Optimal exit choice during highway tunnel evacuations based on the fire locations.

In the case of a fire, the choice of exit in the highway tunnel is strictly limited by fire location, which seriously affects the evacuation time. A spontaneous or disorderly exit choice might result in a decreased evacuation efficiency and utilization rate of exits. In this paper, we propose a strategy to obtain the optimal exit choice based on fire location during highway tunnel evacuations. In our strategy, first, the vehicle distributions and locations of evacuating occupants are determined in the traffic simulation program VISSIM. The evacuation simulation software BuildingEXODUS is employed to obtain the corresponding parameters of the evacuation process and analyze the impacts of different fire locations on the evacuation time. During the analysis, the optimal productivity statistics (OPS) is selected as the evaluation index. Then, the feature points of the crowding occupants are captured by the fuzzy c-means (FCM) cluster algorithm. Next, based on the feature points, the relationship between the location of the fire and boundary of the optimal exit choice under the optimal OPS is obtained through the polynomial regression model. It is found that the R-squared(R2) and sum of squares for error (SSE) of the polynomial regression model, reflecting the accuracy estimation, are 98.02% and 2.79×10-4, respectively. Moreover, different fire locations impact the evacuation time of tunnel entrance and evacuation passageway. This paper shows that the location of the fire and boundary of optimal exit choice have a negative linear correlation. Taking the fire 110 m away from the evacuation passageway as an example, the OPS of our strategy can be decreased by 35.6% when compared with no strategies. Our proposed strategy could be applied to determine the location of variable evacuation signs to help evacuating occupants make optimal exit choices.

Introduction

With the rapid development of highways, traffic safety has become a key concern of the participants and managers [1, 2]. As a bottleneck section of highway, highway tunnels have a higher risk of being the site of crashes than open roads [3]. According to data from Jiangsu Expressway Co. Ltd., China, for a section of provincial highway S38 from Tianwang to Maoshan, the tunnel section accounts for approximately 2.25% of the whole line, while the accident rate was over 10% between 2015 and 2019. Once there is an accident caused by the spontaneous combustion of a heavy truck, in the case of fire, it is easy to spread the fire fast and difficult to escape in the highway tunnel because of closed interior space, limited ventilation and adverse spatial structure [4, 5]. Thus, efficient use of exits during highway tunnel evacuations under the fire environment is an important research topic.

Although lots of attention is paid to the study of exit choice in the building and room [6, 7], there are few systematic studies of exit choice in the highway tunnel. On the one hand, the existing modelling methods of exit choice under the fire environment is concentrated on influencing factors through experiments, without considering the fire locations. For example, Fridolf et al. [8] studied the influence of different way-finding installations on exit choice in smoke-filled tunnels. Furthermore, Ronchi et al. [9] specially conducted the impact of the emergency exit portal complemented with information signs on exit choice. Lovreglio et al. [10] investigated the effect of environmental conditions (presence of smoke, emergency lighting and distance of exit) and social factors (interaction with evacuees) to build a mixed logit model for predicting exit choice. Meanwhile, they applied random utility theory to study the effect of herding behavior on exit choice [11]. Haghani et al. [12] studied spatial distance, congestion level, exit visibility and the size of moving flows on exit choice using under face-to-face interviews and field-type experiments. After that, they studied exit choice of crowds under high and low levels of urgency[13]. However, they mainly concentrated on buildings with a balanced ratio of length to width. And the fire is easier to spread in the highway tunnel. Therefore, more concerns and researches are needed to focus on the exit choice during highway tunnel under fire environment.

Due to the complexity of conducting field research, another method used to model exit choice is simulation modeling. Different simulation models have been established and used by researchers, such as BuildingEXODUS [14] and Pathfinder [15]. Cuesta et al. [16]observed children from 6 to 16 years of age during the evacuation exercises. They used the BuildingEXODUS to simulate the evacuation of children in school and verify the accuracy of the model predictions. Hunt et al. [17]used the BuildingEXODUS evacuation model to represent moving objects and simulated the evacuation of the patients and the staff in hospital. In addition, cellular automata [18, 19], ant colony algorithm [20] and agent-based model [21] are utilized to study influences of different factors on exit choice. These models are used to simulate the spontaneous behavior of human when making exit choice. Moreover, these approaches simulate the real spontaneous exit choice from the theoretical point of view to ensure the accuracy of the simulation.

However, the spontaneous exit choices of evacuating occupants sometimes cause the congested exits because of herding behavior [22]. Therefore, from the perspective of management, manual intervention on exit choice is of great application value to engineering compared with spontaneous exit choice. For example, broadcast and evacuation signs can be used to guide evacuating occupants at different locations to different evacuation exits, which can improve the evacuation efficiency and survival rate of evacuating occupants.

The aim in the study is to determine the optimal exit choice for different fire locations and evacuating occupants at different positions. We propose a strategy to obtain the optimal exit choice during highway tunnel evacuations based on fire location. The research framework includes three parts. First, we use BuildingEXODUS and VISSIM to obtain the parameters about vehicle distributions, locations of evacuating occupants, evacuation process and analyze the impacts of different fire locations on evacuation time. Then, the OPS output from BuildingEXODUS is selected as the evaluation index. Next, FCM cluster algorithm is used to capture the feature points of the occupants. Based on the feature points, the relationship between the location of fire and boundary of optimal exit choice under the optimal OPS is obtained through polynomial regression model.

The main contributions of this paper can be summarized as follows. First, we connect the vehicle distribution of VISSIM with the enclosure and movement modules of BuildingEXODUS to realize the exchanges of simulation data, which include output parameters about vehicle distributions, locations of evacuating occupants and evacuation process. Second, we make use of FCM cluster algorithm to capture the feature points of the occupants, which is valid to describe the characteristics of the crowding occupants, and the feature points could be convenient to calculate. Third, we apply the polynomial regression model to solve the optimal exit choice, which could provide an explicit strategy for evacuating occupants to determine optimal evacuation exit.

The remainder of this paper is organized as follows. Section 2 introduces problem statements. In Section 3, we explain the proposed method in detail. The input parameters and results of model case study in Maoshan Tunnel are analyzed in Section 4. Finally, conclusions are given in Section 5.

Problem statement

In Fig 1, we demonstrate occupants evacuating in the highway tunnel. When a fire breaks out between the tunnel exit and evacuation passageway, evacuating occupants located upstream choose to drive away. The occupants located from fire to evacuation passageway are supposed to escape through the evacuation passageway while the occupants located from evacuation passageway to entrance may choose to escape through the evacuation passageway or entrance.

Fig 1

Occupants evacuating in the highway tunnel.

In the case of fire, the evacuating occupants are stimulated by the crisis environment and subconsciously flee to the entrance in the reverse direction, resulting in the inefficient utilization of evacuation passageway and longer evacuation time [23]. Therefore, it is necessary to study optimal exit choice for occupants during highway tunnel evacuations.

This work focuses on evacuating occupants in different positions to determine optimal evacuation exit. The OPS is used to measure the utilization rate of evacuation passageway and exit [24], and can be expressed as follows: where E denotes the number of used exits, EET is the time of the last occupant evacuating from exit e, and TET represents the total evacuation time, which is equal to the maximum value of EET.

The calculation of OPS includes not only total evacuation time, but also the time of the last occupant evacuating from exit, which reflects the utilization rate and evacuation efficiency. The range of OPS is from 0 to 1. In general, the lower OPS is, the more balanced and higher efficiency of evacuation is. If evacuating occupants ignore one of the exits or gather at the same evacuation exit, the index could reflect the low efficiency of evacuation.

Based on the above analysis, the effects of different fire locations and exit choices of evacuating occupants on OPS can be studied. The relationship between the location of the fire and boundary of optimal exit choice can be built by minimizing OPS. Furthermore, the boundary of optimal exit choice for occupants is determined, i.e., occupants located on the left of boundary evacuate through the evacuation passageway, while occupants located on the right should evacuate through the tunnel exit.

In Fig 2, we establish the mathematical model of occupants evacuating in the highway tunnel. As illustrated in Fig 2, the midpoint of the evacuation passageway is defined as the origin. L denotes the distance from origin to tunnel entrance, and L denotes the distance from origin to fire location, respectively. L means the distance from origin to boundary and L represents the distance from origin to fire location, respectively. To avoid using the specific tunnel length as a constraint interval and increase the applicability of the model, we normalize the data L and L into the same interval, i.e., x = L/L and y = L/L. x describes the ratio of the distance from origin to boundary and that from origin to tunnel entrance, indicating the proportion of occupants evacuating from the evacuation passageway. The larger x is, the more occupants evacuate from the evacuation passageway. y represents the ratio of the distance from origin to fire location and that from origin to tunnel exit, describing how far the fire to the evacuation passageway. The larger y is, the farther the fire is from the evacuation passageway.

Fig 2

Mathematical model of occupants evacuating in the highway tunnel.

Method

Simulation softwares

Due to safety concerns and the complexity of conducting field research, BuildingEXODUS [14] is employed in this study to simulate the evacuation process and evaluate the evacuation efficiency in the tunnel. BuildingEXODUS could repeatedly simulate the evacuation and solve the conflicts between field evacuation and traffic disruption, thereby saving costs and time [25]. It is adopted in modeling and can include occupants, actions, behaviors, disasters, and so forth [26].

The car-following model of VISSIM could easily simulate and extract the trajectories of vehicles[27, 28]. Therefore, the VISSIM is used to display vehicle distributions and operations in real time when the fire occurs [29], determining the locations of evacuating occupants and obstacles in the tunnel. The initial locations of evacuating occupants are calculated based on vehicle distributions. In addition, the initial locations of evacuating occupants and obstacles are associated with the enclosure and movement modules in BuildingEXODUS.

Note that the reliability of the evaluation results partly depends on the accuracy of model input, such as geometric parameters and traffic volumes. Under this condition, the input of model in this study should be carefully calibrated and validated based on monitoring data and literature data.

FCM cluster algorithm

In the process of evacuation, the evacuating occupants tend to choose the most congested exit rather than another exit because they trust the majority of other occupants’ behaviors and consider the congested exit to be the right exit [30]. Thus, the clustering phenomenon easily occurs. To capture the characteristics of the crowding occupants, the FCM algorithm employs the fuzzy theory in the determination of the membership relationship between data elements in the cluster algorithm[31]. It converts the hard division of the data elements in the traditional cluster algorithm with the membership degree between different clusters into soft division[32, 33].

The algorithm is described as follows. The sample space P has N data elements. The number of clusters is C. The objective function of the FCM algorithm J is defined as where m is a fuzziness index, controlling the degree of blurring of the clustering results, u represents the membership of cluster j of sample i, P is the i sample and V denotes the centroid of cluster j, respectively.

The FCM algorithm minimizes the objective function J by iteration. For a sample P, the sum of the membership degrees for each cluster is 1, that is,

To find the minimum value of the Eq (2), the necessary conditions are as follows:

The termination condition of the iteration is shown below: where δ is the error threshold.

The clustering effect is evaluated using the Xie-Beni effectiveness index V [34]. The smaller the index is, the better the clustering effect is. The index V can be expressed as follows:

The FCM cluster method is used to classify different x values based on OPS and y. Then, the regression analysis is carried out for these feature points to research the relationship between OPS and x, y.

Polynomial regression model

Polynomial regression is a kind of linear regression [35, 36]. It is assumed that given N data points, we search an appropriate polynomial of q-order to describe the relationship between independent variable OPS and the dependent variable x, y. Since x and y affect OPS simultaneously, the interaction between x and y is considered to have an impact on OPS. The general form of polynomial regression can be expressed as follows: where ε is the random deviations or residuals, x and y are the independent variables, OPS refers to a dependent variable, β is the coefficient of the polynomial and q is the order of a polynomial, respectively.

Polynomial fitting is used to construct a q-order polynomial, and does not require all data points to be strictly. However, it is expected to pass most of these data points as many as possible, so that the residual error between the estimated parameter value and the actual value can be minimized.

In this paper, the curve fitting tool in MATLAB is exploited to estimate the parameters β. The following step is used to solve the relationship between x and y when OPS = f(x,y) obtains the optimal value in region D (0 < y ≤ 1, 0 ≤ x ≤ 1).

Considering the above analysis, the optimal exit choice during highway tunnel evacuations can be expressed as follows: where i is the location of the evacuating occupants; Z(i) denotes that occupant located i escapes from which evacuation passageway; Z(i) = 0 if evacuating occupants escape from evacuation passageway and 1 if evacuating occupants escape from the tunnel entrance; L refers to the distance from origin to tunnel entrance; L is the distance from origin to tunnel exit; x stands for the ratio of the distance from origin to boundary and that from boundary to tunnel entrance; y is the ratio of the distance from origin to fire location and that from origin to tunnel exit.

Model case study

The highway tunnel

Maoshan Tunnel is one of the earlier highway tunnels built in Jiangsu Province, China. It has been open to traffic for more than 10 years. Because the tunnel is short, it is designed without a ventilation system, resulting in non-flowing internal air. In the case of a fire, the dispersion of smoke tends to reduce the visibility of evacuating occupants and their perception of distance, increasing the difficulty of escaping. Thus, it is of great significance to select a reasonable and effective evacuation exit to improve the efficiency of highway tunnel.

The highway tunnel considered in this study is the westbound tunnel of Maoshan Tunnel, as illustrated in Fig 3. The total length of the tunnel is 585 m. There is one evacuation passageway between the exit and entrance, which is located 235 m from the entrance and 350 m from the exit. The width of the crossing is 1 m, and its height is 2.87 m.

Fig 3

Horizontal and cross section of Maoshan Tunnel.

Model input calibration

A set of simulation parameters is required to ensure the facticity of evacuation under the fire environment in the tunnel. In this study, monitoring data and literature data are collected to calibrate those simulation parameters input into BuildingEXODUS and VISSIM before simulation. Literature data are used to provide fire parameters and evacuation speed to ensure the authenticity of simulation. A previous emergency exercise of Maoshan Tunnel is used to validate the evacuation process in the highway tunnel under fire environment.

Parameters of the fire

The fire is an important input factor that should be carefully calibrated. The parameters of fire determined based on both monitoring and literature data are input into the hazard module of BuildingEXODUS. Assuming that the fire in this study is caused by gasoline combustion. The size of the fire is 4.6×1.7×1.5 m3. The heat release rate is 50 MW and fire growth coefficient is 0.188 [37].

Traffic volumes and vehicle distributions

To study the evacuation under the worst circumstances, the maximum traffic volumes should be taken as the total number of vehicles. The traffic volumes and vehicle types of Maoshan Tunnel at three different time periods, morning peak, flat peak and evening peak were recorded for October 2019 (31 days), as is shown in Table 1. It can be seen that the maximum traffic volume was 1356 veh/h which appeared during the evening peak.

Table 1

Traffic volumes and vehicle proportion at three different time periods.

Time periods	Parameter	Passenger vehicle	Coach	Heavy truck	Articulated vehicle	Traffic volumes
Flat peak	Traffic volumes	571	27	151	32	781
	vehicle proportion (%)	73.1	3.5	19.3	4.1	100
Morning peak	Traffic volumes	1000	53	222	28	1303
	Vehicle proportion (%)	76.7	4.1	17.1	2.1	100
Evening peak	Traffic volumes	1131	26	160	39	1356
	Vehicle proportion (%)	83.4	1.6	10.3	2.8	100

Vehicle distribution is used to describe the spatial location of vehicle in the tunnel, which is another key parameter to determine the initial locations of evacuating occupants. According to the previous emergency exercise of Maoshan Tunnel, the fire alarm rings 30 seconds after a fire occurs. It takes evacuating occupants 30 seconds to escape from the cars after noticing the fire. The maximum traffic volume is input into the VISSIM software to simulate the distribution of vehicles from freely flowing to congested traffic after the fire happens.

The number and initial locations of evacuating occupants

The number of evacuating occupants is determined by the maximum traffic volumes, vehicle proportion, the capacity of different vehicles and passenger proportion of vehicles, and can be expressed as follows: where Q is the number of evacuating occupants; μ1, μ2, μ3, and μ4 denote the vehicle proportion of passenger vehicles, coaches, heavy trucks and articulated vehicles, respectively; m1, m2, m3, and m4 are the capacity of passenger vehicles, coaches, heavy trucks and articulated vehicles; η1, η2, η3, and η4 refer to passenger proportion of passenger vehicles, coaches, heavy trucks and articulated vehicles; n represents maximum traffic volumes.

These parameters in (9) are determined by monitoring data. There are four types of vehicles: articulated vehicles, heavy trucks, coaches and passenger vehicles. The outer dimensions and capacity of different vehicle types are determined by the Chinese Technical Standard of Highway Engineering, as shown in Table 2.

Table 2

The outline dimensions and capacity of different vehicle types.

Vehicle types	Length	Width	Height	Capacity
Passenger vehicle	6	1.8	2	6
Coach	13.7	2.55	4	50
Heavy truck	12	2.5	4	2
Articulated vehicle	18.1	2.55	4	2

The passenger proportion of vehicles is set to 50% for passenger vehicles, 90% for coaches, and 100% for heavy trucks and articulated vehicles. Meanwhile, results of the calculation are rounded to integers. The initial locations of evacuating occupants are calculated by vehicle distributions.

Evacuation speed

The evacuation speed in a dark tunnel is extremely slow and must be determined accurately to assess tunnel fire safety. According to [38], age and gender have little effect on evacuation speed. Evacuation speeds of evacuating occupants are approximately lognormal within a 95% interval (using the 2.5th and 97.5th percentiles of the distribution as endpoints, which are 0.24 and 0.88 m/s with a mean value of 0.49 m/s). In this study, the average speed is 0.49 m/s, which is input into the movement module of BuildingEXODUS to simulate the movement of occupants in the highway tunnel under fire environment.

The analysis is conducted based on the above parameter settings. When a fire breaks out in the evacuation passageway, evacuating occupants cannot evacuate through it. Due to the size of fire, 2 m away from the evacuation passageway (y = 0.05) is selected as the location of the initial fire. To appropriately indicate the trend of OPS, y varies with a step of 0.1. Meanwhile, as the crowd is concentrated, for different y, x varies with a step of 0.05. Finally, the corresponding x, y and OPS are outputted from BuildingEXODUS (S1 Table). The process of solving x, y and the corresponding OPS is shown in Fig 4.

Fig 4

Process of solving x, y and the corresponding OPS.

Results and discussion

The impact of the fire location

The location of fire has a significant impact on the evacuation time [39]. The paper delves into the impact of the distance between the fire and evacuation passageway on the evacuation time of tunnel entrance and evacuation passageway. The results are shown in Fig 5.

Fig 5

Evacuation time and OPS of tunnel entrance and evacuation passageway.

As shown in Fig 5, when y varies from 0.05 to 0.35, the evacuation time for the evacuation passageway and tunnel entrance changes significantly with x. In contrast, as y varies from 0.45 to 1, there is less significant change with x in the evacuation time for the evacuation passageway and tunnel entrance. This indicates that the impacts on the evacuation time in the evacuation passageway and tunnel entrance vary according to different fire locations. The shorter the distance between the location of fire and origin is, the greater the fluctuation of the evacuation time is. As the distance between the location of fire and origin gets farther, the total evacuation time also increases. The potential reason may be that more occupants trapped from fire to evacuation passageway may choose to escape through the pedestrian passageway. This phenomenon could decrease the capacity of the evacuation passageway and increase the total evacuation time.

Optimal cluster number and coordinates

To find the most suitable cluster according to Eq (6), we calculate V of each cluster (S1 Code), as shown in Fig 6, where the number of clusters C varies from 2 to 11.

Fig 6

V of different clusters.

It can be seen from Fig 6 that when clusters number C is 7, V reaches the minimum value, corresponding to the optimal cluster number C*, as shown in Fig 6. Therefore, the optimal cluster number is 7. The corresponding x, y, and OPS coordinates are (0.13,0.25,0.55), (0.05,0.87,0.29), (0.25,0.32,0.16), (0.09,0.55,0.12), (0.40,0.10,0.36), (0.18,0.90,0.53) and (0.45,0.33,0.53), respectively.

The relationship between x and y based on the optimal OPS

The relationship between OPS and x, y is established with the polynomial regression model. In Fig 7, we show the polynomial regression model of x, y and OPS. There is a correlation between x and y. The estimation results of each parameter are shown in Table 3.

Fig 7

Polynomial regression model of x, y and OPS.

Table 3

The results of each parameter.

Parameter	Estimation Value	Confidence interval
β 0	2.45	(2.28, 4.72)
β 1	-9.91	(-13.92, -3.37,)
β 2	-5.52	(-16.96, 5.91)
β 3	12.62	(11.35, 13.60)
β 4	9.71	(-10.36, 29.77)
β 5	3.52	(1.07, 4.11)

From Table 3, the R2 and SSE of the model are 98.02% and 2.79×10−4, respectively. The dependent variable can be estimated from the model, which is generally available with SSE close to 0. Therefore, the regression model under the influence of x and y can be expressed as follows:

The closer OPS is to 0, the more balanced and higher efficiency of evacuation. The relationship between x and y is shown in Fig 8 by drawing the curves of different y when OPS approaches 0.

Fig 8

Different y for OPS = f (x, y).

Since fire detectors in the highway tunnel can find the location of the fire [40], y is known and we take the first and second derivatives of the model OPS = f (x, y) with respect to x in region D (0 < y ≤1, 0 ≤ x ≤ 1), i.e.,

Then, letting to solve the relationship between x and y under the condition of optimal OPS, we obtain

It can be seen in (12) that y and x have a negative linear correlation, which means that the farther the fire occurs away from the evacuation passageway, the closer the boundary is to the evacuation passageway, and the fewer occupants evacuate through the evacuation passageway. This is consistent with the previous study on fire location and evacuation time. Therefore, the optimal exit choice during highway tunnel evacuations can be expressed as follows:

If the fire occurs 110 m away from the evacuation passageway (y = 0.31), x is calculated as 0.27 from Eq (13). Thus, occupants located 63 m away from the origin escape through the evacuation passageway and those who are 63–235 m away from the origin escape through the entrance of the tunnel. The OPS of our strategy is 0.195, while the OPS of a spontaneous and disorderly exit choice is 0.303, lowering OPS by 35.6%. Our strategy not only increases the utilization rate of exits, but also effectively improve the evacuation efficiency when compared with the spontaneous and disorderly exit choice.

According to the results of this study, the location of variable evacuation signs can be adopted to prompt, as shown in Fig 9.

Fig 9

Arrangement of evacuation signs.

Conclusion

In this study, taking a highway tunnel in China as an example, we propose a strategy to choose the optimal exit during highway tunnel evacuations based on the fire location. First, we combine the vehicle distribution of VISSIM with the enclosure and movement modules of BuildingEXODUS to utilize the exchanges of simulation data. Next, the feature points of the crowding occupants are captured by the FCM cluster algorithm when the cluster number is 7. Finally, we apply the polynomial regression model to solve the optimal exit choice based on the feature points. It is found that the R2 and SSE of the polynomial regression model, reflecting the accuracy estimation, are 98.02% and 2.79×10−4, respectively.

Different fire locations are found to impact the evacuation time via evacuation passageway and entrance. The shorter the distance between the location of fire and the origin (the midpoint of the evacuation passageway) is, the greater the fluctuation in the evacuation time is. Under this circumstance, the location of the fire and boundary of optimal exit choice have a negative linear correlation, indicating that the farther the fire occurs from the evacuation passageway, the fewer occupants evacuate through the evacuation passageway. Taking a fire 110 m away from the evacuation passageway as an example, the OPS of our strategy can be decreased by 35.6% compared with no strategy. Overall, the location of the fire is detected by fire detectors in the highway tunnel. The optimal exit choice during highway tunnel evacuations based on fire location is obtained by our strategy, which could provide valuable information for optimal exit choice and arranging variable evacuation signs in the highway tunnel. In our study, vehicle distribution is an important factor that determines the spatial location of vehicle and initial locations of evacuating occupants. Hence, effective information warning and traffic guidance could reduce the number of trapped vehicles and reduce accident casualties. In the future, we could study the optimal exit choice based on the fire locations, considering several evacuation passageways between the exit and entrance. Alternatively, we could study the impact of different fire parameters on the optimal exit choice of the highway tunnel.

All data of OPS, x and y.

(XLSX)

Click here for additional data file.

Code for FCM of data.

(PY)

Click here for additional data file.

13 Jul 2021

PONE-D-21-19452

Optimal Exit Choice During Highway Tunnel Evacuations Based on The Fire Locations

PLOS ONE

Dear Dr. ma,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Aug 27 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Feng Chen

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

3.  We suggest you thoroughly copyedit your manuscript for language usage, spelling, and grammar. If you do not know anyone who can help you do this, you may wish to consider employing a professional scientific editing service.

Whilst you may use any professional scientific editing service of your choice, PLOS has partnered with both American Journal Experts (AJE) and Editage to provide discounted services to PLOS authors. Both organizations have experience helping authors meet PLOS guidelines and can provide language editing, translation, manuscript formatting, and figure formatting to ensure your manuscript meets our submission guidelines. To take advantage of our partnership with AJE, visit the AJE website (http://learn.aje.com/plos/) for a 15% discount off AJE services. To take advantage of our partnership with Editage, visit the Editage website (www.editage.com) and enter referral code PLOSEDIT for a 15% discount off Editage services.  If the PLOS editorial team finds any language issues in text that either AJE or Editage has edited, the service provider will re-edit the text for free.

Upon resubmission, please provide the following:

a) The name of the colleague or the details of the professional service that edited your manuscript

b) A copy of your manuscript showing your changes by either highlighting them or using track changes (uploaded as a *supporting information* file)

c) A clean copy of the edited manuscript (uploaded as the new *manuscript* file)

4. PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager. Please see the following video for instructions on linking an ORCID iD to your Editorial Manager account: https://www.youtube.com/watch?v=_xcclfuvtxQ

5. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This paper proposes an exit selection strategy in highway tunnels under the a fire event. This research is of practical importance for safety management. The logic of this paper is clear, but there are still some issues needed to revise. The comments of the paper are as follows:

(1)The authors used VISSIM to simulate the vehicle distributions when the fire happened. Please explain the advantage of VISSIM.

(2)The research framework should be presented to indicate the methodology and contents.

(3) The rows in Table 2 should be spaced equally.

(4)The flowchart in Figure 4 of the paper is too cumbersome. The authors are advised to streamline it.

(5)The images inserted in the paper are blurry. It is recommended to use clear pictures.

(6)The format of the references should be consistent. For example, the name of a paper with only reference 1 is capitalized.

Reviewer #2: The study investigated an important topic using a simulation based approach. The authors combine two simulation tools for understanding the efficiency of highway tunnel evacuations and assisting the location of evacuation signs. Overall, I found the study interesting and would be of interest to broad audience. A few minor issues need to be corrected before it can be accepted for publication.

1. It is not clear how parameters in VISSIM and BuildingEXODUS are calinbrated, and how the simulation results are validated. The whole calibration process for simulation based study is necessary and the author may want to inlcude more details in supporting information.

2. Only the optimal productivity statistics is used as the index for evaluating the evacuation efficiency. I believe there are many other indexes or metrics that may work in a similar way. The authors should discuss further on the reasoning behind the choice of the index and potentially include the results from other metrics.

3. The figures submitted are of very poor quality. This should be significantly improved. Vectorized images are preferred for scientific publications.

4. The current discussion on major findings is pretty slim. The authors should include more detailed discussion on how the results can inform better evacuation prepration in highway tunnels and what are the major lessons learned from the case study that can be carried over other scenarios. The authors should also discuss the limitations, which is important for simulation based studies.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

6 Aug 2021

Dear reviewers,

Thanks very much for taking the time to review this manuscript. I really appreciate all your comments and suggestions! Please find my itemized responses below and my revisions in the re-submitted files.

Thank you very much for your consideration.

Best regards!

Jianxiao Ma

Reviewer #1:

1. The authors used VISSIM to simulate the vehicle distributions when the fire happened. Please explain the advantage of VISSIM.

Response: Thanks for your suggestions. We have added papers 27 and 28 with more detail about the advantage of VISSIM in the Method section. The specific revises are as follows (Word Lines: 153-156):

The car-following model of VISSIM could easily simulate and extract the trajectories of vehicles[27, 28]. Therefore, VISSIM is used to display vehicle distributions and operations in real time when the fire occurs [29], determining the locations of evacuating occupants and obstacles in the tunnel.

2. The research framework should be presented to indicate the methodology and contents.

Response: Thank you for your questions. We have added more descriptions to indicate the methodology and contents. The specific revises are as follows (Word Lines: 78-87):

The aim in the study is to determine the optimal exit choice for different fire locations and evacuating occupants at different positions. We propose a strategy to obtain the optimal exit choice during highway tunnel evacuations based on fire location. The research framework includes three parts. First, we use BuildingEXODUS and VISSIM to obtain the parameters about vehicle distributions, locations of evacuating occupants, evacuation process and analyze the impacts of different fire locations on evacuation time. Then, the OPS output from BuildingEXODUS is selected as the evaluation index. Next, FCM cluster algorithm is used to capture the feature points of the occupants. Based on the feature points, the relationship between the location of fire and boundary of optimal exit choice under the optimal OPS is obtained through polynomial regression model.

3. The rows in Table 2 should be spaced equally.

Response: Thank you for your questions. In the revised version, we have modified Table 2(Word Lines: 278).

4. The flowchart in Figure 4 of the paper is too cumbersome. The authors are advised to streamline it.

Response: Thank you for your questions. In the revised version, we have modified Figure 4.

5. The images inserted in the paper are blurry. It is recommended to use clear pictures.

Response: Thank you for your questions. In the revised version, we have provided figures of high quality.

6. The format of the references should be consistent. For example, the name of a paper with only reference 1 is capitalized.

Response: Thank you for your questions. In the revised version, we have modified the format of the references.

Reviewer #2:

1. It is not clear how parameters in VISSIM and BuildingEXODUS are calinbrated, and how the simulation results are validated. The whole calibration process for simulation based study is necessary and the author may want to inlcude more details in supporting information.

Response: Thank you for your questions. In the revised version, we have added more descriptions to expand the calibration process. The specific revises are as follows:

A set of simulation parameters is required to ensure the facticity of evacuation under the fire environment in the tunnel. In this study, monitoring data and literature data are collected to calibrate those simulation parameters input into BuildingEXODUS and VISSIM before simulation. Literature data are used to provide fire parameters and evacuation speed to ensure the authenticity of simulation. A previous emergency exercise of Maoshan Tunnel is used to validate the evacuation process in the highway tunnel under fire environment. (Word Lines: 236-241)

The parameters of fire determined based on both monitoring and literature data are input into the hazard module of BuildingEXODUS. (Word Lines: 244-246)

The maximum traffic volume is input into the VISSIM software to simulate the distribution of vehicles from freely flowing to congested traffic after the fire happens. (Word Lines: 260-262)

In this study, the average speed is 0.49 m/s, which is input into the movement module of BuildingEXODUS to simulate the movement of occupants in the highway tunnel under fire environment. (Word Lines: 288-291)

2. Only the optimal productivity statistics is used as the index for evaluating the evacuation efficiency. I believe there are many other indexes or metrics that may work in a similar way. The authors should discuss further on the reasoning behind the choice of the index and potentially include the results from other metrics.

Response: Thank you for your questions. There are some evaluation indexes used to evaluate evacuation efficiency, such as evacuation time. The calculation of OPS includes not only total evacuation time, but also the time of the last occupant evacuate from exit, which reflects the utilization rate and evacuation efficiency. From the perspective of management, the OPS could be used to measure the utilization rate of evacuation passageway and exit. In the revised version, we have added more descriptions to explain the reason for using the index. The specific revises are as follows (Word Lines: 120-121):

The calculation of OPS includes not only total evacuation time, but also the time of the last occupant evacuating from exit, which reflects the utilization rate and evacuation efficiency.

3. The figures submitted are of very poor quality. This should be significantly improved. Vectorized images are preferred for scienific publications.

Response: Thank you for your questions. In the revised version, we have provided figures of high quality.

4. The current discussion on major findings is pretty slim. The authors should include more detailed discussion on how the results can inform better evacuation preparation in highway tunnels and what are the major lessons learned from the case study that can be carried over other scenarios. The authors should also discuss the limitations, which is important for simulation based studies.

Response: Thank you for your questions. In the revised version, we rewrote conclusion section. The specific revises are as follows (Word Lines: 383-394):

Overall, the location of the fire is detected by fire detectors in the highway tunnel. The optimal exit choice during highway tunnel evacuations based on fire location is obtained by our strategy, which could provide valuable information for optimal exit choice and arranging variable evacuation signs in the highway tunnel. In our study, vehicle distribution is an important factor that determines the spatial location of vehicle and initial locations of evacuating occupants. Hence, effective information warning and traffic guidance could reduce the number of trapped vehicles and reduce accident casualties. In the future, we could study the optimal exit choice based on the fire locations, considering several evacuation passageways between the exit and entrance. Alternatively, we could study the impact of different fire parameters on the optimal exit choice of the highway tunnel.

Editor:

1. We suggest you thoroughly copyedit your manuscript for language usage, spelling, and grammar. If you do not know anyone who can help you do this, you may wish to consider employing a professional scientific editing service.

Response: We have employed a professional scientific editing service American Journal Experts（AJE）to provide language editing.

2. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information:

Response: We have added captions of Supporting Information files at the end of your manuscript and updated any in-text citations to match accordingly.

Special thanks to you for your good comments. We have carefully revised and polished the manuscript, please the reviewer to check.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

9 Aug 2021

Optimal Exit Choice During Highway Tunnel Evacuations Based on The Fire Locations

PONE-D-21-19452R1

Dear Dr. ma,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Feng Chen

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: My comments have been addressed by authors. The revised version is suitable for the journal. So my suggestion is Accept.

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

12 Aug 2021

PONE-D-21-19452R1

Optimal Exit Choice During Highway Tunnel Evacuations Based on The Fire Locations

Dear Dr. ma:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Feng Chen

Academic Editor

PLOS ONE