Journal of Structural Engineering

The effects on wood of simultaneous mechanical and moisture loading are studied. A constitutive model is given for the case of uniaxial stress in the longitudinal direction. The model has been quantified on the basis of test results found in literature. The validity of the model was checked independently against other tests and a number of practical load situations were studied. Finally the serviceablity requirements for design are discussed. The calculations show that the model is capable of describing the response of wood with reasonable accuracy. Some differences in behavior are found to depend on moisture cycling. The results of the simulations show that there is a significant influence of strain on the shrinkage and swelling response. An influence of beam dimensions has been found. The relative deflection for a small beam is larger than the deflection for a larger beam. Studies of the effect of varying load show that a temporarily applied load has little effect on the long‐term deflection of structural timber. A simulation of snow‐load variation shows that the long‐term deflections are small in comparison with the case when the characteristic load is assumed to be constant over time.

This paper discusses a variable damper proposed to control bridge response against an earthquake. The variable damper changes its damping coefficient depending on the structural response so that it functions as a passive energy dissipator, a damper stopper at small deck displacements, and a stopper with shock absorber for excessive deck displacements. The effectiveness of variable dampers is studied first by analytical simulations. It is demonstrated that the peak deck displacement and acceleration are reduced to 26% and 44%, respectively, of the bridge response without control. A prototype variable damper with a maximum damping force of 200 kN is developed, and its effectiveness is studied by a series of dynamic loading tests. A series of shake‐table tests is also performed using a model bridge with a deck weight of 390 kN. It is found that providing the prototype variable damper reduces the deck response acceleration and displacement from 376 cm/s² and 9.2 cm to 203 cm/s² and 1.3 cm, respectively. Analytical simulation of the experimental results is made. The predicted deck responses are quite close to the measured ones.

In thin‐walled prismatic members, three different points regulate the cross‐sectional properties. Among these, the centroid and shear center are important for beams with nondeformable cross section, while the rotational centers of plates are important for beams with deformable cross sections. This paper provides methods for determining the rotational centers of plates regarding the distortional problems of thin‐walled, multicell box girders, and proves that the rotational centers of plates can be explained by the same concept as the shear center. The general characters of distortion and location of the rotational center of each plate are derived for typical box girders with a trapezoidal section. A method for resolution of an eccentric load into components of pure bending, pure torsion, and pure distortions is also given.

A new set of effective‐length factors is presented for designing doubly symmetric I‐shaped built‐in cantilever beams against lateral‐torsional buckling when subject to different loading and restraint conditions at the cantilever tip. This refined approach rectifies inherent problems of current solutions caused by overlooking the restraint conditions as well as the limitations that existed in their original derivation. Because of these problems, the effective‐length factors currently available may result in either overly conservative or unconservative designs, depending on the type of problem involved. Also, an interaction buckling design model is suggested for overhanging beams, in which the load is applied only at the cantilever tip. This design model takes into account the ratio of the length of the cantilever span to that of the back span, a significant parameter that has not generally been considered. Finally, a design procedure is given for determining the elastic critical moments of crane‐trolley beams. These analyses are based on a finite‐element model that correlated well with the results of 31 full‐scale tests.

This paper reports on the following: (1) Simple experimental investigations of pultruded fiber‐reinforced plastic (PFRP) material behavior and tests of simple epoxied joints to verify and enhance the reported database; (2) cyclic tests of an attempted PFRP rigid beam‐column connections; and (3) preliminary observations regarding the seismic‐resistance worthiness of PFRP structures. These are simple design issues that must be addressed if PFRP structural shapes are to gain broad acceptance by the design community. It is found that epoxied connections alone cannot fully exploit the high‐strength potential of this material, and that delamination of components is more problematic than the inherent brittleness of the material. The cyclic flexural test of an attempted rigid beam‐column connection has revealed the particular delamination weakness of the flange‐to‐web core of PFRP structural shapes. The construction of rigid beam‐column connections using commonly used steel details is found to be ineffective.

Superstructure design of short‐ and medium‐span highway bridge systems may be conceived as a process of multilevel and multiobjective optimization. Three optimization levels are identified: (1) Level 1—component optimization; (2) level 2—structural configuration optimization; and (3) level 3—overall system optimization. Designs may be optimized by separately or simultaneously considering one, two, or more of the following objectives: cost, prestressing steel or concrete consumption, and superstructure depth. The optimal solution may be found by a sequence of nonlinear programming and sieve‐search techniques. Levels 1 and 2 optimizations identify the best solutions for specific components (precast I‐girders, voided and solid slabs, single‐ and two‐cell box girders) and layouts (for precast I‐girder: one, two, and three; simple or continuous spans). Level 3 optimization selects the overall best system for given bridge lengths, widths, and traffic loadings. The present study results in: (1) A systematic procedure for bridge design; (2) a rational approach to optimization of standard precast sections; (3) direct design aids for selection of optimized bridge systems; and (4) simplified optimality criteria for preliminary design.

A design procedure to determine the capacity against shear failure of webs of prestressed hollow core slabs is presented. The additional shear stresses in webs caused by prestress forces of strands have been taken into account in this procedure. The proposed procedure has been applied to model slabs and also to test specimens. The capability of the proposed procedure to predict the shear failure of web has been demonstrated by the obtained good agreement between the results predicted by FEM and the results predicted by the suggested formula. Agreement was also good between the test observations and the results predicted by the suggested formula. The proposed procedure can more accurately predict the capacity against shear failure of web than the formulas suggested in design codes. The procedure is reasonable and general in nature, and it can be applied to different loading cases, prestress levels, types of cross section, material, and geometric parameters.

The effect of monolith interaction on the overall dynamic response of concrete gravity dams is investigated. For this purpose, a simplified approach is proposed that is capable of including the effect of the longitudinal direction on the response of the critical monolith. The monoliths are represented by beam elements connected by shear links. The effect of hydrodynamic interaction is considered as added mass to the dam structure. The results of this innovative approach are compared to those of a typical three‐dimensional analysis and good accuracy is obtained. It is noted that the fundamental frequency of dams with keyed contraction joints increases when monolith interaction is included. The structure tends to be stiffer and, depending on the input ground motion, may have lower response values than those predicted using typical two‐dimensional analysis. The procedure developed in this study has its limitations but can include the longitudinal effects along the axis of the dam in the response analysis efficiently and fairly accurately.

Nonlinear methods of analysis for space trusses are reviewed. Methods which do not require repeated reformation and reinversion of the stiffness matrix appear to give the greatest computational efficiency. A new method of analysis for static geometrically linear space trusses is presented. The method allows load increments and decrements, nonproportional loading and the simultaneous failure of several members, while retaining the computational efficiency of the “initial stress” method. In addition, the new formulation has the capability of including the computationally difficult internal instabilities caused by rupture of tension members and “chordal displacement snap through” of compression members in their postbuckling regime. The method can be used with both plane and space trusses in addition to geometrically linear fiber element structures. Because of the computational efficiency of the method, it is particularly useful for structural reliability analyses.

An independent review of the bases for the development of resistance factors found in the American Institute of Steel Construction's (AISC) 1986 Load and Resistance Factor Design (LRFD) Specification is provided. The objectives of this paper are to critically evaluate claims found in a recent paper regarding the validity of these factors, to review and investigate the probabilistic approach taken in the development of these factors, and to examine the implications for design. Consideration is given to the reliability levels associated with established φ factors, as well as the implications for design of using a larger dead‐load factor than is currently used. It is shown that, while an analysis of allowable stress design (ASD) criteria indicates little consistency with respect to relative safety for different load combinations, a relatively consistent reliability (about $\beta =2.6$) was assumed in the development of resistance factors for steel beams in flexure. Finally, while improved methods for estimating dead load may be warranted, an arbitrary increase in the dead‐load factor may be overly simplistic and result in excessively conservative designs.

The theory of plasticity and the deformation compatibility equations are used to evaluate the development of lateral confinement of concrete columns, and the resulting increases in strength and ductility. Concrete is modeled as an elastoplastic material following a simple Drucker‐Prager nonassociative hardening model. The lateral reinforcement is replaced by an equivalent sleeve that surrounds the concrete core. The solution is based on the integration of the elastoplastic relations for the concrete core, the transverse steel reinforcement, and the concrete cover. The lateral pressure on the concrete core is calculated based on the compatibility of deformations of the expanding core and the surrounding reinforcement. The ability of the method to predict the response of confined compression members is demonstrated based on numerous published experimental results. The application of the theory is currently restricted to circular ties and spirals. Future studies will address rectangular confinement.

The pseudodynamic test method is an experimental technique that is commonly used worldwide for seismic performance testing. Many studies have shown that the method can provide realistic results, but the results are known to be quite sensitive to experimental errors. There have been efforts to mitigate the effects of experimental errors, but there has been little work to identify parameters that could be used during and after testing to estimate the quality of a given test. Ideally, such parameters could identify experimental problems early enough in a test to allow an orderly system shutdown and evaluation before the specimen is damaged. The parameters could also be used after testing data to ensure the test ran without significant errors, providing confidence in the measured response quantities. Using the results of pseudodynamic tests as exact experimental values without examining some sort of system performance parameters is like using the stress output from a finite‐element program without knowing the mesh that was used to perform the analysis.

Based on Trost‐Bazant stress‐strain response relationship of concrete, an algorithm to predict the final stresses and strains including creep and shrinkage effects in initially cracked structural concrete elements with or without prestress is presented. Necessary expressions that give rise to familiar section modulus concept are developed for the general case of T section with compression reinforcement, nonprestressed tensile reinforcement, and prestressed reinforcement. In addition, the concepts of equivalent effective depth, equivalent eccentricity and equivalent normal force are introduced. Furthermore, the method takes into account the noncoincidence of the neutral axis of stress with the neutral axis of strain. To ease the computational effort, development of design aids in the form of tables is proposed. The numerical examples given in this paper demonstrate the advantages of the method, which are familiarity, simplicity, adoptability, and accuracy that is sufficient for design purposes.

Approximate formulas are presented for use in determining the effective length of steel columns designed to American standards as an alternative to the standard nomogram technique. Their use allows the development of simple computer routines for incorporation into column and frame design programs. In addition, this technical note compares the formulas derived for the British codes of practice with those used in the American Manual of Steel Construction, since they are based on similar theory, and finds that they give significantly different values.