Research Article
Deep Learning Enabled Fault Diagnosis Using Time-Frequency Image Analysis of Rolling Element Bearings
Table 19
Prediction accuracies for 96
96 image inputs.
| | Model | Spectrogram | Scalogram | HHT |
| | MLP flat | 96.7% | 91.7% | 68.0% | | LSVM flat | 95.4% | 84.4% | 71.4% | | SVM flat | 98.7% | 92.1% | 69.0% | | MLP PCA | 96.3% | 97.6% | 85.0% | | LSVM PCA | 87.1% | 74.5% | 65.4% | | SVM PCA | 28.6% | 84.4% | 93.1% | | Architecture 2 | 96.0% | 96.0% | 79.5% | | Architecture 1 | 99.7% | 99.8% | 97.4% | | Proposed CNN architecture | 99.5% | 99.5% | 97.6% |
|
|
