Next Article in Journal
River Flows Are a Reliable Index of Forest Fire Risk in the Temperate Tasmanian Wilderness World Heritage Area, Australia
Previous Article in Journal
A Surrogate Model for Rapidly Assessing the Size of a Wildfire over Time
Article
Peer-Review Record

Customising Evacuation Instructions for High-Rise Residential Occupants to Expedite Fire Egress: Results from Agent-Based Simulation

Reviewer 1: Anonymous
Reviewer 2: Anonymous
Received: 1 March 2021 / Revised: 1 April 2021 / Accepted: 21 April 2021 / Published: 24 April 2021

Round 1

Reviewer 1 Report

Great research topic and well done for the work. As it stands, this paper needs a lot of clarification. Most of my comments below relate to the way the research is described and presented. However, they may highlight shortcomings in the research methodology, as well as some fundamental understanding of the concepts used, most notably BIM and the tools used in the research.

  1. Please describe your work in simple direct sentences.
  2. Review and revise some of the choices of words and phrases to clarify them further and make your points explicit, direct, and easy to understand for your readers.
  3. There have been similar or relate work reported in the literature, but none of these have been mentioned. Consider expanding your literature review to find your research gap. The same research question has been asked many times before, so what sets yours apart from the findings of others?
  4. BIM enables data sharing and collaboration, which is an industry-accepted definition. Please refer to BuildingSMART international for further resources and more definitive details in your revision. In particular, the availability of other data apart from geometry that can be shared with fire engineering simulations, via open standard IFC data exchange.
  5. There have been numerous research reported in literature on the benefits of BIM in Fire Engineering. You should include some of these in your revision, e.g. find keywords such as IFC and fire simulations, IFC and FDS, BIM and fire
  6. There are many grammatical errors in the language. Please check them thoroughly.

More importantly, please consider incorporating the following discussions in your revised paper:

  1. What would be the impact of some evacuees not receiving, ignoring, or simply do not follow the instruction?
  2. Older age group evacuees may need reading glasses to read their cellphones and visibility condition in the evacuation routes may not be ideal for reading instruction. What is the impact of having a more generic instruction broadcasted as well through PA systems?
  3. Where is the location of fire in each scenario of your experiments?
  4. What is the impact of the presence of toxic gases and other fire effects on the ability to evacuate in your experiments? Have you considered performing sensitivity analysis?
  5. It seems that the size of your sample population was determined arbitrarily, although you briefly mentioned a method of determining this, but it is unsure why it was not used.
  6. FDS is another industry de facto simulation tool that also incorporates evacuation model, which had been validated. The evacuation model used in FDS can take into accounts the impact of fire effects.

I look forward to read an improved version of your paper.

 

Author Response

We thank the Reviewer for this constructive review.  An itemised and detailed response to both reviews is in attached file.  

Author Response File: Author Response.pdf

Reviewer 2 Report

 

  1. Lack of grounds for setting evacuation simulation scenarios

The criteria for setting the eight scenarios suggested in this paper should be described.

 

  1. Pre-movement time setting

In this paper, the evacuation delay time is set to 3-5 minutes and 6-8 minutes depending on the presence or absence of sleep. However, since the evacuation delay time can be a problem for the safety of life even by a few minutes, it is necessary to enter an accurate value. In addition, depending on the occupant's location (fire room/non-fire room, fire floor/non-fire floor, etc.), as well as the presence or absence of sleep, there is a difference in the time (perceived time) until the fire is first recognized. Therefore, the evacuation delay time must be set by considering the occupant's location and the presence or absence of sleep.

 

  1. Research methods that do not fit the purpose of the thesis

In the introduction of this paper, it was mentioned that it would conduct research to improve evacuation in high-rise residential buildings by sending personalized guides based on the location of occupants. Based on this, it is expected that research will be conducted focusing on the contents of personalized guides, but unlike the purpose of the study, research on customized guides is omitted, and research to shorten the evacuation time according to the presence or absence of a lift and the location of occupants. Went on. It is not possible to accurately judge what subject this thesis intends to talk about.

 

  1. Walking speed setting

In this paper, an evacuation simulation was conducted for high-rise buildings, and the walking speed at the time of ship evacuation was referenced. High-rise buildings and ships are objects of distinctly different characteristics, and there will be a difference in the walking speed of the occupants accordingly. (ex.Ship-It may shake because it is on the sea) In addition, if you evacuate due to a fire, smoke will be generated in the building, which will cause a difference in walking speed. Considering this, it is necessary to set a walking speed value when smoke occurs in a high-rise building that meets the purpose of this paper. There is inevitably a difference between walking speed on flat ground and walking speed when going down and up stairs. Therefore, simulation should be performed by inputting the appropriate walking speed value on the level and on the stairs.

 

Author Response

We thank the Reviewer for this constructive review.  An itemised and detailed response to both reviews is in attached file.  

Author Response File: Author Response.pdf

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.


Round 1

Reviewer 1 Report

1) Lines 17 & 18, also lines 45 - 47: A fire safety system is not usually installed to prevent a fire from occurring. There are regulatory requirements for fire prevention for certain types of occupancy, the main regulatory goal (in most international fire codes) is to warn occupants early enough so that they have time to escape to a safe place before any evacuation routes become untenable. As correctly stated in Lines 49-51, the secondary goal is usually to protect firefighters when most occupants would have evacuated the building. Some regulatory frameworks also address the protection of neighbouring properties from any structural collapse.

2) Line 43 states that there have been significant changes in fire safety over the past few years. Please highlight which "significant changes" have you observed and discovered from your research? A reference would be useful here. As far as I know, there have been some changes over the last 10 years, but not really significant.

3) Line 56: This statement is only true depending on which country or region of the world you're referring to. This needs some justification. Single storey residential buildings are more common than high-rise buildings in many countries.

4) There are some issues with the numbering in your references. There are many duplications, spelling mistakes, and also formatting issues!

5) Line 96: BIM is multi-faceted, so you should not expect to find ONE definition that fits all. BIM is correctly defined in accordance with which aspect of its application. Refer to the BIM Handbook (Ref [12]). One essential keyword that is missing from your paper is that BIM is intended for "collaboration" (i.e. sharing information) regardless of how it is defined and used.

6) The link to your supplementary materials is broken, so I could not refer to them for review.

7) Lines 597 - 602: ?????

8) There have been several related research reported in the literature, some of which have not been referred to in your paper. More importantly, what is your research contribution? What is your research question?

I'm happy to continue reviewing your draft paper when the above has been addressed and revised.

Look forward to seeing a revised version.

 

 

 

Author Response

The authors thanks the reviewers for their thorough and largely constructive reviews.  The responses below detail how all the points made by rh reviewers were addressed. 

1) Lines 17 & 18, also lines 45 - 47: A fire safety system is not usually installed to prevent a fire from occurring. There are regulatory requirements for fire prevention for certain types of occupancy, the main regulatory goal (in most international fire codes) is to warn occupants early enough so that they have time to escape to a safe place before any evacuation routes become untenable. As correctly stated in Lines 49-51, the secondary goal is usually to protect firefighters when most occupants would have evacuated the building. Some regulatory frameworks also address the protection of neighbouring properties from any structural collapse.

R: The wording in lines 17-18 has been retained as this is part of the Abstract and brevit is important however the more elaborate text in lines 45-47 has been revised according to the Reviewer’s suggestion. Thank you for pointing this out to us

2) Line 43 states that there have been significant changes in fire safety over the past few years. Please highlight which "significant changes" have you observed and discovered from your research? A reference would be useful here. As far as I know, there have been some changes over the last 10 years, but not really significant.

R: Thank you for pointing this out, there have been changes such as emergency management, codes, protection, detection etc… Some examples have now been given.  

3) Line 56: This statement is only true depending on which country or region of the world you're referring to. This needs some justification. Single storey residential buildings are more common than high-rise buildings in many countries.

R: The sentence has been moderated by replacing “are” with “can be”, and: “in urban areas and cities” added - to justify why high-rise buildings are ‘common’.

4) There are some issues with the numbering in your references. There are many duplications, spelling mistakes, and also formatting issues!

R: The manuscript has undergone a thorough proofreading, including references and citations.

5) Line 96: BIM is multi-faceted, so you should not expect to find ONE definition that fits all. BIM is correctly defined in accordance with which aspect of its application. Refer to the BIM Handbook (Ref [12]). One essential keyword that is missing from your paper is that BIM is intended for "collaboration" (i.e. sharing information) regardless of how it is defined and used.

R: That has been corrected.  The passage on the definition of BIM has been revised.  The statement that there is no consensus in the literature about the deifniation of BIM has been followed by a caveat that “This is not surprising…..”.

6) The link to your supplementary materials is broken, so I could not refer to them for review.

R: There is no link to supplementary material, all tables and figures are on the manuscript.

7) Lines 597 - 602: ?????

8) There have been several related research reported in the literature, some of which have not been referred to in your paper. More importantly, what is your research contribution? What is your research question?

R: Done 

I'm happy to continue reviewing your draft paper when the above has been addressed and revised.

Look forward to seeing a revised version.

Reviewer 2 Report

The paper is considerably improved (compared to the initial draft). There are many good aspects and good potentials in the paper. It is well written, and the subject is critically important. There are, however, aspects that the authors can improve before publication of the work.

Major comments:

There's no explanation for the fire behavior and propagation of fire and its products. Further details are needed on the physics as well as computational model. The least requirement would be explaining your assumptions for the case study, such as starting point, fire progression, smoke travelling, visibility, etc. The authors have conducted 12 simulations for each scenario to receive constant results. Yet, it is unclear which occupant parameter is of probabilistic nature, that would cause instability of simulation outputs.The information regarding occupant’s profiles mentioned in the manuscript seems constant. Please elaborate more on which occupant parameters where variable to allow the variation of results between simulations of the same scenario The benefit of BIM in this study is unclear and extremely marginal. It appears that a 3D model would sufficiently serve the objectives. The authors don’t mention the use of building information in fire propagation, nor in any other aspects of the study. A "limitations" section is required to explain the parameters that were not considered in the study such as (but not limited to) Occupants not following the instructions Blockage/damage caused due to fire that could affect instructions Effect of the building material on fire propagation Different occupant densities

Minor comments:

Please review grammatical and spelling mistakes all over (e.g. comma placing in line 15) Line 51-55: revise organization to show that an “acceptable short escape path selection depends on human behavior as well as fire characteristics Line 56-63: revise wording, and revise location as it is out of place Line 65-66: revise wording, “there is a lack of robust” Line 70-72: revise working, unclear Line 128-129: define the “occupant movement manager” Literature Review should be separated from Introduction. The objectives, scope and application must be explicitly articulated in the introduction Line 183 – which “definition of high-rise buildings” was used from literature to consider the building used for the case study as high-rise Fig1- Please change the color of text (from red), since it's not readable ton the dark gray background. Figure 1a & b should be renamed as floor plans and not Revit models Line 250-263: the discussion regarding the accuracy of an evacuation model vs reality, can be reallocated to a “Limitations” section Line 280-284: please clarify if receiving instructions during a fire should affect the pre-movement time or not? Line 299-303: please revise the format Lines 321-328: please clarify the definition of a “room” where the Randomization option of Pathfinder is applied. Does the room mean the entire apartment of an occupant or the room? What is the significance of using this Randomization feature on the results Some recent works which are very closely related to this work are missing from the reviewed literature. The followings are some examples; but the authors must also make sure to cover the other relevant publications.  Mirhadi, et al. 2019 "IFC-centric performance-based evaluation of building evacuations using fire dynamics simulation and agent-based modeling", Automation in Construction Eftekharirad, et al. 2018 "Linking sensory data to BIM by extending IFC – case study of fire evacuation", ECPPM 2018 Sun, et al. 2019 "A BIM Based simulation framework for fire evacuation planning"

Author Response

The paper is considerably improved (compared to the initial draft). There are many good aspects and good potentials in the paper. It is well written, and the subject is critically important. There are, however, aspects that the authors can improve before publication of the work.

Major comments:

There's no explanation for the fire behavior and propagation of fire and its products. Further details are needed on the physics as well as computational model. The least requirement would be explaining your assumptions for the case study, such as starting point, fire progression, smoke travelling, visibility, etc. The authors have conducted 12 simulations for each scenario to receive constant results. Yet, it is unclear which occupant parameter is of probabilistic nature, that would cause instability of simulation outputs.  The information regarding occupant’s profiles mentioned in the manuscript seems constant. Please elaborate more on which occupant parameters where variable to allow the variation of results between simulations of the same scenario The benefit of BIM in this study is unclear and extremely marginal. It appears that a 3D model would sufficiently serve the objectives. The authors don’t mention the use of building information in fire propagation, nor in any other aspects of the study. A "limitations" section is required to explain the parameters that were not considered in the study such as (but not limited to) Occupants not following the instructions Blockage/damage caused due to fire that could affect instructions Effect of the building material on fire propagation Different occupant densities. 

 

R: All major corrections have been done. Limitation second has been added. The authors explained why 12 simulations. The use of BIM has been clarified in several places to refer to ‘3D object-oriented models’. Whereas a non-BIM based 3D model could suffice for agent-based simulation, it would not be advantageous to building owners or facility managers (as suggested by Hackitt Review) who would benefit by integrating the data contained in such models into maintenance and safety strategies.

Minor comments:

Please review grammatical and spelling mistakes all over (e.g. comma placing in line 15) Line 51-55: revise organization to show that an “acceptable short escape path selection depends on human behavior as well as fire characteristics Line 56-63: revise wording, and revise location as it is out of place Line 65-66: revise wording, “there is a lack of robust” Line 70-72: revise working, unclear Line 128-129: define the “occupant movement manager” Literature Review should be separated from Introduction. The objectives, scope and application must be explicitly articulated in the introduction Line 183 – which “definition of high-rise buildings” was used from literature to consider the building used for the case study as high-rise Fig1- Please change the color of text (from red), since it's not readable ton the dark gray background. Figure 1a & b should be renamed as floor plans and not Revit models Line 250-263: the discussion regarding the accuracy of an evacuation model vs reality, can be reallocated to a “Limitations” section Line 280-284: please clarify if receiving instructions during a fire should affect the pre-movement time or not? Line 299-303: please revise the format Lines 321-328: please clarify the definition of a “room” where the Randomization option of Pathfinder is applied. Does the room mean the entire apartment of an occupant or the room? What is the significance of using this Randomization feature on the results Some recent works which are very closely related to this work are missing from the reviewed literature. The followings are some examples; but the authors must also make sure to cover the other relevant publications.  Mirhadi, et al. 2019 "IFC-centric performance-based evaluation of building evacuations using fire dynamics simulation and agent-based modeling", Automation in Construction Eftekharirad, et al. 2018 "Linking sensory data to BIM by extending IFC – case study of fire evacuation", ECPPM 2018 Sun, et al. 2019 "A BIM Based simulation framework for fire evacuation planning" 

R: All minor corrections have been made

Reviewer 3 Report

Thank you for this opportunity to review. While evacuation from high-rise building and evacuation modeling are important subjects on which to publish, this article contains multiple issues that should be addressed before publishing. I have included significant comments below, followed by minor issues.

 

Significant comments/issues:

The authors do not include any of the seminal references required when discussing evacuation or safety of individuals in high-rise structures. They have Kuligowski’s SFPE handbook chapter listed in their reference list – and if they reviewed that chapter, they would see the different authors who study in this field and could have been reviewed/referenced as background for this article. Additionally, the authors state that building occupants “panic” in fires (e.g., line 69 on page 2 or Line 641) or injuries and fatalities in fires are a result of “chaos”, and it is clear from many references in the field (Quarantelli, Fahy and Proulx, Wenger, etc.) that individuals are unlikely to panic in building fire situations (this is also discussed in Kuligowski’s handbook chapter). In fact, fire evacuations are often orderly and the individuals are more likely to help one another out than not. The authors seem to reference BIM or agent-based modeling references throughout the paper, even when they are discussing human behavior. This leaves me wondering if they have a sufficient understanding of human behavior in fire to produce this article.

 

Similar to the comment above, I am left wondering why so many important topics are left out of the Introduction/Literature review. At line 171 (end of the introduction), I did not think that the authors have supported this statement – that, i.e, “…it can be argued that providing customized instructions about egress routes…” They did not discuss any studies on occupant route choice and the factors that affect route choice, for example.

 

The authors mentioned that the simulation strategy was based on Tavares and Ronchi but do not give much more information than that. They should have provided some details in Section 2 on: the scenarios chosen and why, what were the “instructions” given to occupants, what were the model inputs provided to Pathfinder, and how do the model inputs change per scenario? None of this information is provided making this article incomplete.

 

My biggest issue with this article is the discussion of the results. The authors state that they are testing “how different instructions affect evacuation time.” However, after reading the article in full, I realized that they are not testing evacuation instructions. Instead, they are testing the ways in which Pathfinder simulates population splits to various exits in their example building. This is a big difference. Because of this, they are only reporting results on evacuation timing based on the pre-simulation population splits that they provided as inputs to Pathfinder. The authors originally stated that this study was one that could test BIM and the provision of evacuation instructions to occupants, and unfortunately, that is not accurate. I do not understand how this article or study has anything to do with the impact of evacuation instructions provided via BIM technology on evacuation route choice/timing. The concept of evacuation instructions is a deeper concept then just splitting occupants to different routes. I thought that the authors were discussing the impact of providing different types of evacuation information (e.g., warning content and style of the warning) via different channels (e.g., mobile devices) and how that would impact evacuation decision-making. However, as stated earlier, this article is simply about a model’s results from pre-determined population splits. The article turned out to be very different than what I thought and also different than what the authors stated as the original purpose. This is problematic.

 

Also troubling is the way that the authors described their results from the Pathfinder model. Essentially, from the article, I understand that the authors ran 12 simulations a number of different times and got modeling results (Table 3). However, they report these results as if they were actually observing evacuation behavior from a real event. E.g., line 366 states that an exit “had the most occupants exiting through it probably due to familiarization and shortest distance.” Why probably? They should know the underlying reasons for these results because they are directly related to the inputs provided to the model and the model’s underlying algorithm (i.e., how it moves people around the building spaces). Their results have little to do with human decision-making and behavior and everything to do with how Pathfinder simulates occupant route choice based on the population splits pre-determined by the authors. However, the authors consistently state that “occupants…where given basic evacuation instructions” (Line 403) or that they are exploring “the possibilities that some able-bodied occupants may still use lifts…”. Line 419 states that “occupants preferred to use…” or on line 454: “61 occupants who initially decided to use the main stairs decided to use the lift…” rather than reporting that the agents were following Pathfinder’s algorithms.

 

The authors should be very clear up front about the ways in which Pathfinder simulates exit choice. Just from their website (https://www.thunderheadeng.com/pathfinder/pathfinder-features/), one can see that “By default, each occupant (agent) uses a combination of parameters to select their current path to an exit. The parameters include: queue times for each door of the current room, the time to travel to each door of the current room, the estimated time from each door to the exit, and the distance already traveled in the room. The agent responds dynamically to changing queues, door openings/closures, and changes in room speed constraints (simulating smoke and debris). The user can modify the default parameter weights to change the behavior. For example, occupants can neglect queues and only look for the closest exit. Alternately, occupants can be assigned specific goals (such as, go to a location and wait) or specific exits. For example, most occupants of lower floors can be assigned to use stairs, while most occupants of higher floors can be assigned to use elevators.”

 

-Based on the website text, above, the authors should be clear with the reader that Pathfinder is not a true agent-based model in which the agents learn during the simulation such that emergent behaviors could occur. There is an underlying algorithm that Pathfinder uses to “lead” occupants to certain exits – e.g., queue times, travel times, distance already traveled. This is NOT actually based on human behavior or data collected since it is unlikely that people can calculate the distance that they have already traveled or still need to travel or have the capacity to calculate queue times. Also, the user can modify these parameter weights to change behavior. Did the authors use the default weights or modify these in any way? None of these topics are discussed in their paper.

 

-Also, on line 571, the authors state that “the agent-based simulated occupants had calculated the queue time and the distance to be covered in exiting the building, ….” however, neglected to discuss that this is due to the underlying model algorithm. Later, the authors refer to occupants’ behavior as “poor decision making”, without acknowledging that again, this goes back to the inputs that they specified and the model’s algorithms.

 

Because of the many comments that I have listed above, I do not think that the authors have supported their statement on Line 604: “this research has investigated how sending customized instructions to high-rise occupants during fire evacuation could reduce evacuation time…” Because of my comments above, I do not think that they authors can make any conclusions about occupants’ actual behavior or the impact of BIM technology on behavior.

 

 

Minor issues:

The abstract should provide information on what the authors found. The sentences on lines 31-35 are written in future sense (e.g., “could be” or “will attempt”, etc.). Is this what the authors found? After reading these statements, I am left wondering – what did the authors actually find as part of this work? Lines 48-51 – how is [2] an appropriate reference for your work? This is a paper on fire safety for tall timber structures? On Line 55, is [3] the appropriate reference here? The authors are talking about human behavior in fire topics and there are so many seminal human behavior in fire studies that they neglect to reference. I am not sure that the authors understand enough about human behavior in fire based upon their article. On line 126, the authors mention a “study” and then make reference to an NFPA code. This is confusing since a code is not a study. On line 153, the authors make reference to Kuligowski’s handbook chapter, again as “a study”. This chapter is a review of several studies. On line 165, the authors state that in residential buildings, occupants are “often not ready to evacuate.” In which buildings are people ready to evacuate? People often have to go through a series of pre-decisional and decisional processes before they decide to evacuate – and these processes take time to complete. There is often a delay in evacuation response independent of building type. The reference list at the back of the article is not numbered properly.

Author Response

N/A

Round 2

Reviewer 1 Report

A much-improved version. However, a couple of my previous comments have not been responded to adequately, as follows:

The link to your supplementary materials is broken, so I could not refer to them for review.

R: There is no link to supplementary material, all tables and figures are on the manuscript.

Look at your original manuscript line 718 and revised manuscript line 666 for the Supplementary Materials that both point to a broken link!!

7) Lines 597 - 602: ?????

No response to this.

8) There have been several related research reported in the literature, some of which have not been referred to in your paper. More importantly, what is your research contribution? What is your research question?

R: Done

"Done" is not an adequate response to a reviewer's comment. Please elaborate on how you address the comment and help your reviewer by clearly pointing out your changes on the lines on the revised manuscript.

I'm happy to review again if you could respond adequately and appropriately address all matters raised by your reviewers. 

 

Back to TopTop