Damage detection and model refinement of coupled structural systems

This study has investigated the detection of damage and the model correlation of structures that are affected by coupling to either another structure or to another medium, such as a fluid. In the damage detection problem, correlated models from pre- and post-damage are compared. Discrepancies between the two models are considered to be due to damage. In model correlation, the model of a structure is adjusted until the response of the model agrees with the test data.; Currently in finite-element analysis, there exists two types of models: a stress analysis and a dynamic analysis model. In this work, a damage detection model is proposed. A study performed using experimental bridge data indicates that the damage detection model enhances the ability to detect damage in comparison to the other two models.; Component Mode Synthesis (CMS) is a common method of reducing the number of degrees of freedom of a large structural model. In this work, an effective procedure for using direct matrix updating procedures of a CMS model has been developed. The proposed method defines testing configuration requirements for the development of correlated CMS models.; Model refinement and damage detection for structures that are loaded by a fluid have, in the past, been addressed by utilizing data gathered from the structure without the fluid. A damage detection method has been extended to utilize "wet" and/or "dry" data. An important conclusion is that damage detection is still possible when only wet data is available.; Finally, the use of acoustical measurements to detect damage is explored. A structural Frequency Response Function (FRF) is estimated from pressure measurements made some distance from the structure using Nearfield Acoustic Holography. The estimated structural FRF is used to locate damage. Studies performed with analytical data show that the estimated FRF can be utilized to accurately locate damage when the measurements are made close to the structure.