Research On Vision-based Vibration Measurement And Control Of Smart Flexible Structure

Spacecraft’s tasks are becoming more and more complicated in recent years, flexiblestructure is applied more and more extensively as spacecraft is developing towards the trendof low rigidity, flexibility and large scale. Flexible structure is characterized with lowdamping and modal frequency, when it’s subjected to external disturbance or the spacecraft ismaneuvering or steering, it’s easily subjected to vibration which has a low frequency andlarge magnitude and will last for a long time, it degrades the spacecraft’s stability andinfluences the pointing accuracy and many kinds of precise instrument’s normal functioning.Long-term vibration will cause wear of structure’s components and fatigue the structure,which may shorten structure’s life and even make the spacecraft unstable to lead todestructive disaster. So it is necessary to measure, analyze and control the structure’s vibrationeffectively to ensure the normal functioning of spacecraft and flexible structure.Vibration measurement and active control problems are focused on in this study.Considering the piezoelectric smart flexible structure as research objects, research onmodeling with finite element method (FEM), vibration measurement with machine vision andactive control of the smart flexible structure has been carried out, including modeling withFEM, machine vision-based vibration measurement, numerical simulation of on-lineparameter tuning of PD (Proportion-Differentiation) controller with RBF (Radial BasisFunction) neural network and experimental research on machine vision-based vibrationmeasurement and active control. Firstly, according to the finite element theory and Hamiltonprinciple, two kinds of dynamic equation of the piezoelectric smart flexible structure werederived using two-node beam element and four-node rectangle plate element, respectively.Secondly, machine vision-based bending and torsional vibration measurement was studiedand digital image processing method for bending and torsional vibration measurement wasproposed. Then, the derived dynamic equations of the smart flexible structure weretransformed into State-Space representation and numerical simulation study of on-lineparameter tuning of PD controller with RBF neural network was carried out with MATLABsoftware.Experimental system for machine vision-based vibration measurement and active controlof smart flexible structure was established based on the previous theoretical analysis andresearch results of machine vision-based vibration measurement, experimental study ofvibration measurement and active control were carried out. The results demonstrate that thevibration of the smart flexible structure can be measured accurately and damped out quickly and effectively by using the proposed machine-vision based vibration measurement methodand active control strategy, respectively, which validates the reasonableness of the finiteelement model, effectiveness and feasibility of proposed digital image processing method forvibration measurement and on-line parameter tuning algorithm by using RBF neural networkfor vibration control.