SUMMARYIn this paper, the distinct element method is used in order to predict the earthquake response of a multi-drum marble model of a classical column. The results are compared with experimental data for an 'identical' specimen under the same excitation. Both the numerical analysis and the experiments were conducted in three dimensions. The results show that the distinct element method can capture quite well the main features of the response, in spite of the sensitivity of the response to even small perturbations of the characteristics of the structure or the excitation. Attention, however, should be given to the appropriate values of the joint properties to be used. In any case, it seems that the method can be used with conÿdence in the restoration process of ancient monuments, in order to estimate the response to expected earthquake motions.

SUMMARYThe seismic response of free-standing classical columns is analysed numerically through implementation of the distinct element method. Typical sections of two ancient temples are modelled and studied parametrically, in order to identify the main factors a!ecting the stability and to improve our understanding of the earthquake behaviour of such structures.The models were "rst subjected to harmonic base motions. The analysis showed that, for frequencies usually encountered in earthquake motions, intact multi-drum free-standing columns can withstand large amplitude harmonic excitations without collapse. The dynamic resistance decreases rapidly as the period of the harmonic excitation increases. Imperfections, such as initial tilt of the column or loss of contact area due to edge damage, also reduce the stability of the system signi"cantly. The e!ects of such imperfections could be additive and the cumulative e!ect of many imperfections may render deteriorating abandoned monuments vulnerable to earthquakes. The response of more complete sections of the temple, such as two columns coupled with an architrave, did not deviate systematically from that of the single multi-drum column or indeed of the equivalent single block. Therefore, a much simpler single block analysis can be used to size-up the seismic threat to the monument.The model of the column of the Temple of Apollo at Bassae was also tested under recorded earthquake motions by scaling-up the acceleration amplitude progressively until collapse of the column. It was found that the columns are particularly vulnerable to long-period impulsive earthquake motions. A comparison of the instability thresholds associated with harmonic excitations and earthquake motions throws more light onto the dynamic response: it appears that around three cycles of monochromatic excitation at the predominant period of the expected earthquake motions lead to a gross prediction of the stability of a classical column during an earthquake.

SUMMARYA numerical study of the seismic behaviour of a proposed restoration of the Parthenon Pronaos is presented. The column-architrave classical structure was represented by a discrete element model, with the assumption of rigid blocks and frictional joints. Time domain analyses were performed, considering the geometric and material non-linear behaviour at the joints. The deformation and failure modes of drum columns subject to seismic excitations of various types and intensities were analysed. The adverse in uence of drum imperfections on structural safety was examined. A proposal of reinforcement with titanium connections was analysed, and it was found that architrave connections generally improve the response by decreasing the permanent displacements, while titanium dowels between the column drums do not improve signiÿcantly the behaviour and in some cases they may be unfavourable to the safety of the structure.

SUMMARYExperimental results concerning the earthquake response of a marble model of a classical column are reported herein. The model was a 1:3 scale replica of a column of the Parthenon on the Acropolis of Athens, made from the same material as the original. Several earthquake motions, scaled appropriately in order to cause signiÿcant rocking but no collapse of the column, were used as the excitation. The base motion was applied in plane (in one horizontal and the vertical direction) or in space (in two horizontal and the vertical direction), using the shaking table facility at the Laboratory for Earthquake Engineering of the National Technical University of Athens. It was found that the column might undergo large deformations during the shaking, which are not necessarily re ected by the residual displacements at the end of it. For planar excitations, signiÿcant out-of-plane displacements can happen, triggered by the inevitable imperfections of the specimen. It was also veriÿed that the response is very sensitive, even to small changes of the geometry or the input motion parameters. For this reason, the experiments were not repeatable and 'identical' experiments produced di erent results.

SUMMARYA methodology for the performance-based seismic risk assessment of classical columns is presented. Despite their apparent instability, classical columns are, in general, earthquake resistant, as proven from the fact that many classical monuments have survived many strong earthquakes over the centuries. Nevertheless, the quantitative assessment of their reliability and the understanding of their dynamic behaviour are not easy, due to the fundamental non-linear character and the sensitivity of their response. In this paper, a seismic risk assessment is performed for a multidrum column using Monte Carlo simulation with synthetic ground motions. The ground motions adopted contain a high and a low frequency component, combining the stochastic method and a simple analytical pulse model to simulate the directivity pulse contained in near source ground motions. The deterministic model for the numerical analysis of the system is three dimensional and is based on the Discrete Element Method (3D DEM). Fragility curves are produced conditional on magnitude and distance from the fault and also on scalar intensity measures for two engineering demand parameters (EDPs), one concerning the intensity of the response during the ground shaking and the other the residual deformation of the column. Three performance levels are assigned to each EDP. Fragility analysis demonstrated some of the salient features of these spinal systems under near-fault seismic excitations, as for example their decreased vulnerability for very strong earthquakes of magnitude 7 or larger. The analysis provides useful results regarding the seismic reliability of classical monuments and decision making during restoration process.