Parameter identification and model updating for a space station

The need for highly accurate analytical models of space structures is required in order to accurately predict dynamic performance. Owing to the complexity of these structures, a common modeling technique is to use the finite element method. However, the accuracy of the finite element model is limited when compared to test results from modal parameter identification. The lack of accuracy in modal parameters can be a detriment to control system design. One approach to obtain accurate models of space structures is to use models that are extracted directly from on-orbit identification or modal test data. This paper studies the parameter identification and model updating using a space station as research object. First of all, original dynamic model of the space station is established using the finite element method. Then, Observer/Kalman Filter Identification (OKID) and the Eigensystem Realization Algorithm (ERA) are applied to identify the natural frequencies of the system. Finally, model updating work is performed by the design sensitivity analysis. Simulation results show that the identification method presented in this paper is effective in obtaining the natural frequencies of the system, and the updated model of the space station produces significant improvements over the original model.