Real-time hybrid simulation of a shear building with a uni-axial shake table

摘要

Recent investments in earthquake engineering research have produced an array of experimental equipment and testing capabilities worldwide. Laboratories are often equipped with shake tables, ranging from uni-axial tables to six-degree-of-freedom tables to multiple table arrays. These tables are capable of providing interface boundary conditions for substructure real-time hybrid simulation (RTHS). In the simplest case, the lower stories of a shear building are simulated numerically while the upper stories tested experimentally. Even this simple case reveals the challenges of RTHS using shake tables. Shake tables are highly nonlinear devices, making modeling and control a challenging task. Furthermore, the mass of the test specimen is typically large relative to the capacity of the table, leading to substantial coupling of the table and specimen dynamics. These challenges are exacerbated by the loop of action and reaction between numerical and experimental components in RTHS. Any delay or lag in the realization of the desired table trajectory and measurement of the base shear can introduce inaccuracies and instabilities into the loop. This research investigates the challenges of RTHS using shake tables through a simple uni-axial shake table and shear building specimen. A model-based shake table control approach is successfully implemented for online acceleration tracking. A Kalman filter is used to reduce measurement noise in the RTHS loop without introducing phase lag. Numerical and experimental substructures with low damping are selected to demonstrate the robustness of the proposed framework for a challenging RTHS scenario. Even for shake tables with large control-structure interaction and structures with low damping, the proposed framework is robust, reliable, and uses readily available equipment, providing a new experimental tool for laboratories with modest experimental testing capabilities. (C) 2016 Elsevier Ltd. All rights reserved.