Effectiveness of the RBRL isolation system: evidences from seismic tests and numerical simulation

The Rolling-Ball Rubber-Layer (RBRL) system was developed to enable seismic isolation of low-mass structures, such as works of art or special equipment, and is very versatile, a great range of equivalent natural frequencies and coefficients of damping being achievable through the independent choice of the system parameters. The paper presents new results from a previous campaign of shaking-table tests (PSTRBIS ECOEST 2 Project, 1999), related to a superstructure model consisting of two concrete slabs separated by four M16 studs 500mm long, which give a first mode fixed-base response at about 2.5 Hz. In particular, attention is given not only to the global behaviour of the system, which includes the steady-state rolling, but also to its small-deflections behaviour, influenced by the creation of pits in the rubber layer due to its viscoelastic properties. These experimental results are compared to those obtained from numerical simulations, conducted in OpenSees, using a FE fixed-base model previously calibrated using other shaking-table tests performed at fixed-base. These comparisons, isolated (test) versus fixed (model) case, are presented in terms of peak values of acceleration and inter-storey drift, time-history accelerations and displacements and by means of response spectra ratios for both the slabs, and show the effectiveness of the system not only at large displacement but also for small deflections if compared with an equivalent sliding isolation system. Attention is here restricted to uniaxial behaviour. Finally, some considerations are made regarding a possible characteristic frequency of roll-out of the balls from their initial pits.