Fault diagnosis method for the underwater thruster of the unmanned surface vessel based on DRFSN

The unmanned surface vessel (USV), as a crucial autonomous surface platform, plays a significant role in marine environmental monitoring and resource exploration and development. The underwater thruster serves as a crucial power component for USVs, and its malfunction can lead to system instability or even complete failure. Therefore, high-precision fault diagnosis is urgently needed to ensure the safety and stability of the USV system. However, the complexity of the operational environment results in significant interference within the underwater thruster data, making it challenging to extract data features effectively and hindering fault diagnosis. To tackle this challenge, this paper proposes a fault diagnosis model based on the Deep Residual Fusion Shrinkage Network (DRFSN) for the underwater thruster of USVs. The model integrates a fusion feature module into the deep residual shrinkage network to further exploit multi-scale fused feature information, enabling the effective extraction of latent fault features from the data. Experimental validation using real-world underwater thruster fault data shows that the proposed method achieves a training accuracy of 100% and a testing accuracy of 99.1%, demonstrating its ability to accurately diagnose faults in the underwater thruster of USVs.