Finite element model updating using measured modes with generalized inverse techniques

To obtain accurate finite element representation of a structure is the first step towards the optimal design, damage detection and system control, etc. The finite element model updating process involves the merging of the analytical and experimental modal data, obtaining model correlation and determining finite element changes. The key to the model updating process is the proper selection of an "a" set and the expansion of modal vectors as well as the elimination of the expansion effect during model updating.;The expansion technique must preserve the experimental model data while projecting the system changes over the complete model, while the model updating techniques are to be robust identifying model changes despite the expansion effect. A new expansion process that preserves "a" set experimental test data is developed. Two extended inverse localization approaches along with a constrained iterative process, to improve convergence of finite element model changes, are also presented.;The study of a collection of "a" set selection, modal vector expansion and model updating techniques are performed. All techniques used in this dissertation are developed in a consistent fashion, their performances are compared and discussed using various mathematical and simulated models.