ABSTRACT
The motivation for work presented here has been earthquake engineering applications in which a few large capacity actuators are used to counter the effects of disturbances (i.e., ground motion in an earthquake episode), where disturbances typically have a large range of magnitudes. The system
itself is almost always stable, due to structural damping. As a result, the focus is on performance, primarily that of establishing guarantees for the 1/2 gain from the disturbances to the controlled outputs, though other measures are possible-after some standard modifications. The main goal is to reduce the conservatism inherent in many of the techniques that are commonly used in saturation problems. The salient features of the technique presented here include explicit incorporation of the saturation nonlinearity and dependence of the performance guarantees on the actuator capacity (i.e., larger actuators leading to better performance guarantees). This feature can be useful in the design stage for trade-off regarding the actuator size and performance. Furthermore, resulting controllers have higher gains, to better use the available actuator capacity.
