Prediction of Rail Wear Under Different Railway Track Geometries Using Artificial Neural Networks

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. The definition of rail wear remains unclear. In the discussion of Fig.5, for example, the definition and the significance of rail wear were not thoroughly described. And why does the first 10^8 tonnage cause no wear?
2. For the sensitivity analyses in the end of the paper, many similar studies have already been performed through other ways like simulation. What difference does the paper make compared to those studies? And what can people learn or what can industries do from your conclusion?
3. The novelty demonstrated from the paper is not enough. Most of the content are just repeating the existing theories. Overall, the paper placed too much emphasis on ANN and too less on industrial application.

Author Response

Dear Reviewer,
Thank you for your insightful comments. We have addressed them in detail in the Response Letter, clearly indicating the corresponding revisions in the revised manuscript. Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

Reviewer Comments (Minor Revision)

Author Information

The author list ends with "and" without subsequent content; please revise this formatting.
The address for "Accenture GmbH, Balanstrasse 73, bldg 17, 81541 Munich" should include the country name: "Munich, Germany."
Equations

Equations (1)–(6) should be uniformly numbered and right-aligned.
References

Some DOI links in the references are incomplete (e.g., Ref. [1]). Please ensure all DOIs are provided in full.
Language Precision

Improve phrasing for clarity. For example, in the Abstract: "If the rail wear can be calculated..." should be revised to "If rail wear can be predicted..." to align with the study’s objectives.
Terminology Consistency

Unify the abbreviation format for "PSO-ANN" throughout the manuscript (currently, "ANN models" is used inconsistently in some sections).
Methodological Context

Expand the Introduction or Discussion section to include a comparison between PSO optimization and traditional ANN (e.g., convergence speed, avoidance of local optima) to highlight the advantages of the proposed method.
Practical Implications

The Conclusion should discuss the guidance significance of the study for railway alignment design (e.g., how rail wear prediction can inform optimal parameter selection during the design phase).
General Feedback
The paper presents a robust methodology and meaningful contributions to rail wear prediction. The integration of PSO-ANN and Sobol sensitivity analysis is well-executed. However, revisions are required to address the formatting, grammatical, and content-related issues outlined above.

Recommendation: Minor Revisions
The authors are requested to revise the manuscript according to the comments above and provide a point-by-point response to each suggestion in the resubmission.

Author Response

Dear Reviewer,
Thank you for your insightful comments. We have addressed them in detail in the Response Letter, clearly indicating the corresponding revisions in the revised manuscript. Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

The proposed prediction method will be a valuable tool to be applied during the design stage of trackworks to give insight into the rail replacement cost during the lifecycle of the railway line. I would suggest that the authors further develop the prediction algorithm into a complete suite of software to embrace other train operation parameters that may impact rail wear, and an optimization function to guide the designer to adopt the most cost-effective configuration for the curves of new railway lines. A range of values, instead of a single value, should be adopted for each of the parameters to reflect the variations in real-life operations.   My other comments are on Section 2.3 Model Verification :

• Provide information on the type of trackform
• Provide outline details of the metro trains running on the line.
• A range of values for the coefficient of friction should be adopted to reflect the real-life operating environment.
• The difference in rail wear prediction between new wheels and worn wheels. 

   

Author Response

Dear Reviewer,
Thank you for your insightful comments. We have addressed them in detail in the Response Letter, clearly indicating the corresponding revisions in the revised manuscript. Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

It is suggested that the sensitivity analyses in 4.3 should be compared with the simulation results of your own model or other previous works, or this part doesn't seem to have much importance.

Author Response

Dear Reviewer,
Thank you for your insightful comments. In the Response Letter, we have thoroughly addressed your questions and suggestions, clearly indicating the corresponding revisions in the revised manuscript. Please see the attachment.

Author Response File: Author Response.docx