Meta-Narrative Review of Artificial Intelligence Applications in Fire Engineering with Special Focus on Heat Transfer through Building Elements

Round 1

Reviewer 1 Report

The article provides a comprehensive review of recent scientific publications related to the application of Artificial Intelligence (AI) in Fire Engineering, with a specific focus on Heat Transfer through building elements. The authors highlight the growing interest in using AI and Machine Learning techniques in Fire Engineering research and discuss the potential benefits and challenges associated with their adoption.

 

The authors begin by acknowledging the initial delay in utilizing AI in Fire Engineering due to computational limitations. However, they emphasize that recent advancements in computational power have led to increased exploration and incorporation of AI systems in research processes. The review aims to summarize and organize the findings of various scientific studies in order to facilitate further investigation and refinement.

 

The article highlights several key findings from the review. It notes that a large portion of the reviewed research focuses on the combined study of the structural capacity and thermal performance of building elements. While conventional investigation methods such as destructive experiments or finite element (FE) modeling are commonly used, the use of AI methods is still limited. AI is mostly used experimentally, either as a validation tool or for benchmarking previous experiments or FE models.

 Despite the low uptake of AI, the article emphasizes the high potential of AI in Fire Engineering. It suggests that AI can contribute to the development of overarching databases and models, which can reduce the cost, time, and resource requirements of conducting new research. AI also offers opportunities to widen the scope of research and uncover underlying trends in existing data.

 

 The article identifies the need for more targeted research and design principles, as well as the potential for AI to enhance performance-based design codes and regulations. It suggests integrating the basic principles of heat transfer phenomena into AI algorithm topologies to enhance their effectiveness.

 

 In addition to the review findings, the article discusses the process of meta-narrative review and the importance of a consistent and transparent review protocol. It highlights the need for improved presentation of methodologies in scientific studies to ensure reproducibility and clarity.

 

 

The article concludes by outlining future research directions. It emphasizes the importance of establishing appropriate work protocols for the development and implementation of machine learning algorithms in heat transfer research. The authors propose the construction of databases using information from previous experiments, computational fluid dynamics (CFD) analysis, and published literature to train AI models effectively.

Overall, the article provides valuable insights into the current status of AI applications in Fire Engineering, highlighting the potential of AI in advancing research and addressing the challenges in Heat Transfer analysis. The comprehensive review and identification of research gaps pave the way for future investigations and the integration of AI into the field of Fire Engineering.

 

Minor editing of English language required

Author Response

Please see attached file for covering letter and report with detailed answers to all points raised. Thank you very much.

Author Response File: Author Response.pdf

Reviewer 2 Report

The article reviews recent papers published on effects of fire related heating on masonry walls and use of artificial intelligence algorithms to improve model predictions. Writing a review of this topic is an enduring task and the authors have done a good job at describing the methodology used in selecting the articles to review as well as criteria used to further include it in the review process. Overall it is a well written paper, but here are some points that requires some attention

1. One of the important aspect that was not included in the review is the effect of flame color (sooting vs non-sooting) on heat transfer to the surrounding walls. I would imagine the highly sooting flames (yellow) would have high heat transfer through radiation vs the non-sooting flames (blue) the heat transfer through convection would be dominant. This will lead to a change in fire behavior.

2. The article seems a bit biased towards enhancing the use of AI in the fire technology. Being a review article the authors need to acknowledge some drawbacks of the AI technology as well.

3. Line 424, what is Continuous Fluid Dynamics (CFD)? Do you mean to say Computational Fluid Dynamics (CFD)?

English language is fine but could be improved in some places. A through assesssment would be helpful.

Author Response

Please see attached file for covering letter and report with detailed answers to all points raised. Thank you very much.

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper uses the method of meta-analysis to make a review of Artificial Intelligence applications in Fire Engineering with special focus on heat transfer through building elements. Meta-analysis is generally used for quantitative statistical analysis of literature in the field of medicine or biology. The author has applied this method to the field of fire science research, which is a very creative attempt. The new review approach is a good thing, and the author spends a lot of text explaining the advantages and disadvantages of this approach. In this regard, I have a small questions or suggestions that I hope the author can clarify.

1. There are many methods of AI. What are the author's classification methods and standards? What is pure AI and what are other approaches to AI? I think this issue is very necessary to clarify, after all, many laypeople have heard of AI, but are not very clear about what AI is. In fact, some of the studies cited by the author in the article are essentially optimization of finite element analysis algorithms, and also belong to AI methods in a broad sense. Some researchers may be using these methods themselves, but do not know that the method they are using is AI.

2. A very important principle of meta-analysis is PICOS. This is the key to ensure that the meta-analysis is scientific. May I ask how this article, as a topic in the non-medical field, guarantees PICOS?

3. The essence of meta-analysis is quantitative statistical analysis of literature. However, the quantitative part of the literature review in this paper is few or very limited. Quantitative analysis can not be solved by drawing a few bar charts to classify the literature. It is suggested that the author conduct more in-depth qualitative and quantitative analysis on these selected articles, such as the specific working conditions of artificial intelligence algorithm application and the specific effects of artificial intelligence improvement.

Author Response

Please see attached file for covering letter and report with detailed answers to all points raised. Thank you very much.

Author Response File: Author Response.pdf