Bearing Fault Diagnosis Based on Frequency Subbands Feature Extraction and Multibranch One-Dimension Convolutional Neural Network

Table 4

Details of the dataset.

Fault type	Fault size (mil)	Label	Training sample	Test sample	Sample length	Load (HP)
Ball	7	1	90	25	1024	1, 2, 3
Ball	14	2	90	25	1024	1, 2, 3
Ball	21	3	90	25	1024	1, 2, 3
Inner-race	7	4	90	25	1024	1, 2, 3
Inner-race	14	5	90	25	1024	1, 2, 3
Inner-race	21	6	90	25	1024	1, 2, 3
Outer-race@3	7	7	90	25	1024	1, 2, 3
Outer-race@6	14	8	90	25	1024	1, 2, 3
Outer-race@12	21	9	90	25	1024	1, 2, 3
Normal	-	0	90	25	1024	1, 2, 3