WindDefNet: A Multi-Scale Attention-Enhanced ViT-Inception-ResNet Model for Real-Time Wind Turbine Blade Defect Detection

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper presents an improved deep learning-based framework for real-time detection of Wind Turbine Blade defects, which is an actual task for modern renewable energy.

But there are some suggestions for authors:

1) There are Figures, that aren't noted in the text, and it's hard for the reader to understand why these figures are in the appropriate chapter (Fig.1, Fig.2, Fig.3, Fig.6, Fig.8)

2) Some Figures are in the wrong places. It's better to get the Figure after the notes about it, but in the paper there are Figures, that are in another chapter: Figure 1 and Figure 2 (it seems that figures can't be in Introduction and Literature Review), Figure 4 must be in a subchapter 3.1.2, Figures 5,6,7,8 also better to includ in chapters after notes about them.

3) Maybe it is better to name chapter 2 "Literature Review of Wind Turbine Blade Defect"

4) In paragraph 4 (line 1) of chapter 2 it's needed to write "CNN", not "cnn"

5) On page 4 in paragraph 4, you don't need to write the title of the research, its reference number is enough

6) It's needed to describe Table 1 in the text, as it is not clear what exactly this large table shows

7) There are some unclear texts in the paper, for example: "Introduction section" in paragraph 4 on page 4; "Transformer Layers" after equation (14) on page 14, etc.

8) The title of the paper contains the "WindDefNet", so it's better to write the same instead of "WindNet-Defect" on page 7. But in case these are two different elements, it should be written in the text

9) Pay attention to the formatting of the text, there is a lot of bolding, although this is not necessary, and a different font is used.

10) In the title of chapter 3.3.1, it's not understandable about dataset 1, but there aren't  datasets 2 or 3 in the text, so it's better not to write "1" in this title

11) It's not clear in chapter 3.3.2 how has been choosen the datasets and they are used for what? Because there are noted other defects on the pictures, that weren't used for this research. Please check this task

12) In chapter 4.2 it's not clear the second sentence: "The results are in Table 2. Precision measures the proportion of correctly predicted positive observations relative to the total predicted positives". What did the authors mean?

 

Author Response

Many thanks for the time and efforts of the reviewer. pointwise response to the comments has been attached herein. thanks.  

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This paper presents the multiscale attention-based condition monitoring of wind turbine blade fault detections.

1. The abstract needs to be revised in a better way.
2. Line numbers 122 to 138 describe the proposed methodology. However, the reference [28] is mentioned after the description of the proposed methodology. How does the proposed methodology differ from the reference [28]?
3. Figure 9 shows the comparison of different methodologies. The accuracy is almost similar to the other methods.
4. Figure 10 presents the SOTA comparison of the proposed WindDefNet model. The proposed model has 38 classes in total, but the other methods have different numbers. This can be justified with proper analysis.
5. Section 4.3 is meaning less. It's already given in Figure 10.
6. The performance matrix in Table 4 must be analysed using other methods to validate the proposed scheme's efficacy.
7. A detailed analysis of the proposed methodology must be presented.
8. The reference format must be revised according to the journal template.
9. In many places, the font is not similar.

Author Response

Many thanks for the time and efforts of the reviewer. pointwise response to the comments has been attached herein. thanks.  

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This study proposed a method to detect the real time wind turbine blade defect. This paper is well-prepared, and the contribution is clear. The results are sound, and the discussion is throughout. I only have one minor comment regarding the conclusion, which I hope authors can address it before publication.

Please rewrite the conclusion. Too much background information is included, which is unnecessary. Also, please short the summary and organize it in a more logical way.

Author Response

Many thanks for the time and efforts of the reviewer. pointwise response to the comments has been attached herein. thanks.  

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have addressed the comments in a better way. However, I have some clarification in the article. My comments are given below.

1. In the author's response No. 2, the authors mentioned that the reference [28] is the recent one. However, in what way does the proposed work differ from reference [28]? If the reference [28] is more relevant to the proposed work, then I recommend that the authors compare the results with reference [28].
2. In the author's response No. 4, the same number of classes should be used to compare different techniques. Otherwise, the conclusion derived from the analysis is not valid. 
3. Different techniques, such as ResNet, DenseNet, and Inception FPN, have been used to compare with the proposed algorithm. Is there any reference for those techniques that were already applied in the same system? If the answer is YES, then the authors should mention the references in all the figures. 
4. Check the reference according to the journal format
5. The article demands better proofreading. 

Author Response

Many thanks for the reviewer for taking the time to review the manuscript that significantly improvised the presentation and technical aspects of the manusucript. 

Author Response File: Author Response.pdf