ABSTRACT
ABSTRACT: Due to the vulnerability of cable-stayed bridges to dynamic loads such as earthquakes and strong winds, study on vibration control system for the cable-stayed bridges has been of great interest for the past few decades (Dyke 2003). The dynamic response control of the cable-stayed bridges provokes mutually conflicting response control problem. In particular, the strongly coupled model behavior and the complex cable-deck-tower connection configuration further increase the complexities of the response control problem. Viable control strategy for the cable-stayed bridges should thus account for the trade-offs existing between their dynamically induced responses. As one of the gain scheduling scheme (Parlos et al. 2001), variable gain feedback control approach is presented in this paper. The problems with the conventional gain scheduling scheme is: (a) it is difficult to select a set of discrete control gains and determine the proper switching condition among the gains for complex structure such as cable-stayed bridge; (b) degradation of control performance sometimes occurs due to control chattering caused by frequent discontinuous switching. In order to overcome these shortcomings, we present fuzzy-based variable gain approach, which is easy to select a set of discrete control gains, unnecessary to determine the switching condition among the gains, and able to modulate the continuously varying dynamic gain. The proposed technique is a hybrid method with a two-layer topology in which a lower layer consists of several conventional controllers, and a second supervisory layer possesses a fuzzy inference mechanism for endowing the controller with intelligence. Several controllers in the sub-layer are separately responsible for the individual response selected among the multiple target responses of the cable-stayed bridge to be controlled. Thus, each sub-controller is independently designed to efficiently reduce the selected response only. Then, a fuzzy supervisor determines the contribution level of each sub-controller through fuzzy inference mechanism to intelligently manage the overall enhanced control performance of the control system. To investigate the effectiveness of the proposed approach, two kinds of vibration control problems have been dealt with: the seismic response control of cable-stayed bridge and the wind-induced vibration control of cable-stayed bridge under construction. Example designs and numerical simulations of conventional linear quadratic Gaussian (LQG) control system and new fuzzy-based variable gain control system have been performed for the sake of comparison of control performance.
