Research Article
A Novel Method for Diagnosis of Bearing Fault Using Hierarchical Multitasks Convolutional Neural Networks
Table 3
Comparison of diagnosis accuracy with other models in different noise environments.
| Methods | SNR | ā4 | ā2 | 0 | 2 | 4 | 6 | 8 | 10 |
| SVM | 0.663 | 0.724 | 0.735 | 0.761 | 0.812 | 0.822 | 0.823 | 0.826 | BPNN | 0.601 | 0.623 | 0.656 | 0.683 | 0.702 | 0.832 | 0.845 | 0.854 | CNN | 0.651 | 0.687 | 0.787 | 0.811 | 0.832 | 0.856 | 0.864 | 0.886 | LSTM | 0.652 | 0.663 | 0.752 | 0.787 | 0.822 | 0.834 | 0.850 | 0.852 | HMCNN | 0.845 | 0.916 | 0.932 | 0.950 | 0.972 | 0.983 | 0.985 | 0.991 |
|
|