Machine Learning Prediction of Turning Precision Using Optimized XGBoost Model
Round 1
Reviewer 1 Report
The paper is interesting, the illustrations are well selected.
However, I believe that some information should be added.
I ask the Authors to critically address the following issues:
- what cutting tool was used during the tests? (what was the cutting insert edge radius)
- whether each test was made with a new cutting insert edge?
- how were the cutting parameters selected?
- how was the error measured (precision error in Table 3); how were the error values determined with an accuracy of 6 decimal points?
Author Response
Please check the attached file! Thank you so much!
Author Response File: 
Author Response.pdf
Reviewer 2 Report
In this manuscript, the authors have predicted the machining performance at different parameters using machine learning approach. They have found XGBoost is best ML algorithm to predict the performance of turning operation. novelty is not shown in this work. However, a few comments should be taken into account:
1- There are some grammatical and typo errors. Please check the whole manuscript. For instance, “An excessive force may cause vibration of the cutter, which not accelerates the tool wear, but may also seriously degrade the machining quality.”
2- The statement in lines 108 to 112 ( Yet, the literature contains....) is not right as there are several studies on turning operation performance.
3- The novelty of the work is not shown in the abstract and introduction part. Authors can improve the introduction by showing research gaps and the novelty of the work.
4- Why authors are using Taguchi DOE for machine learning. In machine learning algorithms, more and more data at different parameters will be good for prediction?
5- For machine learning algorithms, the experimental data authors used are very less. This will not give a good prediction. Authors should do more experiments.
6- Fig 23 is showing same results for all the models. Why authors are sure that SMOGN-CPSO-XGB model and SMOGN-GWO-XGB model yield the best fit?
7- Authors should show what is the final output of the research. What is the parameter predicted with this method?
Author Response
Please check the attached file! Thank you so much!
Author Response File: Author Response.pdf
Reviewer 3 Report
1. It is suggested to remove the word 'Lathe' in the title (also in the entire manuscript) and to leave the word 'Turning' alone instead since you are mentioning about the process.
2. In my opinion as a reader from machining background, Table 1 seems unnecessary.
3. It is best to add more references in the introduction part (another 5 maybe if any) to the extent in giving readers a wider perspective on the evolution of predictive model in having precise cutting with lathe machine.
4. Figure 11 and 12 can be improved (to make them clear and visible) if they have similar font as Figure 10.
5. It is suggested that the conclusion to be more extensive. You may express those concluding remarks in form on point rather than lump them all in a single paragraph.
Author Response
Please check the attached file! Thank you so much!
Author Response File: Author Response.pdf
Reviewer 4 Report
The paper presents a commendable approach with regards of turning process optimization. However, there are some issues (general and particular) which, in my opinion have to be addressed.
Some general observations:
The description part of the machine tool and the cutting process is very clumsy, as if the authors were not specialists in this field. I suggest they seek the help of a specialist to improve this fact.
The usefulness of the proposed method is not sufficiently argued. Turning cutting regimes can be selected relatively easily from empirical tables and / or from the recommendations of cutting tool manufacturers.
The non-optimized cutting regime does not represent the determining influence on the machining precision, but rather on the roughness of the machined part.
Some particular issues:
Lines 38-39: “One of the most common machines is the CNC lathe, consisting of a high- speed spindle and two ball screws configured perpendicularly to one another and the guide rail.” – the paper will be read by specialists, thus such a simplified (and not so accurate) description of a CNC lathe is unnecessary. Please remove it.
Lines 43-44: “The cutting behavior of lathes is determined by two components: the path and the feed [1].” Please consider “toolpath” instead of “path” and “feed rate” instead of “feed”.
Lines 54-55: “vibration of the cutter, which not accelerates the tool wear” consider replacing it by ”vibration of the cutter, which not ONLY accelerates the tool wear”
Line 138: “precision error” – it a very odd term, please define it and consider replacing it with “machining error” or “dimensional error” or something similar
Length of cantilever cutter – in my experience, there are no studies considering this parameter as important in setting-up the cutting regime for turning operations. Please present some literature references with regards of that or present some arguments backing-up the assumption that this particular parameter should be taken into consideration.
Table 3: How was the so-called “precision error” measured? What dimension of the part was affected by this error? Is it the diameter of the part or the machined length on Z-axis?
The conclusion section is very brief and does not highlight the practical use of the proposed method.
Lines 430-432: “It thus provides a rapid and reliable means of predicting the turning 430 quality or optimizing the turning parameters in industrial settings without the need for 431 time-consuming and expensive trial-and-error experiments.” – this is a qualitative assumption. Please back it up with some quantitative facts and figures.
Author Response
Please check the attached file! Thank you so much!
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Authors answered my all comments.
Reviewer 4 Report
The authors provided sound solutions to every issue pointed by the reviewer. Thus, I now consider that the paper can be published.
