Table 3 Structural parameters of the 1D-CNN
From: Fault Diagnosis Method for Rotating Machinery Based on Multi-scale Features
Layer | Kernel / Stride | Output |
|---|---|---|
Convolutional layer 1 | 6 × 1 × 8/2 × 1 | 1024 × 8 |
Pooling layer 1 | 2 × 1 | 512 × 8 |
Convolutional layer 2 | 6 × 1 × 16/2 × 1 | 256 × 16 |
Pooling layer 2 | 2 × 1 | 128 × 16 |
Convolutional layer 3 | 3 × 1 × 16/1 × 1 | 128 × 16 |
Pooling layer 3 | 2 × 1 | 64 × 16 |
Convolutional layer 4 | 3 × 1 × 32/1 × 1 | 64 × 32 |
Pooling layer 4 | 2 × 1 | 32 × 32 |
Convolutional layer 5 | 3 × 1 × 32/1 × 1 | 32 × 32 |
Pooling layer 5 | 2 × 1 | 16 × 32 |
Convolutional layer 6 | 3 × 1 × 64/1 × 1 | 16 × 64 |
Pooling layer 6 | 2 × 1 | 8 × 64 |
Convolutional layer 7 | 3 × 1 × 128/1 × 1 | 6 × 128 |
Pooling layer 7 | 2 × 1 | 3 × 128 |