FCNet: Deep Learning Architecture for Classification of Fatigue and Corrosion Acoustic Emission Signal

The recent advancement in semiconductor and computing technology has empowered the field of structural health monitoring in many ways. This work introduces a deep learning-based architecture, ‘FCNet,’ to predict acoustic emission signals arising from deformations like corrosion and fatigue crack. The suggested model uses a lightweight framework that takes advantage of the convolutional neural networks to demonstrate the implicit ability of feature identification, which removes the time-consuming stages of feature selection and extraction. The model’s performance was proved using a dataset of 8566 corrosion and fatigue acoustic emission signals. To identify corrosion and fatigue acoustic emission signals, the model attained a 99.7 percent accuracy, demonstrating the efficacy of the suggested model for real-time reliability. The importance of this research for the industry is that it will provide a lethal approach for identifying metal deformation causes and, as a result, reducing accidents.