| System identification and damage detection have received considerable attention recently because of their importance in structural health monitoring. Many civil engineering structures exhibit hysteretic behaviors when subject to severe dynamic loads, such as strong earthquakes. Adaptive estimation methods are needed for on-line identification of non-linear hysteretic structural systems under dynamic excitations. Recently, techniques for parametric estimation and damage evaluation have been proposed in the literatures, however, the on-line identification of non-linear structures is still a challenge. In this thesis, a parametric identification method for non-linear hysteretic systems is presented based on recursive model reference (RMR) adaptive algorithm. In the proposed approach, an adjustable model is established, and the actual system is considered as the reference model. The deviation between the reference model (real system) and the adjustable model is considered as the controller. Through a framework of parametric adjustment, the parameters of the adjustable model can be estimated. When the output of the adjustable model is close to that of the real system, i.e., the deviation is small enough, the parameters of the adjustable model will approach that of the real system. The method presented has the capability of tracking the time-varying parameters and detecting the damages, including the damage location and severity. SDOF and MDOF linear and hysteretic nonlinear system numerical simulations show that the method presented is sensitive to the parametric variations and has good ability against noise. Simulation results for non-linear hysteretic structures demonstrate that the technique presented is effective in detecting the structural damage.The research of this thesis is supported by the National Natural Science Foundation of China under Grant No. 50478037, and by the National Science Foundation of USA (NSF Award CMS-0220027-01).
|
PDF Full Text Request