Research Article
Remaining Useful Life Prediction of Milling Tool Based on Pyramid CNN
Table 5
Performance estimation result of four different networks.
| Testing datasets | Network-1 | Network-2 | Network-3 | PCNN |
| C1 | Score | 40.152 ± 7.486 | 42.934 ± 5.063 | 47.013 ± 4.828 | 53.962 ± 1.813 | RMSE | 29.407 ± 1.712 | 25.122 ± 1.592 | 23.114 ± 1.581 | 19.374 ± 0.923 |
| C2 | Score | 40.274 ± 7.397 | 43.771 ± 4.811 | 48.167 ± 4.765 | 52.217 ± 1.872 | RMSE | 28.903 ± 1.664 | 24.913 ± 1.428 | 23.022 ± 1.412 | 19.139 ± 0.853 |
| C3 | Score | 39.914 ± 8.152 | 43.912 ± 4.702 | 48.369 ± 4.105 | 52.771 ± 1.617 | RMSE | 29.729 ± 1.677 | 24.502 ± 1.437 | 23.216 ± 1.241 | 19.424 ± 0.721 |
| C4 | Score | 40.889 ± 7.109 | 44.068 ± 4.155 | 48.154 ± 3.907 | 54.248 ± 1.712 | RMSE | 28.597 ± 1.402 | 23.662 ± 1.339 | 22.901 ± 1.209 | 19.051 ± 0.804 |
|
|
The bold values express that the PCNN has the best performance.
|