Structural Health Monitoring of Composite Structures using Artificial Intelligence Protocols

This study discusses a structural health monitoring (SHM) system developed to detect the presence of delamination, and predict its location and size in a composite structure. Two structures are considered in this study: a composite beam and a T-joint structure used in ships. Finite element (FE) models of these structures are created, embedded with delaminations, and the strain distribution along the bond-line and surface of the structures is used as a damage characteristic, to get information about the structures' condition. Experimental tests are then conducted to verify the FE model, an excellent corroboration is achieved between the two. Artificial neural networks is then used in tandem with a pre-processing program developed, called the damage relativity assessment technique (DRAT), to determine the presence of the damage and then predict its size and location. This SHM system developed is completely independent of the structures' loading condition and it detected the presence, and predicted the size and location of delaminations with an acceptable level of accuracy.