Shield Tunnel (Segment) Uplift Prediction and Control Based on Interpretable Machine Learning
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis is a very interesting and meaningful article. It is recommended to revise it and publish it directly. The suggestions are as follows:
1. Improve the clarity of Figures 11 and 14, with clear markings in the figures.
2. Further refine the innovative points of the article.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsGeneral Comments: The purpose of this paper is to address the problem of tube sheet uplift prediction and control in tunneling projects, as well as the auxiliary decision-making system for tube sheet uplift control. Firstly, this paper designs a spatio-temporal data fusion mechanism for section uplift to ensure the correctness of the spatio-temporal relationship corresponding to the input data of the prediction model. Then, a prediction model based on XGBoost is established by considering the influence of multiple types of factors on tube sheet bulging.SHAP is used to track the contribution of the influencing factors to the section bulging. Finally, the article designs a decision process for lifting pressure control. The paper in general has a novel research perspective on the pipe sheet uplift problem. The tube sheet uplift control problem is combined with the formulation of shield attitude control objectives to achieve uplift control and optimize shield attitude control objectives. In this paper, tube sheet uplift in the stabilization stage is studied, which is in line with the content of concern in actual construction and has reference significance for the control of tube sheet uplift in the construction site. Considering the standard of this journal, the reviewer suggests minor modifications to this version. The issues raised and recommendations are as follows.
Comment 1: In Section 2, the influencing factors of segment uplift - factors related to geological conditions are more detailed, which can be summarized for readers from several dimensions, and then the frequency of factors under different physical and mechanical properties of soil can be counted.
Comment 2: In the introduction, most literature on factors affecting segment uplift and prediction of uplift are Chinese. Please add some relevant foreign research literature.
Comment 3: Lines 129 and 383 of the article mention the "uplift control target" concept. Please add an explanation on how this target is determined in 3.4.2. It is mentioned in 3.4.2 that the uplift control target is determined based on the uplift prediction value. Please further clarify how to resolve this.
Comment 4: It is suggested to add the research results of the past five years to the references
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for Authors- The manuscript "Shield Tunnel (Segment) Uplift Prediction and Control Based on Interpretable Machine Learning" is well-written. Following improvements are suggested.
-
Methodological Clarification: It is advisable to provide a more detailed explanation of the machine learning algorithms used, particularly how they were selected and their specific relevance to your study. This will make the methodology section more robust and understandable to readers who may not be familiar with machine learning.
- The literature review must be strenghtened by adding some latest studies for AI application in engineering. e.g., "https://www.sciencedirect.com/science/article/pii/S1674775523002226"
-
-
Enhanced Explanation of Statistical Techniques: It is advisabale to add more statistical paramters to judge the accuracy of the developed model.
-
Comparison with Existing Studies: Expand on how your findings compare with existing studies in the field. This will contextualize your results and highlight your study's contribution to the broader research area.
-
Discussion on Limitations: A more thorough discussion of the limitations of your study and the machine learning model would provide a balanced view and suggest areas for future research.
-
Minor editing required.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf