Review Reports

Round 1

Reviewer 1 Report

The authors propose an overview of the main concepts for predictive maintenance where the data are multivariate time series.

The presented approaches are not novel. In fact, the workflow and the techniques illustrated in the paper are commonly used for developing machine learning models for predictive maintenance. Although the level of novelty is not high, I found it very useful to read this paper (I am working on a similar project).  In fact, I don't think the main goal of this work is to present a new approach but to provide a starting point for those researchers that want to apply machine learning for the task of predictive maintenance. If the authors embrace this purpose, then I think they should improve the study of the related work, in order to provide a valid introduction and overview of the state-of-the-art.

[Minor comments]

• I think the last four headers of table 8 are wrong. For instance, "voltmean_3h" should be "voltmean_24h".
• I would prefer the use of the terms "validation set" rather than "development set".

Author Response

Point 1: The presented approaches are not novel. In fact, the workflow and the techniques illustrated in the paper are commonly used for developing machine learning models for predictive maintenance. Although the level of novelty is not high, I found it very useful to read this paper (I am working on a similar project).  In fact, I don't think the main goal of this work is to present a new approach but to provide a starting point for those researchers that want to apply machine learning for the task of predictive maintenance. If the authors embrace this purpose, then I think they should improve the study of the related work, in order to provide a valid introduction and overview of the state-of-the-art.

 

Response 1: We thank the reviewer for the very useful comments. In fact, the purpose of this paper is to make clearer an approach to the application of Machine Learning framework and techniques to maintenance data. In the short time allowed to improve the text, we introduce two sentences, lines 72-77 and 105-113 that we hope improve the text. In these two sentences, new references were added to complete the previous ones.

 

Point 2: I think the last four headers of table 8 are wrong. For instance, "voltmean_3h" should be "voltmean_24h".

I would prefer the use of the terms "validation set" rather than "development set".

 

Response 2: We thank the reviewer for these comments that help us to make the correction of mistakes and improve the text. The changes and corrections were done.

Author Response File: Author Response.docx

Reviewer 2 Report

Dear Authors,

thank you for the possibility to read your article. I have found a very interesting topic of your manuscript which is in the range of the journal's scope. I regard “Application example” in the Section 3 as the most valuable in the paper. However, there are some issues that, in my opinion, should be improved. I recommend a major revision and offer some improvement suggestions that can perhaps increase the readability and quality of the article.

1. The main disadvantage of the paper is a lack of research results carried out in this area into so far. I would recommend to present them in Section 2 based on the current literature. Merely referring to a literature review in another article (lines 88-96) is not sufficient.
2. Besides, the literature research gap should be much more identified and described.
3. The value of the article would be much greater if the authors introduced an extension of discussion elements and comparisons with the results of other studies at the end of section 3 or as a separate section.
4. The novelty of the reviewed paper is not clearly presented. Authors should briefly highlight their own achievements, findings and the main contributions of the manuscript both at the end of the introduction and in the conclusion section.
5. Finally, the directions of the further research should be mentioned in the conclusion section.
6. Detailed remark: figures and tables must be renumbered. There are two “Figure 1.” and two “Table 5.”

Author Response

Point 1: The main disadvantage of the paper is a lack of research results carried out in this area into so far. I would recommend to present them in Section 2 based on the current literature. Merely referring to a literature review in another article (lines 88-96) is not sufficient.

 

Response 1: We thank the reviewer for the very useful comments. In fact, the purpose of this paper is to make clearer an approach to the application of Machine Learning framework and techniques to maintenance data. In the short time allowed to improve the text, we introduce two sentences, lines 72-77 and 105-113 that we hope improve the text. In these two sentences, new references were added to complete the previous ones.

 

Point 2: Besides, the literature research gap should be much more identified and described.

 

Response 2: We thank the reviewer for the very useful comments. As already mentioned in Point 1, we introduce two sentences, lines 72-77 and 105-113 that we hope improve the text. In these two sentences, new references were added to complete the previous ones. Also, we call attention to the difficulties in approaching the very extensive literature that has been published recently, but which is applied to particular case studies and in which the comparability of results is almost impossible to do.

 

Point 3: The value of the article would be much greater if the authors introduced an extension of discussion elements and comparisons with the results of other studies at the end of section 3 or as a separate section.

 

Response 3: We thank the reviewer for the very useful comments. As we have already mentioned in the previous Response 2, it is very difficult to compare the results, as the data sets are different and the feature engineering applied can influence the results very much.

 

Point 4: The novelty of the reviewed paper is not clearly presented. Authors should briefly highlight their own achievements, findings and the main contributions of the manuscript both at the end of the introduction and in the conclusion section.

 

Response 4: We thank the reviewer for the very useful comments. In the short time allowed to improve the text, we introduce in the new sentence, lines 72-77, the main reason why we have decided to perform this work.

 

Point 5: Finally, the directions of the further research should be mentioned in the conclusion section.

 

Response 5: We thank the reviewer for the very useful comments. We introduce one final sentence, lines 537 – 542, where we state the future steps of our research work.

 

Point 6: Detailed remark: figures and tables must be renumbered. There are two “Figure 1.” and two “Table 5.”

 

Response 6: We thank the reviewer for these comments that help us to make the correction of mistakes and improve the text. The changes and corrections were done.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Changes made in the article are satisfactory and have greatly improved the quality of the manuscript. I recommend to publish the paper.