Tool Wear Monitoring with Artificial Intelligence Methods: A Review
Round 1
Reviewer 1 Report
The 32-page manuscript is a review on the use of artificial intelligence for tool wear monitoring. While interesting, it fails to provide hindsight on the subject. The reviewer recommends rejecting the paper for now and possibly resubmitting it after a deep revision.
The paper's contents are relevant but fail to provide an in-depth analysis of the identified papers. Instead, it mainly gathers results and methodologies. Some meta-statistics and graphs presenting the proportions of use of each input variable or IA model would be appreciated. If, as the abstract mentions, "the purpose of this research work is to provide an overview..." the authors need to go beyond stating the results of each paper and provide an in-depth, critical analysis of the current trends.
To improve the paper in this regard, the authors could:
- move Tables 6 and 7 to appendices, as they mostly show raw data: instead, provide in-depth analysis and hindsight on the proportion of each variable and technique.
- provide graphs showing the proportions of each technique or input variable over time
- reduce the proportion of the article dedicated to presenting the monitoring variable since this matter has already been extensively covered in the literature (see, for example, Siddhpura, A., Paurobally, R. A review of flank wear prediction methods for tool condition monitoring in a turning process. Int J Adv Manuf Technol 65, 371–393 (2013). https://doi.org/10.1007/s00170-012-4177-1, which the authors cite)
- likewise, the AI techniques used in tool monitoring are extensively covered in the literature, and their description can be reduced.
- the space gained by reducing these descriptions can be used to provide actual and critical analyses of the identified papers' results.
- update the review, given that the review ends in 2020 and significant new results have been published in the last few years.
- use most of the current conclusion as the results, then add a discussion section that compares the results with recent reviews.
The paper fails to discuss the results of their review in light of other reviews on the matter and thus does not discuss the added value of the study compared to previous reviews. Recent examples of such reviews are:
- Pimenov et al. (2022) Artificial intelligence systems for tool condition monitoring in machining: analysis and critical review. Journal of Intelligent Manufacturing https://doi.org/10.1007/s10845-022-01923-2 (which the authors cite but do not comment in comparison to their results)
- Colantonio et al. (2021) A Systematic Literature Review of Cutting Tool Wear Monitoring in Turning by Using Artificial Intelligence Techniques. Machines https://doi.org/10.3390/machines9120351
- Serin et al. (2020) Review of tool condition monitoring in machining and opportunities for deep learning. Int. J. Adv. Manuf. Technol.
- Mohanraj et al. (2020) Tool condition monitoring techniques in milling process—A review. J. Mater. Res. Technol.
To improve the paper in this regard, the authors could identify previous reviews, clarify their contribution compared to these other reviews, and compare and discuss their results.
In addition, here are further questions and remarks on the contents:
- In the introduction, the study is justified, but no information is provided on what makes the quality of a workpiece (surface roughness, residual stresses, dimensional accuracy?) and how it relates to the wear of the cutting tool. This relationship is essential to understanding the need for such a review, yet it is subject to some debate in the literature.
- There is little justification for the search strategy. From Table 1 and Section 2.1, it seems like the research question is initially broad, then refined until the number of articles falls within an acceptable number, contrary to the PRISMA-P concept, which suggests that the objectives and research question are identified first.
- Please identify which databases were used for the research (Google Scholar, Scopus, Web of Science, other?).
- Consider distinguishing between tool state forecasting (estimate of the tool state in a future time, e.g., RUL estimate) and tool state estimation (estimate of the tool state at the current time). Make sure, throughout the paper, to avoid confusing measurement and estimation. In this framework, at line 251, it is said that Figure 3 shows "the steps to obtain online measurement of tool wear": it is not a measurement but rather an estimate or a prediction.
- Please provide summarizing tables or graphs for the performance of AI algorithms. Section 4.5.4.1. is too full of numbers to make sense to the average reader.
- It is unclear whether optical measurements are considered in the review. In the fifth research level (Table 1), it is written that optical measurements are not considered, yet they are mentioned at lines 166, 194, and 220.
The paper should be carefully proofread.
There are just plain sentences written in Italian. E.g., lines 740-742.
Some words are likely not the ones the authors intended to choose. For example, in line 18, "revision" likely should be "review". There are other examples throughout the paper.
Some phrases make little sense and should be rephrased for clarity. E.g., lines 67-88, 122-124, 130-139, 192-195, etc.
Some words are unnecessarily capitalized. For example, the title, and the words "online" and "offline" in lines 113-115.
Author Response
Please see the attachment author-coverletter-2949466.v1.docx
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This paper provides an overview of the development of tool wear monitoring through AI in the past decade. This work has been well presented and has a great attraction for researchers in the field of mechanical manufacturing. The following questions should be revised before future publication.
1. Abstract: the abbreviation of PRISMA-P should be defined for the first time of this paper. The description of “provide an overview of the last 10 years” is not accurate enough because some of the publications list in the literature are not ten years ago.
2. Introduction: What is the purpose of writing this review paper? What is the difference from other similar review papers? It is not enough to simply list the papers related to AI inspection of tool wear in the past decade. Besides, paragraph writing is too scattered, it is recommended to merge.
3. Some lately researches on tool wear control are missed. For instance, two papers on tool wear control in drilling and milling by ultrasonic vibration (https://doi.org/10.1016/j.cja.2022.12.009, https://doi.org/10.1016/j.jmrt.2022.12.054) also mentioned the topic of “premature failure or breakage of one or more tools leads to material loss, machine downtime, which affects products productivity, quality and lead-time”. Both of them can be added to enrich your literature review.
4. Some annotation can be added to figure 3: Online measurement scheme.
5. Intelligent machining has unique manufacturing advantages in tool wear monitoring. What are the bottlenecks in industrial applications? Why are there no large-scale applications now? Please analyze and summarize in detail.
6. Improve the conclusion. It should be shortened. Besides, the novel findings can be further refined.
The quality of the language is ok. In general, the language could be refined carefully with the help of a native English speaker.
Author Response
Please see the attachment author-coverletter-29451302.v1.docx
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The authors produced a revised version that takes into account most comments from the reviewer. Overall, the improvements are good and make the paper more original and better prepared for publication. A few final comments are:
- The analysis of the relative importance of variables and techniques is relevant and contributes positively to the originality of the paper.
- The comparison made with recent reviews' findings in the conclusions is insightful and brings to the quality of the paper.
- The quality of figures 4-23 could be perfected by having an even, white background matching the background of the paper. It is unclear why they sport a red square on the bottom left corner.
Overall, the Authors make clear responses to the reviewers' remarks and valid points to comments that they chose to overturn.
There is still plain text in Italian at lines 577-586. Here and there, the English language still needs more proofreading, but the general level is now acceptable.
Author Response
Please see the attachment author-coverletter-30482691.v1
Author Response File: Author Response.pdf
