Accurate Prediction of Tunnel Face Deformations in the Rock Tunnel Construction Process via High-Granularity Monitoring Data and Attention-Based Deep Learning Model
Round 1
Reviewer 1 Report
Dear authors,
you have written very interesting article according to my opinion. The importance of monitoring and predicting movement is well explained and documented in detail. Also, the models to perform time series prediction on the monitoring data are well presented and confirmed on real monitoring data. The article also shows the tunnel monitoring system in detail, with sensors installation, location, and the entire monitoring scheme.
Below are a few comments and technical remarks to consider improving the manuscript:
In Introduction section (line 46 – line 52) monitoring with a total station is mentioned and authors state the limitation of total station measuring frequency to a few times a day. Robotic total station which is essentially an advanced model of the total station don’t have this limitation and can be used for monitoring displacements during tunnel construction. If the limitations of the total station are mentioned, authors should also mention the robotic total station (RTS) which is basically the same instrument.
In line 473 authors mention for the first time measurements with a total station in the Tongluoshan tunnel and compare the results with the filtered monitoring data. If possible, describe the performed measurements used for comparison and achieved results.
Figure 7 – if possible, increase the font of date and time values as in the figures 10 and 12.
Figure 10 – text box “Monitoring time (a)” covers the axis of the graph with date and time values
Figure 12 – on figure (c) transformer model on deformation axis “-“ is missing next to – 10 deformation value
Best regards.
Author Response
The authors would like to express their gratitude to the esteemed reviewer. Constructive suggestions have helped us improve the paper. In light of the comments, we have assessed the manuscript thoroughly and revised it carefully. Detailed responses to all the comments can be found below. For clarity, specific responses (regular font) to the review comments (italic font) are presented immediately after the review comments. The line numbers refer to those in the pdf file of the revised manuscript. The revised content is marked with yellow colour in the revised manuscript.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The prediction of deformation in rocks during the tunnel construction process is quite important, therefore this article, the technique presented and the methodology (sufficiently validated) seem very interesting and useful for this purpose, especially in complex geological conditions. I think it is ideal to be published in your magazine.
Author Response
The authors would like to thank the esteemed reviewer for the encouraging comment. Based on the comments from the other reviewers, changes have been made to improve the validation of the results. Revised content is marked with yellow colour in the revised manuscript.
Reviewer 3 Report
It seems that the tunnel's displacement data were trained through an artificial intelligence model, and comparison between the actual monitored data and the predicted data was attempted. I'd like to hear your opinion on a few things:
1. The application of artificial intelligence to the tunnel field is seen as a fresh attempt. However, since the number of datasets for applying the AI ​​technique is not very large, the learning effect is not expected to be large. What is your opinion on this?
2. And in the test data(Figure 12), there doesn't seem to be much agreement between monitoring and predicted values. What's your opinion on this?
3. In general, in the case of displacement, which indicates the degree of deformation of the tunnel, it is known that it is sensitive to even a small displacement that can be measured down to the millimeter unit. If there is a difference between the predicted displacement and the actual measured value according to the application of the artificial intelligence technique, please explain whether the deviation is engineeringly acceptable in terms of tunnel safety.
Please reflect the answers to the above questions appropriately in the relevant part of the text.
Author Response
The authors would like to express their gratitude to the reviewer. We have assessed the manuscript thoroughly and revised it carefully in light of the comments. Detailed responses to all the comments can be found below. For clarity, specific responses (regular font) to the review comments (italic font) are presented immediately after the review comments. The line numbers refer to those in the pdf file of the revised manuscript. The revised content is marked with yellow colour in the revised manuscript.
Author Response File: Author Response.docx
Round 2
Reviewer 3 Report
Thanks for your reply.
