Abstract | ||
---|---|---|
In order to solve a series of problems such as complex structure and low training efficiency in traditional deep learning, a fault diagnosis method of rolling bearing based on graph convolution neural network is proposed. Firstly, the convolution layer of neural network is constructed based on graph convolution, and the first-order ChebNet is used to optimize the network model, so as to improve the operation efficiency of the model. Secondly, aggregate the convoluted node information of each layer, and add the features of each layer as the global features of the original graph to achieve effective and accurate feature extraction. Compared with the traditional neural network, the proposed method significantly reduces the complexity and computing time and the network model can still maintain high accuracy when using unbalanced data sets. Through comparative experiments, it is proved that the model has strong feature extraction ability and higher training efficiency, and can still perform well in dealing with the data set with unbalanced sample. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1007/978-3-031-13870-6_16 | INTELLIGENT COMPUTING THEORIES AND APPLICATION (ICIC 2022), PT I |
Keywords | DocType | Volume |
Fault diagnosis, Graph convolution, ChebNet, Rolling bearing, Deep leaning, Unbalanced sample | Conference | 13393 |
ISSN | Citations | PageRank |
0302-9743 | 0 | 0.34 |
References | Authors | |
0 | 2 |