基于因果解耦的域自适应滚动轴承故障诊断

黄爱颖,李晓辉,孙淑娴,朱逸群

Causal disentanglement-based domain adaptation for rolling bearing fault diagnosis

Aiying HUANG,Xiaohui LI,Shuxian SUN,Yiqun ZHU

表 2 不同方法在12种迁移任务中的轴承故障诊断准确率

Tab.2 Bearing fault diagnosis accuracy of different methods in twelve transfer tasks

迁移任务 A/% BASE DDC[9] CORAL[10] DANN[11] CDAN[13] SFDA[14] 本研究 [1]-[2] 96.53 ± 0.12 95.71 ± 0.41 94.52 ± 0.51 96.43 ± 0.23 97.59 ± 0.13 96.83 ± 0.79 98.63 ± 0.07 [1]-[3] 92.84 ± 0.49 95.48 ± 0.25 92.34 ± 1.10 96.24 ± 0.53 99.17 ± 0.17 98.63 ± 0.84 99.39 ± 0.13 [1]-[4] 88.91 ± 0.93 92.19 ± 0.58 91.56 ± 2.08 97.82 ± 0.36 98.32 ± 0.21 96.25 ± 0.66 98.69 ± 0.19 [2]-[1] 98.82 ± 0.07 98.96 ± 0.43 98.33 ± 0.24 98.65 ± 0.20 99.10 ± 0.26 98.25 ± 0.45 99.89 ± 0.06 [2]-[3] 98.56 ± 0.21 97.79 ± 0.26 98.82 ± 0.09 99.37 ± 0.18 99.53 ± 0.06 97.22 ± 0.76 99.64 ± 0.09 [2]-[4] 92.67 ± 0.18 93.48 ± 0.91 94.72 ± 1.61 99.15 ± 0.32 98.97 ± 0.21 98.62 ± 0.47 99.23 ± 0.05 [3]-[1] 96.31 ± 0.42 96.99 ± 0.72 97.53 ± 0.14 92.85 ± 0.71 99.12 ± 0.13 96.09 ± 0.77 99.15 ± 0.08 [3]-[2] 97.15 ± 0.29 98.51 ± 0.44 99.49 ± 0.11 95.47 ± 0.93 98.87 ± 0.22 97.53 ± 0.49 99.92 ± 0.03 [3]-[4] 98.88 ± 0.06 99.22 ± 0.34 98.28 ± 0.07 99.64 ± 0.09 99.95 ± 0.13 98.85 ± 0.39 100.0 ± 0.00 [4]-[1] 81.09 ± 1.13 84.35 ± 0.81 88.37 ± 0.13 88.34 ± 0.45 92.23 ± 0.18 95.62 ± 0.42 97.42 ± 0.12 [4]-[2] 84.56 ± 0.75 89.61 ± 0.74 90.43 ± 0.38 87.48 ± 0.73 91.43 ± 0.61 93.26 ± 0.72 95.34 ± 0.21 [4]-[3] 95.93 ± 0.53 96.49 ± 0.63 97.03 ± 0.16 97.42 ± 0.16 98.81 ± 0.09 97.31 ± 0.59 98.82 ± 0.16