Damage Detection of High-Speed Railway Box Girder Using Train-Induced Dynamic Responses
Round 1
Reviewer 1 Report
1. In the figure 6 more explanations are needed for the “ the up train side” and “the down train side”
2. In the figure 5: For the damaged state: a total of 3x3x6x4=216 different working conditions. Detailed comments should be given to understand this part
3. How do you choose the sensors positions in the studied bridge ? and why ?
4. In the lines 214-216 “The UIC good and UIC bad in Figure represent the low-interference spectrum and high-interference spectrum respectively, which are widely used in high-speed railways and ordinary railways in Europe” does the cited paper explaining the UIC values or what was your aim will citing it ?
5. In the lines 206-207 can you explain the background of the numerical analysis ? in terms of accuracy, how could it express or replace the simulation damage conditions through field experiments ?
6. As remark after reviewing this paper, we would suggest you to add a section to clearly explain the framework of the proposed method to help the readers easily understand the main aim of the study.
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
This paper about a damage detection method based on the high-speed railway train-induced responses considering the bending and torsion coupling effect of box girder. paper is well written but need to address following issues:
1. Author should add how the coupling effect and torsional effect is incorporate in the study.
2.What is the co-relation of bending and torsional effect in section 3.4 and it's justification.
3.Paper only addressed about L/4 and L/2 data what out 3L/4? Also, representation of L/4, L/2 and 3L/4 is not mention in paper.
4.What is train quality in table 1?
5.Why the value of DF is different in different sensor and how it can be helpful for the damage detection?
6.Verification of proposed model in section 3.3 and it's relation with bending and torsional coupling is not clear.
7.Also, mention merits of proposed model for damage index.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
This study proposed a damage detection method to identify damage in a simply supported box-girder. The method is investigated using a finite element model of the structure of interest. Prior to considering this article for publication, the following comments shall be addressed:
1- Please spell out the acronyms on their first mention, such as MLP (line 65), ANN (line 79), and UIC (line 214).
2- Line 101 states “The FEM verified by experiments was used …”. The paper, however, does not include empirical data/evidence. Please explain or modify this part accordingly.
3- Line 103 mentions the “vertical track instability”. Please define this term and explain what it physically refers to.
4- Please also describe how this vertical track instability is implemented in the FEM.
5- In Eq. (1) please define “Y” in the text.
6- Line 131, what is parameter m? Please introduce before using.
7- Please add references from which the equations provided in section 2 are taken, whenever applicable.
8- Please introduce parameters S, H, and w in equations 6 and 7. Also, explain what they physically show.
9- Since Mahalanobis squared distance (MSD) is another widely used metric to perform the outlier analysis, please consider comparing your ICDF approach with MSD.
10- Line 174 specifies that the beams are prestressed. Please describe how this prestress effect is implemented in the ANSYS model.
11- Please provide FEM parameters in the manuscript, such as the material properties as well as the spring and damping coefficients used to model the concrete and track supports respectively.
12- Please reword the sentence in line 183. It is not very clear.
13- Line 198 explains Figure 3 after the introduction of Figure 4 in line 196. Please introduce Figure 3 before describing Figure 4.
14- In Figure 6, what are the three solid black squares under the girder? If not relevant, please delete them.
15- Based on lines 226 to 229, the first two PCs are explaining 80% of the variations. However, since EOV seems not to be included in the FEM, what do these 2 PCs show? Please elaborate a bit on the physical meaning of the removed and remaining PCs.
16- Totally 36 sensing locations were investigated. However, the question is that if the damage is at L/4, can instrumenting the girder at L/2 identify the damage? or the sensors need to be at the same location as the damage.
17- Based on the answer to the question above, what would be the optimum number of sensors to identify any potential damage along the entire length of the bridge? Please add more material to the manuscript based on your thoughts about this question and the previous one (comments 16 and 17).
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The authors incorporated all the comments and now the paper is clear for the potential readers.
Reviewer 3 Report
No Comments.
