Anchor-Damage Event Recognition Based on FBG sensors and CNN-BiLSTM

The safety monitoring of submarine cables is of increasing concern in the research community. In this study, Fiber Bragg Grating (FBG) strain sensors and FBG accelerometer were used to monitor the health status of cables. A method for identifying and localising submarine-cable risk events based on a multi-input Convolutional Neural Networks-Bi-directional Long Short-Term Memory (CNN-BiLSTM) neural network was proposed. The training results of CNN BiLSTM network showed that the total recognition accuracy of the method for eight types of events can reach 97.68%. The results of this study should enrich and improve the theory and methods for identifying and locating submarine-cable risk events and provide support for the development of submarine-cable safety and health-monitoring technology.