System identification methods for damage detection and assessment

摘要

Damage in a structure can be inferred from a reduction in a constitutive property of a parameterized finite element model between two time separated inferences, assuming that the baseline parameter values are known from an earlier test. Because data are usually sparse, an adaptive parameter grouping scheme is necessary to localize damage in the structural system. At each stage of the localization process one must solve a nonlinear parameter estimation problem. This paper outlines an algorithm for executing damage detection using modal test data.