Journal of Earthquake Engineering

Probabilistic seismic hazard is usually assessed by means of computer codes using seismogenic sources, parametric catalogues, seismicity rates and attenuation relationships. All these ingredients are conditioned by expert judgement that influences the final results. Even the attenuation relationships, though strictly based on experimental data, are considered a weak point due to the difficulty of modelling the interaction between seismic energy radiation and site response, and because earthquakes do not usually repeat themselves according to one theoretical model. Recently, methods making wide use of site intensity data have been developed in regions such as Italy, where the observed seismic history at selected sites is quite exhaustive. We analyzed these observations to assess seismic hazard at about 600 sites. We used a probabilistic counting technique, integrating the observations (when necessary), with computed shakings obtained from a logistic-type attenuation model. The results were then compared with the estimates provided by the recent seismic hazard map of Italy, compiled according to the traditional probabilistic seismic hazard approach. The match shows significant differences for some sites. A tentative explanation which seems to point to three alternatives is provided: (1) The mismatch between the two methodologies might appear because the stationary assumption has a poor fit with reality (at least in certain areas); (2) Some sites show a response that is systematically different from the average values predicted using attenuation relationships; (3) The definition of seismogenic zones leads to a bias in the seismicity rate estimate.

This paper addresses the problem of the behaviour of underground tunnels subjected to longitudinal earthquake excitation. State-of-the-art solutions, which are relative to infinite length tunnels in the stationary case, with linear models for both the ground and the structure, are first recalled. These models are either static or dynamic; results generally indicate large strains, often incompatible with reinforced concrete structures, and small crack widths.

For finite length tunnels, in this study the dynamic solution for the linear case is developed and it is shown that maximum strain is an increasing function of the tunnel length so that, in many situations, the behavior of the lining is satisfactory also under the aspect of crack formation. When the nonlinear behaviors of the tunnel and soil are taken into account, segments of reinforced concrete tunnels, after cracking, may undergo considerable relative displacement, undesired both for serviceability and ultimate limit states. After review of the constitutive relationship for the ground, a parametric study on the crack openings for an example tunnel is presented. It is shown that crack widths computed with linear static analyses underestimate the real value for tunnels embedded in stiff grounds and that the lining longitudinal reinforcement can reduce crack width so as to cope with serviceability and ultimate limit states.

A constitutive model for predicting the cyclic response of reinforced concrete structures is proposed. The model adopts the concept of a smeared crack approach with orthogonal fixed cracks and assumes a plane stress condition. Predictions of the model are compared firstly with existing experimental data on shear walls which were tested under monotonic and cyclic loading. The same model is then used in the finite element analysis of a complete shear wall structure which was tested under a large number of cyclic load reversals due to earthquake loading at NUPEC's Tadotsu Engineering Laboratory. Two different finite element approaches were used, namely a two-dimensional and a three-dimensional representation of the test specimen. The ability of the concrete model to reproduce the most important characteristics of the dynamic behaviour of this type of structural element was evaluated by comparison with available experimental data. The numerical results showed good correlation between the predicted and the actual response, global as well as local response being reasonable close to the experimental one.

The city of Port Vila, Vanuatu, is located in one of the most active seismic regions on earth. Earthquakes are felt frequently and, due to very rapid plate convergence rates, return period of large earthquakes (M>6) in the New Hebrides Benioff zone can be less than 10 years. Even though Port Vila does not lie on an identified seismic fault zone, strong motions by nearby earthquakes have to be expected due to the city's geographical location close to the plate boundary of New Hebrides convergence zone. An accurate estimation of the seismic ground motion across the city is of prime importance for urban developments and mitigation of earthquake risk. Following many examples of monitored strong earthquakes in the current century, it is evident that the local site effects may have a dominant contribution to the intensity of damage and destruction. In this study we focussed on the first stage of associating site effects and seismic hazard by preparing a microzonation map for Port Vila. The seismic microzonation of the city has been carried out to provide a detailed map of the zones that exhibit site effects in terms of resonance frequencies and approximated amplification of the ground shaking. Having in mind that these data will be used in improving building design to sustain strong ground motions, our analysis is limited to the frequency band of 1–10 Hz, corresponding to the expected resonance of different types of buildings in Port Vila. The Nakamura technique has been used to estimate site amplification effects from single station noise recordings. Interestingly, excluding one site located on an old dump zone, the amplification factors at about the 100 sites surveyed in Port Vila remain below 3 with an average well below 2 in the 1 to 10 Hz frequency band. These results suggest that there is no significant Vs velocity change in consequently layered material and that the uppermost sedimentary layers in the surveyed down town area are relatively thin. These observations are in agreement with the mapping of limestone terraces throughout Port Vila area. However, both the surface geology and results from seismic zonation indicate a thicker (up to several tens of meters) sedimentary cover around the Bauerfield airport and in the Méle terrace zone. Low resonance frequencies (around and below 1 Hz) and amplification factor of the order of 5 were observed over this large area, immediately outside Port Vila. Any building development in this area should take these results into account.

In this paper, the responses of multidegree of freedom (MDOF) structures on sliding foundations, subjected to harmonic or random base motions, are investigated taking into consideration the variations of friction forces. The variation of friction force is a consequence of variation of friction coefficient, which depends on such parameters as relative velocity and the existing pressure. Modelling of the friction force under the foundation raft is accomplished by using a fictitious rigid link with a rigid-perfectly plastic material. This results in identical equations of motion for the sliding structure, both in the sliding and nonsliding (stick) phases and considerably decreases the required number of time steps for the nonlinear analysis. Since the force in the link is of constant value, to consider the varying friction force, a compensatory force, which is the difference between the exact friction force and the constant force in the rigid link, is applied to the foundation raft. A model of variable friction coefficient for Teflon–steel interfaces is used for the assessment of the method and the results are compared with existing literature, through which, the capability of the method is illustrated. It is shown that by using exact model of friction lower values for the superstructure responses are predicted compared with those obtained by using Coulomb friction model. Furthermore the effect of the stiffness of the structure on the differences between the results of the two models is also studied.