Assessment of a rolling isolation system using reduced order structural models

• Rolling isolation systems (RISs) are used to protect equipment inside buildings.
• Hysteretic condensed models of building structures are proposed and verified.
• In-building responses of lightly- and heavily-damped RISs are analyzed.
• Peak responses and their distributions are determined via time-history analyses.
• Installation guidelines for RISs are deduced, and important risk factors are identified.

This paper examines the performances of lightly- and heavily-damped rolling isolation systems (RISs) located within earthquake-excited structures. Six steel structures of varying height and stiffness are selected so as to represent a range of potential RIS installations. Computation models of representative frames from each of the six structures are reduced through dynamic condensation and assembled with models for biaxial isotropic hysteretic behavior within each floor. A novel reduced order modeling approach is presented in this paper. The method combines a dynamic condensation of a linear-elastic frame with the inelastic-push over curve for a detailed elastic–plastic frame model and a novel bi-axial hysteretic model for the net inter-story inelastic behavior. The reduced inelastic model combines stiffness and mass matrices from the reduced linear model with the bi-axial inelastic floor model, and is subsequently fit to push-over curves from the detailed hysteretic model. The resulting reduced order model has three coordinates per floor and provides a much simpler model for simulating the floor responses of inelastic structures. The resulting inelastic structural models are isotropic in plan and uniform along the height. Suites of recorded ground motions representative of near-fault and far-field hazards are scaled and inputted into these hysteretic reduced models. The bidirectional floor responses at varying heights are then applied to experimentally-validated models of lightly- and heavily-damped RISs. Empirical cumulative distribution functions of peak isolator responses (relative displacement and total acceleration) for the two systems are compared, from which installation guidelines are presented.