Influence of Smoke Exhaust Volume and Smoke Vent Layout on the Ceiling Centralized Smoke Exhaust Effect in Tunnel Fires

Round 1

Reviewer 1 Report

In this work, the impact of smoke exhaust volume and smoke vent layout on the smoke exhaust effect in tunnel fires were investigated by numerical simulation. The novelty of the numerical work is high and the study of the smoke exhaust effect is well done. Here are some comments,

1.       Why do you set some measuring plane around the exhaust vent? How to use these measured data?

2.       The details for explaining the setup of the smoke exhaust vents and smoke exhaust volume should be labeled in Fig. 2.

3.       How to measure the mass flow rate of gas passing through each smoke vent?

4.       More explanations should be given on the mass flow rate of smoke passing through each smoke vent for different exhaust volumes shown in Fig. 8.

5.       What are the main different effects between the increasing smoke exhaust volumes?

Author Response

Thanks a lot for you and the reviewers’ comments and suggestions for our manuscript (fire-2860185). We provide this cover letter to explain, point by point, the detailed responses to you and the reviewers’ comments and the revisions in the manuscript as follows. In order to make the changes easily viewable for you and the reviewers, in the revised manuscript, the revisions are marked with red color. We hope the revised paper would satisfy you and the reviewers.

Comments 1: Why do you set some measuring plane around the exhaust vent? How to use these measured data?

Response 1: Thank you for pointing this out. We agree with this comment. Therefore, we have added the explanation in the section of “3.2 Influence of smoke exhaust volume”, on page 7,

“Smoke flows from the fire source to both ends, and fresh air flows from both ends to the fire source. Therefore, the flow difference in these two directions is equivalent to the amount of smoke and fresh air exhausted passing through the smoke vent, respectively.”

The mass flow rate slice can measure the mass flow rate passing through the cross-section in a certain direction, which exactly meets the above requirements.

Comments 2: The details for explaining the setup of the smoke exhaust vents and smoke exhaust volume should be labeled in Fig. 2.

Response 2: We agree with this comment. Therefore, we have revised Fig. 2 in the section of “3.1. Smoke spread and temperature profile over time”, on page 5.

Comments 3: How to measure the mass flow rate of gas passing through each smoke vent?

Response 3: We agree with this comment. Therefore, we have added the method of measuring the mass flow rate of gas passing through each smoke vent in the section of “3.2 Influence of smoke exhaust volume”, on page 7,

“The total gas passing through the smoke vent is obtained by slicing the mass flow rate inside the smoke vent. The smoke and fresh air passing through the smoke vent is obtained by the flow difference, which is measured by the mass flow rate slicing set before and after the smoke vent.”

Comments 4: More explanations should be given on the mass flow rate of smoke passing through each smoke vent for different exhaust volumes shown in Fig. 8.

Response 4: We agree with this comment. Therefore, we have added the corresponding explanation in the section of “3.2 Influence of smoke exhaust volume”, on page 8,

“It is evident from Figure 8 that the change in smoke exhaust volume has a stronger effect on the upstream smoke vents than the downstream ones. With a growth in the smoke exhaust volume, the smoke passing through each vent upstream shows an increasing trend, while the smoke downstream shows a decreasing trend. This phenomenon is attributed to the single-side smoke exhaust of the tunnel, where an increase in the smoke exhaust volume leads to more smoke being drawn upstream, subsequently decreasing the utilization of the downstream smoke vent.”

Comments 5: What are the main different effects between the increasing smoke exhaust volumes?

Response 5: We agree with this comment. Therefore, we have added the corresponding explanation in the section of “4. Conclusions”, on page 11,

“The smoke control effectiveness is significantly influenced by changes in the total smoke exhaust volume. Increasing the smoke exhaust volume allows for better control of smoke spread distance, smoke ceiling temperature, and visibility in the tunnel, however, blindly increasing the smoke exhaust volume may lead to more smoke vents being overwhelmed or failing.”

Author Response File: Author Response.docx

Reviewer 2 Report

The impact of smoke exhaust volume and smoke vent layout on the smoke exhaust effect in tunnel fires with ceiling centralized smoke exhaust system are investigated by numerical simulation methods. The length of smoke distribution, smoke temperature under the ceiling vertical visibility, and the exhausted smoke mass flow rate are described and analyzed. Below the threshold, increasing the smoke exhaust volume could achieve better control of smoke spread distance, smoke ceiling temperature, and visibility in the tunnel. In cases of small smoke exhaust volumes, changes in smoke vent numbers can obviously influence the smoke control effect. When the smoke exhaust volume is excessive, altering smoke vent numbers has a minimal impact on smoke exhaust. This research would be used for engineers to estimate the smoke control  effect in tunnel fire. The authors should revise the following points in order to better express the research results.

1.       As for single-side ceiling centralized smoke exhaust mode, is it possible to adopt a smoke exhaust strategy based on the location of the fire source and the number of asymmetric smoke exhaust outlets on both sides?

2.       As for  Fig.6,  it can be seen between  the range from -90m  location to 90m location ,the ceiling  smoke temperature ae higher than 100℃. While  beyond -90m,  smoke temperature decrease to 50℃, the smoke exhaust effect in this area is good. Can we consider this area as an evacuation safety zone?

3.       The authors should compare the research results of this article with similar published research results or experimental results in order to demonstrate the rationality of the numerical simulation research results.

Moderate editing of English language required

Author Response

Thanks a lot for you and the reviewers’ comments and suggestions for our manuscript (fire-2860185). We provide this cover letter to explain, point by point, the detailed responses to you and the reviewers’ comments and the revisions in the manuscript as follows. In order to make the changes easily viewable for you and the reviewers, in the revised manuscript, the revisions are marked with red color. We hope the revised paper would satisfy you and the reviewers.

Comments 1: As for single-side ceiling centralized smoke exhaust mode, is it possible to adopt a smoke exhaust strategy based on the location of the fire source and the number of asymmetric smoke exhaust outlets on both sides?

Response 1: Thank you for pointing this out. We agree with this comment. Therefore, we have added the corresponding explanation in the section of “2. Research Methodology”, on page 3,

“This study focuses on the smoke control performance of Tunnel A, a connecting passageway within the Luao Road urban interconnected tunnels, of Haicang Evacuate Passage Project, Xiamen. A 350 m segment from Tunnel A is selected with the arched cross-section. In the actual environment of the tunnel, when a fire occurs on the tunnel, the smoke is mainly discharged from the vertical shaft on one side of the tunnel, resulting in a situation of single-sided smoke exhaust.”

For the suggestions proposed by the reviewer, the content considered is more comprehensive, but it differs from the actual situation of the study.

Comments 2: As for Fig.6, it can be seen between the range from -90m location to 90m location, the ceiling smoke temperature ae higher than 100℃. While beyond -90m, smoke temperature decreases to 50℃, the smoke exhaust effect in this area is good. Can we consider this area as an evacuation safety zone?

Response 2: We agree with this comment. Therefore, we have added the corresponding explanation in the section of “3.2. Influence of smoke exhaust volume”, on page 8,

“Meanwhile, it can be seen from Figure 6 that after passing through the smoke vent 90m upstream and downstream of the fire source, the smoke temperature significantly de-creases. The smoke exhaust effect at this smoke vent is good. Considering the length of tunnel smoke spread and visibility, it can be considered that this area has reached the evacuation safety zone.”

Comments 3: The authors should compare the research results of this article with similar published research results or experimental results in order to demonstrate the rationality of the numerical simulation research results.

Response 3: We agree with this comment. Therefore, we have added the corresponding comparison in the section of “2.2 Grid sensitivity analysis”, on page 4,

“At the same time, the numerical simulation results under a 0.200m grid were compared with previous experimental data in the article [24,25]. The figure demonstrates the relationship between the dimensionless temperature and the dimensionless distance. It is evident that the numerical simulation results are consistent with previous experimental results. Finally, to balance operating speed and accuracy, the model grid size is selected as 0.200 × 0.200 × 0.200 m.”

Author Response File: Author Response.docx

Reviewer 3 Report

The authors have discussed an important topic of smoke exhaust from in tunnel fires. But there are many drawbacks to the current study that require further attention

1. The study presented is this article is based on numerical modeling. As such proper validation of the numerical model being used needs to be presented. Comparison against some existing simulations or experiments is suggested.

2. Details of combustion model used to simulate the fire are missing.

3. How do you simulate smoke? Is is considered a passive or is there a two-way coupling with the flow field?

4. What kind of model is used for the smoke particle nucleation?

5. Do you consider the effect of thermal radiation? If so what radiation model is being used?

6. What are the boundary conditions being used at various boundaries?

7. What are the y+ values at the tunnel walls? If the near wall flow is not being modeled (or resolved) properly it will affect the flow field significantly, hence affecting your results.

8. You tested different mesh resolutions, why not show the results at these resolutions? Did you see grid convergence as you refine the mesh?

9. How is the smoke exhaust volume measured? For the setup where total number of vents are changing, how do you keep the smoke exhaust volume constant?

English language looks okay for most parts. Minor updates required.

Author Response

Thanks a lot for you and the reviewers’ comments and suggestions for our manuscript (fire-2860185). We provide this cover letter to explain, point by point, the detailed responses to you and the reviewers’ comments and the revisions in the manuscript as follows. In order to make the changes easily viewable for you and the reviewers, in the revised manuscript, the revisions are marked with red color. We hope the revised paper would satisfy you and the reviewers.

Comments 1: The study presented is this article is based on numerical modeling. As such proper validation of the numerical model being used needs to be presented. Comparison against some existing simulations or experiments is suggested.

Response 1: Thank you for pointing this out. We agree with this comment. Therefore, we have added the corresponding comparison in the section of “2.2 Grid sensitivity analysis”, on page 4,

“At the same time, the numerical simulation results under a 0.200m grid were compared with previous experimental data in the article [24,25]. Figure demonstrates the relationship between the dimensionless temperature and the dimensionless distance. It is evident that the numerical simulation results are consistent with previous experimental results.”

Comments 2: Details of combustion model used to simulate the fire are missing.

Response 2: We agree with this comment. Therefore, we have added the corresponding details of combustion mode in the section of “2.1 Fire Scenario”, on page 2,

“FDS of this article uses a combustion model based on the mixing-limited, infinitely fast reaction of lumped species. The combustion is mixing-controlled and that the reaction of fuel and oxygen is infinitely fast. This combustion model uses the single mixture fraction as its conserved scalar and is conventionally called as a mixture fraction model. The mixture fraction model obtains encouraging results in certain fire scenarios.”

Comments 3: How do you simulate smoke? Is considered a passive or is there a two-way coupling with the flow field?

Response 3: We agree with this comment. Therefore, we have added the corresponding explanation in the section of “2.1 Fire Scenario”, on page 3,

“When you do a fire simulation, FDS automatically creates two output files that are rendered by Smoke view as realistic looking smoke and fire. By default, the output quantities are the “DENSITY” of “SOOT” and “HRRPUV” (Heat Release Rate Per Unit Volume).”

Comments 4: What kind of model is used for the smoke particle nucleation?

Response 4: We agree with this comment. Therefore, we have added the corresponding explanation in the section of “2.1 Fire Scenario”, on page 3,

“The movement of Lagrangian particles over the course of a time step is calculated using an analytical solution and remains stable regardless of the time step used by the flow solver.”

Comments 5: Do you consider the effect of thermal radiation? If so, what radiation model is being used?

Response 5: We agree with this comment. Therefore, we have added the corresponding explanation in the section of “2.1 Fire Scenario”, on page 3,

“The influence of thermal radiation near the smoke vent mainly comes from the smoke. Considering that the preset smoke and CO production have limited effects, this study ignored the effects of smoke and carbon monoxide production.”

Comments 6: What are the boundary conditions being used at various boundaries?

Response 6: We agree with this comment. Therefore, we have added the boundary conditions in the section of “2.1 Fire Scenario”, on page 3,

“The boundary of the open ends and smoke vent is specified as “OPEN”, which denotes a passive opening to the outside. The internal lining of the tunnel is specified as the boundary condition of ‘‘CONCRETE’’ and its conductivity, specific heat and density are 1.8 W/(m∙K), 1.04 kJ/(kg∙K) and 2280 kg/m3. Install an “OBATACLE” at each smoke vent to block and set “CONTROL” to automatically disappear after 120 seconds, which is used to simulate the opening of smoke vents. Install a “VENT” on one side of the smoke exhaust duct, with the surface set to “EXHAUST”. It automatically actives at 120 s, which is used to simulate the start of the smoke exhaust fan.”

Comments 7: What are the y+ values at the tunnel walls? If the near wall flow is not being modeled (or resolved) properly it will affect the flow field significantly, hence affecting your results.

Response 7: We agree with this comment. Therefore, we have added the corresponding explanation in the section of “2.1 Fire Scenario”, on page 3,

“FDS uses wall models or wall functions for near wall treatment, aim to mimic the sudden change from molecular to turbulent transport close to the walls, so there is no need to resolve the otherwise computationally expensive region of the near-wall flow-field.”

Comments 8: You tested different mesh resolutions, why not show the results at these resolutions? Did you see grid convergence as you refine the mesh?

Response 8: We agree with this comment. Therefore, we have added the corresponding comparison in the section of “2.2 Grid sensitivity analysis”, on page 4,

“Considering the suggestions of the FDS user guide [20], the range of , the ratio of the fire characteristic diameter to mesh size, ought to be between 4 and 16. Considering the fire size  as 20 MW, the calculated is 3.0 m, so the range of mesh size should be guaranteed to be between 0.187 m and 0.750 m. The computation speed increases with grid size, but accuracy suffers as a result [23]. The temperature distribution under the tunnel ceiling with different grid sizes was compared in this article, and the smoke exhaust system is not turned on at this time. The results indicate that the decrease in grid size leads to a gradual convergence of the temperature of the longitudinal tunnel roof. When the grid size is not greater than 0.200 m, the temperature difference becomes almost negligible.”

Comments 9: How is the smoke exhaust volume measured? For the setup where total number of vents are changing, how do you keep the smoke exhaust volume constant?

Response 9: We agree with this comment. Therefore, we have added the method of measuring smoke exhaust volume has been added in the section of “3.2 Influence of smoke exhaust volume”, on page 7,

“The total gas passing through the smoke vent is obtained by slicing the mass flow rate inside the smoke vent. The smoke and fresh air passing through the smoke vent is obtained by the flow difference, which is measured by the mass flow rate slicing set before and after the smoke vent. Smoke flows from the fire source to both ends, and fresh air flows from both ends to the fire source. Therefore, the flow difference in these two directions is equivalent to the amount of smoke and fresh air exhausted passing through the smoke vent, respectively.”

And the explanation about keeping the smoke exhaust volume constant have been added in the section of “2.1 Fire Scenario”, on page 3,

“Install a “VENT” on one side of the smoke exhaust duct, with the surface set to “EXHAUST”. It automatically actives at 120 s, which is used to simulate the start of the smoke exhaust fan.”

We determine the total exhaust volume by setting different exhaust values for this vent.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The authors carefully revised paper according to the comments of reviewers.  I recommend that this paper can be accepted. However, the paper needs to supplement the technical requirements for centralized smoke exhaust in existing tunnel design specifications.

Minor editing of English language required

Author Response

Comments 1: The paper needs to supplement the technical requirements for centralized smoke exhaust in existing tunnel design specifications.

Response 1: Thank you for pointing this out. We agree with this comment. Therefore, we have added the explanation in the section of “1. Introduction”, on page 1,

“According to China National Standard “Code for Fire Protection Design of Buildings” [7] and “Guidelines for Design of Ventilation of Highway Tunnels” [8], centralized smoke exhaust is a special aspect of smoke exhaust technology, which is an effective method for controlling fire smoke. This involves setting up a certain smoke vent number and installing a smoke exhaust duct in the longitudinal direction of the tunnel. When the tunnel appears fire, the smoke vent in the designed area is activated to quickly and effectively discharged the smoke into the tunnel ceiling exhaust duct, while fresh air is provided from the tunnel's two open ends. This creates a certain longitudinal wind speed in the tunnel.”

Author Response File: Author Response.docx

Reviewer 3 Report

The authors have done a good job at revising the article, but there are still some issues that have not been fully addressed.

1. This has been resolved upto a certain extent but not completely due to two reasons. First, the figure that is included in the cover letter is not included in the revised manuscript. Second, while validating the setup needs to match the experiments as closely as possible. Here, the experiments have HRR of 0.851 and 1.5 MW. Why did you choose to run the simulation with HRR of 20 MW for validation? Why not choose operating condition of one of the experiment?

2. Resolved.

3. The question is about the smoke transport model in FDS and not the rendering of smoke for output. Let me clarify, are smoke particles modeled as a passive or do they have an impact on the flow field (Eulerian sources due to presence of smoke particles)?

4. The question is about the nucleation of smoke particles and not the trajectory. What kind of model is used for generating smoke particles from fire?

5. The comment dose not explain the thermal radiation model being used.

6. Resolved.

7. Resolved.

8. Grid convergence seems fine. Please include the figure highlighting grid convergence in the revised manuscript.

9. Resolved.

Author Response

Comments 1: This has been resolved up to a certain extent but not completely due to two reasons. First, the figure that is included in the cover letter is not included in the revised manuscript. Second, while validating the setup needs to match the experiments as closely as possible. Here, the experiments have HRR of 0.851 and 1.5 MW. Why did you choose to run the simulation with HRR of 20 MW for validation? Why not choose operating condition of one of the experiment?

Response 1: Thank you for pointing this out. We agree with this comment. Therefore, we have added the figure in the section of “2.2 Grid sensitivity analysis”, on page 4,

In addition, in comparison with the experiment, we made the temperature dimensionless, eliminating the influence of HRR and cross-sectional size differences. Therefore, it is reasonable to choose fire experiments with HRR of 0.815 and 1.5MW for comparison.

Comments 3: The question is about the smoke transport model in FDS and not the rendering of smoke for output. Let me clarify, are smoke particles modeled as a passive or do they have an impact on the flow field (Eulerian sources due to presence of smoke particles)?

Response 3: We agree with this comment. After carefully reviewing the “FDS User Guide” and “FDS Technical Reference Guide”, we added the corresponding explanation in the section of “2.1 Fire Scenario”, on page 3,

“FDS uses the Lagrangian particle model, which only considers the bidirectional coupling between gas and particles. This means that each particle interacts separately with the carrier fluid. The momentum lost by particles will be added to the fluid, and vice versa.”

Comments 4: The question is about the nucleation of smoke particles and not the trajectory. What kind of model is used for generating smoke particles from fire?

Response 4: We agree with this comment. After carefully reviewing the “FDS User Guide” and “FDS Technical Reference Guide”, we added the corresponding explanation in the section of “2.1 Fire Scenario”, on page 3,

“The mass transport equations are used for smoke nucleation process in this article (the mass transport equations make no distinction between a single or lumped species). The conversion of fuel combustion is a simple problem of performing matrix multiplication. Fuel is usually a single type of gas, but air and products are often referred to as " lumped species ". A lumped species represents a mixture of gas species that transport together.”

From a numerical model perspective, a lumped species can be regarded as smoke particles produced by the combustion of individual species.

Comments 5: The comment dose not explain the thermal radiation model being used.

Response 5: We agree with this comment. After carefully reviewing the “FDS User Guide” and “FDS Technical Reference Guide”, we added the corresponding explanation in the section of “2.1 Fire Scenario”, on page 3,

“In this article, the default gray gas model of FDS is used for thermal radiation analysis. This model assumes that all radiation amounts are almost uniform throughout the entire spectrum, and the radiation intensity at all frequencies is the same. Its reliability has been confirmed in previous articles [21].”

Comments 8: Grid convergence seems fine. Please include the figure highlighting grid convergence in the revised manuscript.

Response 8: We agree with this comment. Therefore, we have added the corresponding figure in the section of “2.2 Grid sensitivity analysis”, on page 4,

Author Response File: Author Response.docx