Robust control system for smart structures

• Main Author: Xie, Lihua.
• Other Authors: School of Electrical and Electronic Engineering
• Format: Research Report
• Language: English
• Published: 2009
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation::Control engineering
• Online Access: http://hdl.handle.net/10356/17211
• Institution: Nanyang Technological University

This project involves the study of active vibration control of the Stewart Platform using robust control and adaptive filtering approaches. The system is first studied and examined. Robust adaptive filtering algorithms for active vibration control are then considered. A robust control design based on sensitivity minimization is also studied. The desired system with an effective controller is required to be able to achieve more than 20dB of vibration reduction. Experiments are conducted to verify the controller performance and the results are presented. About 20 dB to 30 dB of attenuation is achieved in real-time experiments for vibration frequency ranging from 60 Hz to 220 Hz. Unlike the feedforward adaptive filtering systems, adaptive feedback algorithm only utilizes the error signal to generate a control signal to counteract the vibration disturbance signal. The reference signal is regenerated from the error signal and fed back to the filtered-X LMS adaptive filter. In the algorithm, control signals to the actuators are also fed into the FIR filter model of the secondary path of each actuator. The sum of the outputs of the filters is then subtracted from the error signal to get an estimate of the disturbance signal. The error signal and the estimated disturbance signal are fed back into the controller. The controller is the filtered-X LMS adaptive filter whose coefficients are adjusted to approach the inverse model of the secondary path of each actuator. The secondary path model is estimated offline in order to reduce computation burden. Note that accurate modeling of secondary path is extremely important for the stability of the algorithm. FIR adaptive identification of the secondary path of each actuator is performed. Nonlinear modifications have been made on the controller to minimize the effect of the modeling error.