چكيده به لاتين
Abstract:
Multiplicity of the required sensors in the centralized control of large-scale structures, leads to excessive data transmission and the resulted system lags. On the other hand, decentralized control, as an alternative approach, offers sub-optimal responses. In this thesis, to compensate for the mentioned deficiency, an innovative estimation procedure is presented. In this method, any substructure is equipped with an embedded Luenberger observer, which estimates the whole states of the structure based on the available ones. Then, the actuator commands are locally calculated by the estimated states. To evaluate the proposed method, fully decentralized control of a 3-story frame structure and semi-decentralized control of a 20-story one are firstly investigated. Furthermore, a comprehensive comparison is made between the proposed method and the alternative approach based on the useful Kalman filter. The performance of the presented method is also evaluated considering the structural uncertainties, the noisy signals, and the probable time-delay.
Keywords: decentralized control, Luenberger observer, Kalman filter, uncertainty, state estimation
