Digital Twin-Based Intelligent Safety Risks Prediction of Prefabricated Construction Hoisting
Round 1
Reviewer 1 Report
Thank you for submitting your paper to Sustainability.
If I revise some of the following points that I suggest, it would be high-quality research.
In the title, "risk prediction" can be interpreted as a broader meaning, so I recommend you modify it to "safety risks prediction ".
In lines 67-68, the authors said that safety risk prediction results will be shown to managers through the platform. However, I think it is also necessary to show at the construction site worker such as hoist drivers, safety managers, or related work labors.
I can’t find the relevant information in the manuscript, thus additional related descriptions are required.
In line 247, Sa deleted (released twice)
The authors applied to only one case, but there seems to be a limit to its applicability in the practical fields.
Therefore, it is necessary to describe an additional reason (or rationale) for applying to one case in 5.1.
Also, in the conclusion section, the authors describe the case study's limitations.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
KEY STRENGTHS
The subject of this article is ambitious: it presents a use of the application of digital twins (BIM and IoT) to safety on construction sites. It could be said that it tries on the one hand to raise a framework for the application of digital twins in risk management in construction in the lifting of prefabricated elements; and on the other hand to present a model to predict the risk of accidents. That is digital twins and machine learning. Each of them could be a subject for an article in themselves treated in more depth.
The last part of the article uses a case study to illustrate the demonstration.
MAJOR SHORTCOMINGS
- One of the basic principles of a scientific publication is that the documentation provided should be traceable. From the reviewer's point of view, many aspects such as identification of factors, specific data, function generation, etc., are not verifiable.
- The two main objectives of the article are not always developed in accordance with each other.
- This document can be consulted by researchers from different backgrounds (construction management, modeling, risk estimation). For non-experts in applied mathematics, a better explanation of the definition of the model parameters and the definition of the decision function would help to sort out all the information provided. There are sentences with repeated information and yet parameters whose definition does not appear anywhere.
Background and concept. Many references are presented. It is necessary to better specify the relationship between the numerous references used and the objectives of this article. It is difficult to relate the decisions made in the development of the article, such as the selection of the SVM algorithm with the justification provided in the state of the art. There is a certain separation between this section and the justification of the proposed digital twin frameork and the risk prediction model.
Digital twin framework for building hoisting risk prediction framework
The proposal in Figure 1 is very general; it is far from the specific proposal of the article and is more appropriate for the introductory part. If it is kept, more emphasis should be placed on the specific parts for the inclusion of the use of "risk prediction", such as the twin data platform. The way it is stated is very general and in some arrows the directionality is not well understood. It is suggested to merge Figure 1 and 3 into a framework more adapted to the problem.
It is recommended to better document Figure 2. As documented traceability and data sources are important. For example, in the environmetal block it could be discussed whether temperature and wind direction are factors to be considered.
4.1. Collection and transmission of hoisting information
This part of the article (Figure 5) is rather generalist and not very novel. It is suggested to use it to particularize the general information with the equipment actually used in the example of Tianjin Jintangyuan Construction Site.
4.2. Hoisting risk prediction based on multi-source heterogeneous data
As it was mentioned before, a better explanation of the definition of the model parameters and the definition of the decision function would help to sort out all the information provided. From the state of the art “Support vector machine (SVM) algorithm has a good performance in small sample classification operation” is nor inferred. Why radial kernel function is the best for this application?
Case study
Pg. 15. It would be helpful to better detail the nature of the actual data acquired at the Tianjin Jintangyuan Construction Site versus simulated data. Any examples of data capture? Need to attach sample size information and some examples of actual data. Out of the 130 samples some real data should be listed. When listing factors, each and every one of them must be indicated; do not end sentences with "and so on". The selected parameters are argued with "expert consultation", or the information is traced with bibliography or the methodology for the collection of expert data is described.
In the general description of the structure, what exactly does assembly rate is 37.5% in line 342 mean? Is there a change from an initial state?
At the end, to present the case, everything is reduced to overcome a tension value in the sling or in the element (maybe the anchorage?) Should first tell lifting process, critical points, how the critical values are measured or calculated, i.e. present first the type of accident, then the main factor and then the secondary factors. Does the accident occur because the sling breaks or the anchor breaks? According to the document the ultimate stress of concrete should be the critical factor.
Conclusions.
It is identified that the risk factor variables are not sufficiently rigorous. Given that it is a very relevant part of the article, they should be presented with more rigor and thus give greater traceability to the document.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
Some issues and concerns about the manuscript:
1) It is not clear the specific contributions (novelty) of the proposed approach, include this in introduction, and conclusion sections.
2) The literature review (analysis of previos works) does not follow a systematic procedure. How it was collected?
3) What are the benefits and limitations of the approach compared to conventional risk prevention/prediction in the construction sector? describe it in detail please
4) Practical implications using the Digital Twin approach? please describe the potential impacts in sustainability (environment, economic, social).
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 4 Report
The article is interesting and relevant to the field.
The structure, length and the selected methods are suitable for analysis and clearly described. The article has practical value. The authors of the article proposed safety risk management method to predict real-time risk and deduce risk evolution law in prefabricated construction hoisting.
I have the following comments for the authors:
- The article has to be prepared in accordance with the requirements of the journal (it is incorrect according to the template requirements).
- The reference list is up-to-date but should be checked. Reference [36] was provided incorrectly (the authors described it in the wrong way). Should be:
Glatt, M., Sinnwell, C., Yi, L., Donohoe, S., Ravani, B., & Aurich, J. C. (2021). Modeling and implementation of a digital twin of material flows based on physics simulation. Journal of Manufacturing Systems, 58, 231-245.
- Please check all literature list. Similar mistakes occur in other references like [41]
Should be: Tsai, Y. H., Wang, J., Chien, W. T., Wei, C. Y., Wang, X., & Hsieh, S. H. (2019). A BIM-based approach for predicting corrosion under insulation. Automation in Construction, 107, 102923.
- There is a wrong citation in the text of some literature sources: for example Zhou, C. et al. [53] and Zhou, 128 Y., et al. [54], Should be [52] and [53]
Author Response
Please see the attachment.
Round 2
Reviewer 2 Report
No comments
Reviewer 3 Report
The comments were correctly addressed in this new version of the manuscript
