Exploiting multiple reference models for adaptive control of flexible structures

Flexible structures like tall buildings are often equipped with active control systems which allow the reduction of structural vibration induced by wind and earthquakes. Wind gusts and earthquakes are different in terms of intensity level, frequency and duration of the excitation. In particular, wind induced vibrations are usually moderate and require a continuously operating control system whose main goal is to provide occupants’ comfort in serviceability conditions. Conversely, earthquakes of significant intensity are rarely experienced by the structure and may produce structural damage that leads to deterioration in structural stiffness. The use of Model-Reference Adaptive Control (MRAC) has recently been considered for structural applications for its capability of dealing with systems’s uncertainties and time dependence of parameters. The adaptive controller has two loops: the inner loop consists of an ordinary feedback control process while the outer loop adjusts the controller parameters through an adaptation rule in order to minimize the difference between measured output and model output. In this paper the use of MRAC in conjunction with a Lyapunov-based adaptation rule is investigated for the response mitigation of a tall building equipped on top with an active device and subjected to both wind and seismic hazards. Since the required structural performances in case of wind and earthquake-excited vibrations are different, in order to optimize the control effectiveness, a modified MRAC algorithm based on multiple reference models (M-MRAC) is proposed in which the switch between the targets is performed based on the measured feedback information. The main advantage of the proposed methods is its capability of providing power saving and limitation of the peak control force. Parametric analyses allow to identify proper threshold levels for the switching condition and optimal reference models providing a compromise between safety and economy. Results of the numerical analyses on a benchmark tall building show the effectiveness of the control strategy for several loading conditions