The Formation of a Flame Front in a Hydrogen–Air Mixture during Spark Ignition in a Semi-Open Channel with a Porous Coating
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors1. The research work has good practical application prospects;
2. The article is relatively short and should not be published as an "article". I suggest changing it to 'communication';
3. It is best to unify the multiple shooting time intervals in Figures 2-4, but this has little impact on the discovery of experimental patterns. The author obtained very ideal research results;
4. The control of flame propagation speed has always been a research hotspot, which is of great significance for protecting public life and reducing property losses. Therefore, this article is worth publishing as soon as possible.
Author Response
We are grateful to the reviewers for their valuable comments! We have tried to respond to all comments. Thanks to the reviewers, we have corrected a number of errors and shortcomings.
Sincerely, authors.
Responses to the first reviewer's comments are highlighted in green.
- The article is relatively short and should not be published as an "article". I suggest changing it to 'communication';
- After corrections, we increased the length of the manuscript to 3,800 words, not counting the bibliography (+ 990 words).
- It is best to unify the multiple shooting time intervals in Figures 2-4, but this has little impact on the discovery of experimental patterns. The author obtained very ideal research results;
- Thank you for the comment! Undoubtedly, we have previously tried to do as the reviewer suggests. The number of photographs necessary for an objective understanding is about 16. The movement of the flame front to the bottom wall is more dynamic than further movement along the surface and axis of the channel. Therefore, we divided the sequence of photographs into two blocks. The first three photographs are with a small interval until the flame front approaches the lower wall and the last three photographs are with a different interval for flame moving along the surface. There are different time scales here, so to reduce the total number of identical photos. We indicated this in the revision (Section 3, Lines 141-145).
Reviewer 2 Report
Comments and Suggestions for AuthorsFormation of a flame front in hydrogen-air mixture during spark ignition in a semi-open channel with a porous coating
73 - What’s the window/walls made off ? Lime-glass ? Quartz ?
98 - Is the polyurethane foam changed between all shot repeat ?
99/100 - How was the degree of porosity and pore size assessed for either material ? What’s the porosity definition ? 1-V_steel/Vtotal ?
Do you also have the average diameter of the steel wool ?
111 - Why is the FOV borders distorded ? Are the windows warped ?
122 - How was the flame speed measured ? Can you also clarify the direction it is measured, e.g. rew or green arrow on Figure 2. Looks like op-end from Figure 6. Also why do some of them stop at 30 mm ? Why is the x axis limit different between ER 0.4 and 1.0 ? Are those shot averaged values for each time ?
137 - Does that mean the flame quenches in the polyurethane ?
It looks like the steel wool is not as contained as the foam (seeing on Figure 4).
169 - “method” not “methodic”
The authors are reporting results from the experimental investigation of flame propagation in a premixed mixture in a semi-open tube equipped with a poreous material on the wall positioned oposite the ignition source.
The results show that adding the poreous material yields an acceleration of the flame speed, compared with a quasi-constant velocity without it.
The origin of the acceleration is unclear and not discussed by the authors. It is clear that less poreous materials, here the steel wool yields smaller acceleration than lower porosity (polyurethane). The flame accelerates for several factors that include laminar to turbulent transition, acceleration of the gas upstream of the flame from diaplacement effect, flame front instabilities. The authors evaluates the flame front instabilities structure size but not the authors parameters. The poreous material although it is of very high porosity values acts as a flow restriction. The control experiment comparison with a non-poreous material is required to evaluate the effect of the complete flow restriction just by reducing the effective flow cross-section.
Could the authors also evaluate the Reynolds number upstream of the flame front to rule out a turbulent transition ?
Besides answering the above question. Before considering publication, I think the author need to consider the addition of this second control experiment, with the addition of a lot more detail on the poreous material used in the experiment.
Author Response
We are grateful to the reviewers for their valuable comments! We have tried to respond to all comments. Thanks to the reviewers, we have corrected a number of errors and shortcomings.
Sincerely, authors.
Responses to the second reviewer's comments are highlighted in yellow.
73 - What’s the window/walls made off ? Lime-glass ? Quartz ?
- Windows were made of optical quartz glass (Section 2, Line 88)
98 - Is the polyurethane foam changed between all shot repeat ?
- No, polyurethane does not change. A reference was provided to photographs from our previous publication. Steel wool was used for only one experiment, it was changed before each experiment. We added comments to the revision (section 2, Lines 114-116, Lines 118-119).
99/100 - How was the degree of porosity and pore size assessed for either material ? What’s the porosity definition ? 1-V_steel/Vtotal ?
Do you also have the average diameter of the steel wool ?
- The porosity of the polyurethane used was ε = 95–98%. The porosity of steel wool was 99%. The porosity of polyurethane was determined by the displacement of water. And the porosity of steel wool was determined using the weight and density of the steel and the dimensions of the sample. The pore density of polyurethane was determined statistically in several linear directions on the surface of the sample. The transverse size of polyurethane and steel wool fibers was determined with a micrometer. For polyurethane with a pore density of 10 ppi, the transverse size of the fibers was 200 μm; with a pore density of 80 ppi, the fiber size was 20 μm. The average transverse size of the steel fiber was 30 μm. (Section 2, Lines 119-130, + Table 1)
111 - Why is the FOV borders distorded ? Are the windows warped ?
- The field of view is partially blocked by sealing rubber inserts, which are located outside and do not affect the flow inside the chamber.
We indicated this in the revision (Section 3, Lines 145-147).
122 - How was the flame speed measured ?
- The velocity was calculated based on the time interval between frames. The error in determining the flame front velocity was less than 3%. (Section 3, Lines 169-174)
Can you also clarify the direction it is measured, e.g. rew or green arrow on Figure 2. Looks like op-end from Figure 6.
- The red direction (I) corresponds to the direction towards the open end of the channel, the blue arrow (II) corresponds to the direction towards the bottom coated wall, and the green arrow (III) corresponds to the direction towards the closed end of the channel.
Comments are given in the text and in the caption to Figure 2.
Section 3, Lines 219-222.
Also why do some of them stop at 30 mm ?
- We are grateful to the reviewer for his comment! We gave some velocities depending on the pixels rather than mm when processing photographs. At the same time, the discussion remained true in the manuscript. We corrected the plots and gave the dependence on the dimensional value - on the distance from the spark gap in mm. This is our technical error.
Why is the x axis limit different between ER 0.4 and 1.0 ?
- We made the plots with the same axis limits.
Are those shot averaged values for each time ?
- Yes, these values are the average. We added to the manuscript that the average deviation for velocities was about 10%:
Figure 5 shows the average velocities calculated based on five experiments. The average deviation for velocities was about 10%. The maximum deviation did not exceed 20%. Figure 5c shows typical deviation for the velocities. (Section 3, Lines 169-174.)
137 - Does that mean the flame quenches in the polyurethane ?
- The movement of the flame front in the visible part of the channel is registered. Combustion inside the polyurethane is not under consideration. (Section 2, Lines 93-94.)
It looks like the steel wool is not as contained as the foam (seeing on Figure 4).
- The reviewer is absolutely right. Unlike foam, steel wool does not have a structured shape and consists of a large number of fibers. We wrote:
As can be seen from Figure 3c and Figure 4e, the steel wool has residual deformation and elasticity, so it is difficult to position it perfectly evenly. (See Section 3, Lines 154-155.)
169 - “method” not “methodic”
- It was corrected.
The control experiment comparison with a non-poreous material is required to evaluate the effect of the complete flow restriction just by reducing the effective flow cross-section.
- Experiments were added in a narrow solid channel to compare the effect of narrowing the channel. (Section 3, Lines 150-153. + Figure 4f + Lines 199-211)
As can be seen from Figure 5a,b, the change in flame front velocity when placing a steel impermeable plate (empty 30 mm) is different from other velocities. In this case it reaches its maximum value (20-22 m/s). This may be due to the fact that the combustion products and the unburned mixture do not shift in the transverse direction relative to the channel axis inside the porous layer. An increase in the pore density of the coating leads to a decrease in permeability for the transverse flow of the unburned mixture and combustion products, which, on the contrary, increases their average velocity relative to the channel axis. When studying the influence of pore density on the propagation of the flame front, the value of hydraulic resistance can be taken into account, which was studied in detail for polyurethane in [Kuleshov 2022]. It is assumed that at the maximum value of hydraulic resistance, combustion products and unburned mixture spread to a lesser extent inside the porous layer, which subsequently leads to the acceleration of the flame front in the free space above the porous surface.
Could the authors also evaluate the Reynolds number upstream of the flame front to rule out a turbulent transition ?
- We estimated the Reynolds numbers and inserted the discussion into the manuscript.
Section 3, Lines 247-262
Besides answering the above question. Before considering publication, I think the author need to consider the addition of this second control experiment, with the addition of a lot more detail on the poreous material used in the experiment.
- We wrote the characteristics of polyurethane.
The porosity of the polyurethane used was ε = 95–99%. The porosity of steel wool was 99%. The porosity of polyurethane was determined by the displacement of water. And the porosity of steel wool was determined through the weight and density of the steel and the dimensions of the sample. The pore density of polyurethane was determined statistically in several linear directions on the surface of the sample. The transverse size of polyurethane and steel wool fibers was determined with a micrometer. For polyurethane with a pore density of 10 ppi, the transverse size of the fibers was 200 μm; with a pore density of 80 ppi, the fiber size was 20 μm. The average transverse size of the steel fiber was 30 μm. Geometric characteristics of polyurethane and hydraulic resistance coefficients are presented in Table 1.
Table 2. Geometric parameters and hydraulic resistance of polyurethane.
|
Pore density |
Porosity |
Pore size |
Hydraulic resistance |
|
10 ppi |
99% |
2.5 mm |
0.46*l [Kuleshov] |
|
20 ppi |
98% |
1.3 mm |
1.28*l [Kuleshov] |
|
45 ppi |
96% |
0.6 mm |
1.75*l [Kuleshov] |
|
80 ppi |
95% |
0.3 mm |
5.28*l [Kuleshov] |
l – is the total length of the foam
Section 2, Lines 121-132 + Table 1.
Reviewer 3 Report
Comments and Suggestions for AuthorsJournal: fire
Manuscript number: fire-2721848
Title: Formation of a flame front in hydrogen-air mixture during spark ignition in a semi-open channel with a porous coating
The paper is interesting although some issues should be addressed before publication. Major revision is suggested to further improve its quality. Specific suggestions are provided below.
1. What is the biggest flash point in this article?
2. The conclusion need to be rewritten. The conclusion and is not focused enough and the texts are too few.
3. What are the advantages of Polyurethane foam? Why choose it?
4. What is the difference between Figures 2, 3, and 4?
5. Some tables can be added as appropriate.
Comments on the Quality of English Language
Reduce some of your spoken language.
Author Response
We are grateful to the reviewers for their valuable comments! We have tried to respond to all comments. Thanks to the reviewers, we have corrected a number of errors and shortcomings.
Sincerely, authors.
Responses to the third reviewer's comments are highlighted in blue.
The paper is interesting although some issues should be addressed before publication. Major revision is suggested to further improve its quality. Specific suggestions are provided below.
- What is the biggest flash point in this article?
- We expanded the abstract. We also expanded the introduction and purposes of the work.
- Qualitative features of the deflagration of the flame front at ER 0.4, consisting of disturbances resembling small balls of flame, were discovered, in contrast to the flame front at ER 1.0. The sizes of these disturbances significantly exceed the analytical values for the Darrieus–Landau instability.
- A lot of research is currently being carried out flame front propagation in lean mixtures. In contrast to deflagration flames propagating in near-stoichiometric mixtures, in lean and ultra-lean hydrogen mixtures the structure of the flame front can be a system of disturbances or multiple discrete combustion sources, the dynamics of which significantly depends on the initial concentration of hydrogen in the mixture. In different works they were called cup, caps, small balls of flame, vortex rings [Bohm, Ronney].
- The purpose of this work was to determine the propagation velocity of a hydrogen-air flame during ignition in a channel in the presence of porous coatings and to determine disturbances at the lean flame front.
- The conclusion need to be rewritten. The conclusion and is not focused enough and the texts are too few.
- Conclusions were rewritten.
This work investigates the process of hydrogen-air mixture flame propagation when a spark is ignited in a channel coated with steel wool or polyurethane foam. Using the Schlieren method and a high-speed camera, images of the flame front were obtained, graphs of the velocities of the flame front were plotted in 3 directions, and the sizes of disturbances at the front were determined. The main findings are summarized as follows.
Placing steel wool leads to an increase in the flame speed to 10-12 m/s relative to an empty channel (4 m/s).
The velocity of the flame front towards the closed end and down is approximately three times lower than towards the open end.
The average velocity of the flame front for the stoichiometric mixture increases with increasing pore density.
The placement of coatings of various types did not lead to significant changes of distributions of the sizes of disturbances in the same mixture.
The flame front in mixtures with ER= 0.4 consists mainly of disturbances resembling small flame balls, in contrast to the flame front in stoichiometric mixture (ER = 1.0). The sizes of these disturbances significantly exceed the analytical values for the Darrieus–Landau instability.
- What are the advantages of Polyurethane foam? Why choose it?
- In the combustion mode presented in this work, polyurethane is chemically inert and elastic. Therefore, it can be used to attenuate explosions and detonations when the initial temperature is below the melting point. Compared to porous metals and ceramics, cost is also an advantage.
Section 2, Lines 130-133.
- What is the difference between Figures 2, 3, and 4?
- Figure 2 shows photographs of the flame front propagation in the empty 20*40 mm channel for a stoichiometric hydrogen-air mixture with ER = 1 (Figure 2a) and for a lean mixture with ER = 0.4 (Figure 2b).
- Figures 3 and 4 show photographs of the flame front in a channel with a porous coating for two mixture compositions: for a mixture with ER = 0.4 (Figure 3) and with ER = 1.0
- Some tables can be added as appropriate.
- In accordance with the discussion and comments of the second reviewer, we added Table 1 with some characteristics of polyurethane including the hydraulic resistance, which affects the lateral gas displacement. (Lines 134-135)
We have also added Table 2 with the values of the maximum velocities in this channel. (Lines 211-216.)
Reduce some of your spoken language.
- A native speaker checked our revised manuscript.
Reviewer 4 Report
Comments and Suggestions for AuthorsThe manuscript describes the results of experiments on flame propagation in hydrogen-air mixtures in a semi-open channel with different porous coatings. Flame dynamics was investigated with the help of high-speed shadow video recording. It was found a pronounced influence of coating on flame propagation velocity. This velocity increase in comparison with empty channel. The collected results demonstrates new features of hydrogen-air combustion and will be interesting for specialists. . The manuscript could be published in special issue of Fire journal devoted to hydrogen combustion.
Some questions should be resolved before final submission:
1. In the lines from 125 to 127 “The maximum velocity is achieved using a coating with a pore density of 80 ppi, and the minimum velocity of 21 m/s is recorded in an empty channel without coating.” 21 m/s probably is maximum velocity? The sentence should be rearranged.
2. Why in the figure 5a the curves break off at different distances? What was camera field of view?
3. In the lines from 161 to 164 the authors make a comparison of flame front speed with laminar burning velocity. Generally, the visible velocity of flame is connected with laminar burning velocity via expansion ratio.
Author Response
We are grateful to the reviewers for their valuable comments! We have tried to respond to all comments. Thanks to the reviewers, we have corrected a number of errors and shortcomings.
Sincerely, authors.
Responses to the fourth reviewer's comments are highlighted in purple.
- In the lines from 125 to 127 “The maximum velocity is achieved using a coating with a pore density of 80 ppi, and the minimum velocity of 21 m/s is recorded in an empty channel without coating.” 21 m/s probably is maximum velocity? The sentence should be rearranged.
- We are grateful to the reviewer for his accurate remark! This is our technical error.
Lines 177-178.
- Why in the figure 5a the curves break off at different distances? What was camera field of view?
- We also responded to another reviewer. This is our technical mistake when we inserted the здщеы depending on the pixel number and not the length. To avoid such an error, we inserted pictures in PNG format and attached *.eps files.
- In the lines from 161 to 164 the authors make a comparison of flame front speed with laminar burning velocity. Generally, the visible velocity of flame is connected with laminar burning velocity via expansion ratio.
The reviewer is absolutely right. The visible velocity depends on the burning velocity and on the degree of expansion. Three directions were considered in the manuscript. If the unburned mixture moves towards the open end, then towards the closed end or in the downward direction this displacement can be neglected. We have clarified this in the new manuscript.
It is worth noting that the visible velocity of the flame front towards the open end of the channel is determined not only by the laminar burning velocity, but also by a thermal expansion of the combustion products. Therefore, the velocity of the flame front towards the open end is on average several times higher than the above values.
Lines 242-246.
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThanks for the modification, that's a lot clearer.
As expected the porous materials stand between the two full steel plates inserts. So the acceleration is likely a result of the partial constriction.
A few minor comments and correction to add prior to publication.
Figure 5 should state the direction where the velocity is measured.
L104: "Each shot was repeated five times to achieve reproducibility.", would be better read with "ensure" instead of "achieve" (in red). You want to evaluate the reproducibility.
L164 "The velocity was calculated based on the time interval between frames."
Was the pixel position selected with some kind of algorithm or was that manual ? It's fine either way, just write it down.
L288 (typo) "increase" not "increae"
L299 missing a dot at the end.
Author Response
We are grateful to the reviewer for the review!
Sincerely, authors.
Responses are highlighted in green.
Figure 5 should state the direction where the velocity is measured.
- “Figure 5. Flame velocities in stoichiometric (a,b) and lean (c,d) hydrogen-air mixture, calculated towards the open end of the channel (direction I)”
L104: "Each shot was repeated five times to achieve reproducibility.", would be better read with "ensure" instead of "achieve" (in red). You want to evaluate the reproducibility.
- It was corrected.
L164 "The velocity was calculated based on the time interval between frames."
Was the pixel position selected with some kind of algorithm or was that manual ? It's fine either way, just write it down.
- In this work, we determined the front positions manually. We have a software application for determining the trajectory of the flame front both along the channel axis and in its average position. However, in this particular case we had software shortcomings, so we determined it manually.
We wrote:
“The pixel position of the flame front was determined manually by the position of the most contrast point in the selected direction.”
L288 (typo) "increase" not "increae"
- It was corrected.
L299 missing a dot at the end.
- It was corrected. The text was removed to Section 2.
Reviewer 3 Report
Comments and Suggestions for AuthorsAccept in present form.
Author Response
We are grateful to the reviewer for the review!
Sincerely, authors.
Reviewer 4 Report
Comments and Suggestions for Authors A revised version of the manuscript may be published in the journal Fire.
Author Response
We are grateful to the reviewer for the review!
Sincerely, authors.
