Research On Identification Of Inertia Moment And Centroid Position For Satellite With Rotating Payload

With the continuous development of space science,a new type of rotating scan imaging satellite has emerged.Such satellites carry large mass rotating payload connected by magnetic bearings.The payload rotates during the imaging process,so there will be dynamic and static imbalance interference,which affects the imaging quality and the control index of the whole satellite.In order to achieve high-precision and high-stability control of the satellite,it is necessary to identify the moment of inertia and the centroid position of the rotating payload.The main research contents of this paper are:First of all,according to the particularity of the magnetic bearing connection,the dynamic modeling of the magnetic bearing rotor and the satellite is studied.The magnetic bearing force model and the air gap-pose relationship model have been derived.And based on this,the dynamic model of the magnetic rotor the satellite are established respectively by applying Newton’s law and the principle of velocity variation.Then,the identification method of the moment of inertia and the centroid position of the magnetic rotor under the fixed base is studied.A magnetic rotor pose estimator that does not rely on the measurement information of the star sensor is designed considering the star sensor failure during rotation.The estimator can also compensate for the gyro installation deviation.Through the analysis of the rotor dynamics model,a step-bystep identification strategy for the static and dynamic imbalance of the magnetic rotor is proposed.Based on EKF,the estimation algorithms of rotor centroid position and moment of inertia are designed respectively.And the effectiveness of the pose estimator and parameter estimation algorithm is verified by simulation analysis.Next,the same identification problem is studied considering the satellite is operating in orbit.Due to the coupling effect of the satellite platform and the rotating payload,the step-by-step identification of dynamic and static imbalance cannot be realized.Therefore,a synchronous identification strategy for the dynamic and static imbalance is proposed.The SVD-UKF algorithm is used to avoid the non-positive definite covariance matrix in the traditional Cholesky decomposition.Therefore the parameter identification and speed estimation can be completed simultaneously.Finally,the effectiveness of the parameter estimation algorithm is verified by simulation.Finally,aiming at the problem of whether the estimated parameters can converge to the true value,the analysis method of parameter identifiability and the design method of parameter identification excitation signal are studied.Combining the Normal Matrix and singular value decomposition,a method for analyzing the identifiability of each parameter is proposed.The mutually orthogonal excitation signals are parameterized to transform the continuous input signals into a series of sets of sinusoidal basis functions of amplitude and phase,enabling them to be optimally solved by particle swarm algorithms.Finally,the correctness of the parameter identifiability analysis method and the effectiveness of the excitation signal design method are verified by simulation.