Floor Response Based Earthquake Action Mechanism and Seismic Demand of Curtain Wall System of Shanghai Tower

A scaled structural model of Shanghai Tower was constructed in accordance with the philosophy of similitude law. Several natural and synthetic earthquake records were selected considering the characteristics of the main tower, the site soil condition and the prime frequency ranges, etc. The records were scaled corresponding to the similitude coefficients, and input to the shaking table to excite the scaled structural model fixed on the table. The floor responses of the model were achieved during the selected earthquake record excitations. For comparison, a finite element model of the prototype main tower was created, and time history dynamic analyses were also conducted in order to get the characteristics of the floor responses. Peak floor acceleration distributions along the height of the structure were acquired in terms of the floor acceleration magnification (FAM) curves. The results showed that the acceleration responses of the lower floor levels were small under minor level earthquake excitations, while in the higher floor levels, especially near the top crown of the tower were very serious. Floor spectral analysis showed that the prime frequency ranges of the floor responses had relative obvious effects to the curtain wall system attached to it. Inter-storey drift responses were also analyzed through the shaking table test results, and the typical inter-storey drift ratios (IDR) of the crown were discussed. Finally, in order to investigate the deformation behaviour of the curtain walls under the IDR time history excitations, a simplified FEM model with the radial steel pipes and the elliptical steel rods was created and analyzed, the results showed that the radial steel could exceed elastic range under the deformation originated from the fierce floor responses.

The case study shows that floor response based seismic demand evaluation of the curtain wall system of Shanghai Tower is applicable and efficient.