Predicting Model of Dual-Mode Shield Tunneling Parameters in Complex Ground Using Recurrent Neural Networks and Multiple Optimization Algorithms

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors used an intriguing methodology and framework to construct a deep-learning-based model to estimate dual-mode shield tunneling parameters in complex terrain. The writers also employed an actual case study based on a project in China. Here are some additional recommendations and comments that may help the article get accepted:

1-In the introduction part, the innovation and main contribution should be clearly highlighted.

2- I believe the authors should reinforce the state of the art even more since the references list appears to be rather restricted.

3- Why did the authors select LSTM over other machine learning techniques? This must be justified.

4- The optimization algorithms' parameters should be listed!

5- Why did the authors select that particular splitting for training and testing the algorithms?

6- The performance evaluation is limited to a few metrics; please see <Modeling the nonlinear behavior of ACC for SCFST columns using experimental data and a novel evolutionary algorithm> and <Simulation of the ultimate conditions of fiber-reinforced polymer confined concrete using hybrid intelligence models> for more metrics and graphical comparisons.

7- The conclusion should address the study's limitations and future directions.

Comments on the Quality of English Language

There are numerous typos and grammatical errors discovered, necessitating English proofreading. 

Author Response

Dear Editors and Reviewers:Thank you for your letter and for the reviewers’ comments concerning our manuscript entitled “Predicting model of dual-mode shield tunneling parameters in complex ground using recurrent neural networks and multiple optimization algorithms” (ID: applsci-2787495).Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have studied comments carefully and have made correction which we hope meet with approval.

1. In the introduction part, the innovation and main contribution should be clearly highlighted.

Thanks to the reviewer for pointing out this problem, it has been revised in the article.Based on the left line project of Liuxiandong Station-Baimang Station Tunnel of Shenzhen Metro Line 13, this paper obtains a large number of time series characteristic parameters based on the data acquisition system, eliminates abnormal data through the isolated forest algorithm, and optimizes the original shield parameters with the improved mean filtering algorithm.Considering the influence of stratum conditions on tunneling parameters from three dimensions of surrounding rock grade, tunnel depth-span ratio and soft-hard composite ratio, an LSTM model integrating four super-parameter optimization algorithms is established. Combined with dropout algorithm and five-fold time series cross-validation, the two shield tunneling modes of EPB and TBM and the propulsion speed under different strata are predicted and analyzed, which provides feasible guidance for intelligent control of dual-mode shield tunneling process.

2. I believe the authors should reinforce the state of the art even more since the references list appears to be rather restricted.

We will strengthen technical research, make a model more in line with engineering practice, and verify it in other shield tunneling projects.

3. Why did the authors select LSTM over other machine learning techniques? This must be justified.

The data of shield tunneling are time-series, and the data are purely dependent. Traditional machine learning algorithms ( such as BP neural network, random forest, etc. ) cannot capture the time-series value between data. The LSTM network has an internal gating mechanism, which enables it to effectively capture and retain information from past inputs. At the same time, the generalization ability of the LSTM neural network is stronger. It is a deep learning algorithm widely used in the market and more suitable for engineering needs.

4. The optimization algorithms' parameters should be listed!

Thank you to the reviewer for pointing out the problem, which has been changed in the text and marked.

5. Why did the authors select that particular splitting for training and testing the algorithms?

We establish the model mainly to explore the real-time operation law of the tunneling speed under the unsupervised training model during the tunneling process of the dual-mode shield, and train the EPB / TBM two modes of data together. In the process of data processing, in order to consider more composite strata, the obvious feature segment is selected as the input feature.

6. The performance evaluation is limited to a few metrics; please see <Modeling the nonlinear behavior of ACC for SCFST columns using experimental data and a novel evolutionary algorithm> and <Simulation of the ultimate conditions of fiber-reinforced polymer confined concrete using hybrid intelligence models> for more metrics and graphical comparisons.

This study mainly considers MAE, RMSE, MAPE and model running time as the overall evaluation index of the model, which is to consider that the field application not only has accuracy requirements but also time requirements. MAE, RMSE and MAPE express the prediction error and accuracy of the model from different perspectives, and have met the evaluation requirements as a whole. More evaluation indicators will be added in the follow-up study.

7. The conclusion should address the study's limitations and future directions.

In this study, the mechanical properties of rock have not been considered, and the model is not refined enough. In order to obtain a refined model that is more generalized and more in line with engineering practice, future research will consider more input of tunneling parameters and rock mechanics parameters.

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors.

I reviewed the proposed article titled: Predicting model of dual-mode shield tunneling parameters in complex ground using recurrent neural networks and multiple optimization algorithms.

I think the article is interesting but needs corrections.

Incidentally, I am not authorised to judge English grammar and spelling, but I think it would be useful to check it, e.g. tunneling - tunnelling.

It is necessary to clarify the following terms such as tunnelling, tunnelling speed, and tunnelling parameters. In the explanation, it is necessary to state the authors' understanding of what each term includes.

In equations (1), (2), (3) and (4), it is necessary to explain the meaning of Wf, bf, Wi, bi, Wc, bc, W0 and b0 in the text.

Describe the individual classifications of rock mass.

Check in line 331 whether C(n) is the same as c(n) below in line 334, or whether C(n) refers to c(n) from equation (7).

Explain the meaning of s(x, n) from equation (8).

Explain equations (9) and (10) and the meaning of the symbols in the equations.

Explain equations (15), (16) and (17) and what the individual symbols in the equations mean.

In the conclusions, add how the hypothesis put forward in the article will be verified in practise. In view of the quality of the article, confirmation of the implementation of the model would be desirable.

Kind regards.

Author Response

Dear Editors and Reviewers:Thank you for your letter and for the reviewers’ comments concerning our manuscript entitled “Predicting model of dual-mode shield tunneling parameters in complex ground using recurrent neural networks and multiple optimization algorithms” (ID: applsci-2787495).Those comments are all valuable and very helpful for revising and improving our paper, as well as the important guiding significance to our researches. We have studied comments carefully and have made correction which we hope meet with approval.

1. It is necessary to clarify the following terms such as tunnelling, tunnelling speed, and tunnelling parameters. In the explanation, it is necessary to state the authors' understanding of what each term includes.

Thank the reviewer for pointing out the grammar and spelling problems, which have been modified in the article. Also in this article, the term is explained in Section 4.1.

2. In equations (1), (2), (3) and (4), it is necessary to explain the meaning of W_f, b_f, W_i, b_i, W_c, b_c, W₀and b₀ in the text.

In the formula, W_f, W_i, W_c, W_o are different calculation matrices ; b_f, b_i, b_o and b_c are the bias terms of the three gated units and the cell state.Has changed, and reflected in the text.

3. Describe the individual classifications of rock mass.

This study mainly considers the influence of composite strata and surrounding rock classification, and does not refine the problem of rock sampling and classification, and the follow-up research will be deepened.

4. Check in line 331 whether C(n) is the same as c(n) below in line 334, or whether C(n) refers to c(n) from equation (7).

C(n) is written wrong, should be c(n), this is our mistake. The average path length of a binary tree constructed by c(n) for n samples has been unified and modified in this paper.

5. Explain the meaning of s(x, n) from equation (8).

S(x, n) is the anomaly index of the tree which records the training data of x in n samples.Has changed, and reflected in the text.

6. Explain equations (9) and (10) and the meaning of the symbols in the equations.

Has changed, and reflected in the text.

7. Explain equations (15), (16) and (17) and what the individual symbols in the equations mean.

Has changed, and reflected in the text.

8. In the conclusions, add how the hypothesis put forward in the article will be verified in practise. In view of the quality of the article, confirmation of the implementation of the model would be desirable.

The established model primarily aims to explore the real-time operational patterns of tunneling speed in the dual-mode TBM excavation process under unsupervised training models and provide a foundational model for intelligent decision control.In practice, the data in the tunneling process is input into the model in real time, and the model is trained and the results are output in real time, which plays a guiding role in the field.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

No further comments 

Author Response

Thank reviewers for their comments.

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors.

I reviewed the proposed article titled: Predicting model of dual-mode shield tunneling parameters in complex ground using recurrent neural networks and multiple optimization algorithms.

I have a minor comment on the authors' response to comment 3 from the previous revision.

Authors’ response:

3. Describe the individual classifications of rock mass.

 

This study mainly considers the influence of composite strata and surrounding rock classification, and does not refine the problem of rock sampling and classification, and the follow-up research will be deepened.

Reviewer's note:

Since rock mass classification standards vary from country to country, in describing the presentation of each rock, it is only necessary to explain what each classification covers, e.g. hardness, strength, toughness, difficulty of excavation, geological description... It was not my intention to talk about rock samples and research. From the description that the rock belongs to Grade III, for example, I find it difficult to imagine what this means for tunnelling because, as already mentioned, classifications can vary from country to country. You can also specify the standard according to which the classification was made (with a corresponding reference).

Kind regards.

Author Response

Thank you very much to the reviewer for pointing out this issue, which has been corrected and added the following explanation.

According to the national standard ' Code for geotechnical engineering investigation of urban rail transit ' ( GB50307-2012 ) Appendix F, the geotechnical construction engineering classification of each rock and soil layer revealed by this investigation is carried out.