Seismic behaviour of ancient colonnades interacting through a diaphragmatic roof

In ancient monuments the space between the colonnade of the peristyle and the colonnade of the prostylon (prostasis) is usually covered with a marble coffered roof. The same applies in the case of the Parthenon of the Athenian Acropolis. This current numerical study aimed to establish whether this roof, though it was just resting on the colonnades without being attached to them, accounted for a diaphragm providing some kind of interaction between the independent colonnades during seismic events as it was intuitively expected. The case of the west pteron roof of the Parthenon was considered, under the scenarios of its restoration. This initial study provides the guideline for attempting a more thorough study on the rehabilitation of the columns for both colonnades, considering the actual preservation state due to a plethora of damages the temple suffered in its long course of existence of approximately 2500 years. Three models were considered for the numerical analysis. In all of them the actual layout of the peristasis colonnade was considered along with the prostasis colonnades and part of the north south and west wall of the cella. The first model accounts for the current state of the monument, i.e. without the west pteron roof. The second model accounts for the restoration of the grid of marble beams that hold the marble coffers of the roof. The third model accounts for the full restoration of the coffered roof. The numerical model includes the actual number of architectural blocks without damages (rigid blocks) and well fitted together (closed joints). The base of the structure was considered rigid and uniform in the level of the pavement. For the analysis the discrete element method was used (the 3DEC 5.0 code by Itasca), according to which every architectural block is considered as individual and rigid, while all deformation occur in the joints (interfaces) of the blocks. The method used has been validated with actual accelerograms from the accelerographs’ network established on the rock of Acropolis and on the Parthenon in particular. For the analysis of the structure 9 appropriate actual accelerograms of strong ground motions -suitably amplified to designate the worst case scenario for a return period equal to more than 2 thousand years- were used. The time-histories records of critical architectural members’ displacements were compared and evaluated for each seismic event investigated separately. The investigation focused on crucial members’ behaviour such as capitals, architraves, pediment blocks, the roof beams and the roof coffers. Among the conclusions of the current study it can be ascertained that there is an inherit sensitivity of such models on the frequency content of the ground excitations used, which is depicted by the lack of systematic members’ behaviour and the inconsistency between maximum and permanent displacements with the seismic excitation. In particular architectural blocks a systematic behaviour seems to remain constant in all the investigated ground motions. For example the tendency of the outer peristasis capitals to lean outwards at the angle was obvious. Furthermore, the general behaviour of the lengthening of the façade demonstrated by the opening of the joints at the architrave level and at the pediment blocks level was observed. As for the displacements, the models confirm that the roof tends to restrain the overall displacements of the structure, especially where the maximum displacements appear towards the centre of the façade, whereas the grid of beams scenario tends to amplify the displacements of the upper blocks of the structure.