| The crucial requirement for up-to-date knowledge of structural integrity of in-service structures has prompted the development of many structural health monitoring strategies. In such strategies, critical data concerning the health state of the structure is recorded on a continuous or periodic basis through a sensoring system. The data is then processed and interpreted using certain algorithms in order to detect anomalies and damages within the structure.;The integrity of the proposed health monitoring technique and its effectiveness will be demonstrated through a series of experimental and numerical case studies. These include the identification of single and multiple corrosion areas in a pipeline, detection of cracks and qualification of their severity in a pipeline's girth weld, as well as detection of cracks of various depths and at different locations along a cantilever beam. The results will prove the applicability of the proposed methodology as a robust and effective tool for structural health monitoring purposes.;The effects of various parameters such as boundary conditions, impact locations and impulse hammer tip stiffness on the predicted results were also experimentally investigated and discussed. Furthermore, some refinements to the methodology are suggested. These include modification to the method of frequency analysis of vibration signals to establish the most sensitive frequency bands to damage, the analysis of different IMFs, and the influence of the instantaneous phase.;The study presented in this dissertation utilizes the Hilbert-Huang Transform (HHT) and presents a novel structural health monitoring technique based on dynamic characteristics of structures. In this technique, the decomposition part of the HHT, known as the Empirical Mode Decomposition (EMD), is employed, by which the free vibration signatures of the structure are decomposed into a collection of oscillatory modes, called the Intrinsic Mode Functions (IMFs). In the developed methodology, the free vibration response of the structure is monitored by piezoceramic sensors or a laser Doppler vibrometer. The deviations in structural integrity, measured from a healthy-state baseline, are reflected through a damage index established based on the energy of the IMFs of the signals collected before and after occurrence of damage. |
