Modal Parameters Identification of Bridge Structures from GNSS Data Using the Improved Empirical Wavelet Transform

Round 1

Reviewer 1 Report

• In the manuscript, the authors present their studies on using an improved data analysing (signal processing) method (which is called "Improved Wavelet Transform") aimed at SHM systems of bridges. The problem is clearly explained as well as the performed experiments are fairly and comprehensively shown. Nevertheless (except my forthcoming remarks and suggestions regarding both the merit and text editing), my general concern is whether such material refers to "Remote Sensing". Unfortunately, I cannot see any similarities with remote sensing at all. Structural Health Monitoring, both with GMS (geodetic monitoring systems), belong directly to the Civil Engineering and - in their current form presented by the authors (RT-GNSS and accelerometer) do not represent remote sensing (RS) methods! In that mean, RS-methods on engineering objects refer instead to photogrammetry, which is not the subject in this particular case.

  Regarding that, I've got a severe problem with recommending the paper for publishing. Should the authors have submitted their text to more civil engineering journals, the text could have proceeded with some revision. Here, I would recommend changing the journal. In my opinion - more suitable would be ex. "Sensors". Please consider that!

  Besides, I would also suggest considering the following hints:

   - please check the whole text in terms of missing spaces and typos; for example, prevalently citations lack spaces between preceding words (abbreviations) - as it is in Line 39, 42, 47, 50, 61, 69, 78, and so on;

  • Line 101: it would be necessary to explain using the Yule-Walker algorithm; why exactly this algorithm? How is it based in the literature etc.? I know, it might be evident for the authors and other professionals but for a wider audience reading the Remote Sensing journal, probably not;
  • Line 130 - 161: the authors give the Empirical Wavelet Transform method principles with its improved form. From the text, it seems that they have been the inventors of it.  I would recommend either putting relevant citations and references to the literature or - if the authors have invented it - to underline that fact more clearly!
  • Line 178: the method using the Pearson correlation coefficient would need justification; 
  • Line 205: if the equation had been taken from the others, it needs relevant references;
  • Lines 216-218: it only proves that the text fits other signal-processing journals (or "Sensors");
  • Line 255: choosing the criteria needs justification and references;
  • Line 264: the word "improved" would need some justification; on which criteria have the authors based their improvement, and how did they decide about it?
  • Table 1: RMSE needs a unit too, in my opinion;
  • Line 282: what kind of sensors have been used? What are their parameters? Please provide it at least in the references;
  • Section 4 ("Field experiments"): for me, it is a crucial point of the whole article because it represents the practical field verification of the proposed approach; hence, the location of the sensors would definitely need a sketch presenting it! For the surveyors, the geometry of control networks play a crucial role and decides about the total accuracy etc.!
  •  Lines 364-366: the sentence informs that there have been two GNSS-receivers but only one antenna (Leica AT504). It is probably not the case. The choke-ring AT504 antenna belongs to the reference station established in the vicinity of the bridge. In such a case, it lacks information about the receiver used for the base station;
  • Lines 366-269: 2012 - it was a pretty long time ago - I would definitely give the information about the monitoring intervals within that period (the performance of GNSS systems may be dependent on time due to the development of particular satellite systems which stand a part of the whole GNSS);
  • Line 374: please provide information about potential obstacles etc. - at least, please write the sketch of your control network geometry;
  • Lines 377-379: does it mean that both receivers observed the same antenna? Please excuse me, but I don't entirely understand your intention;
  • Lines 384-386: what conditions took place that time in terms of surveying with GNSS?
  • Is your choice for GNSS equipment used anyhow justified? 
  • Figure 18 is not entirely visible - maybe its resolution should be higher?
  • Line 422: I am wondering why the pure RTK using a reference station located nearby gave (according to your information) worse results than the network solution (NRTK)... It should be checked or at least, anyhow explained;
  • Line 426: we have to be aware that many error sources influence GNSS measurement; hence, I would consider the equipment and the external conditions (plus geometry of a control network!);
  • Lines 479-481: the RT-GNSS-surveying is generally limited by the accuracy (typically, it represents a cm-level); however, in regular monitoring geometries, it is possible to filter the data and obtain (statistically) a millimetre-level. Nevertheless, the measurement uncertainty of GNSS never reaches sub-millimetres, so I would be careful with providing readers with such information;
  • Line 493: choosing the objects needs justification.

  All and all, I do appreciate the authors' valuable work, and I am encouraging you to submit the material for publication after considering my remarks. Thank you!

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Manuscript ID: remotesensing-1313226

Title: Modal parameters identification of bridge structures from GNSS data using the improved Empirical Wavelet Transform

Authors: Zhen Fang, Jiayong Yu, and Xiaolin Meng

 

In this manuscript authors present a procedure based essentially on the improved Empirical Wavelet Transform (EWT) aimed at denoising GNSS data and identifying modal parameters of bridge structures. Moreover, to validate the proposed methodology, 1) three groups of simulation data were used, 2) a field experiment with three sets of GNSS receivers and an accelerometer was carried out on the Wilford suspension bridge in Nottingham, UK.

 

In my opinion in its present form the proposed manuscript doesn’t fit the standards of MDPI-Remote Sensing and improvements are needed. In the following points are summarised the suggested corrections:

 

As general remarks:

• I suggest to the authors the revision of the language by a native English speaker, cause many sentences in the body of the manuscript seem syntactically incorrect.
• The manuscript is 22 page long, I think it is better to shorten it. In particular, the section “1. Analysis of a simulated data” can be shortened (thee groups of simulation data are proposed, in particular, the first test seems trivial), conversely some more technical information of the GNSS instruments used should be added in the section “4. Field experiments”, this is very important (see title: from GNSS data).
• The use of acronyms it is excessive, and the reader gets lost.

As specific indications:

• At line 49 (section: Introduction), I think that is necessary to add the verb “are”: “and frequency domains, are becoming more”.
• At line 109 (Introduction), I think that is better to replace the word “fellows” with “follows”.
• At line 127 in the section Methodology perhaps the acronym NR is used in place of NF?
• The sentence from Line 146 to 147 of the Methodology is not clear and needs to be reorganised.
• At line 156 a₁ – a_p are indicated in the text while in the equation (5) the a_i parameter is indicated.
• The sentence from line 162 to 163 needs to be reorganised
• The sentence from line 168 to 170 in not clear.
• In the sentence from line 182 to 183, the IMFK is intended as threshold? The authors should explain better.
• The sentence from line 208 to line 210 should be rewritten.
• The sentence from lines 216 to 218 is not clear.
• The sentence from line 219 to line 222 needs to be rewritten.
• In the section “Numerical studies” the sentence from line 233 to line 237 needs a revision.
• At line 364 of the section “Field Experiments” authors should indicate some more characteristics of the GNSS receivers used.
• At line 376 authors should indicate the characteristics of the 3 experimenters used (add a Figure).
• The two sentences from line 447 to line 451 of this section needs to be reorganised.
• At line 477 the displacements are indicated with two decimal places (10.10 and 10.40) but authors declare a millimetre-level accuracy. I suggest to indicate all the displacement values in the text with one decimal place.
• In the “Discussion and Conclusion” section at line 500, I think that the authors intended the “field experiments” and not the “filed” ones.
• The sentence form line 506 to 507 needs reorganisation/rewriting.
• At line 508 a millimetre-level accuracy is declared by authors, but the displacement indicated in the text with two decimal places is better than hundredth of a millimeter (10.10, see point 16).

 

Following the corrections indicated above I suggest a major revision of this manuscript.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Reviewer’s Report on the manuscript entitled:

Modal parameters identification of bridge structures from GNSS data using the improved Empirical Wavelet Transform

The authors proposed an improved Empirical Wavelet Transform (EWT) based procedure to denoise GNSS data and identified modal parameters of bridge structures. The paper is generally suitable for publication in Remote Sensing; however, I have several comments that need to be addressed as listed below.

General Comments:

The introduction section should be improved:

Line 53. There are several other robust time-frequency decompositions that need to be included in the Introduction, such as

The continuous wavelet transform (CWT) for identification of noise modal in GNSS time series https://doi.org/10.3390/rs10101611 with software package freely available at https://atoc.colorado.edu/research/wavelets/

The least-squares wavelet analysis (LSWA) for modal parameter identification https://doi.org/10.1088/1538-3873/abaf04 whose software package is publicly accessible in the online portal of the National Oceanic and Atmospheric Administration (NOAA): https://www.ngs.noaa.gov/gps-toolbox/LSWAVE.htm

The software package also includes the antileakage least-squares spectral analysis tool (ALLSSA) that can accurately estimate the signal frequencies, very useful for random noise attenuation.

And the weighted wavelet analysis (WWA) for processing GNSS position time series https://doi.org/10.3390/rs12060992

 

Line 510. What type of noise? Gaussian random? There are various types of noise (color noise).

The Conclusion part must be separated from the Discussion part (please also consult with the MDPI guideline). Please add what was proposed, its application, limitation, and future direction in separate paragraphs. The authors should also discuss how the presence of trends, jumps in the trend component (usually observed in GNSS time series) can affect the results. The limitations of the method should also be discussed. A brief paragraph for future direction can also be added in the light of the recent methods that I mentioned above (CWT, LSWA, ALLSSA, WWA).

 

Comments of the Figures:

Figure 5(b), Figure 14(b), and Figure 16. All the x-axes are the same, so I suggest the authors attach all the panels below each other and use a common x-axis below the figure. This way there will be more space and the y-axis can also be enlarged for a better display. The title of each panel (e.g., ewt1, ewt2, etc.) can then be placed inside each panel.

Figure 18. Again, the x-axis is the same and the two panels can be attached. See for example Figures 5 and 7 in the LSWA paper that I mentioned above.

Figure 23. Please show the difference between NRTK and ACC in a new panel attached below to aid comparison.

Minor Editorial Comments:

Line 14. Please remove the comma after spectrum.

Line 25. Please remove the ‘dot’.

Line 125. Please define the abbreviation NExT because one may confuse it with “next”.

Equation (1) please have n as the subscript of omega (w_n).

Lines 143, 144. “Where” should be “where”

Equations (5)-(8) the comma should be inserted after the equation.

Figure 6. The x-axis should typically be in ascending order (1,2,3,4, …,9) not (9,8,7, ….,1).

Line 279. Please define the abbreviation NEEDS.

Line 498. Please start a new sentence after “order”.

Lines 514 and 519. Please avoid starting the sentences with “And”.

Line 515. At the end of the sentence: Please mention “in the experiment presented herein”

Adding an acronym table at the end of the manuscript can be very useful.

Thank you for your contribution

Regards,

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Dear Authors,

Thank you very much for your extensive answers to my questions/suggestions formulated in my review. I accept all of them as well as I appreciate your work. I do not have any objections, so in my opinion - after a thorough linguistic check, the text can be published. Hence, my recommendation is a minor revision without the necessity for another reviewing cycle. 

I wish to congratulate you on your interesting studies.

Regards,

Author Response

Thank you for your wishes. 

Following your comments, a minor revision has been carried out. The changes in the revision include:

1. The resolutions of the figures have been increased to 300 dpi.
2. The structure of the table #1 and #2 has been revised slightly.
3. Some grammar problems have been corrected, like singular & plural, abbreviation, and spelling.
4. The style of references has been checked carefully, like the abbreviation of authors, journal title, the page numbers or article ID of a reference.

 

Reviewer 2 Report

The manuscript has been greatly improved.
Now is ready for publication on Remote Sensing.

 

Author Response

Thank you very much for your helpful suggestions. According to your comments, we learned the shortcomings of the manuscript, and then corrected these questions finally.

Reviewer 3 Report

I would like to thank the authors for addressing my comments. The manuscript looks better now. I have a few minor comments to consider:

The figures look much better now. Just make sure that the resolution of the figures is at least 300 dpi. Almost all the figures have poor resolutions compared to the first version that I reviewed, but I suppose it is due to the highlighting changes and reformatting. Please make sure they are nice and readable with high quality and resolution (at least 300 dpi).

Line 236. it should be least-squares, not least-square

Would be nice to enlarge the numbers on the x-axis and y-axis in Figure 10.

Line 505. Grammar. Further research needs to ... not need

References. If the page numbers of a reference are not provided, then please write the article ID instead. For example:

Line 534 Reference #6:

Please replace "2020, 31" with "2020, 31, 125102"

Line 546 Reference #12:

Please replace "2020, 132" with "2020, 132, 114504"

Line 559 Reference #18:

Please replace "2014, 12" with "2014, 12, 1450044"

Please double-check all the references, formulas, and tables. Please carefully proofread the article before publication.

 

Thank you for your contribution

Regards,

Author Response

Please see the attachment.

Author Response File: Author Response.docx