System Invariant Method for Ultrasonic Flaw Classification in Weldments Using Residual Neural Network

Round 1

Reviewer 1 Report

The paper was interesting.The author simply makes a spectrum analysis, and then uses the neural network to analyze. However,  the spectrum analysis can basically get a conclusion. The author should analyze why do by using the neural network in the introduction

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Dear Authors,

The manuscript "System invariant method for ultrasonic flaw classification in weldments using residual neural network" by Jinhyun Park and coworkers discusses the use of residual neural network in nondestructive-ultrasonic evaluation of weldments for classification of main flaws. The Authors analyzed a very large amount of data (5,839 measurement results), which concerned five types of defects (crack, lack of fusion, slag inclusion, porosity, and incomplete penetration) occurring in weldments. It is a very practical research with clear industrial application. However, in my opinion some parts of the article should be improved, and thus major corrections should be implemented before considering the manuscript for publication.

Evaluation of the paper general remarks:

1. The Abstract section should present quantitative results and not only the most important qualitative results and/or generic considerations. The specific research results contained in the article are missing. Therefore, significant improvements are expected in this part of the manuscript.
2. The Authors have written "To solve this problem, many studies on ultrasonic testing to determine the presence, type, shape, and size of flaws using artificial intelligence are being conducted". What studies? In this part of the article the Authors should add the references and few studies should be described in details.
3. Line 56 - Table 1, Why the table is cut? Some important information about ultrasonic transducer are missing. 
4. In line 68 the Authors use the personal form ("…they were insignificant ..."). This is not correct in high-quality articles. It suggests modifying this part of the article. Please use passive verb or different verb in active voice. Please check the entire text of the manuscript.
5. Line 71 and 75 - Please read the instructions on how to describe the references in the main text of the article in the authors' guide and change it. Currently, the references are not in line with the journal requirements. In instruction for Authors is written "References must be numbered in order of appearance in the text (including citations in tables and legends) and listed individually at the end of the manuscript...".
6. Line 108 and 109 - Authors have written "For more information on the modern training procedures, weight parameter 108 selection, activation, and loss functions, see Refs. [6–9]." In my opinion, this is an important point of the article, which should be described in more detail, especially publications 6-9. This will allow for a better presentation of the topic and determination of the novelty of the authors' publications in relation to the quoted references. The same remark for line 117-119, 130, 159.
7. Line 240 and 252 - why there is two times the same title of the subsection?
8. The test results available in the literature should be specifically referred to Authors results. When the results are not discussed and conveniently supported by the open literature, questionable conclusions are obtained. Currently, the article looks more like a  report from tests than a scientific article. Significant improvement in the description of the results is required.
9. Research articles should present the directions of further research. I suggest adding one paragraph in the Summary chapter.
10. The article contains a number of abbreviations. I recommend adding a list of abbreviations in the article along with a detailed explanation of each, for example Res-Nets - Residual neural networks. The article will be more readable for the reader.

Specific remarks/editorial comments/typos:

• line 56, 61, 66, 89, 90 and others - the dot is missing at the end of the capture, please check entire manuscript and add the dots.
• line 75 - there is "described in Refs [1, 2]." and should be described in [1,2].
• line 148 - the colon is missing at he end of the sentence.
• Figure 9 - the (c) in the picture should be bold, is it? please check it.
• References - some lines (i.e. 316, 318, 321 and others) in the References should be aligned, please modify and reorganize this section of the manuscript.

In my opinion the article requires the above major changes. I hope these suggestions can help to improve the quality of this paper.

I wish you all the best.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The paper can be accepted  .

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Dear Authors,

thanks for addressing the comments. However, in my opinion, the manuscript needs further improvement as not all issues have been modified.

The main issue concerned the comparison of the results of the Authors' research with the test procedure and results presented in the literature. It was not about extending/modifying the description of own research (in Summary), but comparing the specific results obtained with the results of other researchers - Point 8 of the Author's response. This manuscript needs significant improvement at that point.

• line 106 - the dot is missing at the end ot the sentence, the same remark for line 291 and 308,
• line 125 - two dots?
• line 152 - at the end of the sentence is dot, before for table caption Author's did not use the dot, line 107, 83, 78, 73, 305. Please use one style - the dot should be at the end of the sentence.

I wish you all the best.

Author Response

Please see the attachment.

Author Response File: Author Response.docx