DC-NNMN: Across Components Fault Diagnosis Based on Deep Few-Shot Learning

Table 3

Description of the training datasets.

Frequency (kHz)	Position	Fault categories (total numbers)	Fault size (mils)	Load (hp)	Number
12	FE	BF, IF, OF@3, OF@6, and OF@12 (5)	0.007	0	90
12	FE	BF, IF, OF@3, and OF@6 (4)	0.014	0	90
12	FE	BF, IF, and OF@6 (3)	0.021	0	90
12	FE	BF, IF, OF@3, OF@6, and OF@12 (5)	0.007	3	90
12	FE	BF, IF, and OF@3 (3)	0.014	3	90
12	FE	BF, IF, and OF@3 (3)	0.021	3	90
12	DE	BF, IF, OF@3, OF@6, and OF@12 (5)	0.007	0	90
12	DE	BF, IF, and OF@6 (3)	0.014	0	90
12	DE	BF, IF, OF@3, OF@6, and OF@12 (5)	0.021	0	90
12	DE	BF and IF (2)	0.028	0	90
12	DE	BF, IF, OF@3, OF@6, and OF@12 (5)	0.007	3	90
12	DE	BF, IF, and OF@6 (3)	0.014	3	90
12	DE	BF, IF, OF@3, OF@6, and OF@12 (5)	0.021	3	90
12	DE	BF and IF (2)	0.028	3	90
48	DE	BF, IF, OF@3, OF@6, and OF@12 (5)	0.007	0	90
48	DE	BF, IF, and OF@6 (3)	0.014	0	90
48	DE	BF, IF, OF@3, OF@6, and OF@12 (5)	0.021	0	90
48	DE	BF, IF, OF@3, OF@6, and OF@12 (5)	0.007	3	90
48	DE	BF, IF, and OF@6 (3)	0.014	3	90
48	DE	BF, IF, OF@3, OF@6, and OF@12 (5)	0.021	3	90
12	DE	N (1)	—	0	90