Journal of Earthquake Engineering

The indiscriminate use of the point source model can grossly overestimate near-field predictions of peak and spectral velocities and accelerations and, that a planar model is a much better source approximation. The relatively large variability of ground motions predicted from point source models (say, 0.27 on the log of acceleration) cannot be reduced without an improvement of our knowledge of the associated source location, attitude and thickness of the non-seismogenic zone, information that has to be assumed by the engineer a priori.

An efficient use of beam-to-column connections improves the seismic performance of semi-continuous braced steel frames, if joints may undergo large plastic deformations without forming brittle collapse mechanisms. Remarkable benefits are associated with this design philosophy. Therefore, accurate and suitable methods for the prediction of the nodal response under cyclic reversal loading must be defined for use in design.

This paper proposes two different approaches for estimating the behaviour of top-and-seat cleated connections under cyclic reversal loading. The first is based on the monotonic moment-rotation curve, while in the second method, the connection response is built up by considering the cyclic behaviour of the relevant components. For each of these approaches two models are presented and validated on the basis of experimental studies on top-and-seat cleated connections tested by the author and by other researchers.

Proposed in this paper are two analytical models for predicting the inelastic response of unreinforced brick masonry infills in reinforced concrete frames subjected to monotonic and reversed cyclic loading. The first model is based on the traditional diagonal strut concept, while the second one is a simple isoparametric element with shear deformation only. All the essential characteristics of the hysteretic behaviour of the panel, including strength and stiffness degradation, pinching and slippage, are explicitly taken into account. The models are implemented in a general-purpose program for the inelastic time-history analysis of structures, and are used for studying the seismic behaviour of typical multistorey frames with various arrangements of infill panels, including structures with an open ground storey. The results of the analysis are in agreement with both experimentally observed behaviour and with experience regarding seismically damaged buildings.

A 14-storey reinforced concrete building and the underlying soil were instrumented with strong motion accelerographs to study the amplification of seismic waves, the soil-structure interaction and the structural response to earthquakes of buildings founded on soft soil, such as the lake-bed area of Mexico City. A detailed investigation of the geotechnical properties of the site and of the structural characteristics of the building was performed, including the measurement of the shear wave velocity in the different soil layers. The instrumentation consists of 14 strong motion accelerographs, two of them underground, one in free-field and 11 in the building. In the first three years of operation of the instrumentation, six sets of records were obtained from weak and moderate earthquakes. The main features of the seismic response of the building and its interaction with the soil were derived from the analysis of seismic recordings. Ambient vibration tests were also performed to determine basic dynamic properties. The accuracy of different models of the structure and of its interaction with the soil are discussed. In general terms, the building proved to be rather flexible in both horizontal directions, and the deformations due to soil-structure interaction were not very significant. Vibration periods of the building were found to increase significantly with the intensity of the ground shaking, indicating the non-linear behaviour of the structure, even under small levels of excitation.

In earthquake resistant design of RC frame buildings, capacity design of columns in flexure is applied to eliminate the possibility of storey sway mechanisms and to spread the inelastic deformation demands and energy dissipation throughout the structure. The paper considers two alternative column capacity designs: the conventional, full capacity design of columns relative to the beams, and the relaxed one allowed by Eurocode 8 depending on how much the seismic action controls the flexural capacity of beams. Twelve RC frame buildings, designed in detail according to the two capacity design alternatives, are nonlinearly analysed under spectrum-compatible motions applied separately in the two horizontal directions and scaled to intensity from once to twice the design ground motion. In both design versions the slab participation to the beam tension flange is considered either as in the design calculations — including those of capacity design — i.e. very little, or as expected in reality, i.e. very significant. In most cases the dynamic response to the design-level motion is found to be nearly elastic, due to the overstrength of materials and members and to the "understress" of the structure due to cracking-induced softening. At higher motion intensities the effect of column capacity design and of the slab participation to the beam flexural capacity is not dramatic: column inelasticity and some light damage cannot be prevented by relaxed or full capacity design, while the large participation of the slab to the beam negative moment capacity does not overly distort the strength balance between beams and columns or the seismic response of the structure. Under any circumstances, column plastic hinging does not lead to a storey sway mechanism.

Wave propagation phenomena in soils can be experimentally simulated using centrifuge scale models. An original excitation device (drop-ball arrangement) is proposed to generate short wave trains. Wave reflections on model boundaries are taken into account and removed by homomorphic filtering. Propagation is investigated through dispersion laws. For drop-ball experiments, spherical wave field analysis assuming linear viscoelasticity leads to a complete analytical description of wave propagation. Damping phenomena are examined and evaluated using this description.