Development of clustering and classification software system for structure health monitoring
Structural health monitoring (SHM) is a process to detect damage(s) in engineering structures and to determine the damage locations and types. Typically, damages in a structure are detected by using measurements collected from sensors built for that purpose. This research is multidisciplinary in nature aimed towards improving some of the Air Force system performances by employing; signal processing techniques, advance clustering and classification methods and, decision fusion theories. The main objective of the work was to develop SHM software system for real-time monitoring of aircraft structures. The research utilized experimental data obtained from experiments conducted on aluminum structure fixed with bolts. The experiment consisted of four bolts used to fix the aluminum plate, one piezoelectric actuator to excite the structure and three piezoelectric sensors to collect vibration data from the structure. First, advanced signal processing technique was used to extract frequency features from the vibration data. The linear predictive coefficient (LPC) transformation was used to generate unique set of features that can be used for clustering and classification of the status of the structure's health. Second, a fuzzy c-mean clustering algorithm was used to group the extracted sets of features into homogeneous classes of similar features. Third, neural network was used as the decision fusion technique to: resolve any conflict, reduce the level of uncertainty, and produce trustful decision with high level of confidence in the decision made by the clustering and classification algorithm. Finally, the different process algorithms were integrated into one operational software system that can perform structural health monitoring in real-time. Testing and evaluation results of the developed SHM system will be reported and presented in this thesis.
Abdalla Hasan Al-Salah,
"Development of clustering and classification software system for structure health monitoring"
ETD Collection for Tennessee State University.