Machine Learning-Based Prediction of Dynamic Responses of a Tower Crane under Strong Coastal Winds
Round 1
Reviewer 1 Report
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
The paper presents an application of machine learning to predict dynamic responses of tower cranes due to typhoon. It shows some interesting results and is of interest of researchers in wind and structural engineering. The paper therefore can be considered for publication. However, revision is required with the following comments and suggestion.
1. Introduction, Page 2, references for the statistics of the crane damages should be provided.
2. Introduction, the literature review should provide critical comments on the relevant past studies instead of listing them only.
3. Introduction, page 3, paragraph “In addition, current research …”, the statement in the last sentence is not quite true. The simulations presented in [21-31] are not a result of the rapid Monte Carlo methods and computer technology. Also, the two parts grouped from [21-29] and [30-31] are unclear and convincing. I suggest revising the sentence. Furthermore, most of these studies were for the Atlantic Ocean. I would be good to add past studies for Chinese coast and Northwest Pacific region; for example:
https://doi.org/10.5194/nhess-20-1617-2020
https://doi.org/10.1016/j.jweia.2020.104441
4. Equations (1)-(2) should be provided with references.
5. Figure 9 should be presented clearer so that numbers and graph can be readable.
6. Section 4.1. The definition and description of the coordinate systems are unclear. It would be good to provide a figure in which the relationship between the axes, the angles and the structure is shown.
7. The monitoring duration of about 1.5 year does not look long enough, in my opinion. Also, only 22 days of data were used for the validation of the model which seems short to me. Please justify the choice of the durations.
8. The validation of the model is not convincing to me as shown below. This is very important to make sure the results shown later is reliable.
i) using the only metric such as Mean Square Error is not reliable as shown in many studies in the literature. I suggest using more metric to validate the model.
ii) It is not enough to use only the maximum displacements for comparison. For a given day, the maximum displacement provided by the IoT system and that provided by the proposed model can occur at different time. Therefore, it is necessary to have the comparison of other data (e.g. instantaneous wind speed) from the two sources.
iii) Comparing the maximum displacements in time between the two sources is not enough. It is necessary to compare the data in different manners, for example probability and cumulative distributions of wind data.
Following studies are good references to address those points.
https://doi.org/10.1016/j.engfailanal.2021.105933
https://doi.org/10.1016/j.jweia.2020.104441
9. The study were conducted for a specific crane subjected to a specific typhoon. Therefore, the limitation of the study should be highlighted in the Introduction and Conclusion.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper presents an innovative real-time machine-learning-based monitoring system for the response of tower cranes to extreme meteorological events.
The paper is well written and can be recommended for publishing in JMSE, provided that the following minor revisions are made:
A general comment: The reviewer suggests including a table explaining all acronyms and abbreviations used in the paper at the end of paper text.
[Page 2]: ‘Gu and Guo [10] applied Ansys software…’ – The reviewer suggests providing links to official websites of Ansys and other software mentioned in the paper in footnotes to the paper.
[Page 2-3]: The second paragraph in the ‘Introduction’ section is very long. It would be better to re-organize it as a few shorter paragraphs.
[Page 3]: ‘…the difference between the response in a Class A land-form and a Class B landform was small and insignificant.’ – Please, add a brief explanation for non-expert readers on what Class A / Class B landforms mean.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The authors accurately addressed all the reviewer's concern.
Author Response
Thanks to the reviewers for their approval.
Reviewer 2 Report
The author has addressed most of my comments. However, the are a few points in their responses needed improvement and/or correction. Specifically,
Authors’ response 5
Figure 8 is not acceptable as many numbers are not readable with the black background and small size. Also, the Chinese text should be replaced by English text.
Authors’ response 8
As no further validation metrics are investigated, adding the reference [42] is not helpful. Therefore, it should be removed. Also, the limitation to this point should be emphasized in the paper.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
