Development of Deep Convolutional Neural Network with Adaptive Batch Normalization Algorithm for Bearing Fault Diagnosis

Table 4

Recognition rate of the ABN-CNN model under noise environment.

Convolution kernel size	SNR (dB)
	−4 (%)	−2 (%)	0 (%)	2 (%)	4 (%)	6 (%)	8 (%)	10 (%)
16	81.87	90.34	95.57	98.45	99.04	99.45	99.77	99.74
24	87.21	93.96	97.28	99.08	99.66	99.82	99.85	99.90
32	89.83	95.16	97.96	99.23	99.52	99.65	99.73	99.88
40	90.92	95.97	98.25	99.33	99.54	99.77	99.80	99.90
48	91.69	96.39	98.43	99.47	99.70	99.81	99.83	99.85
56	92.68	96.52	98.65	99.37	99.76	99.73	99.85	99.88
64	92.56	96.83	98.81	99.56	99.65	99.87	99.88	99.95
72	92.36	96.33	98.55	99.23	99.60	99.79	99.80	99.87
80	92.35	96.80	98.63	99.44	99.61	99.76	99.89	99.83
88	92.63	97.06	98.79	99.52	99.63	99.73	99.87	99.80
96	92.65	97.10	98.79	99.61	99.62	99.82	99.84	99.81
104	92.45	96.59	98.59	99.53	99.62	99.75	99.80	99.90
112	91.84	96.21	98.64	99.42	99.69	99.75	99.90	99.90
120	92.13	96.38	98.73	99.45	99.63	99.72	99.85	99.86
128	91.90	96.62	98.60	99.44	99.55	99.71	99.84	99.84
The average	92.68	97.10	98.79	99.61	99.76	99.87	99.90	99.95