基于空洞卷积和增强型多尺度特征自适应融合的滚动轴承故障诊断

韩康,战洪飞,余军合,王瑞

Rolling bearing fault diagnosis based on dilated convolution and enhanced multi-scale feature adaptive fusion

Kang HAN,Hongfei ZHAN,Junhe YU,Rui WANG

表 3 12k驱动端轴承数据在不同噪声下8种模型的准确率

Tab.3 Accuracy rates of 8 models under different noises for 12k driving end bearing data

模 型 A SNR=4 SNR=6 SNR=8 SNR=10 SNR=12 MSCNN-LSTM[14] 0.6646±0.0145 0.8195±0.0148 0.8892±0.0300 0.9236±0.0333 0.9706±0.0177 WDCNN[15] 0.5968±0.0106 0.7578±0.0364 0.8665±0.0171 0.9499±0.0094 0.9701±0.0172 MSCNN[16] 0.7605±0.0352 0.8549±0.0443 0.9195±0.0226 0.9493±0.0251 0.9774±0.0104 DC 0.6346±0.0214 0.8108±0.0158 0.8900±0.0139 0.9410±0.0110 0.9754±0.0091 Resnet[17] 0.7485±0.0111 0.8654±0.0096 0.9336±0.0051 0.9703±0.0064 0.9904±0.0017 DRSN-CS[18] 0.7928±0.0151 0.8665±0.0193 0.9423±0.0019 0.9755±0.0107 0.9875±0.0059 MA1DCNN[19] 0.8019±0.0374 0.8744±0.0229 0.9076±0.0140 0.9537±0.0232 0.9836±0.0122 DC-MAFFM 0.8808±0.0009 0.9346±0.0073 0.9639±0.0091 0.9848±0.0065 0.9992±0.0010