Monitoring and prediction of fatigue crack growth in S690 steel using acoustic emission and DPGMM clustering

Fatigue damage is a significant concern for fixed offshore platforms and there is growing interest in using structural health monitoring (SHM) technology for continuous monitoring and evaluation of fatigue damage through the use of sensors permanently attached to the structures. In this study, the acoustic emission (AE) technique is employed to monitor the fatigue crack growth of S690 steel, which is commonly used in offshore platforms. The relationship between the temporal evolution of AE signals and crack length is established. The Dirichlet process Gaussian mixture model (DPGMM) is used to assign AE signals into a dynamically sized set of clusters. Based on the clusters corresponding to the crack growth, a mathematical relationship between fatigue crack growth rate (FCGR) and AE absolute energy rate (dE/dN) is developed to predict crack extension. The outcomes provide a new approach for fatigue crack length estimation of offshore platforms based on the AE technique.