Journal of Structural Engineering

Progressive collapse is a situation in which a localized failure in a structure, caused by an abnormal load, triggers a cascade of failure affecting a major portion of the structure. Several buildings have collapsed in this fashion in recent years, and the possibility of progressive collapse is a source of continuing concern. Several alternative methods to deal with the problem of design for the prevention of progressive collapse are reviewed. A computer analysis program capable of tracing the behavior of framed structures through collapse is explained. Of particular note is the capability to remove selectively any member in the structure and determine if collapse will result. Several examples using interactive computer graphics techniques in applying the collapse resistant design procedures are presented.

Although classical suspension bridges have been designed for over 70 years using the deflection theory, it is difficult to obtain from literature a clear understanding of their structural action. The all‐important “gravity stiffness” is isolated from the modifying effects due to cable extensibility, tower stiffness, and deck bending stiffness. Graphs of maximum gravity stiffness deflections and changes of slope due to concentrated and distributed loads are presented, and formula for the modifying effects are given. It is found that with flexible decks simple formula exist for the maximum bending moment and shearing force in the deck due to concentrated and distributed loads.

Hurricane Frederic of 1979 caused wind damage to buildings, power lines, and trees in the coastal areas of Alabama, Florida, and Mississippi. Wind speeds were recorded in Hurricane Frederic at sufficient number of stations to establish maximum ground level wind speed regimes in affected communities. The established wind speeds are fastest‐mile wind speeds at 10 m height in flat, open terrain. Thus, established values are similar in character to those employed in building codes. Evaluations of the performance of buildings reveal that fully engineered buildings performed well, even when wind speeds were sligritly above code specified values. One common damaging mode was breakage of glass windows by windborne debris. There was significant structural damage to pre‐engineered and marginally engineered buildings, including some frame collapses in wind regimes corresponding to or slightly in excess of code specified values. Damage to these types of buildings was attributed to “weak links” and was influenced by limited redundancies common to other building types. Residences performed satisfactorily from a life‐safety point of view, but sustained extensive economic losses.

An exploratory study is made of the possible application of active control systems to tall buildings in along‐wind motion. An analytical procedure is developed to compute the statistical properties of the structural response of a tall building implemented by an active control system and excited by a random wind flow which is stationary in time and nonhomogeneous in space. The problem is formulated in terms of transfer matrices and the random vibration analysis has been carried out. A numerical example of a 40‐story building is given to illustrate the significant reduction of the acceleration response using either an active mass damper or an active tendon control system.

The propagation of blunt smeared crack bands through a finite element mesh in plain or reinforced concrete. structures is studied. A previous work, which demonstrated (approximate) independence of results from the chosen mesh size, is extended by demonstrating the same for the changes in mesh orientation with regard to the crack band and to the steel bars. Bond slip of bars must also be considered. This is satisfactorily accomplished by extending the previously proposed concept of free bond slip length for the cases of skew meshes and skew bars. In the presence of reinforcement, the previous formula for the equivalent strength giving approximately the same results as the energy criterion must be expanded by a term which corresponds to the third term of the asymptotic expansion of displacements near the tip of a sharp crack. The energy criterion is shown to give consistent results for different types of fracture specimens. However, the equivalent strength calibrated for the center‐cracked tensile specimen leads to a greater error (about 15%) when applied to bent specimens. Thus, the similar equivalent strength approach is acceptable only if fine meshes are used although correct convergence appears to occur as the mesh size tends to zero.

The deformation and crack width in concrete walls of slabs, plates, panels, and shells reinforced by a regular rectangular net of steel bars and subjected to in‐plane (membrane) internal forces is analyzed taking into account the frictional‐dilatant behavior of rough interlocked cracks, and for the dowel action of bars at crack crossings. The tension‐stiffening effect, i.e., the restraint of the bars between the cracks due to their embedment in concrete, is also taken into account. Numerical computer studies are carried out. Reinforcement designs obtained from equilibrium conditions alone on the basis of either the classical frictionless approach or the recent frictional (slip‐free) approach are compared in terms of the resulting crack widths. It is found that the use of frictional equilibrium design based on a low friction coefficient (0.75) leads to a much smaller crack width than the classical frictionless design when the reinforcement is laid in a direction which significantly deviates from the principal internal force direction. The influences of bar diameter and crack spacing on the crack width are also determined. The deformation analysis in which the frictional dilatant behavior of cracks, the dowel action, and tension stiffening are neglected leads to rather different values for the crack width.

A finite element analysis is made of the bending moments and shear forces in a floor slab coupling a pair of laterally loaded plane shear walls in order to give guidance on the design of the slab. Contour diagrams and design curves showing the variation of critical bending moment factors are presented to facilitate a rapid evaluation of the slab stresses induced by coupling action. Based on an appraisal of the interactive shear forces between wall and slab, a design procedure is suggested for checking against punching shear failure in the slab.

Full‐scale testing of double‐row bolted‐web connections was performed on coped and uncoped specimens. Parameters considered were the end and edge distance, bolt arrangement, and the type of hole (standard or slotted). All of the specimens failed in a block shear mode with factors of safety lower than predicted by the 1978 American Institute of Steel Construction (AISC) specification. An elastic finite element model was used to study the experimental behavior and develop a modified block shear failure model which shows good agreement with the test results. The finite element analysis also indicated the compression bending stresses at the cope to be twice that predicted by normal beam theory.

A method of evaluating various acceptance test plans for concrete is presented. This method uses operating characteristic curves and a definition of what is good concrete and what is an acceptable level of consumer's risk. The results presented in the paper are intended to illustrate the behavior of various acceptance testing procedures with a standard deviation of 500 psi (3.5 MPa). What is good or bad is not a part of the test but a criteria that all parties involved with concrete production have agreed to use. Once this agreement is reached, the operating characteristics for various testing plans should be evaluated to determine which tests provide the best balance between consumer and producer risk.

A procedure is developed for estimating speeds with which postulated missiles hit any given set of targets in a nuclear power plant or similar installation. Hit speeds corresponding to probabilities of occurrence of ${10}^{-7}$ are calculated for a given nuclear power plant under various assumptions concerning the magnitude of the force opposing missile takeoff, direction of tornado axis of translation, number and location of missiles, and size of target area. The results of the calculations are shown to depend upon the parameters: $C_{D}A/m\text{,}$ where $C_D=\mathrm{drag}$ coefficient, $A=\mathrm{projected}$ area, $m=\mathrm{mass}$ of missiles; and the ratio, k, between the minimum aerodynamic force required to cause missile takeoff, and the weight of the missile.

The usefulness and simplicity of an existing stiffness matrix method is emphasized by presenting it without its relatively complex proofs, and with simple examples. The method ensures that no natural frequencies are missed in undamped free vibration problems, and that no critical load factors are missed in buckling problems. This is maintained even when using the ‘exact’ member equations obtained by solving the appropriate differential equations. The method applies to all structures for which ‘exact’ member equations exist. The published applications are surveyed. They include rigidly jointed plane frames, space frames, and prismatic assemblies of flat or curved plates which are isotropic or anisotropic, and subjected to in‐plane normal or shear stresses. They also cover efficient multi‐level substructuring, design instead of analysis, damping, forced vibrations, elastic‐plastic collapse and programs for large computers, small computers, and programmable pocket calculators.

Current methods of steel‐bridge design, including the AASHTO Load Factor Design method, are based on elastic analyses with only limited implicit recognition of the ability of such members to adjust for. the effects of yielding. The Autostress method, which is an extension of the Load Factor method, explicitly considers these effects at Overload and Maximum Load. At Overload, autostresses and automoments caused by local plastic deformations assure elastic behavior; i.e., shakedown, after initial passages of heavy permit loads. Limit‐state criteria for riding quality and concrete cracking are considered. At Maximum Load, the strength of continuous bridge members is determined with a new plastic‐design method for noncompact members. The new design criteria can be used for the design of new bridges and for the rating of existing bridges.

Cable‐stayed bridges and suspension bridges are classified by the degree of anchoring: (1) Cable‐stayed bridges, which are self, partially, and fully anchored; and (2) suspension bridges, which are self and fully anchored. The bridges have been modelled in a simple fashion. With this modelling, equations for the cost of the cables and of the horizontal structural steel have been developed. A comparison is made of the structural efficiency of the different kinds of bridges. The effects of the selfweight, of the different mechanical properties, and cost of the different materials employed in the bridges are included. Partially anchored fan‐type cable‐stayed bridges are found to be the most efficient for spans up to 6,000 ft (1,846 m).

Minimum‐weight designs of statically indeterminate elastic beams subject to given loading are provided. Constraints are prescribed on flexural and shear stresses, and on the maximum deflection. The resulting mathematical optimization problems are solved by using concepts from Differential Game Theory and Variational Calculus. The game problem under stress constraints is reduced to a minimax one. The solution procedure is illustrated by several examples of statically indeterminate beams. The optimal designs are compared with prismatic and linearly tapered beams. Suggestions are made as to the extension of the procedure to other indeterminate structures, such as beamcolumns.

Most stability analyses of building frames are based on the assumption that the beams and columns are rigidly connected. However, many beam-to-column connections exhibit nonrigid, flexible behavior. This paper presents a method for modifying the stability analysis for flexibly-connected frames. The procedure is demonstrated for a simple portal frame, and observations are made on the effects of connection stiffness on frame buckling capacity.

The predictability of the National Defense Research Council (NDRC) equations to estimate penetration depth in concrete structures has been reevaluated using the presently available test results on missiles similar to those expected in a nuclear power plant. This study confirms the industry notion that the NDRC equations overpredict the penetration depth. In a probabilistic formulation like the one routinely used in the nuclear industry, the use of this deterministic penetration depth is overconservative and unrealistic. The NDRC equation is updated statistically using the presently available test results. This statistical NDRC equation significantly improves the predicted penetration depth. However, the form of the NDRC equation may not be ideal. A new relationship is proposed here to estimate the penetration depth by introducing a dimensionless impact factor. The proposed relationship is better than the NDRC or even the statistical NDRC equation. The proposed relationship should be used in the future to estimate the penetration depth for concrete structures.

The exact stability matrix approach is used to study the internal force distribution effect on the elastic critical load and subsequently on the failure load of planar frameworks. In particular, three options are considered. The first option utilizes an assumed axial load distribution in the structure, while in the second option an iterative procedure is employed to find the exact distribution of the axial load. The third option accounts for second order effects. A parametric study is conducted to determine the elastic critical load of various rigid-jointed frames by these three options. Also, the aforementioned options are compared on the basis of accuracy and efficiency.

An equation for predicting the elastic buckling stress of orthotropic plates is presented. The plate analyzed is simply supported along both transverse edges and elastically restrained along both longitudinal sides. Also, the plate is subjected to compressive stress that varies linearly in the direction of loadings and shear stress.

An approach to the estimation of cladding loads in hurricane-prone regions is presented which is both statistically rigorous and convenient for practical application. A computer program is described for carrying out such estimates from data pertaining to wind speeds in hurricanes, blowing from each of the 16 compass directions, and the aerodynamic coefficients corresponding to each of these directions. Stored in the program is information on hurricane wind speeds used in the development of the American National Standard A58.1-1982 wind map. The input to the program consists of the geographical location of the building, its distance from the coastline, the building height, the terrain roughness, and the aerodynamic coefficients measured in the wind tunnel. The program is available on tape from the National Technical Information Service.

Quasi-steady theory can be used for the prediction of fluctuating wind pressure coefficients acting on the edge strip of flat or low-sloped roofs and the ridge strip of steeper roofs of low buildings. Comparisons with experimental data indicate that the quasi-steady model, expressing the root mean square pressure coefficients as functions of mean pressure coefficients and the turbulence intensity at eave height, predicts the fluctuating wind pressures for wind directions normal to the roof edge; for oblique wind directions the model holds for flat or low-sloped roofs but only for the edge strip away from the corners. Many older test results dealing only with mean pressures may now be used with the model suggested to yield root mean square pressure coefficients. The latter can lead to the prediction of expected peak pressures, which are necessary for design, if an appropriate load probability distribution is assumed.

A graphical approach for the design of timber beams is presented. The proposed design charts and the method enable the selection of allowable bending stress that meets the bending, deflection, and shear requirements recommended by the National Forest Products Association. A number of illustrative examples are shown to show the step by step design approach. The method is applicable for both glulam and sawn lumber beams.