Gain-scheduled adaptive control of a hybrid structure

Abstract

Structures undergo large deformations when subjected to strong earthquakes, and hence exhibit nonlinear behavior. While passive base isolation systems are effective for protecting seismicaly-excited structures, they are limited to low-rise structures because for tall structures uplift forces may be generated in the isolation system, leading to instability failure. When active control systems are used as the only seismic control device, the forces required for control may be very large. In recent years the combined use of passive/active (hybrid) control systems has been proposed. Any model-based control system is only as good as the structural model utilized for its design. As the structure enters the inelastic region, its structural properties may change, thus deviating considerably from their original estimated values. This may lead to poor performance and instability of the controlled structure. Therefore, a robust controller capable of handling nonlinear structures is necessary. The purpose of this study was to develop a robust controller for a six-story, base-isolated office building. This was accomplished by developing several linearized models of the structure using a validated nonlinear finite element model. A dynamic controller was designed for each of the linearized models based on the H,, -optimal control method. A performance study was conducted in order to investigate the robustness of the H. compensators. Structural responses of the controlled structure utilizing two active bracing system configurations were studied in order to evaluate the benefits of using multiple active members. The effects of time-delay and parameter variations on the performance of the H. compensators were studied. The trade-off between compensator bandwidth and robustness was also investigated. Once compensators for each linear structure were designed, scheduling was performed. This was accomplished by fitting the gains of the compensators to the ductility of the system and the disturbance (scheduling variables) resulting in a global nonlinear compensator. The performance of this compensator was investigated via transient response simulations. Since this is the first known attempt to use gain-scheduling to control seismically excited civil engineering structures, it is anticipated that this study will significantly impact the way control problems are approached in nonlinear structures.