Shock and Vibration

Research Article

Compound Fault Diagnosis for Gearbox Based Using of Euclidean Matrix Sample Entropy and One-Dimensional Convolutional Neural Network

Table 4

Parameters of CNN based on LMD.
Layer (tyPF)Output (ShaPF)Param
ReshaPF_1 (ReshaPF)(None, 6656, 1)0
Conv1_1 (Conv1D)(None, 3325, 16)144
Conv1_2 (Conv1D)(None, 1663, 16)2064
Conv1_3 (Conv1D)(None, 832, 16)2064
Max_pooling1d_1n (Maxpooling1)(None, 208, 16)0
Conv1_4 (Conv1D)(None, 104, 64)4160
Conv1_5 (Conv1D)(None, 52, 64)16448
Max_pooling1d_2 (Maxpooling1)(None, 26, 64)0
Conv1_6 (Conv1D)(None, 13, 256)65792
Conv1_7 (Conv1D)(None, 7, 256)262400
Max_pooling1d_3 (Maxpooling1)(None, 3, 256)0
Conv1_8 (Conv1D)(None, 3, 512)262656
Conv1_9 (Conv1D)(None, 3, 512)524800
Max_pooling1d_4 (Maxpooling1)(None, 1, 512)0
Global_average_pooling1d_1(None, 512)0
Dropout_1 (dropout)(None, 512)0
Dense_1 (dense)(None, 6)3078
Total params: 1143606. Trainable params: 1143606. Nontrainable params: 0.