Adaptive Filter Vibration Control For Smart Piezoelectric Structure

Piezoelectric smart structure active vibration control methods and techniques isan interdisciplinary forefront research topic, which is involved in advanced materials,control theory, mechanical analysis, mathematical modeling, scientific computing,experimental technology, et al. And it has important scientific and academicsignificance and project application value.As an important research development direction for piezoelectric smartstructure active vibration control, adaptive control strategy is a current researchfocus. Though great progresses have been made for adaptive filter control technology,more research is needed for engineering realization and specific implementation ofadaptive controller design and algorithm process, as well as system stability andcontrol performance analysis.The research in this dissertation is based on National Natural ScienceFoundation research project. Taking a piezoelectric smart frame structure, whichsimulates model near space vehicle, as the experimental model object, thisdissertation is focused on adaptive filter vibration control methods andimplementation algorithms. Structural dynamics analysis method and optimalpiezoelectric sensors and actuators placement strategy is investigated. Experimentalplatform and integrated measurement and control system is developed to verify thefeasibility and effectiveness of the methods and techniques. The dissertation can besummarized as three parts, structural dynamics analysis and optimal piezoelectricpatches placement, adaptive filter vibration control method and controller design,experimental platform development and experimental validation analysis. Thecontribution of this dissertation is as follows:(1) Taking piezoelectric smart frame as experiment object, distributedpiezoelectric sensors and actuators are pasted on the structure surface. With theconstructed experimental piezoelectric smart near space vehicle mode, experimentaladaptive filter control system is developed.(2) Traveling-wave analysis is used for dynamic analysis of piezoelectric smartstructure, while finite element analysis is introduced to analyze the optimalplacement of piezoelectric sensors/actuators. Taking the main component beam units of the framework, the objective target function is given. Particle swarmoptimization method is used to optimize the target function. Optimal placement ofthe distributed piezoelectric sensor/actuators is achieved.(3) As classical adaptive filtered-X LMS algorithm needs a reference signal,that is related to the excitation signal, there are defects for engineering application.An improved filtered-X algorithm with reference signal extraction is presented byconstructing reference signal from vibration residual signal with the controller itselfdirectly. Simulation analysis and experimental verification shows that the improvedcontrol algorithm, not only realizes reference signal extraction but also has fastconvergence rate and good control performance.(4) In the implementation of the classical adaptive filter-X LMS algorithm,control channel model parameter identification is very important. Though offlineidentification strategy could gain the channel model parameters, but it is not practicalfor engineering application. An adaptive filter active vibration control method withonline control channel model identification is proposed. The basic idea is tointroduce a random noise signal into the control output, while the FIR filter is usedfor real-time online identification, filtered-X control algorithm is used as the controlalgorithm. Simulation analysis and experimental verification shows that theproposed controller design method and its implementation algorithm have goodcontrol performance, which laid foundation for further practical research.(5) Since the transfer function of the filtered-X controller is an all-zero structure,it does not consider the feedback of the output signal. But this effect can not beignored in the actual system. As the transfer function of the filtered-U structurecontaining poles and zeros, the instability problem of vibration feedback can besolved to a great extent. Reference signal extraction and online control channelidentification is investigated respectively. And filtered-U vibration control algorithmwith online identification and reference signal extraction is proposed. Simulationanalysis and experimental verification shows that the proposed control method andits implementation algorithm has good control performance.(6) The structural vibration active control platform for piezoelectric smart frameis developed, as well as control system software. Experimental analysis andverifications for the proposed methods and algorithms are done. Research andexperimental verification shows that the multi-channel adaptive filter vibration control methods have strong adaptive ability, which are able to track the changes ofsystem parameters and external disturbance responses. Although the convergencerate of the proposed adaptive filter vibration control algorithms are somewhatslower, than the classical filtered-X, and filtered-U algorithm, the controlperformance is good. Meanwhile the practicality and engineering applicability of theadaptive filter method is improved by the proposed methods and algorithms.