A two-tier SHM framework for nonlinear identification of friction pendulum isolators in base-isolated buildings

This paper presents a structural health monitoring (SHM)–oriented framework for the nonlinear identification of friction pendulum isolators in a real base isolated school building. Based on a parametric investigation carried out for this study, which identifies the high-speed friction coefficient and the stiffness parameter as the most influential variables governing isolator deformations and base shear, the proposed identification strategy focuses on the high-speed friction coefficient while assuming the linear stiffness properties to be known from independent calibration using ambient vibration data. A detailed three-dimensional finite element model is paired with a condensed representation characterized by a limited number of degrees of freedom, where the isolation system is replaced by a single equivalent FP isolator. A genetic algorithm–based identification procedure is first applied to the full finite element model, accurately recovering the family-wise high-speed friction coefficients and reproducing target accelerations and isolator deformations with very small discrepancies, even in the presence of realistic Gaussian noise. The same optimization framework is then applied to a condensed model with only few degrees of freedom, where the isolation system is represented by a single equivalent FP isolator and the search space is reduced to two decision variables, one for each horizontal direction. The results show that the simplified model retains sufficient accuracy, yielding response histories comparable to those of the full model while requiring only about 6% of the corresponding computational time. This makes the condensed model suitable for rapid event-by-event updating, whereas the detailed FE model is reserved for less frequent in-depth analyses when significant changes are detected.