Journal of Structural Engineering

Present methods of fatigue research tend to concentrate on predicting the endurance of the structural component when it is subjected to a combination of cyclic loads. This has led to methods of design that assume the strength of the component is independent of the applied cyclic loads and that are therefore of limited use in assessing the remaining performance of an existing structure. It is suggested that design is more concerned with predicting the strength of a structure after a period of cyclic loading, which is the converse of present fatigue design techniques. A design procedure has therefore been proposed that places the emphasis on predicting the strength of a structural component after it has been subjected to cyclic loading. In this new design procedure, the strength of a structure has been linked to the fatigue loads so that this procedure can also be used in the analysis mode to estimate the residual strength and residual endurance of an existing structure.

The purpose of this paper is to investigate experimentally the local buckling strength and interaction strength between local and overall buckling of closed polygonal section steel columns. A total of 23 specimens was tested to failure under axial compression loading. Seventeen specimens were stub columns with regular rectangular, pentagonal, hexagonal, heptagonal, and octagonal sections. The others were medium‐length columns with regular rectangular, hexagonal, and octagonal sections. An empirical formula based on the compression tests is presented to predict local buckling strength and interaction strength between local and overall buckling strengths of closed polygonal steel columns. Edge strain at the maximum stress‐and‐strain energy is an important standard to estimate the load‐carrying capacity of a member same as the maximum stress. A comparison of edge strain and strain energy with equivalent width‐thickness ratio is included. They show a similar tendency and decrease with the increase of equivalent width‐thickness ratio. Also, they decrease with the increase with the equivalent slenderness ratio.

Elastic moduli for a specimen of Norway spruce (Picea abies) have been evaluated by dynamic tests and corresponding computer calculations. Modal testing of a wooden beam of structural size resulted in modal properties for 13 different modes of vibration. Corresponding modes of vibration were calculated using a finite element model based on Timoshenko beam elements with constant homogeneous properties. Similarity between experimental and theoretical values for eigenfrequencies for the fundamental axial mode of vibration required a value for Young's modulus $E=11\text{,}444\mathrm{MPa}\text{.}$ Corresponding matching of the fundamental frequency for torsional vibration yielded $G=1\text{,}074\mathrm{MPa}\text{.}$ The calculated eigenfrequencies for higher modes of vibration correspond very well to the experimental values. It should be noted specifically that the eigenfrequencies for edge‐wise bending modes correspond well, although the value for the shear modulus was based on a torsional mode matching.

An experimental study aimed at understanding the deformational behavior of conventionally reinforced steel fiber concrete beams in pure bending is reported in this paper. One group of beams has steel fibers dispersed in the entire volume of the beam and the second has fibers dispersed over half the depth of the beam on the tension side. A comparative study of the deformational characteristics of these beams has been made. Half‐depth fiber inclusion, requiring only half the quantity of fibers of full‐depth inclusion, is found to be equally effective in improving the deformational behavior of beams. Thus, by such modes of inclusion of fibers, an economical and efficient use of expensive steel fibers can be realized.

Reliability analysis of a jacket platform presently operating in the North Sea is performed. The limit states correspond to design criteria for each structural component. For the tubular columns, beam buckling, shell buckling and maximum allowable stress are considered. For the tubular joints the punching shear failure mode is analyzed. A slightly redesigned version of the jacket is also considered for assessment of the inherent reliability level of the relevant design code. Latin hypercube sampling is employed for simulating vectors of realizations of basic variables. Probability distributions of extreme dynamic load effects are subsequently identified based on a sequence of response analyses. Reliability analysis is performed for selected structural components by FORM/SORM techniques. For a representative set of structural elements, the random variables influencing the load effects are ranked by means of results from an uncertainty analysis. Similarly, importance measures obtained as by‐products of the reliability analysis identify key variables of the failure functions. Finally, sensitivity of reliability measures to perturbations of basic distribution parameters are investigated by repeated analyses with revised input data.

In 1991, Ylinen's column equation was adopted for the design of wood columns in the United States. Ylinen originally derived this equation as a plastic buckling criterion, using a nonlinear compressive stress‐strain law as the starting point. The stress‐strain law contained a parameter $c⩽1\text{,}$ which controlled the degree of nonlinearity. Linearity was achieved at $c=1\text{.}$ This same parameter appeared in the final column equation. At $c=1\text{,}$ the column equation reduced to elementary theory, i.e., to a perfectly plastic‐perfectly elastic failure criterion. Column design equations are viewed in this paper as interaction equations between two modes of failure: crushing and buckling. Ylinen's column equation is found to result from the addition of a cross‐product term to the linear interaction equation. This enables us to view c as measuring the degree of departure from the assumptions of elementary theory, including such phenomena as inhomogeneity of material, imperfections of shape, and accidental eccentricity of loading, in addition to non‐linearity of the stress‐strain law. Column design equations from wood design codes in Europe and Canada are compared and discussed from this point of view.

An ongoing program to measure silo pressures and frictional wall loads in full‐size silos has been conducted over the past 17 years. The paper summarizes some of the more recent findings of the work. These include observations of the nature of pressures and strains in silos during filling and the changes that occur at the start of emptying. Pressures and strains for silos may be nonuniform around the perimeter, and are dependent on thermal effects. Suggestions are made as to how these observations could be incorporated into design procedures for silos.

Limited experimental results were reported in the literature on the behavior of prestressed composite girders subjected to negative bending moment. This paper examines experimentally the behavior of prestressed composite steel‐concrete girders. The steel beams were all welded plate girders. Five composite specimens were tested to study the various aspects of prestressed composite girders, including the effects of construction stages, prestressing sequence, tendon type, and compactness of the plate girders on their structural performance. Load versus midspan deflection, strains in the concrete deck and steel girder, and the force increase in the prestressing tendons due to the applied loads were measured and studied. The test results showed that prestressing a composite girder in the negative moment region increases its stiffness by preventing the cracking of concrete under service loads; enlarges the elastic range; and increases the ultimate capacity due to the effective use of high‐strength tendons.

A method for the structural analysis of prestressed composite steel‐concrete girders was studied in this paper. The deflections, forces in the prestressing tendons, and strains in the steel beam and concrete slab of composite girders were computed throughout the entire loading range up to failure. Equations are provided for the calculation of the yield and ultimate load capacities of the girders. The developed analytical models were based on the incremental deformation method. The results of the analytical study were compared with test results of several girders. Reasonably good correlations between analytical and experimental results were obtained. Also, the results showed that a substantial increase in the yield and an increase in the ultimate load capacities can be achieved by adding prestressing tendons to the composite girders and prestressing them. It was determined that the most effective construction sequence for prestressed composite girders in negative moment regions is to posttension the composite girders with tendons in the concrete slabs.

Recent International Institute of Welding/Eurocode 3 (IIW/EC3) recommendations for welding rectangular hollow‐section (RHS) truss connections are shown to be too simplistic since they do not consider the special aspects of a weld connector at a gapped or overlapped truss connection. Furthermore, these weldment recommendations are generally conservative and not consistent with the design approach of national welding and structural codes in that weld size requirements are related to the full strength of members with no consideration of actual member forces. This paper reviews codified formulas for the design of individual fillet weldments in steel made by arc welding. Next, design recommendations pertinent to fillet weldments in rectangular hollow‐section truss connections are reviewed and shown to result in quite different weld proportions. The paper concludes with design examples and thereby offers preferred design approaches. A companion paper deals with experimental work used to verify the effective weld length concept for flexible rectangular hollow‐section connections necessary for the weldment design examples.

Laboratory‐based investigations were conducted to assess the performance of and evaluate the design of weldments around the perimeter of rectangular hollow‐section (RHS) web members at their connection to the chord member of a truss. Experimental work was carried out by subjecting a connecting weldment to a tension load, initially within an isolated connection specimen and then within a truss environment. These research projects entailed the testing of 29 isolated connections and two large‐scale, 12.2‐m (40.0‐ft) and 12.0‐m (39.4‐ft) span, simply supported Warren trusses. The results of these tests have been used to evaluate international design recommendations for weldments and thereby produce preferred design approaches that utilize an effective‐length concept for the weldment. For welded trusses, with gapped K‐ (or N‐) or 50% overlapped K‐connections, and comprising rectangular or square hollow structural section (HSS) members, it has been shown that the strength of the welded joint can be related reasonably well to the axial force in the connected web member. Any local or secondary bending moments in the connection are thereby accommodated in the design of the welds by means of a weld effective length.

The flexural behavior of reinforced concrete beams containing steel fibers is investigated in the present study. An experimental program was set up and nine reinforced concrete beams have been tested, including two series of singly reinforced concrete beams and one series of doubly reinforced concrete beams. The fiber contents of reinforced concrete beams for each series were 0%, 1%, and 2% by volume. The load was applied in four‐point flexural loading condition. The deflections and the steel and concrete strains were automatically measured at the various locations. The crack widths and crack spacings at each loading step were observed during the loading process. It was found from these measurements that the crack widths increase almost linearly with the increase of steel stress and the crack widths at the same loading stages are greatly reduced as the content of steel fibers increases. The present study also indicates that the ductility and the ultimate resistance are remarkably enhanced due to the addition of steel fibers. The design implication of fiber‐reinforced concrete beams is also discussed. A method for incorporating fiber effects in the flexural analysis of singly reinforced concrete beams is redrived, slightly modified and used to analyze the singly reinforced concrete beams. A new method to analyze doubly reinforced concrete beams containing steel fibers is derived and proposed for future use. It was found from this study that the neglect of fiber contribution may considerably underestimate the flexural capacity of fiber‐reinforced concrete beams. The present study allows more realistic analysis and design of reinforced concrete beams containing steel fibers.

The anchorage of fiber‐reinforced plastic (FRP) tendons in pretensioned prestressed concrete (PC) members may cause the formation of cracks as a result of tendon wedge effect and mechanical interlocking during prestress force transfer. This paper presents experimental results obtained with a braided epoxy‐impregnated aramid FRP tendon having a 16‐mm nominal diameter and initial prestress force of 125.6 kN. Minimum concrete cover was 52 mm. End‐zone reinforcement and tendon partial blanketing were used to determine their effect on concrete stress field intensity and on cracking due to prestress force transfer. It was found that steel stirrups do not prevent cracking, but they limit its propagation. A carbon FRP coil of small pitch inhibits crack formation, but the concrete tensile strain remains high. Partial blanketing is the most effective solution, causing only a limited increase in transfer length (approximately 20%). An analytical solution to the experiment was obtained by using the finite element method, assuming a partly cracked elastic concrete ring model. This proved to be an effective tool for predicting the occurrence of cracks when using tendons of different size and/or initial prestress force.

Lag screws are frequently used to fasten relatively thin, often metal, cleats to wood structural elements. Design values for engineered lag‐screw connections have been based on proportional limit loads derived from limited test data. A new allowable stress design specification and a draft load and resistance factor design (LFRD) specification for engineered wood structures incorporate theoretically based yield equations to calculate connection strength. These equations, based on a set of assumed yielding actions, offer designers much greater freedom and guidance in checking dowel‐type connection strength than has previously been the case. This paper shows the development of yield models and simplified equations for the lateral strength of lag‐screw connections with wood and metal side plates. Experimental data from several sources are presented in verification of the models. Calibration of the yield equations to traditional safety levels for use in codification is described.

The behavior of boxed double‐angle bracing members subjected to out‐of‐plane buckling under severe cyclic load reversals is investigated. Six full‐size hinged‐end and three fixed‐end bracing members made of double angles of A36 steel with welded stitches and end gusset plates were tested under large cyclic deformations representative of severe earthquakes. The effects of section configuration, stitch spacing, and end fixity were studied. The tests showed that boxed angles had a behavior much superior to the conventional back‐to‐back angles. Smaller stitch spacing in boxed angles increased the first buckling load and enhanced the postbuckling performance. Smaller stitch spacing, in general, and continuous stitch plate, in particular, significantly enhanced the performance of boxed angles by reducing section distortion. End fixity resulted in a twofold increase in the energy dissipation capacity and cyclic peak compression loads of the bracing members.

The purpose of this paper is to present a theory that incorporates a special treatment of warping in the free vibration analysis of continuous curved thin‐walled girder bridges. The theory uses an additional degree of freedom for curved beam elements with thin‐walled sections to capture the effects of warping. The ability to use flexible piers or hinge supports with the curved‐girder warping theory is included. Formulation of both element stiffness and mass matrices is discussed. The element stiffness matrix is determined by inverting the flexibility matrix. Two example calculations using the theory are presented. The examples include a three‐span, continuously curved bridge with two flexible piers and a three‐span, continuously curved bridge with four flexible piers. The results obtained are compared to results based on other theories. Effects of warping, noncoincidence of shear center and centroid of section, flexibility of piers, and radius of curvature are shown.

The results of an experimental investigation of the behavior and compression strength of a slender steel tube under prestressing force are described. A series of tests were conducted using 20‐ft (6.09‐m) long straight or circularly bent steel pipes of 1.315‐in. (33.4‐mm) and 1.9‐in. (48.3‐mm) external diameter. Prestressing tendons were passed through the pipes, and anchorage devices were placed at both ends, allowing prestressing force to be introduced by means of a jack. Separate tests were conducted on test samples (both straight and curved) where tendons were loosely passed through the interior of the pipes and on samples where tendons were provided with rings spaced at intervals of 1.0 ft (30.5 cm) and 1.5 ft (45.7 cm), securing almost an concentric position of the tendons in the pipes' interior or limited eccentricity $e=1/35$ of 1.315 in. and $e=1/50$ of 1.9 in. diameter of pipe. The relaxation and its effect on the prestressed regular steel pipe was also measured.

In the last two or three decades, several methods have been proposed for the design of columns in steel rigid frames under gravity loads; one of them was proposed by Yura in 1971. In this note it is shown that Yura's method, adequate when applied to ordinary regular structures, can be overconservative in special cases, as it does not take adequately into account several important features of the problem. A new method is presented, based on the determination of the lateral stiffness of the frame or story under consideration, which takes into account both PΔ and P° second-order geometrical effects. The method is general in scope, does not require the use of K factors, and has the advantage that the frame or story first-order lateral stiffness can be easily computed with the information given by all commercial computer programs for structural analysis. The proposed method is compared with design methods available in the literature through its application to several numerical examples.

A finite element model capable of tracing the nonlinear response of coupled shear walls to earthquake motions is proposed. The walls are idealized as an assembly of quadrilateral elements with three degrees of freedom, two translational and one-rotational, at each corner, and nonlinearities, such as cracking, crushing of concrete, and yielding of steel, are included in the model. The coupling beams are idealized as line elements, and the nonlinearities are confined to element ends through inelastic experimentally based moment-rotation relationships. Allowance has been made for stiffness degradation and bond slippage. The rotational degree of freedom to the plane stress element makes the compatibility of rotations at the coupling-beam-wall interface possible, the evaluation of coupling beams rotation straightforward and the mathematical model elegant. The model was applied to trace the response of two coupled shear walls tested elsewhere and the results achieved were in good agreements with experimental results.