Research Article
The Cable Fault Diagnosis for XLPE Cable Based on 1DCNNs-BiLSTM Network
Table 2
The structural parameters of the 1DCNN-BiLSTM network.
| Layers | Parameters | Data shape |
| Input | | (8004, 1) |
| Conv1 | Fitters = 16, kernel size = 3, stride = 1, activation = “tanh” | (8002, 16) | Conv2 | (8000, 16) |
| MaxPooling1 | Pool size = 3 | (2666, 16) |
| Conv3 | Fitters = 32, kernel size = 3, stride = 1, activation = “tanh” | (2662, 32) | Conv4 | (2660, 32) |
| MaxPooling2 | Pool size = 3 | (888, 3) |
| Conv5 | Fitters = 64, kernel size = 3, stride = 1, activation = “tanh” | (886, 64) |
| MaxPooling1 | Pool size = 3 | (295, 64) |
| BiLSTM1 | Hidden = 128, activation = “tanh” | (256) | BiLSTM2 | (256) |
| Softmax | Classifier | (4) |
|
|