ABSTRACT
ABSTRACT: Various damage assessment algorithms based on system identification (SI) schemes have been proposed during the last few decades. In these algorithms, the stiffness parameters of structures are estimated by minimizing the least squared errors between measured and calculated responses of structures, and damage status of structures are assessed based on the estimated stiffness parameters. Therefore, a stable SI scheme is essential for reliable damage detection. Unfortunately, however, SI is a type of inverse problem that suffers from ill-posedness. The ill-posedness of inverse problems is characterized by the discontinuity and non-uniqueness of solutions, which are caused by noises in measurements and the sparseness of measured data, respectively.
