| The models identified by the Data-based Mechanistic (DBM) black-box method can not only explain the data in a statistically efficient, parsimonious manner, they also provide an acceptable interpretation of the system under study. So it has been widely utilized in the black-box identification, especially the identification of the nonlinear and non-stationary systems. On the other hand, it seems not quite easy to use the conventional methods to identify the dynamic characteristic of the honeycomb sandwich structure, which is a main structure of the modern satellite, due to its complex construction. The purpose of this thesis is to use this powerful method to identify the dynamic characteristic of the honeycomb sandwich structure.The paper first generalized the basic characteristic of the DBM approach in the context of black-box nonlinear identification. Then the characteristic and estimation approaches of two important Time Variant Parameter (TVP) and State Dependent Parameter (SDP) models are presented and the efficiency of the DBM method is illustrated by three examples based on the analysis of Duffing Equations, Lorenz Equations and Time-Variant Equation, in contrast with other methods. The results show that although DBM method can fit the systems with high precision, avoid over-fitting and accurately reflect the time-variance and state-dependency in the systems, it may also reduce the efficiency of the whole method due to the exponentially increasing calculating numbers with the increase of the number of the states. And genetic algorithm is introduced to improve the search efficiency. Finally, the improved method is utilized to identify data acquired by vibration test of a honeycomb sandwich structure with different magnitudes, locations or vibration directions, and the identification results are used to analyze the characteristic of the structure in order to guide the further study of satellite structure's design.
|
PDF Full Text Request