Size of a laboratory model slab is chosen so that this assumption may be checked for validity. A load schedule is used to study strain distribution for a variety of tendon and load patterns. Values of strain at gage locations as predicted by a finite element analysis are compared to the actual data collected. Reasonably good correlation between laboratory data and the computer program allows increased confidence in numerical simulation. Following testing and analysis of elastic loads, the slab is loaded to impose a punch‐through shear failure. Shear formulas recommended by the American Concrete Institute and other researchers are compared with experimental failure and reserve capacity loads.

The wood industry is developing a load and resistance factor design (LRFD) specification for wood construction. This effort provides an opportunity to reevaluate and restructure the current bases for design with wood, and to develop a specification using concepts of probability‐based limit‐states design. Wood design must take into account the fact that wood is a natural material with large variations in mechanical properties, sensitive to the rate of application and duration of structural loads. These complications are addressed by modeling structural loads as random processes, rather than as random variables, and by postulating limit states that reflect the possibility of failure by either progressive damage accumulation or overload. Current treatment of duration of load effects in the national design specification is found to be inconsistent from a uniform reliability viewpoint. Practical LRFD criteria that are consistent with a desired reliability measure can be developed to reflect the time‐dependent nature of wood behavior.

The behavior of an eight‐story, reinforced concrete apartment building that suffered severe structural damage during the 1985 Chile earthquake is investigated. A series of linear and limit analyses are described in an attempt to identify the cause of the collapse. The results indicate that global response parameters, such as base‐shear strength and mean drift ratio, are insufficient to explain the observed damage. The building had been designed with strength and stiffness characteristics comparable to other buildings that survived the earthquake with light to moderate damage. Investigation of the behavior of individual members leads to the conclusion that the building collapsed after the longitudinal reinforcement fractured in a first‐story wall. This form of brittleness is related to under‐ rather than overreinforcement. The observed failure and data from laboratory tests demonstrate that minimum amounts of longitudinal reinforcement should be established in seismic‐design requirements for structural walls.

Further results on the frictional properties of Teflon-steel interfaces in sliding bearings are reported. Values of the breakaway (static) and sliding coefficients of friction of woven Teflon in contact with polished stainless steel for sliding velocities of up to 10 in./sec (254 mm/s) and bearing pressure of 6,500 and 10,000 psi (44.9 and 69 MPa) are reported. These values are sufficiently low to be of interest in applications of seismic base isolation. Furthermore, the effect of dwell of load on the breakaway coefficient of friction of sheet type Teflon-steel interfaces is studied in a single test in which the load is maintained on the specimen for 20 months prior to testing. The test demonstrated no difference in the breakaway friction in comparison to a test with only one-half hour of load dwell.