GNGTS 2014 - Atti del 33° Convegno Nazionale

GNGTS 2014 S essione 2.3 327 taking into account the identified optimal value of elastic stiffness ( K X = K Y = 194.2 MN/m) and the inelastic mechanical behavior provided by the manufacturer (Fig. 1b). An elastic link with stiffness K z = 7631 MN/m was used along the vertical Z -direction. The structural solution of the bridge with traditional constraint scheme, in contrary, included two supports on each abutment and two supports on the pier. All the supports were located exactly in the same position of the isolators. Seismic analyses were developed according to the Nonlinear Time-History Analysis implemented in the software package SAP2000 and suggested by NTC2008. Seven groups of earthquakes – inclusive of X , Y , Z components – were taken into account. Earthquakes were defined by synthesized accelerograms randomly selected and representative of ground motions expected in the area of the bridge. Fig. 3e shows an example of the acceleration time- histories and corresponding pseudo-acceleration response spectra for the usual 5% structural damping, soil type B (average velocity of shear waves V s,30 comprised between 300 and 800 m/ s), peak ground acceleration a g = 0.35 g and importance factor 1.3 (e.g. construction of strategic relevance). Experimentally identified damping ratios were considered in the structural analysis (Tab. 1). A time length of 50 s and a step-resolution of 5×10 −4 s were used in the numerical analysis. Main results are summarized in Tab. 3, in the form of maximum values of support reactions at the abutments and at the pier (average value of maximum reactions obtained for each group of ground motions). Reaction forces are proposed for the optimal 2D-OPT model and the traditionally restrained 2D-FIX models. Based on collected predictions, it is clear that the global seismic force transmitted by the superstructure to the elevation support components is about the 30% of the corresponding value for the bridge with traditional constraint setting. In addition, base isolation system allows to distribute seismic force effects an all the vertical resisting elements, avoiding the concentration of forces on a single structural component. Maximum shear displacements of isolators under earthquake resulted equal to 72 mm and 84 mm along the X and Y directions respectively. Finally, as expected, dynamic analysis showed that the incursions of the isolators into the inelastic range during ground motions are significant, and that these devices are responsible for the energy dissipation through the hysteretic behavior and for rising the period of the lower vibrating modes of the bridge. In Tab. 3, it can be also seen that improper mechanical calibration of seismic isolators would provide inaccurate estimations, and possibly unsafe seismic predictions, for the studied bridge. For the 2D model with nominal shear stiffnesses K X = K Y (2D-NOM), for example, predictions of maximum reaction forces would result in discrepancies up to ≈11%, compared to the 2D-OPT model. Tab. 3 – Seismic analysis of Dogna bridge. Comparison between maximum reaction forces F X , F Y , F Z on (i) single support and (ii) single isolator, respectively. X = transversal direction; Y = longitudinal direction; Z = vertical direction. ∆= 100×( F (ii) – F (i) )/ F (i) ; ∆ 2D = 100×( F (OPT) – F (NOM) )/ F (NOM) . Traditionally supported bridge (i)          Isolated bridge (ii)      Isolated bridge (ii) Δ 2D [%] 2D-FIX 2D-NOM 2D-OPT [kN] [kN] [kN] [kN] Δ [%] FX Pier 6978 867 -87.6 841 -87.9 3.1 Abutment 3161 844 -73.3 829 -73.8 1.8 FY Pier - 766 - 858 - -10.7 Abutment 8562 911 -89.4 878 -89.7 3.8 FZ Pier 15101 9980 -33.9 10304 -31.8 -3.1 Abutment 10095 9422 -6.7 9505 -6.0 -0.9 Conclusions. In the paper, dynamic characterization of the base-isolated Dogna bridge was carried out using harmonic vibration tests and FE analyses. A refined 3D FE-model of the bridge was firstly calibrated, based on natural frequencies and modal shapes extracted from