Effect of friction on the re-rentring capability of sliding bearings with curved surfaces

A complete understanding of the mechanical response of sliding bearings with curved surfaces, like e.g. the Friction Pendulum SystemTM, has not yet been achieved. Among the arguments under debate is their ability to present small residual displacements at the end of the ground motion, and the possible accrual of displacement in presence of an initial offset displacement. Current re-centring criteria provided in the standards were indeed formulated for low-damping bilinear hysteretic systems, and their suitability for curved sliding bearings, which present an inherent nonlinear behavior due to the effect of friction, has been never deepened. In this research, the re-centring capability of curved surface sliding bearings is investigated in shake table tests performed on a structural mock-up consisting of a single story steel frame, scaled at one-third length and isolated with four bearings. Three bearing surfaces lining materials were considered in the study, with design coefficient of friction of 2%, 5% and 10% respectively. A detailed characterization of the sliding behavior of the lining materials was performed on a custom biaxial bench and the variation of the coefficient of friction with velocity was assessed over a range from 1 to 200 mm/s. The base isolated mock-up was tested under a set of seven natural ground motions in two series of tests: in the first series, the bearings were undeformed at the beginning of each ground motion time history, whereas in the second series an offset displacement was imposed before the run of the ground motion. The results of the tests were analyzed in terms of maximum and residual displacement of the isolation system. Additionally, the effect of the static coefficient of friction of the lining material on the response of the isolation system, the level of ground acceleration that promotes the breakaway, and the acceleration induced in the superstructure were assessed. It is eventually concluded that the re-centring criteria provided in the codes may be not conservative in presence of high friction forces developed by the isolators; an offset displacement may affect the maximum displacement during the earthquake for system with low re-centring capability; the amplification of ground acceleration through the isolated structure may be even substantially affected from the static coefficient of friction developed before the breakaway.