Machine Learning in CNC Machining: Best Practices

Round 1

Reviewer 1 Report

This paper illustrates five best practices and learning, gained from building an ML system to detect tool wear in metal CNC machining. Here are some suggestions for improving the quality of this paper:

1) In the Abstract, it is unclear what are the practical values of this paper. Please revise the Abstract to highlight the main contribution of this work.

2) Fig. 1 can be improved. It is not a standard procedure of a machine learning algorithm.

3) It is suggested to introduce the types and mechanisms of wear. At the current stage, there is no unitive standard for wear classification. The authors can refer to a review of vibration-based gear wear monitoring and prediction techniques & fault mechanism and dynamic modeling of planetary gear with gear wear & modern tribology handbook.

4) Some recommendations for future work can be given at the end of the Conclusion.

Author Response

Please see attached Word document.

Author Response File: Author Response.pdf

Reviewer 2 Report

 

Improve the quality of the abstract, Content needs to be further refined and enriched.

Research gap and problem definition and/or objectives aren't discussed qualitatively. Should relook into it.

Figure 1 is not clear, please provide the link in between the two modules.

Literature review is very poor. Not done systematically. Would have kept in the Table form. Instead of compiling the reviwed data. Not acceptable.

Provide the information about the flank wear rate cutting and its parameters.

In table number 4, some references are missing, please add the reference.

Provide the X-Y axis references in Figure 8 and 9.

 Not qualitative. What is the take home message? Future scope? Give some point about the future scope of the work.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

In the following I provide a few suggestions on how I think the paper can be further improved:   
- The difficulty of overfitting as a key problem in machine learning should be adressed in more detail.
- It would be good to give some information about the source of the discussed "five best practices". Are they based on own practical experience or theoretical considerations?
- The title could be adapted to include the objective of tool wear detection by using Machine Learning methods. With the current title, we would expect a comprehensive paper on various application scenarios. Also, in the abstract, this objective could be highlighted a little more  

Some minor changes:
- Figure 8: Please write out the abbreviations for the different algorithms used from the scikit library in full
- Line 33: "is beneficial to all engaged" --> gramatically wrong
- Reference 30: "Proceedings of the Proceedings..."
- References 32, 35 and 38: "Proceedings of the" is doubled

Author Response

Please see attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

This paper has been improved by addressing the comments from reviewers. I agree to accept it.