Journal of Structural Engineering

Transverse prestressing of bridge decks is an attractive concept with substantial benefits in both economy and improved durability. This paper summarizes a series of interrelated physical tests and computer analyses which were conducted to provide necessary information for the development of design criteria for transverse prestressing of bridge decks. It addresses such important design areas as the effective distribution of edge prestressing force across a bridge slab as affected by both diaphragm and girder restraints; the behavior of transversely prestressed decks under typical wheel loadings; and the effect of posttensioned strand spacing on the distribution of horizontal slab stresses near the deck edge. The principal attention is focused on slab and girder bridges. Experimental verification of the finite element analysis program developed for use in the research allowed substantial parameter studies to be carried out with reasonable confidence.

A series of fabric‐reinforced concrete shells are tested for buckling under uniform normal pressure in an air chamber. The test results are compared with previous experimental results for steelreinforced concrete shells, and are used to provide a tentative design procedure for fabric‐reinforced spherical shells. The design procedure is illustrated with an example, comparing results with the results obtained by design methods for steel‐reinforced shells described by the American Concrete Institute (ACI) Committee 344 and by S. J. Medwadowski and others in 1979. Experiments to maximize buckling stiffness per unit material cost in the shell cross section are summarized. Practical limitations on shapes of fabric‐reinforced shells are described as well as their buckling sensitivity to variables in construction procedure.

Structural dynamic responses imposed by large base excitation and the effectiveness of an intermediate constraint in elevator counterweight systems are studied. A nonlinear contact finite element with nonlinear stiffness and amping approximated by polynomial functions is developed to simulate the dynamic contacts in elevator counterweight systems. A passive vibration control device—intermediate constraint—is proposed and evaluated. Analysis results show that the number of consecutive contacts increases as the excitation frequency decreases; large contact force can occur in current design practice, and constraining the motions of two rails by intermediate ties can reduce the contact load by 40%.

Finite‐element models of reinforced‐concrete beam‐columns often fail to behave in a stable manner near the point of maximum resistance. Simple numerical examples presented indicate that the conventional displacement formulation when used at the section or member level is unable to establish solutions associated with softening behavior. Thus, such analysis models are often limited in their applicability or necessitate the use of unrealistic material laws. A mixed finite‐element scheme is proposed herein that exhibits stable numerical behavior even when critical regions become ill‐conditioned. It is demonstrated that enforcing equilibrium within the member during state determination provides the necessary constraint to obtain a stable solution. Examples are presented to demonstrate the reliability and realism of this approach for members subjected to generalized excitations even where severe deformation softening occurs.

Angle members in steel lattice towers are currently individually designed by approximate formulas that take into account geometric and material properties, as well as conditions of eccentricity and restraint at end joints. Full‐scale tests of towers show that current design formulas are not always appropriate and that failures often involve a group of members (subassembly) rather than individual members. An alternate is proposed herein to full‐scale testing for determining the ultimate strength of angle members or subassemblies. It proposes to model angle members by nonlinear plate elements located in the legs of the angles. Such a detailed modeling allows for the occurrence of local buckling, lateral‐torsional buckling, eccentricity of connections, localized plasticity, etc., A plate element appropriate for modeling steel angles is discussed in detail, and several examples are presented to demonstrate the usefulness of the concept.

An experimental study is presented on the interaction of local and lateral‐torsional buckling of thin‐walled I‐shaped beams. Welded beams with four different cross‐sectional sizes are tested under a concentrated load at the center of the simply supported span. The flange width‐thickness ratios are 14.7 and 17.5, and those of the web vary from 62.3–96.0. The slenderness ratios about the minor axis are approximately 50, 70, and 100. For the beams with intermediate length, the combined failure mode of local flange and lateral‐torsional buckling is obtained, while the long beams fail only by lateral‐torsional buckling. Though local flange buckling with a short wavelength occurs only near the span center of the beams, the ultimate strengths are considerably lower than those of stocky beams. By comparing the experimental ultimate strength and the nominal design strength of the specifications in North America, it is found that the design approaches based on the effective width concept give reasonable estimates for the test results.

In precast large-panel multistory buildings, prefabricated jointed panels form the shear wall to carry the lateral loads due to wind, etc. The in situ vertical and horizontal joints form slip surfaces and the interaction between the precast panels are provided by shearing action at the joints. In this paper, plane stress and interfacial finite elements have been used to represent the precast wall panels and the in situ joints. The effect of the in situ joint properties has been studied in depth as regards the overall flexibility and stresses in the prefabricated shear wall. It was observed that very stiff vertical joints between the panels makes the shear wall stiff like monolithic cantilever walls, whereas loosely jointed prefabricated walls are more flexible. It was also observed that the effect of horizontal joints at the floor levels is only marginal under normal circumstances. Failure load of a typical prefabricated shear wall has been determined from the nonlinear analysis of the wall assembly assuming Mohr-Coulomb's failure criterion for the precast panels and nonlinear shear stiffness of the joints.

The current AITC recommended procedure for the design of pitch and tapered curved (PTC) glulam beams is an iterative one that requires initial approximation of certain parameters. Trial designs are checked for compliance with the design criteria (i.e., allowable stresses, deflections, and dimensional limits), and revisions are made as necessary. The resulting design is structurally satisfactory, but generally the procedure does not converge to the most efficient design based on material usage. In the present study the design of a PTC glulam beam is formulated as a structural optimization problem. Minimization of material volume is the design objective. Solutions are obtained for representative values of top slope, span length, superimposed loads, and allowable stresses. The results are intended to: (1) Evaluate certain AITC recommendations regarding trial values of parameters; (2) suggest better approximations where necessary; and (3) identify the design constraints controlling the optimum solution. In addition, several shortcomings of the current procedure are identified, and recommendations are made for correcting them.

The response of axisymmetric shells to internal localized impulsive loadings that produce large deformations and material nonlinearities is analyzed. Also, simplified methods for the prediction of the strain ductility of these shells are presented. These simplified methods are developed by idealizing the structure as a single‐degree‐of‐freedom model. Large deformation effects are included, and an elastic‐perfectly plastic material model is used. Several impulsive problems are analyzed both by using a finite‐element and the simplified solution. The agreement between the two results indicates that the simplified methods provide an accurate result at a fraction of the cost of the finite‐element solution.

A statistical analysis of the moment capacity of a tubular reinforced‐concrete (R/C) section is performed. The moment capacity of the cross section is estimated utilizing a second‐order stress‐strain relationship for concrete. This study followed the American Concrete Institute (ACI) code in assuming that the ultimate moment‐capacity limit state is reached when the concrete strain attains a value of 0.003. For the serviceability limit state, it is assumed that the sections fails to perform satisfactorily when the first steel reinforcement bar yields. The capacity of a cross section may vary due to variabilities in section properties, material strengths, and uncertainty associated with the prediction model. A systematic analysis of the propagation of uncertainties associated with these factors and their contribution to the overall uncertainty in the moment capacity is obtained based on the first‐order secondmoment (FOSM) and Monte Carlo simulation techniques. A closed‐form expression is derived for the uncertainty in the moment capacity following the FOSM approach. The FOSM approach provides results that are in good agreement with the Monte Carlo simulation. The results of a parametric study suggest that, for the example chimney, the coefficient of variation of the moment capacity is insensitive to the variation in the steel ratio. The analysis presented in this study would aid in determining resistance factors for probability‐based design of the tubular reinforced‐concrete members.

Haunched beams are used in building construction for architectural and sometimes economical reasons. The American and British codes do not give any formulas for the design of haunched beams under shear. However, the German DIN code and some other European codes give a formula for calculating the equivalent shear stress. A formula is proposed by the writer, which gives the ultimate shear force of reinforced‐concrete haunched beams under shear in terms of concrete strength, stirrups contribution, and dowel action. The proposed formula is a modification of the American Concrete Institute (ACI) code expression for shear design. The modification takes into account the angle of haunch inclination and the dowel contribution. Ten beams of T‐cross section and different haunch inclinations were tested, and analyzed using the proposed formula to test its suitability for each beam. The correlation of results was satisfactory.

Results of an experimental investigation on a single‐span cylindrical shell roof subjected to vertical support settlement are presented. Due to rotations associated with vertical support settlement, the diaphragm tries to move horizontally in the transverse direction by as much as 0.20 of the vertical support settlement. If the support conditions of the shell are such that this movement in the transverse direction is permitted, the shell can withstand large‐support settlements without appreciable additional stresses. On the other hand, if this movement is prevented because of practical support conditions, large stresses both in the edge beam and the shell are developed. A practical example under different sets of support conditions is analyzed. Support conditions for a statically and externally determinate cylindrical shell roof that would permit the shell to undergo large vertical support settlements without any appreciable stresses in the shell are proposed.

A formulation of a failure‐conditioned reliability index of correlated multivariate normal distribution is developed to estimate simultaneous failure probability and, finally, system reliability. The expression is based on the evaluation of failure‐conditioned mean and standard deviation. It is simple in expression and effective for all ranges of correlation coefficients. Although the equation is just an approximate one, it appears that the errors are within the allowable limit, as shown in the examples.