Structural identification: Analytical aspects

Analytical aspects of structural identification are discussed considering different types of models, different model calibration spaces, and different types and amounts of experimental information. Three-dimensional finite element models and 2D grid models of an operating highway bridge are developed. The models were calibrated by taking advantage of an extensive database of experimental information accumulated from the bridge. Different types of models of varying sizes permit comparison of automated optimized calibration and manual calibration approaches. The experimental data collected characterizes both the local and global behavior of the structure. Modal data acquired with high temporal and spatial resolutions, and instrumented monitoring data describing local and global responses from static and moving truck-load configurations were used to investigate the requisite quantity, quality, and attributes of experimental data that should be considered for analytical modeling and for choosing a model calibration strategy. It is shown that modal updating and calibration must be carried out in both modal and flexibility spaces, and further, that a certain amount of experimental local and global behavior information must be included to accomplish this task. In spite of the difficulties associated with analyzing and calibrating the large 3D finite element models, they performed better in representing the overall properties and responses of the structure. While the 2D grid models are more suitable for automated calibration, given their smaller size and faster computational attributes, their validity in representing the properties of the structure needs further exploration.