|
|
Publications
Li. H., Peng, W.W., Adumene, S. and Yazdi, M. (2023), Using global average pooling convolutional siamese networks for fault diagnosis of planetary gearboxes, Intelligent Reliability and Maintainability of Energy Infrastructure Assets, H. Li, W.W. Peng, S. Adumene & M. Yazdi (Eds.), Springer, Switzerland, pp. 73-91.
Planetary gearboxes have been maturely applied to a broad of scenarios in industry, and fault diagnosis is of great need for their design issues and operation and maintenance activities. This chapter introduces a Global Average Pooling-based Convolutional Siamese Network (GAPCSN) for the fault diagnosis of planetary gearboxes, which can cope with limited data situations. Initially, a convolutional layer with a wide convolutional kernel is used in the feature extraction module to improve the feature extraction capability of the model. Subsequently, a maximum pooling layer and a global average pooling layer are designed for the dimensionality reduction of extracted features and reducing the number of parameters of the network model. Euclidean distance is applied to quantify feature vectors to improve the classification capability of the model. The fault diagnoses of planetary gearboxes is carried out to validate the proposed model carry out the model validation, and the results show that GAPCSN performs better than other existing models in fault diagnosis of planetary gearboxes under limited data. Overall, the introduced fault diagnoses model, GAPCSN, contributes to condition-based maintenance and predictive maintenance of complex systems such as planetary gearboxes in energy infrastructure assets.
If you did not manage to obtain a copy of this paper: Request a copy of this article
For information about all CENTEC publications you can download: Download the Complete List of CENTEC Publications
|