Nature of seismic control force in acceleration feedback

Abstract

Active control methods have been applied to numerous civil structures in recent years. A significant amount of the research on active control methods has been based on full-state feedback using the linear quadratic regulator (LQR) control algorithm; because measurement of the full state (i.e., the displacements and velocities of all DOFs) is difficult, such algorithms are impractical in full-scale implementations. Output feedback strategies based on measured acceleration at a limited number of points has been proposed and validated. However, a thorough understanding of the dissipative nature of the associated control forces and the way in which these forces protect the structure have been elusive. This paper considers the hysteric behavior of the control forces produced by the widely employed linear quadratic Gaussian (LQG)-based acceleration feedback control strategies and explores parallels with the pseudo-negative stiffness (PNS) semi-active damping systems. Numerical simulation studies carried out on one-story and three-story buildings with active bracing show that the LQG-based algorithms are quite versatile and can produce controllers with a variety of behaviors, including the PNS, depending upon the control objectives chosen. Copyright © 2012 John Wiley & Sons, Ltd.