Composite Active Vibration Control For Smart Piezoelectric Structure Using Feed-forward Compensation

The vibration in the structures not only affects the performances but also causes fatigue damages,so how to solve the vibration problems has become increasingly important. But the traditionalstructures and control methods are difficult to solve these problems. The piezoelectric smart structureis an active smart structure which is composed by bonding the piezoelectric elements on the basalstructure without changing the strength properties of basal structure. This structure has manycharacteristics, such as sensing and actuation functions, excellent electromechanical couplingcharacteristics, light quality and reliability, etc., which is an effective structure in the vibration control.But it is tough and has a variety of uncertainties, couplings and unknown external excitations, whichbring great challenges to the active vibration control of the piezoelectric smart structure. Soresearching the easy implement methods to further reduce the vibrations are the urgent research topicsfor the researchers.As for the adverse effects caused by various excitation disturbances in the piezoelectric smartstructure, some feed-forward compensation based active vibration control methods are researched inthis dissertation to improve the vibration suppression ability of the structure. The main works aresummarized as follows:(1) According to the constitutive equation of the piezoelectric element, the basic theory of thestructural vibration and the couplings between the piezoelectric actuator/sensor and the basal structure,two state-space models based on piezoelectric patches and accelerometer as the sensors respectivelyare got by using the finite element method.(2) There are various internal and external disturbances including the spillover, harmonic effect andmodel errors in the active vibration control system. These disturbances have a great influence on thevibration control performance and stability of the closed-loop system. So a composite controller basedon disturbance observer (DOB) is presented. Firstly, DOB is used to estimate the external and internaldisturbances and the estimated value is used for the feed-forward compensation design to suppress thedisturbances combining with the feedback controller LQR. Secondly, in order to improve thevibration suppression performance of LQR, a chaos optimization method is proposed to automaticallytune the weight matrices of LQR. The all-clamped panel is used to verify the proposed method. Theexperiment results demonstrate that the proposed method has a good vibration suppressing propertyand stability. In the multi-modal active vibration control, there exist output superposition, the control input coupling and collocated sensor/actuator. To this end, an amended DOB based multi-modalcomposite active vibration control is developed. The simulation and experiment results demonstratethe excellent vibration suppressing property.(3) Because of the non-collocated of sensors/actuators, the phenomenon of phase hysteresisinevitably exists in the structure. A composite vibration control method with two degrees of freedombased on the amended DOB is proposed. And the stability and performance of the strategy areanalyzed. The acceleration sensor high-frequency noise is also considered in this dissertation. Besidethe amended DOB, a high-pass filter is introduced. So the effects caused by all the low-frequencydisturbances and the high-frequency noise can be eliminated. The all-clamped stiffened panel smartpiezoelectric structure is used to verify the strong ability in suppressing the internal and externaldisturbances, acceleration sensor high-frequency noise and the phenomenon of phase hysteresis.(4) In order to overcome the weakness of the output superposition and the control input coupling ofthe multi-modal active vibration control, a multi-modal linear active disturbance rejection compositeactive vibration control strategy which is independent of the mathematical model of the structure isproposed. First, the output superposition and the control input coupling of other modes are consideredas the lumped disturbances. The linear extended state observer (ESO) of each individual loop canestimate the lumped disturbances, and remove the effects by feed-forward compensation. And then thelinear PD controller for each single mode is designed, respectively. The experiment results followingthe several situations on the all-clamped stiffened panel experimental platform demonstrate that theproposed method not only effectively suppresses the vibration excited by first two resonancefrequencies of the stiffened panel, but also has the capability of suppression fluctuation caused byuncertain factors.(5) In order to overcome the weaknesses of the original second order active disturbance rejectioncontrol (ADRC) for the active vibration control, several ADRC compensation control methods arepresented. First, it is inevitable that the ESO has the estimation errors of the disturbances and stateswhile the disturbances change, so an ADRC based on state estimation error compensation is proposed.The estimation error of the ESO is used to compensate the ADRC to reduce the disturbances andstates estimation burden of the ESO and to improve the effects of the active vibration control. Theexperiment results with four situations on the all-clamped panel experimental platform demonstratethe effectiveness, practicality and strong anti-disturbance ability of the proposed control strategy.Second, in order to solve the phenomenon of phase hysteresis in the multi-modal vibration control ofthe stiffened panel structure, an output predictor is introduced to help the original second order ADRCactive vibration controller to cancel the adverse effects caused by time delay. Experimental results demonstrate that the presented approch is an effective approach for suppressing multi-mode vibration.Third, a kind of linear composite active vibration control strategy combining a second order ADRCwith an acceleration feedback controller is designed through analyzing the stability condition of theacceleration feedback closed-loop system. And the stability and superiority of the strategy areanalyzed theoretically. The experiment results demonstrate that the proposed strategy not onlyeffectively suppresses the vibration excited by the sinusoidal excitation and the external disturbancesof the stiffened panel, but also can suppress the fluctuation caused by uncertain factors.