Journal of Structural Engineering

A hybrid method, consisting of spectral and probabilistic analysis techniques, is proposed to evaluate the dynamic response characteristics of offshore platforms considering nonlinear wave kinematics and free‐surface fluctuation effects. The effect of nonlinear wave kinematics and intermittent loading is that the wave loading is non‐Gaussian. The response of platforms to such non‐Gaussian loading will be non‐Gaussian. A combination of spectral and probabilistic analyses is developed to estimate the response up to the fourth moment. The probability distribution of the response is estimated as a mixture of Gaussian distribution and shifted exponential distribution. With the help of an example, it is shown that the free‐surface fluctuations have a significant effect on the response. The skewness and kurtosis are shown to increase with an increase in significant wave height. The probability distribution is shown to deviate from the normally assumed Gaussian distribution, particularly at the tails commonly found in full‐scale measurements.

A fatigue analysis procedure is proposed for offshore platforms where the response is non‐Gaussian due to nonlinear wave kinematics and free‐surface fluctuations. The non‐Gaussian distribution of the stress process is considered to be a mixture of Gaussian and shifted exponential distributions. The level crossings of this process are estimated considering it a translation process, and the probability density function of the peaks is estimated numerically. Using Palmgren‐Miner's rule for fatigue damage accumulation, the total cumulative damage over the long‐term wave climate is estimated. It is shown that at higher sea states, where the effect of intermittent loading is significant, the probability density function of the peaks of the stress deviates significantly from the normally used Rayleigh density. The assumption of Rayleigh density for peaks at higher sea states is shown to be unconservative, and the unconservatism increases with an increase in significant wave height. Since field observations confirm that the dynamic response of offshore platforms is non‐Gaussian, the method proposed could be used to check the fatigue design of offshore platforms to avoid overestimating the fatigue life.

The behavior of partially prestressed composite beams consisting of a concrete slab supported by a steel beam is analytically evaluated. The prestress is introduced in the negative moment regions when the slab acts compositely with the beam. The effect of this scheme on crack prevention in the negative moment region and savings in material due to superior structural action are considered. This is the first of a two‐paper set. In the second paper the results of tests on a model continuous beam, of the type considered here, is presented and compared with the analytical predictions of this paper.

The behavior of a two‐span composite beam, consisting of a concrete slab shear connected to a steel beam and prestressed near the central support region, was studied experimentally. The test beam was 37 ft (11.28 m) long with two spans 18 ft (5.49 m) in length each. The test results were compared with predicted values. The advantages of partial prestressing of composite beams in terms of performance under service loads and economy are further established.

A state‐of‐the‐art review of recent developments in the area of numerical structural optimization over a period of approximately five years, 1980–1984, is presented. Recent reviews, techniques and civil engineering design applications are covered, as well as problems with stability, frequency and shape constraints; a diverse list of papers, books and computer programs is cited.

Load and resistance factors are used to reduce the probability of overloading, understrength, or both to acceptable levels. The ACI Building Code specifies the use of resistance factors for beam‐columns in the following ways: (1) Resistance factors are applied to cross‐section strength; and (2) resistance factors are applied to critical buckling strength used in moment magnification due to slenderness effects. Probabilistic examination of ACI resistance factors as used in (1) was reported in an earlier study and is available in the literature. This paper examines the ACI resistance factors as employed in (2) and suggests a value for design office use for computing moment magnification of concrete slender columns.

Presented are the results of analyses on ring‐stiffened cylinders under combined axial compression and external pressure by rigid plastic mechanism approaches and finite elements. The shell geometries examined lie in the range used for primary members of floating offshore platforms. For each geometry, a series of stiffener sizes is analyzed to determine the limiting stiffener area required to prevent failure in the general collapse mode. The results are presented in the form of design charts. Comparisons between the present results and Det norske Veritas (DnV), European Convention for Constructional Steelwork (ECCS), and American Society of Mechanical Engineers (ASME) design codes have highlighted areas where these codes need to be reassessed.

The dynamic response of a simply supported box girder bridge due to a vehicle moving across the span is presented. The effect of the randomly varying dynamic force generated by the vehicle moving over a rough bridge deck is considered in the analysis. A comparison is made of results obtained by modeling the vehicle and vehicle‐pavement interaction as a single unsprung mass and as a concentrated time‐dependent force. A statistical analysis of the results shows that peak impact factors for deflection, moments, and shear are considerably greater than those normally prescribed in bridge codes, especially when the vehicle is moving at high speeds. The effect of vehicle mass is also more significant at higher speeds.

A basic theory for the analysis of buckling behavior of sheathed wood walls is developed herein. The formulation is based on a simplified model of the wall. A governing differential equation that assumes a linear load‐slip relationship for the connection between stud and sheathing is derived and solved. The theory developed is used for investigating the contribution of the sheathing to the overall stiffness and strength of walls. The partial composite action between studs and sheathings produces a higher strength than do estimated values provided by current design methods.

A model to predict the ultimate capacity reliability of wood systems is developed. The model is analogous to Markov cumulative damage models. System ultimate capacity is defined as the failure of two adjacent members. The reliability of a 12‐member wood system under a snow load history is examined. Practical systems exhibit a 50‐yr ultimate capacity reliability greater than or equal to the reliability with respect to first‐member, first‐cycle failure. This reliability level is suggested as a possible future design criterion. The model is used for snow load histories that produce both stationary and nonstationary behavior over the structure life. Potentially unconservative assumptions used in the model development are explicitly stated and discussed.

Previous studies of the strength of steel columns are reviewed, with particular attention being given to the role of the column as part of a frame. The concept of partially restrained (PR) construction in the draft LRFD Specification is shown more appropriately to describe the behavior of frames incorporating most of the normal forms of field bolted connections. Despite several recent theoretical studies of end‐restrained columns, complementary experimental data against which the analyses may be checked and refined are virtually nonexistent. A series of full‐scale tests on I‐shaped, nonsway subassemblages comprising a column attached via various types of semirigid connections to 4 beams is described. Both direct column load and independent beam loading were used with the response of the column, connections, and beams being monitored. The subassemblage was constrained to eliminate out‐of‐plane action. Preliminary study of the results clearly demonstrates the large amounts of end restraint transmitted by even the most flexible connections. When this is taken into account in determining column strength through a modification to the SSRC G‐factor approach, acceptable estimates of the test values may be obtained from the normal interaction formulas approach to beam‐column design.

A procedure is presented which is suitable for determining statistically based load factors for use in loading codes, without consideration of the statistics of the material of construction. This procedure involves investigating the probability of exceeding the factored design loads during the service life of the structure. The probability of exceeding the design load is defined in terms of a load index. In the development of limit states material codes the material factors can subsequently be determined using a full reliability analysis. The load index implied in the American loading code ANSI A58.1–1982 and i the American concrete code ACI 318–1983 are investigated.