| Popis: |
Proposed large space structures have many characteristics that give an advantage to modeling their structural dynamics by a set of partial differential equations (PDE's). By far, the most effort put forth in studying system identification and control to date has been based on finite element models. Despite its acceptability, such lumped parameter models have inevitable deficiencies primarily because of their complexity. Both approaches are complementary and each should be used where it is most effective. But the fact is that the distributed parameter approach has not been developed to the level of maturity as for the lumped parameter approach. Motivated by these reasons this dissertation conceived a new approach to system identification procedure on the basis of partial differential equation theory. The essential difference of the two approaches is in utilizing different mathematical models. The equivalent continuum model characterized by corresponding PDE's is designed to represent the real NASA Mini-Mast truss in terms of its global dynamic properties. A simplified maximum likelihood estimator has been successfully extended for distributed parameter estimation. The proposed procedure has been verified by both computer simulations and by analyzing actual test data. The agreement between the measured and estimated responses demonstrates that the approach is effective. Even though the evidence provided by this dissertation is significant, further research work is required to make mature the distributed parameter approach. It is clear that the implementation of the distributed parameter approach is feasible and promising for the system identification applications to large space structures. |
