Fault Identification in Membrane Structures Using the Hilbert Transforms

Round 1

Reviewer 1 Report

This study proposed to use the Hilbert transform to realize damage detection in membrane structures, and experimentally validated its effectiveness. The paper shows more or less impact on the development of damage detection in membrane structures. However, there are quite a lot of issues that should be addressed to improve the quality of the paper.

1. Inappropriate article title. This study used the spectral moments with generalized and fractional Hilbert transform for damage detection. However, in the title, only the generalized Hilbert transform is mentioned.
2. For both results of fractional and generalized Hilbert transforms, i.e., Figs. 5 and 6, the mentioned additional frequencies cannot be observed. The reviewer suggests the authors replot these figures to highlight the named additional frequencies indicating damage occurrence by setting a small range in the z-axis (amplitude).
3. Page 5. “The filters designed 182 for this puropuse are refered in the paper as F160 (first natural frequency) and F240 (sec-183 and natural frequency) respectivly.” (1) Please check the typos first; (2) The reviewer has no idea what are the mentioned F160 and F240. Are the filtered results of the measurement response using a low pass filter with 160 Hz and 240 Hz, respectively? Please provide corresponding accurate descriptions of this. In addition, please present an example of the time history of the measurement result by SLV, and the corresponding F160 and F240.
4. Considering the excitation frequency ranges from 20-500 Hz, the reviewer wonders how the authors get the results below 20 Hz. Noted that the response below 20 Hz cannot be excited under such a limited-bandwidth excitation; in this scene, those responses under 20 Hz should be measurement noises from the view of random vibration. Please explain it.
5. The reviewer would like to know the merits of the proposed detection method compared with other frequency-domain methods? In addition, please provide more information about the essential theory of why Hilbert transform can be used for detecting damage. We cannot suppose that all the readers understand the theory; thus, a brief description of it is necessary.
6. In this study, the authors did not present any numerical results to support the effectiveness of the proposed method. As far as the reviewer’s concerned, the detection performance of the proposed method is not so well. For example, in Fig. 8, the reviewer cannot clearly determine the damage-induced change law. Therefore, numerical validation will be helpful.
7. The reviewer does not understand the essential theory of “Above all, the moments that indicate the damage are the first spectral moment at the first vibrational form and the second moment at the second vibrational form.” Please give a brief explanation.
8. The description “In contrast to other methods, the presented fault detection method gives qualitative diag-nostic information allows for evaluation of thechnical state without reference to historical data(Fig.9).” (1) Check the typo. (2) It is noted that the results of the undamaged state are required in this method; the reviewer wonders why the authors mentioned no reference data are required. Please explain it.
9. Literature review. (1) The other detection methods for membrane structures should be given a brief review. (2) The review on Hilbert transform for damage detection is not comprehensive.
10. Others.

(1) Abstract. “This paper presents the results obtained for damaged and nondamaged membrane struc-tures”. It is likely an incomplete sentence.

(2) Page 2. “Section the second” should be “Section two”?

(3) Figs. 5 and 6. The caption should be F240 instead of F420.

(4) Page 5 “…work of A. Venkitaraman and C.S. Seelamantula [27]..” should be “Venkitaraman and Seelamantula [27]”? Please check the guideline for the authors in Sensors for accurate citation style.

(5) What is the meaning of “dla” in Eq. (3)?

(6) There are quite a lots of typos in this study. Please read through the draft carefully and correct them.

(7) Typically, if the abbreviation appears less than five times in the study, we should not use the abbreviation, such as SLV.

(8) Noted that Fig. 3 has been used in the previous study, reusing a same picture from other paper might be not appropriate if no permission is allowed. Please use other photos showing the SLV device.

Author Response

Dear Sir, 
thank you very much for your comments on our work.
Kind regards,
Aleksandra Waszczuk-Młyńska

Author Response File: Author Response.pdf

Reviewer 2 Report

This submission needs clarity in its presentation more than anything else.  At one level for example it is never really clearly stated what the damage is in the diaphragm structure.  Has some been removed? I s it a scratch?  Is it an added patch? At another, the motivation for even considering the Hilbert transform is never made clear - and this is essential in order to encourage potential readers to have any interest in the text.  What are the insights that the Hilbert transform facilitates and how do these compare with the more familiar approaches - of which the Fourier transform springs immediately to mind.

 

So - with clarity in both presentation and motivation, the submission may be of some interest to readers, but in its present form, for this reviewer reader interest is minimalised!

Author Response

Dear Sir, 
thank you very much for your comments on our work.
Kind regards,
Aleksandra Waszczuk-Młyńska

Author Response File: Author Response.pdf

Reviewer 3 Report

In this work is proposed a  structural health monitoring method which is based on the application of the generalized Hilbert transform (GHT), for detecting faults in membrane structures; the GHT is implemented to identify the fault-related features.

The proposal is interesting but there are some issues that have to be addressed.

1. The novelty and contribution are not clear and is mandatory to highlight both in the abstract and at the end of the introduction. Also consider to write the Abstract in a single paragraph.
2. Condition monitoring strategies are important in order to ensure the proper operation or working of systems (in general) in many fields, i.e., industry (machinery), civil constructions (buildings), among others. In this sense, I suggest to include in the Introduction a brief discussion about the key role that condition monitoring strategies play in different fields. Consider to include some of the following references: https://doi.org/10.1109/TII.2020.2973731 ; https://doi.org/10.3390/s21051558
3. I suggest to change Section 3 for Section 2 and viceversa, and for the Section "The membrane structure and experimental setup" please include a brief description about how the experiments were performed.
4. Figure 4 is not clear and its description doesn´t match with all the contain of Section 4. Thus, if in Section 4 is presented the proposed method, the section requires to be ordered and separated from the results.
5. A specific section of Results is missing
6. The Conclusion Section is composed by several paragraphs, I suggest only include two or three paragraphs in the Conclusion section and I also suggest to improve and order the main ideas that are intended to be exposed in such section.

Author Response

Dear Sir, 
thank you very much for your comments on our work.
Kind regards,
Aleksandra Waszczuk-Młyńska

Author Response File: Author Response.pdf

Reviewer 4 Report

1. The authors mentioned in the last line of the abstract that “This paper presents the results obtained for damaged and non damaged membrane structures”, which is very vague and does not tell the authors any additional information. Please rephrase it.
2. I do not seem to find the contributions of this work in the first section. Please make sure you add a clear list of contributions. For example, this sentence in line 62 becomes the first contribution and so on.
3. How does the paper relate to the journal? I do not see this part.
4. There are so many typos in the paper such as line 68 “appli-cations” 
5. Section 3 called theoretical background whilst it contains a lot formal modelling. Please change the title into something that fits better with the contents of this section.
6. I really struggle to find out what the novelty of the presented work and what the authors are trying to achieve in this article. You probably need to motivate your work properly.
7. Figures 5 and 6 are too small to be checked for accuracy.
8. Figure 8 uses various values among the plots, look to me unreliable.
9. I am just wondering whether the authors have been able to read figure 9 before submission. I really cannot see the bottom part of it.
10. I am sorry for being possibly harsh, the I just feel that the paper is yet to be ready for publication and needs a lot more work

Author Response

Dear Sir, 
thank you very much for your comments on our work.
Kind regards,
Aleksandra Waszczuk-Młyńska

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have not answered most of the last-round comments/questions. Although this paper shows more or less impact on the developing membrane-structures damage detection, the following comments are worth being addressed before a further decision. Please response the comments directly instead of runaround or mentioning something irrelevant.

1. The title has not been modified as the authors mentioned/answered. In this scene, the last-round comment response is likely unprofessional and not serious.

2. The authors did not address the quality issue of Figs. 5 and 6. The reviewer cannot figure out the mentioned characteristics in the figures.

3. Comment 4 in the last review round. Noted that the excitation frequency range was 20-500 Hz. The authors explained that the low-frequency results are related to the signal process. Are the results below 20 Hz caused by noise? The reviewer wonders whether the results below 20 Hz can be removed. Furthermore, the low-frequency results in Figs. 5 and 6 are much more significant than those above 20 Hz. Please explain it.

4. Comment 5 in the last review round. The authors did not clearly present the merits of the proposed method compared with others.

5. Comment 8 in the last review round, the reviewer feels confused about the authors’ definition of reference data. Noted that the qualitative or quantitative analysis is not the reviewer’s concern.

In the reviewer’s opinion, if the monitoring data in the basic/non-damaged state is required, they are typically known as reference data to compare with others obtained in the damaged state. In practical applications, the reviewer wonders whether the monitoring data in the basic state is required or not. If yes, then the proposed method should be a reference-data-required method.

Author Response

Dear Reviewer,
thank you for all comments in the attachment you will find the answers.
Kind regards

Author Response File: Author Response.pdf

Reviewer 2 Report

This latest version is now adequate for publication.

Author Response

Dear Reviewer,
thank you very much for your review.
Kind regards

Reviewer 4 Report

The authors have not provided a response letter to explain how they have addressed the raised issues from the previous round. This is a major issue. I can only see that the authors left this message in their submitted report "thank you very much for your comments on our work."

Author Response

Dear Reviewer,
thank you for all comments in the attachment you will find the answers.
Kind regards

Author Response File: Author Response.pdf

Round 3

Reviewer 1 Report

The paper can be accepted, considering that most of the comments have been addressed.

Just for the authors’ information:

(1) The authors do not answer the last question relating to the reference data in the last review round. The reviewer would like to know whether the method is a reference-data-required approach or not.

(2) Please reply to the reviewer point-by-point in your future comment response. The comment responses are a bit messy.

Author Response

Thank you once again for the submitted review and the remarks that are extremely valuable to us.

The problem of the reference system in our measurements is of particular importance. The comment of the reviewer is accurate and in the considered example the role of the reference system is played by our analytical model. Of course, in fact, such a role can also be played by the measurement of an undamaged system. We fully agree with this remark.

Reviewer 4 Report

This version of the paper looks a little bit better that the previous ones. However, the conceptual modelling of the proposed solution is very poor. The authors have only made the formal model available and ignored the conceptual model completely. I urge the authors to pay more attention to this part if they were given another chance to revise the paper. 

Author Response

Thanks again for your review. The comments regarding the need for a broader presentation of the conceptual model used in the publication are of particular importance to us.

We are sorry, but the basis for our analysis and selection of the method of measuring and analyzing the signal was the analytical model of the membrane vibration, starting from the model of the undamaged membrane through the analytical model of the vibrations of the damaged membrane. Using the analytical solutions received, we found that the information about the damage will be included in its modulated signal amplitude. Hence the choice of the Hilbert transforms and their analysis in terms of sensitivity to changes in the modulated amplitude.