PSI - Issue 47

A. Vrouva et al. / Procedia Structural Integrity 47 (2023) 521–534 Vrouva et al. / Structural Integrity Procedia 00 (2019) 000 – 000

524

4

Fig. 5. Preservation state of the backing wall blocks before restoration works. Photogrammetric survey the upper surface of the 3 rd course of the backing wall’s blocks, Photo archive Acropolis Restoration Service,S. Gesafidis photography, D. Mavromati, S. Kyparissi photo grammetric development. The blocks of the backing wall though suffered the most devastating damages as can be seen in Fig. 5, provoked mainly by the 2 nd century AD fire, put by the Eruli. The fire caused severe cracking to the inner part of the isodomic wall. Although the fa çade of the blocks seems intact a multifracture texture exist s in depth affecting about half of each block’s width (in Skaris et. al., 2019). The mechanical properties of the burnt marble are compromised and so is its bearing capacity. It can be noted that the anchorage areas of the transversal clamps were significantly fractured. Most of the blocks were restored in situ with the use of grout injections and titanium reinforcement and only the necessary dismantlments occurred in the backing wall. The restoration of those blocks bared strict limitations due to the preservation state itself. The attainable after restoration bearing capacity of those blocks was considered reduced according to findings of in situ investigation by endoscopic instrumentation. The anchorage area of the bearing transversal connectors was also considered irreparably damaged. 3. Seismic response of the structure In order to assess the seismic risk and the forces that would be induced to the connectors the construction was modeled using DEM analysis (Itasca, 2003). The blocks are assumed as rigid blocks and all deformation takes place at the joints, where elastic contacts are assumed in the normal (only in compression) and the tangential direction. Appropriate values for the joint parameters were considered, based on previous studies for similar structures (Dasiou et.al., 2009a and 2009b, Psycharis et al., 2000, Papantonopoulos et al., 2002, Psycharis et al., 2003). The connectors (clamps and dowels) incorporated in models were modeled with nonlinear springs placed at the positions of the connectors in the original structure. Their properties derived from experimental studies conducted for the axial and the shear behavior of metal connections in marble and concrete (Psycharis – Mouzakis, 2012, Vintzileou – Papadopoulos, 2001, Vintzileou – Toumbakari, 2004, Vintzileou et al., 2006, Pasiou et al. 2019). The damping and period parameters of the models were calibrated using real accelerograms obtained by an installed accelerometric array in collaboration with the Geodynamic Institute of the National Observatory of Athens and the Acropolis Restoration Service (Kalogeras, 2020). The calibration procedure consisted of using as input the base registration and comparing the output on the top of the column with its respective registration (Fig. 6). The Magoula 19/07/2019 earthquake registrations were used and the parameters of the model were defined as damping equal to 3% and period Τ=6sec. The calibrated model was then used to establish the design forces acting on the connectors between the two parts of the tympanon (i.e. orthostates and backing wall). The seismic events selected were chosen to correspond to the worst case earthquake that could have occurred in Athens in more than 2000 years (Psycharis I., 2015). Five real accelerograms were considered with appropriate amplification factors (see Table 1) and various dimensioning schemes for the connections were evaluated (see Fig.7).