Blind Prediction on the Seismic Response of a 5 Storey Steel Frame Building on Triple Friction Pendulum Isolators

Abstract

This paper details an entry into the 2011 Blind Analysis Contest hosted by the Center for Urban Earthquake Engineering (CUEE), Japan. The contest was a part of the larger collaborative NEES TIPS project and the EDefense Seismic Isolation Test Program. The testing program subjected a full-scale five story steel frame building to a series of past earthquake motion using the shake table in Miki City, Japan. The first shake table test was conducted with the building situated over triple friction pendulums (TFP). Later, the test was repeated without the base isolation devices. Contest participants were challenged to estimate key building response parameters during these two rounds of shaking with only as-designed information, specifically, without any knowledge on the as-built performance to simulate the experience of an authentic design project. The University of Auckland entry was developed using SAP2000. The notable features of the finite element model included; i) the use of membrane elements connected at frame centrelines to simulate a flexible diaphragm, ii) the simulation of composite floor action through an tuned effective width, iii) the use of SAP2000 fibre hinges as a lumped plasticity approach to model the nonlinear composite action, and iv) the inception of a simplified nonlinear spring arrangement to model the response of the TFPs. A comparison of the blind predictions and the actual experimental results is presented. The results showed that the blind prediction matched two of the three key response parameters very well but not so well in storey drift angles. Good time histories comparison suggests that errors were made in the post analysis calculation or a erroneous definition of drift angle.