Investigations On Some Key Technologies For Three-direction Vibration Isolation Smart Structures

Combined with smart structure technology, some key technologies for three-dimension vibration isolation were investigated in this dissertation, including new type actuator development, model identification for active vibration control system, generalized feed forward adaptive control algorithm and decoupling control strategy. Experiments for vibration isolation platform based on piezoelectric smart structure were also carried out for validation. The details are as follows:First, as the frequency response range of some individual actuators could not meet the demand of wide-band actuation, a metal-piezoelectric cylinder structure was designed to compensate the limitation of magneto-electric actuators for high frequency actuation, while a novel magneto-electric actuator was developed to overcome the disadvantage, being not capable to output large displacements when used in low frequency, of piezoelectric actuators. Thus, a new type combined actuator whose output displacement and response frequency were expended. Through theoretical analysis and finite element simulation, the impedance characteristics and the relationship among the magneto-electric force, the excitation current and the magnets position were derived. Afterwards, the influence of the magnetic/air gap on the actuation force was deduced. The mathematical relationships of the force, strain and structural parameters of the metal-piezoelectric actuator were given. And the frequency response of the compound actuator was tested.In practical applications, the combination of uncertain factors, such as errors in manufacture, inaccuracy of measurement, would have bad influence on the model identification for vibration system, therefore, an grey-model identification method, using some ideas from grey model theory, for the control system is proposed, in which the weighted accumulation of the sampling data is adoptedAdaptive feed forward control algorithm was widely used in active vibration control. By employing the poles of the controlled structure to construct orthogonal basis functions, a generalized adaptive feed forward algorithm was proposed in this dissertation. It has the advantages of both FIR and IIR filters. The FIR algorithm is always stable since it does not have output feed-back; however, it needs a sufficient number of coefficients to represent a more complex dynamical system, thus would increase the computational burden substantially, and may also cause a slow convergence. Contrary to FIR, the algorithm based on IIR, matching poles and zeros of a physical system by output feedback, could achieve the same performance as the FIR structure with much small number of coefficients, however, it was not unconditionally stable. The algorithm for three-dimensional isolation should be a multi input multi output one, so a decoupling strategy must be considered if the isolated structure could not decoupled by itself. The problem of coupling effect on secondary path and the coupling between the actuator output signals and the reference signals were discussed in this dissertation. According to the decoupling strategy, the vibration control system for vertical tail could suppress the vibration in flight successfully. The power spectrum of the single frequency vibration signal could be reduced by 35dB.Finally the experiments on three-axial vibration isolation platform using piezoelectric structure were performed. Experimental results show that the algorithms proposed are effective. Compared with FXLMS algorithm, generalized feed forward control algorithm has better performance and stability in band-limited vibration control.