Research and application of marine crane gearbox fault diagnosis based on multispectral attention and EfficientNet algorithm

This study aims to solve the problem of data fusion and feature weight allocation in the fault diagnosis of the gearbox of marine cranes, and to improve the accuracy and efficiency of fault diagnosis through the optimization algorithm, so as to ensure the safe and stable operation of cranes and reduce the safety hazards of shipbuilding enterprises. In this study, the temperature and vibration data of the gearbox of marine crane are firstly collected by temperature and vibration integrated sensors, and the continuous wavelet transform is used to realize multi-feature fusion and generate the fusion time-frequency diagram. Then, the dataset is expanded by ProGAN network to cope with the problem of unbalanced data distribution. On this basis, the multispectral channel attention module and the EfficientNet fusion algorithm are proposed to improve the weight assignment coefficient of key features. In the latter half of the network, the traditional swich activation function is replaced with hard-swich to further improve the detection accuracy and efficiency of the model. Validated by the test set, the improved model proposed in this study has achieved significant results in the fault diagnosis of reduction gearboxes of marine cranes. The recognition accuracy of the model reaches 99.28%, which is a 3.91% improvement compared to the traditional EfficientNet algorithm. This result quantitatively proves the superior performance of the fusion algorithm in gearbox fault diagnosis. This study successfully improves the accuracy and efficiency of gearbox fault diagnosis in marine cranes by adopting a series of innovative methods, such as temperature and vibration integrated sensor, continuous wavelet transform, ProGAN network expansion data set, and multispectral channel attention module with EfficientNet fusion algorithm. This research result is of great significance to ensure the safe and stable operation of cranes and reduce the safety hazards of shipbuilding enterprises, and also provides a useful reference for the fault diagnosis of similar equipment.