An AI-based approach to structural damage identification by modal analysis

Abstract

Flexible-structure damage is presently addressed by a combined model- and parameter-identification approach which employs the AI methodologies of classification, heuristic search, and object-oriented model knowledge representation. The conditions for model-space search convergence to the best model are discussed in terms of search-tree organization and initial model parameter error. In the illustrative example of a truss structure presented, the use of both model and parameter identification is shown to lead to smaller parameter corrections than would be required by parameter identification alone.