Table 8 Performance comparion for training bearing 1_1 using CNN with the different number of cells.

From: Light convolutional neural network by neural architecture search and model pruning for bearing fault diagnosis and remaining useful life prediction

Cell count

Test loss

R-square

Predicted RUL (\(\times\)10 s)

Actual RUL (\(\times\)10 s)

RUL error percentage

Training time (s)

2

0.117

0.995

3000

2793

7.40%

168

3

0.229

0.991

2642

2793

5.40%

176

4

0.264

0.989

2660

2793

4.80%

190

5

0.503

0.980

2792

2793

0.04%

193

6

0.185

0.993

2752

2793

1.47%

212

Back to article page