Real-time fatigue crack prediction using self-sensing buckypaper and gated recurrent unit

Aircraft is regarded as a collection of modern technologies from throughout all industries. However, it is inevitable to develop defects during its service life. In general, the aircraft has a periodic maintenance period, and is inspected according to a well-established process, for example non-destructive testing. However, the maintenance requires massive time and cost. If an unexpected defect occurs due to external environments before the maintenance cycle returns, it is impossible to prevent subsequent damage. This study proposes a novel real-time fatigue crack prediction method using self-sensing carbon nanotube buckypaper and deep learning algorithm. Carbon nanotube buckypaper was fabricated by the wet method. The physics-informed gated recurrent unit was used to predict real time crack growth. The physics-informed deep learning model accurately predicted the fatigue crack length. The results showed that the proposed method is promising in detecting the real-time fatigue crack growth of aircraft structure.