Vibration Control Of Large Space Structures Based On Model Identification

With the increasingly complex tasks undertaken by spacecraft,the spacecraft structure tends to be large-scale and flexible.Due to the lack of atmospheric damping and other factors in the space environment,the flexible vibration caused by the position and orientation adjustment of large spacecraft is difficult to attenuate.In the long run,it will even cause structural damage or mission failure of the spacecraft,and reduce the service life of the spacecraft.Therefore,it is urgent to suppress the vibration of the flexible structure of large spacecraft.Due to the complex in-orbit space environment and other uncertain factors,the dynamic model parameters of the large-scale space structures determined by finite element simulation or ground test still have large errors compared with the actual parameters.Therefore,the vibration control of large-scale space structures based on model identification is studied in this paper.Firstly,the dynamic modeling of solar panels of the Dongfanghong-3 satellite was studied,and the dynamic model containing piezoelectric material was established in the finite element software ANSYS.Furthermore,the Modal Cost Analysis criterion is used to reduce the order of the dynamic model,and the accuracy of the reduced model is verified by the bode diagram and dynamic response of the model before and after the reduction.Secondly,the model identification is studied by using input and output data combined with the eigensystem realization algorithm.The effects of input data and the number of observation points on the model identification results are compared.The results show that the model is more accurate under continuous random input data and multi-point information.The error of natural frequency is less than 1% under multi-point information,and the dynamic response curves of the identification model and the real model almost coincide.Finally,the vibration control algorithm and simulation research of large space structures based on model identification is carried out.Combining the classical LQG control theory with model identification,an LQG vibration control method based on model identification is proposed.Through the design of the excitation signal and the measurement signal,the ERA algorithm is used to complete the high precision identification of the state-space model of large space structure,and the linear-quadratic optimal controller(LQR)and Kalman filter are designed by using the identification model.Firstly,the influence of single-point and multi-point information on filtering results is compared.The results show that the filtering value obtained by multi-point information is more accurate.Then,the vibration control simulation under pure bending and bending-torsion coupling is designed.The results show that the controller based on model identification can effectively control the vibration under these two conditions,and it is quickly controlled.Finally,compared with the traditional controller design based on a finite element model,the results show that the controller design based on model identification is better than the controller design based on a finite element model,which also verifies the effectiveness of the controller design based on model identification.