Evaluation of Equivalent Linear Methods for Seismic Response Analysis of Base Isolated Structures

Abstract

Base isolation system protects structures against severe earthquakes by inserting isolators with low lateral stiffness between supporting points and superstructure so that most of deformation occurs at isolators. In the design of base isolation system, idealized bilinear models have been generally used to represent the hysteretic behavior of isolators such as high damping rubber, lead-rubber bearings and friction pendulum. In various design standards and codes including IBC 2015, FEMA-440 and Eurocode 8, equivalent linear (EL) methods using equivalent viscous damping and equivalent stiffness are permitted for seismically isolated structures when specific conditions are met. In this study, estimation accuracy of five existing EL models are evaluated for the multi-degree-of-freedom structure isolated with lead-rubber bearings. Analyses including wide range of ductility ratio were carried out by considering post-to-yield stiffness ratio and isolated period as variables. Earthquake data consist of 20 sets of recorded or simulated data provided from FEMA. Relative displacement of the base floor and absolute acceleration of roof floor are the main indexes of the structure response. The result indicates that several EL methods estimate the relative displacement of base floor almost accurately with error ratio less than 5 percent. Relative displacement is the index that affects the structural damage, so that EL method is quite reliable method for the design in the aspect of the isolators performance. On the other hand, EL method significantly underestimates the response of the absolute acceleration which can affect the damage of the non-structural member and equipment inside of the structure. It is demonstrated that modified EL method includes structure mode is required in further researches.