Depth-wise separable convolutional neural-network-based intelligent chatter  monitoring for thin-walled polish grinding

Zhao, Yuan; Zhu, Chunxia; Xu, Guofa

doi:https://doi.org/10.5194/ms-16-615-2025

Articles | Volume 16, issue 2

https://doi.org/10.5194/ms-16-615-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/ms-16-615-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 16, issue 2

Research article

|

22 Oct 2025

Research article |

| 22 Oct 2025

Depth-wise separable convolutional neural-network-based intelligent chatter monitoring for thin-walled polish grinding

Yuan Zhao, Chunxia Zhu, and Guofa Xu

Cited articles

Arnold, N. R.: The mechanism of tool vibration in the cutting of steel, ARCHIVE Proceedings of the Institution of Mechanical Engineers 1847–1982, 154, 261–284, 1946.

Altintas, Y.: Analytical Prediction of Three Dimensional Chatter Stability in Milling, JSME International Journal Series C Mechanical Systems, Machine Elements and Manufacturing, 44, 717–723, https://doi.org/10.1299/jsmec.44.717, 2002.

Altintaş, Y. and Budak, E.: Analytical Prediction of Stability Lobes in Milling, CIRP Annals, 44, 357–362, https://doi.org/10.1016/S0007-8506(07)62342-7, 1995.

Altintas, Y. and Weck, M.: Chatter Stability of Metal Cutting and Grinding, CIRP Annals, 53, Arnold and N, R.: The mechanism of tool vibration in the cutting of steel, ARCHIVE Proceeding, 619–642, https://doi.org/10.1016/S0007-8506(07)60032-8, 2004.

Cao, H. R., Lei, Y. G., and He, Z. J.: Chatter identification in end milling process using wavelet packets and Hilbert-Huang transform, International Journal of Machine Tools & Manufacture, 69, 11–19, https://doi.org/10.1016/j.ijmachtools.2013.02.007, 2013.

More articles (20)

Short summary

To enhance chatter monitoring in polishing and grinding, we propose a model fusing deep separable convolutional neural networks (DCNN) and Gated Recurrent Unit. Signals are preprocessed via optimized variational mode decomposition (VMD) and wavelet denoising. The squeeze excitation deep separable convolutional neural networks gated recurrent units (SE-DCNN-GRU) model employs depth-wise convolution for multi-scale feature extraction. Experiments achieve 98.8 % accuracy for titanium alloy parts, offering a robust solution for complex conditions.