219 15 9MB
English Pages 242 [238] Year 2020
Fluid Mechanics and Its Applications
Nikolai M. Rubtsov Boris S. Seplyarskii Michail I. Alymov
Initiation and Flame Propagation in Combustion of Gases and Pyrophoric Metal Nanostructures
Fluid Mechanics and Its Applications Volume 123
Series Editor André Thess, German Aerospace Center, Institute of Engineering Thermodynamics, Stuttgart, Germany Founding Editor René Moreau, Ecole Nationale Supérieure d’Hydraulique, St Martin D’heres Cedex, France
The purpose of this series is to focus on subjects in which fluid mechanics plays a fundamental role. As well as the more traditional applications of aeronautics, hydraulics, heat and mass transfer etc., books will be published dealing with topics, which are currently in a state of rapid development, such as turbulence, suspensions and multiphase fluids, super and hypersonic flows and numerical modelling techniques. It is a widely held view that it is the interdisciplinary subjects that will receive intense scientific attention, bringing them to the forefront of technological advancement. Fluids have the ability to transport matter and its properties as well as transmit force, therefore fluid mechanics is a subject that is particulary open to cross fertilisation with other sciences and disciplines of engineering. The subject of fluid mechanics will be highly relevant in such domains as chemical, metallurgical, biological and ecological engineering. This series is particularly open to such new multidisciplinary domains. The median level of presentation is the first year graduate student. Some texts are monographs defining the current state of a field; others are accessible to final year undergraduates; but essentially the emphasis is on readability and clarity. Springer and Professor Thess welcome book ideas from authors. Potential authors who wish to submit a book proposal should contact Dr. Mayra Castro, Senior Editor, Springer Heidelberg, e-mail: [email protected] Indexed by SCOPUS, EBSCO Discovery Service, OCLC, ProQuest Summon, Google Scholar and SpringerLink
More information about this series at http://www.springer.com/series/5980
Nikolai M. Rubtsov Boris S. Seplyarskii Michail I. Alymov •
•
Initiation and Flame Propagation in Combustion of Gases and Pyrophoric Metal Nanostructures
123
Nikolai M. Rubtsov Merzhanov Institute of Structural Macrokinetics and Materials Science Russian Academy of Sciences Moscow, Russia
Boris S. Seplyarskii Merzhanov Institute of Structural Macrokinetics and Materials Science Russian Academy of Sciences Moscow, Russia
Michail I. Alymov Merzhanov Institute of Structural Macrokinetics and Materials Science Russian Academy of Sciences Moscow, Russia
ISSN 0926-5112 ISSN 2215-0056 (electronic) Fluid Mechanics and Its Applications ISBN 978-3-030-57890-9 ISBN 978-3-030-57891-6 (eBook) https://doi.org/10.1007/978-3-030-57891-6 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2021 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
“Man’s quest for knowledge is an expanding series whose limit is infinity, but philosophy seeks to attain that limit at one blow, by a short circuit providing the certainty of complete and inalterable truth. Science meanwhile advances at its gradual pace, often slowing to a crawl, and for periods, it even walks in place, but eventually it reaches the various ultimate trenches dug by philosophical thought, and, quite heedless of the fact that it is not supposed to be able to cross those final barriers to the intellect, goes right on.” Stanislaw Lem
Preface
The number of publications in various fields of science related to combustion processes has been growing every day. This leads to the occurrence of rather specialized scientists and, therefore, produces particular difficulties for a researcher to access the value of his work. In this book, the issues are examined, which have not been considered in our previous books The Modes of Gaseous Combustion (2016, Heat and Mass Transfer, Springer International Publishing) and Key Factors of Combustion, From Kinetics to Gas Dynamics (2017, Springer Aerospace technology, Springer International Publishing) where only gaseous combustion was included; the book Ignition and Wave Processes in Combustion of Solids (2017, Heat and Mass Transfer, Springer International Publishing), which deals with basic approaches to understanding of solid substances combustion presented in contemporary literature. All these books are intended to allow a reader orientating combustion science confidently. The book is aimed at consideration of the essential problems of combustion science; it is complimentary to previously published ones, mentioned above, and presents the results obtained by the authors in 2017–2020. This book focuses on the new data on combustion processes having practical applications and includes both fire safety issues in the development of flame arresters and issues in the use of noble metals in hydrogen recombiners for NPP, as well as in catalytically stabilized (CS) combustion technology; the establishment of basic principles of production of metal nanostructures, namely nanopowders of metals and compact products made of them, with the preservation of the unique properties of nanoproducts. This book discusses the effects of gas-dynamic processes and chemical kinetics on the modes of hydrocarbon oxidation processes, including catalytic oxidation on noble metals; catalytic oxidation of hydrogen and deuterium on platinum, palladium, ruthenium and rhodium; new regularities and novel theoretical ideas on combustion of a number of metal nanopowders and compact samples made of them. We wanted also to show how interrelated seemingly different areas of combustion science are. The book focuses also on the experimental investigation into the interrelation of kinetics and gas dynamics in gas combustion. Some modern problems in the area of gas combustion, as well as the methods allowing calculation and estimation of the vii
viii
Preface
conditions of ignition, and flame propagation on the basis of experimental results obtained by the authors of the book are considered as well as new theoretical grounds of combustion of solid nanostructures. However, we have recently shown that the concepts of the classical macroscopic theory of a thermal explosion are quite applicable to the massives of nanoobjects. Therefore, the theoretical analysis in this book is carried out from a unified position of the classical combustion theory. The book focuses on the application of classical combustion theory to the problems on combustion of both gases and solids, stability and preservation of the unique properties of nanopowders and compacts made of nanopowders; the experiments and theory are based on the original works of the authors. The book may be useful for undergraduate and postgraduate students and qualified scientists in the area of experimental studies of combustion processes. In the book, it has been experimentally found that a flame of dilute natural gas– oxygen mixtures does not penetrate through the central opening of a confuser, but it penetrates only through the central opening of a diffuser, even if there were additional openings on the cone elements. The numerical modeling performed using compressible dimensionless reactive Navier–Stokes equations in a low Mach number approximation made it possible to qualitatively interpret the results. The results obtained by the visualization of flame penetration through orifices of different shape are important for the solution of explosion safety problems for volumes of complex geometry such as production floor areas, combustion chambers and chemical reactors. It is established that both minimum diameter of a central opening, through which the flame of the diluted methane–oxygen mix can penetrate, and minimum pressure of flame penetration decrease with an increase in the number of openings. It was experimentally shown that the penetration limit by the diameter of the orifice for a single asymmetrical obstacle (1200 °C) in the process are observed only after the first obstacle, i.e., after turbulization of a
8
1 Gas-Dynamic Factors in Combustion Processes
Fig. 1.3 High-speed filming of FF propagation through the combined obstacle. 15.4% CH4 + 30.8% O2 + 46% CO2 + 7.8% Kr. Initial pressure is 180 Torr. a Interference filter 430 nm, b interference filter 520 nm, c interference filter 590 nm. 300 frames/c. The number on a shot corresponds to a serial number of a shot after the initiation moment
gas stream. The obtained result means that the experiment allows separating in time and space “cold” and “hot” flame kinetic regimes in a single experiment using flow turbulization. The result is also important for verification of numerical models of methane combustion. Therefore, the experiment on the penetration of a flame through obstacles unambiguously shows that gas-dynamic factors, for example, flame turbulization, can determine kinetics peculiarities of combustion, for instance, the transition of the low-temperature regime of hydrocarbon combustion to the high-temperature mode. Summarizing the earlier studies before presentation of the new data obtained by the authors, we will provide some published results that will be needed in the course of the further consideration. The evidence is obtained for the occurrence of the ignition of diluted stoichiometric methane–oxygen mix (total pressure up to 200 Torr) behind a single opening at the transition of the laminar flow to the turbulent one rather than after a delay period of ignition [44]. The features of flame penetration through rectangular openings in comparison with circular ones are experimentally investigated with the use of both color speed cinematography and visualization of gas currents by the illumination of fine powder with a laser sheet [2]. Flame penetration through circular and rectangular openings in methane–oxygen mixtures is shown in Fig. 1.4. As is seen, the “flame jump” in case of the rectangular opening is comparably short and the first spot of ignition is observed near the obstacle surface, as distinct from the flame front propagation through a round opening, behind which a long “flame jump” occurs. An initially undisturbed submerged axisymmetric or plain jet is formed in the gas behind an obstacle (Fig. 1.4a, b, respectively).
1 Gas-Dynamic Factors in Combustion Processes
9
Fig. 1.4 a High-speed filming of FF propagation through the circular opening of 25 mm in diameter, b high-speed filming of FF propagation through the rectangular opening of 10 mm width and 65 mm in length (the slit is placed vertically). Initial pressure is 170 Torr. 15.4% CH4 + 30.8% O2 + 46% CO2 + 7.8% of Kr. Initial pressure is 180 Torr. The figure on each frame corresponds to a frame number after the discharge
After a contact of the flame with the obstacle, the primary ignition centers (the local volumes containing both the active centers of combustion and the gas heated to combustion temperature) arise in this submerged jet. Seemingly, these primary centers move in the submerged jet during the delay period (induction period) of ignition; then, the ignition takes place. Let us roughly estimate the time t of the movement of a primary center in the submerged jet in the approximation of incompressible flow [44]. For the axial velocity component vL in a plane-parallel stream, vL /v0 = 1.2/sqrt(0.1x/L − 0.41); and in an axisymmetric stream, vR /v0 = 0.96/(0.07x/R − 0.29). Here, v0 is the flame velocity at the very moment of contact of FF with the obstacle, x is a coordinate, L is the width of a rectangular slit, R is the radius of the circular opening, and numerical values X are empirical parameters from [44]. Then, tL = 0.41 dx/(vL /v0 ) and, respectively, X tR = 0.96 dx/(vR /v0 ). The t R /t L relation [44] for equal values of upper integration limit X = 10, makes ~4 and for X = 3 makes ~2. It means that in case of the planeparallel stream, the primary center travels a given distance (say X) for the time, which is much less than in the case of the axisymmetric stream and, therefore, during the delay period of ignition, the primary center will move away farther from the obstacle than in the case of the axisymmetric stream. On the other hand, the experiment shows (compare Fig. 1.4a, b) that the ignition behind the opening occurs earlier when the flame passes the rectangular opening than in the case of the circular opening. It means that if the length of the “flame jump” were determined by the delay period of ignition, it would be smaller for the circular opening contrary to the experiment. The explanation was as follows. Recently, the coordinates of laminar and turbulent transition in a submerged jet for both flat and round streams at different values of the Reynolds number by means of visualization and the measurements by a heat-loss
10
1 Gas-Dynamic Factors in Combustion Processes
anemometer were determined. It was shown that the length of the laminar part in flat streams is considerably (by 2–5 times) less than in the round one. It gives the grounds to assume that the length of the “flame jump” in the submerged jet formed after the opening is determined by the time of occurrence of the transition from the laminar flow to the turbulent one rather than the time of the ignition delay period in the flammable mixture. Therefore, the length of the “flame jump” after the opening in an obstacle is mostly determined by the time of occurrence of the transition from the laminar flow to the turbulent one rather than the time of an ignition delay period. Evidently, the role of gas dynamics will increase under conditions of cumulation; this will especially manifest itself at flame penetration through conical obstacles.
1.1 Penetration of the Laminar Flames of Natural Gas–Oxygen Mixtures Through Conical Obstacles The hazards of methane explosions and flame deflagrations still represent a threat for chemical plants, mining tunnels, pipes, etc. Accidental fires in process industries can cause enormous losses of human lives and capital [45–48]. A challenge is to eliminate and reduce the consequences of accidental fires and explosions in pipes. To achieve that goal, accurate data concerning model setups are required to understand the characteristics of methane explosions in pipes [49] and in production floor areas, combustion chambers and chemical reactors, which represent volumes of complex geometry. The modeling of turbulent premixed explosions involved in deflagrating flames inside a confined chamber remains a challenging problem particularly with respect to the adequate representation of the burning rate and the structure of the reaction zone. In hazardous gas explosions typically found in process industries, the number and location of openings in obstacles are the parameters (along with the shape of an obstacle, blockage ratio, size, etc.) that affect the severity of such explosions [50, 51]. However, the influence of the parameters is poorly understood. In low-speed turbulent combustion applications, the low Mach number approximation of reactive Navier–Stokes equations is an appropriate basis for a qualitative simulation [2, 32]. The heat release generated by combustion creates flow instabilities in the forms of buoyancy and gas expansion, which in turn intensifies a natural transition from laminarity to turbulence. Turbulence also enhances combustion by increasing the mixing process. In our consideration, a premixed flame propagates first as a laminar spherical front wrinkled by obstacles and becomes a turbulent flame propagating at higher velocities. We have earlier shown that the flame of a natural gas–oxygen mixture does not pass through a cone with the only opening in its center located as a confuser, but the flame readily penetrates through the diffuser. It means that, under the same conditions, the limit of penetration of dilute methane–oxygen flame through a diffuser is markedly less than in the case of a confuser; therefore, the confuser is the most effective flame
1.1 Penetration of the Laminar Flames of Natural Gas …
11
arrester [52]. There are no published data on flame penetration through the cone with the central opening if there are additional openings on the elements of the cone. In addition, we have found that maximum pressure and maximum acoustic intensity are much greater for the plain obstacle with several openings as compared with that with a single opening [53]. The observed pattern is likely a reflection of the fact that two openings and more, especially three ones, are more effective turbulizers than a single one. In this paragraph, we studied the flame propagation through the cone oriented both as a diffuser and as a confuser with round openings with additional ones on the elements of the cone. The numerical modeling of the observed regularities using compressible dimensionless reactive Navier–Stokes equations in the low Mach number approximation was performed. We also considered the means of improving the models taking into account relative contributions of gas dynamic and chemical factors in combustion, which can be used for the simulation of flame propagation through the obstacles of different geometry. The experiments were performed with stoichiometric methane–oxygen mixtures diluted with CO2 and Ar (the reagents of chemically pure grade) at initial pressures of 100–200 Torr and an initial temperature of 298 K in a horizontal cylindrical quartz reactor 70 cm in length and 14 cm in diameter. A pair of spark ignition electrodes was located at the butt-end of the reactor. The reactor was fixed in two stainless steel gateways at the butt-ends, supplied with inlets for gas pumping and gas inflow and a safety shutter, which swung outward when the total pressure in the reactor exceeded 1 atm [1, 2, 44]. Plastic funnels (d = 14 cm) with a central opening and two openings (17 mm in diameter each) on the elements of the cone [the opening angles of the funnels were 55° and 83° (Fig. 1.5)] were oriented both as a diffuser and a confuser and placed at the center of the reactor. The obstacle was fixed in the reactor with a foam ring (see Fig. 1.5). The combustible mixture (15.4% CH4 + 30.8% O2 + 46% CO2 + 7.8% Ar) was prepared (CO2 was added to decrease a flame front velocity and to enhance the Fig. 1.5 Conical obstacle with three openings (opening angle, 83°)
12
1 Gas-Dynamic Factors in Combustion Processes
quality of filming, Ar was added to diminish the discharge threshold). The reactor was filled with the mixture up to a necessary pressure. Then, spark initiation was performed (the discharge energy was 3 J). The speed filming of ignition dynamics and flame front (FF) propagation was carried out from the side of the reactor [1, 2, 44] with a Casio Exilim F1 Pro color high-speed digital camera (frame frequency, 600 s−1 ). Simultaneous detection of radicals CH (A1 D– X2 P) at 431 nm [42] was carried out with the use of the second Casio Exilim F1 Pro high-speed movie camera equipped with a 430 ± 15 nm interference filter (40% transmittance at 430 nm). A video file was stored in computer memory and its timelapse processing was performed. The pressure change in the course of combustion, was recorded by a piezoelectric gage synchronized with the discharge. Acoustic oscillations were recorded with a Ritmix-sensitive (up to 40 kHz) microphone. The audio file was stored in computer memory and analyzed with the Spectra Plus 5.0 software package. The representative experiments of high-speed filming of FF propagation in the combustible mixture at an initial pressure of 165 Torr through the cone obstacle oriented as a confuser and as a diffuser are shown in Fig. 1.6 for the cone opening angles of 55° and 83°. Under our conditions, the flame always penetrated only through the central opening of the diffuser. At the same time, the flame penetrated only through lateral openings of the confuser at the cone opening angle of 55° (Fig. 1.6a, frames 21, 22 and Fig. 1.6c, frame 19). In this case, the combustion was accompanied by a loud and sharp sound and the shutter swung outward. Note that flame propagation under conditions of the diffuser is not accompanied by a sharp sound effect, and the shutter did not swing. Figure 1.7 shows the time dependences of acoustic amplitude for flame propagation, illustrating the aforesaid, in the reactor containing the confuser (Fig. 1.7a) and the diffuser (Fig. 1.7b). With an increase in the value of the opening angle, the flame begins to penetrate through the central opening of the cone (Fig. 1.6c, frames 19, 20). The opening diameters in the conical obstacle are significantly smaller than the minimum diameter of flame penetration through a plain obstacle with a single central opening (20 mm [53]). Therefore, in assessing the fire safety of a room or confinement with several openings, the minimum size of the single opening should not be used because, with an increase in the number of openings, the minimum size decreases. In case of a plain obstacle with three openings (the opening angle is obviously 180°), the flame penetrates through each of the three openings [53]. In our case, a conical cavity influences the development of combustion by the occurrence of reflected acoustic waves with stagnant zones and the interaction of these waves with the initial combustion front, which has generated waves; maximum pressure is, thus, implemented at some distance from the cone top. The complexity of the process provides, in particular, the observed feature that with decreasing the opening angle in the obstacle with a central opening at a certain value of this angle the flame does not penetrate through the opening at all [52] regardless of the existence of additional openings on the elements of the cone.
1.1 Penetration of the Laminar Flames of Natural Gas …
13
Fig. 1.6 High-speed filming of FF propagation through the cone obstacle with a central opening and two openings on the elements of the cone: a confuser (cone opening angle, 55°); b a diffuser (cone opening angle, 55°); c confuser (cone opening angle, 83°); d confuser (cone opening angle, 83°), 430 nm interference filter is placed before the camera; e diffuser (cone opening angle, 83°); f diffuser (cone opening angle, 83°), 430 nm interference filter is placed before the camera. Initial pressure is 165 Torr. The figures on each frame correspond to the frame number after initiating discharge
Conversely, with an increase in the opening angle in the obstacle, its shape tends toward the plain one, at which the flame penetrates through each of the three openings at the same time [53]; therefore, at a certain value of this angle, the flame will penetrate through all openings. This is shown in Figs. 1.6c and 1.8b. Numerical modeling based on compressible dimensionless reactive Navier– Stokes equations in the low Mach number approximation showed a qualitative consent with experiments (see the Abstract of this chapter). The problem was solved by finite element analysis with the FlexPDE 6.08 package [37]. Initiation condition was taken as T = 10 on the boundary of the channel; there was a conical obstacle with two additional openings on the elements of the cone in the center of the channel. Boundary conditions (including the obstacle) were C x = 0, C y = 0, n = 0, u = 0, v = 0, ρ x = 0, ρ y = 0, and convective heat exchange T t = T − T 0 . A chemical reaction was presented by a single first-order Arrhenius equation. The
14
1 Gas-Dynamic Factors in Combustion Processes
Fig. 1.7 Time dependence of acoustic perturbation amplitudes at flame propagation in the gas mixture at an initial pressure of 165 Torr. a Confuser, corresponding frames from Fig. 1.6a are shown; the center of each frame corresponds to the current time; b diffuser, corresponding frames from Fig. 1.6b are shown; the center of each frame corresponds to the current time
Fig. 1.8 Qualitaive calculation of flame penetration through a conical obstacle. Change in dimensionless temperature for flame propagation. a A confuser for a simple Arrhenius reaction, the opening angle is 100°; b a confuser for a simple Arrhenius reaction, the opening angle is 150°; c a diffuser for a simple Arrhenius reaction, the opening angle is 100°. The scale of dimensionless temperature is presented to the right
1.1 Penetration of the Laminar Flames of Natural Gas …
15
results of qualitative calculations of flame penetration through the conical obstacle as both the confuser and diffuser are shown in Fig. 1.8. The calculations are in qualitative consent with the experiments shown in Fig. 1.6; namely, the flame penetrates through the diffuser (Fig. 1.8c) and does not propagate through a central opening of the confuser (Fig. 1.8a) with the smaller opening angle (100°). At the larger opening angle (150°, Fig. 1.8b), the flame penetrates through all the three openings in the confuser in a qualitative agreement with experimental data (see Fig. 1.6). In the case of a plain obstacle with three openings (one of them is a central one, and the opening angle is 180°), the flame penetrates through each of the three openings [53]. Evidently, the qualitative consideration (a single Arrhenius reaction instead of a complete chemical mechanism, two-dimensional modeling, etc.) does not allow one to obtain the angle, at which the flame begins to penetrate through the central opening of the confuser. In addition, such a qualitative difference from the process of flame penetration through a plain obstacle with the central opening indicates a noticeable role of interaction of acoustic fluctuations in the reactor containing a conical obstacle with the propagating front of combustion even for subsonic flames. Therefore, regardless of the qualitative consideration, we took into account the main features of flame propagation through a conical obstacle with additional openings on the cone elements. Namely, the flame does not penetrate through the central opening of the confuser, but it penetrates only through that of the diffuser even if there are additional openings on the elements of the conical obstacle. Note that the analysis of a three-dimensional model is necessary for the quantitative description of the flame penetration through a conical obstacle. At the same time, the results of the two-dimensional modeling are in qualitative agreement with experimental data. In addition, the results obtained by the visualization of flame penetration through orifices of different shape are important for the solution of explosion safety problems for volumes of complex geometry such as production floor areas, combustion chambers and chemical reactors.
1.2 The Features of Penetration of Methane–Oxygen Flames Through Flat Obstacles with Several Openings Influence of different obstacles located in the volumes, filled with combustible mixture, on the propagation of FF is investigated for a long time [43, 53]. It is known that if the composition of a gas mixture is far from ignition limits, the velocity of flame front (FF) in the presence of obstacles can increase to supersonic values [15, 17]. The most prominent aspect in the investigation of these accelerated flames is caused by problems of explosion safety [54]. In a breakdown of fire safety, a certain amount of flammable gas can be released into the ambient air. The resulting explosive mixture can endanger the integrity of the vessel, reactor, mine, etc. The power of the explosion depends on the shape of the confinement and number of openings in it,
16
1 Gas-Dynamic Factors in Combustion Processes
namely doors, windows and air vents. Although the global characteristics of the flame acceleration have been investigated by various authors [1, 15, 17, 43, 54, 55], the data basis obtained by locally highly resolved measurement methods, determining process variables like density, temperature, velocity and species concentration, are still rather scarce. The data are very important for the solution of the problems of explosion safety as well as for the validation of computer codes, simulating these accidents. In accordance with a concept of a limit of flame propagation on the diameter of the single opening in an obstacle, there is a critical value of the diameter of the opening, below which the flame does not penetrate through the opening. However, we earlier showed that a diluted methane–oxygen flame penetrates through a closemeshed grid, i.e., through the obstacle consisting of a large number of openings of a very small diameter [24]. It means that the number of openings in an obstacle influences on a limit of flame penetration through an obstacle. This matter, which is directly relevant to explosion safety, is not in essence considered in the literature. We have recently experimentally revealed that at the penetration of FF through obstacles, gas-dynamic factors, for example, turbulization of a flame, show a noticeable feedback with combustion kinetics [24, 44]. It was established in [24] that FF after a single obstacle does not occur in close proximity to an obstacle, the primary center of ignition is observed far from an obstacle surface (“flame jump,” see Abstract of this Chapter). It was shown that the use of a mesh sphere as an obstacle leads to an increase in the length of “jump” of FF behind an obstacle in comparison with a round opening. It is shown that two or more obstacles of both spherical form and flat form considerably suppress FF propagation. It was shown that the length of the “flame jump” after the opening in an obstacle is mostly determined by the time of occurrence of the transition from the laminar to turbulent flow rather than the time of ignition delay period. It was experimentally shown that at the penetration of a flame through obstacles gas-dynamic factors, for example, flame turbulization, can determine the kinetic peculiarities of combustion, for instance transition of low-temperature hydrocarbon combustion to a high-temperature mode [38]. Due to the complexity of branched chain combustion processes and the geometry of containment, the propagation of a flame and the resulting warming-up cannot be simulated with suitable accuracy. The compressible reactive Navier–Stokes equations can be simplified and used to modeling a non-isothermal flow only if the flow with a low Mach number is assumed [1, 56, 57]. In low-speed turbulent combustion applications, the variable-density low Mach number approximation of the Navier– Stokes equations is an adequate basis for simulation. Nevertheless, any comparison of experimentally detected flame front propagation with a result of numerical modeling is credible only in a qualitative aspect, e.g., on propagation velocity of the boundary of initial and reacting gas, as well as on the shape of this border. The present paragraph is focused on the establishment of the regularities of the influence of the number of openings in a flat obstacle on a minimum diameter and minimum pressure of flame penetration through the obstacle. Flame propagation in the stoichiometric mixtures of methane with oxygen diluted with CO2 and Ar at initial pressures of 100–200 Torr and 298 K in a horizontal
1.2 The Features of Penetration of Methane–Oxygen Flames …
17
Fig. 1.9 Experimental installation. 1—reactor, 2—gateway, 3—vacuum gasket, 4—safety shutter, 5—electrodes, 6—spark power unit, 7—Casio Exilim F1 Pro digital camera, 8—obstacle, 9— interference filter
cylindrical quartz reactor 70 cm in length and 14 cm in diameter was investigated. The reactor was fixed in two stainless steel gateways at the butt-ends supplied with inlets for gas pumping and blousing and a safety shutter, which swung outward when the total pressure in the reactor exceeded 1 atm [24]. A pair of spark ignition electrodes was located near the left butt-end of the reactor [24]. Thin obstacles 140 mm in diameter with two circular openings 15 mm in diameter and three circular openings 12 mm in diameter were placed vertically in the center of the reactor (Fig. 1.9). In a few experiments, the obstacle with two symmetrical openings was equipped with two reservoirs where iron nanopowder was placed. Iron nanoparticles, obtained by the method of chemical metallurgy [58], which were blown out of the reservoir through an opening with a gas flow at flame propagation from the left to the right, were ignited with a methane flame. Thus, burning iron nanoparticles visualized the gas flow during combustion. In some experiments, a complex obstacle consisting of the obstacle with two symmetrical openings and a flat obstacle with a single opening of 15 mm in diameter located within 90 mm from the first obstacle was used. The combustible mixture (15.4% CH4 + 30.8% O2 + 46% CO2 + 7.8% Ar) was previously prepared; CO2 was added to enhance the quality of filming by decreasing FF velocity; Ar was added to diminish the discharge threshold. The reactor was filled with the mixture to necessary pressure. Then, spark initiation was performed (discharge energy, 1.5 J). Speed filming of ignition dynamics and FF propagation was carried out from the side of the reactor with a Casio Exilim F1 Pro digital camera (frame frequency, 600 s−1 ) [1, 2, 44, 58]. The interference filter 435 nm (40% filter factor, the halfwidth 15 nm) was applied to select the emission of CH (A1 –X2 ) at 431 nm [42]. The video file was stored in computer memory, and its time-lapse processing was performed. Acoustic oscillations were recorded with a Ritmix-sensitive microphone (up to 40 kHz). The audio file was stored in computer memory and analyzed with the Spectra Plus 5.0 software package. Chemically pure reagents were used.
18
1 Gas-Dynamic Factors in Combustion Processes
We approximately illustrated some features of penetration of the flame through the obstacles under investigation by means of numerical modeling using compressible dimensionless reactive Navier–Stokes equations in low Mach number approximation [33], which describe flame propagation in a two-dimensional channel. The equations showed a qualitative consent with experiments [1, 2]. The solution of the problem was carried out by finite element analysis with the package (FlexPDE 6.08, 19962008 PDE Solutions Inc. [37]). The simple chain mechanism (see Abstract of this chapter) was used. Initiation condition was taken as T = 10 on the left boundary of the channel; there was an obstacle in the channel. Boundary conditions (including the orifice) were C ξ = 0, u = 0, v = 0, ρ ξ = 0 and n = 0 (I type of boundary conditions) only on the obstacle surface, as well as a convective heat exchange T t = T − T 0 where ξ is the dimensionless coordinate. Initial density ρ 0 and consequently initial pressure P0 (see Abstract of this chapter) have been chosen to provide the lack of flame penetration through the single central opening of the same width. We earlier determined the minimum diameter of flame penetration through an obstacle with a single central opening, which in our installation made 20 mm [24], the minimum pressure of flame penetration through this opening made 170 Torr. It should be noted that a safety shutter did not swing outward under these conditions, i.e., the maximum pressure in the reactor was lower than 1 atm. We will use these values for comparison, in the further discussion. The results on high-speed filming of FF propagation in the combustible mixture (see above) at initial pressure 155 Torr through a flat obstacle 14 cm in diameter with two asymmetrically arranged openings 15 mm in diameter are presented in Fig. 1.10a. In Fig. 1.10b, the interference filter 435 nm was placed in front of the lens of the camera; i.e., CH radical distribution during flame propagation was visualized. As shown in Fig. 1.10b, the “flame jump” is longer for the opening, which is located closer to the surface of the reactor. The result of calculations in Fig. 1.10c is in agreement with this experimental fact. Flame propagation is accompanied by a characteristic sound, and a safety shutter opened indicating that the maximum pressure exceeded 1 atm. Dependence of acoustic intensity for a flat obstacle with a central circular opening at initial pressure 170 Torr at initial pressure 170 Torr is shown in Fig. 1.10a, and the dependence for a flat obstacle with two asymmetric openings at initial pressure 155 Torr is shown in Fig. 1.11b. The typical frames of high-speed filming of FF propagation in the combustible mixture at initial pressure 155 Torr through a flat obstacle 14 cm in diameter with two symmetrical openings 15 mm in diameter placed in 55 mm from the center of the obstacle are presented in Fig. 1.12a. In Fig. 1.12b, the interference filter 435 nm is placed in front of the lens of the camera; i.e., CH radicals’ distribution during flame propagation is also recorded. Flame propagation was accompanied by a sharp sound, and a shutter opened indicating that the maximum pressure exceeded 1 atm.
1.2 The Features of Penetration of Methane–Oxygen Flames …
19
Fig. 1.10 High-speed filming of FF propagation through a a flat obstacle 14 cm in diameter with two asymmetrically arranged openings 15 mm in diameter (the first is located in 35 mm from the center; the second is in 55 mm from the center); b the same obstacle, interference filter 435 nm, is placed in the front of the camera. Initial pressure is 155 Torr. The figure on the frame corresponds to a frame number after discharge. 600 frames/s. c Calculation of the process of flame propagation through the obstacle: change in dimensionless concentration of initial substance for n = 0 on the mesh (I type of boundary conditions)
Fig. 1.11 Dependence of the acoustic intensity on the number of openings: a a flat obstacle with a central circular opening 20 mm in diameter. Initial pressure is 170 Torr; b a flat obstacle with two asymmetric openings (Fig. 1.9) 155 Torr
The sequences of frames of high-speed filming of the flame propagation in the mixture under investigation at initial pressure 150 Torr through a flat obstacle 14 cm in diameter with three openings 12 mm in diameter (placed in 55 mm from each other, one opening is in the center) are presented in Fig. 1.13a. In Fig. 1.13b, the interference filter 435 nm is placed in front of the camera; i.e., CH radicals’ distribution during flame propagation is detected. Flame propagation was also accompanied by a sharp sound, and a shutter opened indicating that the maximum pressure exceeded 1 atm.
20
1 Gas-Dynamic Factors in Combustion Processes
Fig. 1.12 High-speed filming of FF propagation through a a flat obstacle 14 cm in diameter with two symmetrically arranged openings 15 mm in diameter (55 mm from the center); b interference filter 435 nm is placed in the front of the camera. Initial pressure is 155 Torr; c the same obstacle, each opening is equipped with a reservoir with iron nanopowder; d calculation of the process of flame propagation through the obstacle: change in dimensionless concentration of initial substance for n = 0 on the orifice (I type of boundary conditions)
Fig. 1.13 High-speed filming of FF propagation through a a flat obstacle 14 cm in diameter with three openings 12 mm in diameter placed in 55 mm from each other, one opening is in the center; b the same obstacle, interference filter 435 nm is placed in the front of the camera. Initial pressure is 150 Torr; c numerical calculation of the process of flame propagation through the obstacle: the change in dimensionless concentration of initial substance for n = 0 on the mesh (I type of boundary conditions)
As shown in Fig. 1.13a, flame penetrates only through two openings; therefore, the conditions of Fig. 1.13 a can be considered limiting for this obstacle. As is shown in Figs. 1.10, 1.12 and 1.13, the diameters of openings for the obstacle in the figures are significantly less than the minimum diameter of flame penetration through an obstacle with a single central opening (20 mm, see above). The total
1.2 The Features of Penetration of Methane–Oxygen Flames …
21
Fig. 1.14 High-speed filming of FF propagation through a complex obstacle containing a a flat obstacle 14 cm in diameter with two symmetrically arranged openings 15 mm in diameter (similar to Fig. 1.11) and the second flat obstacle with a single opening 15 mm in diameter located within 90 mm from the first obstacle. Initial pressure is 155 Torr; b calculation of the process of flame propagation through the complex obstacle: change in dimensionless concentration of initial substance for n = 0 on the mesh (I type of boundary conditions)
pressure of flame penetration is also less than the minimum total pressure for flame penetration through a single opening (170 Torr). However, maximum pressure and maximum acoustic intensity (see Fig. 1.11) are much greater for the obstacle with two openings. The observed pattern is quite likely a reflection of the fact that two openings and more, especially three ones, are more effective turbulizers than a single one. The bear witness to that fact are the results of calculations of the process of flame penetration through the obstacle, which are presented in Figs. 1.10c, 1.12d and 1.13c, which qualitatively agree with the experimental results, are shown in Figs. 1.10a, b, 1.12a, b, c and 1.13a, b correspondingly. We recall that the conditions of calculations were chosen in such a way that flame penetration through the single central opening of the same width as in Fig. 1c did not occur. In Fig. 1.14a, typical frame sequences of flame penetration through a complex obstacle containing a flat obstacle 14 cm in diameter with two symmetrically arranged openings 15 mm in diameter (similar to Fig. 1.11) and the second flat obstacle with a single opening 15 mm in diameter located within 90 mm from the first obstacle at initial pressure 155 Torr are presented. As is seen, the flame penetrates through both obstacles, though if we use a flat obstacle with a single central opening 15 mm in diameter instead of the obstacle with two openings, the flame will not penetrate through that complex obstacle at all. It means that preliminary flame turbulization with the first obstacle under conditions of Fig. 1.14 provides flame penetration through the second obstacle. It can be concluded that at assessment of a fire safety of the room or confinement with several openings one should not use the value of the minimum size of the single opening, because at an increase in the number of openings the size sufficient for flame penetration decreases. Notice that in reality the openings in the rooms are not always symmetrical. Therefore, we will consider the asymmetric case below.
22
1 Gas-Dynamic Factors in Combustion Processes
1.3 Interaction of Laminar Flames of Natural Gas–Oxygen Mixtures with Planar Obstacles with Asymmetrical Openings In many different practical applications, it is necessary to account for the possible consequences of a large-scale accidental release of flammable gas or vapor. The risk of a large explosion represents a principal hazard, which must be adequately assessed as an integral part of the design process for any new installation or operating procedure [59]. Making reasonable predictions of the effects of an accidental explosion is a difficult task due to the complexity of the physical and chemical processes involved. There is a lack of reliable scaling laws which could be used to link small-scale experimental tests to full-scale explosion behavior [59–62]. Full-scale experimental testing is often impracticable or prohibitively expensive, or both. Current industrial practice in explosion modeling generally relies on simplifying the process plant geometry in a rational manner while retaining the principal physical features as far as possible. These models should be very quick run on standard desktop computers, and hence, it is possible to analyze a large number of explosion cases and technological room designs in a very short time. The governing equations for the simulation of gas explosions are the reactive compressible Navier–Stokes equations. However, it is necessary to have trustworthy experimental data when modeling explosions in pipes [49], in technological rooms, combustion chambers and chemical reactors, which are the volumes of complex geometry. Modeling of turbulent deflagration combustion of mixed in advance gas mixtures in the closed volume of complex geometry is a challenging problem, especially due to the need of adequate representation of a reaction zone; it assumes a certain experiment. In this regard, the modeling in small volumes at pressures lower than atmospheric one is very desirable to anticipate the effects expected on a large scale; in addition, it is much less expensive [50]. Besides, such modeling allows determining whether a natural experiment on a large scale in atmospheric pressure might pose a danger to the integrity of the installation and to personnel lives. In explosions of gas taking place in processing industries, the number and location of openings in obstacles along with others, for example, an obstacle form, the size of a relative opening, etc, are the key parameters, which determine the intensity of such explosions [32, 51]. In literature, there is very limited information on the contribution of these parameters. We have shown earlier [24] that the features of penetration of diluted methane– oxygen flames through planar and spherical obstacles are determined for the most part by gas-dynamic processes; thus, the kinetic mechanism of combustion also qualitatively influences on the flame penetration process. For example, flame turbulization can determine the kinetics of combustion, for instance, a transition of lowtemperature hydrocarbon combustion to a high-temperature mode [38]. We showed that the flame after a single obstacle does not occur in the immediate vicinity of an obstacle; the primary center of ignition is observed within a certain distance from
1.3 Interaction of Laminar Flames of Natural Gas–Oxygen …
23
an obstacle surface (flame jump). Two or more close located flat obstacles considerably suppress the flame propagation; the flame jump length after the opening in an obstacle is mostly determined by the time of occurrence of a transition from laminar to turbulent flow rather than the time of an ignition delay period. This paragraph is aimed at identifying patterns of flame penetration through flat obstacles arranged in such a way that the openings of neighboring obstacles are not coaxial in relation to one another. Experimental data were modeled to determine the influence of the obstacles on the flame penetration process as compared to the flat obstacles with the openings arranged along the reactor axis. It is also relevant to provide the safe arrangement of the openings between the neighboring technological rooms. The results of the experiments are compared to combustion simulations. In the experiments, flame propagation in the stoichiometric mixtures of methane with oxygen diluted with CO2 and Ar at initial pressures of 150–200 Torr and initial room temperature in a horizontal cylindrical quartz reactor 70 cm in length and 14 cm in diameter was investigated. The reactor was fixed in stainless steel gateways at the butt-ends, which were supplied with gas inlets for pumping and blousing, and a safety shutter, which swung outward when the total pressure in the reactor exceeded 1 atm [38]. Two spark ignition electrodes were placed near the left butt-end of the reactor (Fig. 1.15). Thin obstacles 140 mm in diameter circular openings 22 mm, 15 mm and 13.5 mm in diameter were placed vertically at the center of the reactor in the ways shown in Fig. 1.15. In the majority of the experiments, the openings in the obstacles were located asymmetrically relative to a reactor axis (40 mm from the center); one, two and three obstacles with the same diameter of the opening were placed in the reactor (Fig. 1.15). In the case of two and three obstacles in the reactor, the distance between them was 40 mm. In some experiments, the first obstacle toward the direction of flame propagation had one symmetrical opening of the appropriate diameter (Fig. 1.16d, left). The combustible mixture (15.4% CH4 + 30.8% O2 + 46% CO2 + 7.8% Ar) was previously prepared; CO2 was added to enhance the quality of filming by decreasing
Fig. 1.15 Experimental installation (1) quartz reactor, (2) stainless steel gateway, (3) Viton seal, (4) stainless steel shutter, (5) spark electrodes, (6) spark ignition circuit, (7) high-speed color movie camera, (8) rail, (9) obstacle (Fig. 1.16c), (10) interference filter 435 ± 15 nm
24
1 Gas-Dynamic Factors in Combustion Processes
Fig. 1.16 High-speed filming of FF propagation through a 2 circular asymmetrical openings 22 mm in diameter, upper frame sequence—no optical filters, lower frame sequence—interference filter 435 ± 15 nm; b 3 circular asymmetrical openings 22 mm in diameter, no optical filters; c an asymmetrical opening 15 mm in diameter, upper frame sequence—no optical filters, lower frame sequence—interference filter 435 ± 15 nm; d three circular asymmetrical openings 15 mm in diameter, no optical filters; e one circular symmetrical opening and two circular asymmetrical openings 15 mm in diameter, no optical filters. Initial pressure is 180 Torr. The figure on a frame corresponds to a frame number after discharge. The configurations of obstacles (flame propagates from left to right) are shown on the left of the figure
FF velocity; Ar was added to diminish the discharge threshold [1, 2]. The reactor was filled with the mixture to necessary pressure. Then, spark initiation was performed (discharge energy, 1.5 J). Speed filming of ignition dynamics and FF propagation was carried out from the side of the reactor with two Casio Exilim F1 Pro digital cameras (frame frequency, 600 s−1 ) [38]. The interference filter 435 nm (40% filter factor, the half-width 15 nm)
1.3 Interaction of Laminar Flames of Natural Gas–Oxygen …
25
was applied to select the emission of CH (A1 –X2 ) at 431 nm [38]. The video file was stored in computer memory, and its time-lapse processing was performed. Chemical pure reagents were used. Flame quenching is considered using blocking ratios (BRs) (BR = 1 − (d/D)2 , where d and D denote the orifice and tube inner diameters, respectively) of a single orifice with a round opening. In the work, the obstacles with openings of BR = 0.975 (diameter 22 mm), 0.981 (diameter 19 mm) and 0.989 (diameter 15 mm) were examined. First, the limit value of the diameter of the single central opening was determined. It made 20 mm for our installation at total pressure 180 Torr in accordance with Sect. 1.2. The results on high-speed filming of flame front propagation in the combustible mixture described in the experimental at initial pressure 180 Torr through flat obstacles 14 cm in diameter with differently arranged openings are presented in Fig. 1.16. In Fig. 1.16a, typical frame sequence of flame penetration through two circular asymmetrical openings 22 mm in diameter (the value is above the penetration limit) is shown; upper frame sequence corresponds to the lack of optical filters. In the lower frame sequence, the interference filter 435 ± 15 nm is placed in front of the lens of the camera; i.e., CH radical distribution during flame propagation is visualized. Further, we showed that the flame penetrates through three circular asymmetrical openings 22 mm in diameter as well (Fig. 1.16b). It was also shown that the flame penetrates through two circular asymmetrical openings 19 mm in diameter, but it does not penetrate through three circular asymmetrical openings 19 mm in diameter. In Fig. 1.16c, typical frame sequence of flame penetration through a single asymmetrical opening 15 mm in diameter is shown; the upper frame sequence corresponds to the absence of optical filters. In the lower frame sequence, the interference filter 435 ± 15 nm is placed in front of the lens of the camera. As is seen, the value of the diameter is lower than the critical diameter of flame penetration limit through a single central opening; i.e., the flame penetrates through the asymmetrical opening of a subcritical diameter as compared to symmetrical case. In Fig. 1.16d, the typical frame sequences of high-speed filming of flame front propagation through three circular asymmetrical openings 15 mm in diameter are shown. It should be noted that the flame does not penetrate through two asymmetrical openings 15 mm in diameter. However, one can provide the penetration by changing the first obstacle with the asymmetrical opening 15 mm in diameter by the obstacle with a symmetrical opening 15 mm in diameter. In Fig. 1.16e, typical frame sequences of flame penetration through a complex obstacle consisting of one circular symmetrical opening and two circular asymmetrical openings 15 mm in diameter are presented, which illustrate the aforesaid. We conclude that it was experimentally shown that the penetration limit by the diameter of the orifice for a single asymmetrical obstacle (97% zirconium) generates hydrogen gas upon the oxidation reaction with steam [65]. The extensive quantities of H2 and steam generated in the BWR system create high pressure and temperatures. This may lead to the reactor failure. Thus, the removal of excess amount of H2 is necessary and extensive research is required to overcome the highly undesirable phenomenon in nuclear industries. Improvement of passive catalytic hydrogen recombiners for the removal of H2 could avoid such risks
28
1 Gas-Dynamic Factors in Combustion Processes
[66]. Further, large-scale fuel cell plants are being developed for power generation for stationary, transportation and portable applications [67]. Often, fuel cells require an excess amount of H2 to achieve stable voltage. The elimination of excess H2 from the exhaust flow is required for safe operation [68]. In addition, residential central heating can be accomplished by hydrogen or natural gas combustion boilers (HCB or NGCB). However, direct hydrogen combustion in NGCB causes NOx emissions because of high operating temperatures. Catalytic hydrogen combustion boilers operate at relatively low temperatures and are able to generate heat without unwanted CO2 and NOx emissions [69]. For H2 combustion reaction, the catalysts should possess properties such as oxygen storage capacity and thermal stability; these should be able to provide H2 oxidation without explosion. That can be achieved using noble metals. The metals have high adsorption capability of H2 and O2 at low temperatures [70]. Moreover, understanding H2 and O2 behavior on the catalyst surface is important for understanding the mechanisms of many commercialized processes such as preferential oxidation and H2 combustion. In catalytic combustors used in gas turbines, the fuel burns at more moderate temperatures to reduce NOx emissions. However, platinum-based catalysts are not effective enough with methane; these eliminate a small fraction of methane contained in the exhaust gases under normal lean combustion operating conditions. Since these Pt-based oxidation catalysts are not very efficient, Pd catalyst can provide higher methane conversion [71]. Nevertheless, the peculiarities of catalytic action of noble metals have been under discussion. In the work [72], it was found for thermal ignition of 2H2 + O2 mixes, that at pressures up to 180 Torr at 288 °C over Pd foil, the catalytic activity of Pd surface is much higher than over Pt foil. The activity of Pd foil expresses itself both in the occurrence of local ignition centers on the foil, from which combustion wave propagates, and in the dark catalytic reaction of consumption of the flammable mixture. In the paper [73], the experimental value of effective activation energy of the ignition process is evaluated as 3.5 ± 1 kcal/mole that is characteristic of surface processes. The work described in this paragraph is performed with a view to accessing the activation energy of the dark reaction. We have shown in [74] that under certain conditions Pd catalysts can suppress the developing flame propagation in diluted methane–oxygen mix due to the high efficiency of Pd surface in the reactions of termination of active centers of combustion. Therefore, even under conditions of high turbulence, kinetic factors can determine combustion regularities. The paragraph is aimed at an establishment of specific features of oxidation of hydrogen and methane over platinum and palladium at low pressures (70–200 Torr). A series of experiments were performed with stoichiometric gas mixtures 2H2 + O2 . A quartz reactor of 4 cm in diameter and 30 cm long heated up with an electric furnace, which temperature was controlled by means of a thermocouple, was used. The reactor was supplied with a removable quartz window at its butt-end. Pd or Pt wires 0.3 mm in diameter, 150 mm long were placed in the reactor (Fig. 1.19a). The pumped and heated reactor was filled with the gas mixture to necessary pressure. Before each experiment, the reactor was pumped down to 10−2 Torr. Total pressure
1.4 Catalytic Activity of Platinum and Palladium in Gaseous …
29
Fig. 1.19 a Experimental installation for the study of the dark reaction under static conditions. 1—quartz reactor of 4 cm in diameter and 30 cm long heated up in the electric furnace, 2—heater, 3—thermocouple, 4—Pd wire, 5—vacuum gauge VDG-1, 6—optical window; b experimental installation for the study of initiated combustion. 1—quartz cylindrical reactor, 2—stainless steel gateway, 3—silicone laying, 4—stainless steel shutter, 5—spark electrodes, 6—power supply, 7— movie Casio Exilim F1 Pro cameras, 8—sensitive microphone “Ritmix,” 9—Pt/Pd cylinder of 3 cm in length inserted into a planar obstacle 14 cm in diameter, 10—interference filter 430 nm
in the reactor was monitored with a vacuum gauge with an indicator; its readings were recorded with a color digital camera Nikon 1. Flame propagation in stoichiometric mixtures of methane with oxygen diluted with CO2 or Kr at initial pressures in the range of 100–200 Torr and 298 K was investigated in the pumped out horizontally located cylindrical quartz reactor 70 cm in length and 14 cm in diameter. The reactor with blunt ends was fixed in two stainless steel frames, supplied with inlets for gas injection and a safety shutter, which swung outward when the total pressure in the reactor exceeded 1 atm. The obstacle was a Pt or Pd cylinder 40 mm in length made of foil 0.3 mm thick and inserted into a vertical well-fitting planar discus 14 cm in diameter (Fig. 1.19b). A pair of spark ignition electrodes was located near the left blunt end of the reactor [75]. The reactor was filled with the mixture up to necessary pressure. Then, spark initiation was performed (the discharge energy was 1.5 J). Speed filming of ignition dynamics and flame front (FF) propagation was performed from the side of the reactor [44] by color highspeed digital Casio Exilim F1 Pro cameras (frame frequency 600 s−1 ). Simultaneous detection of radicals CH (A1 D–X2 P) at 431 nm [44] was carried out by two highspeed movie cameras, one of which was equipped with a 430 nm interference filter.
30
1 Gas-Dynamic Factors in Combustion Processes
Fig. 1.20 a Dependencies of total pressure of the 2H2 + O2 mix on time in the installation: 1—150 °C, 2—200 °C, 3—270 °C, 4—300 °C; b Arrhenius plot of the dependencies (a)
The video file was stored in computer memory, and its time-lapse processing was performed. The combustible mixture (15.4% CH4 + 30.8% O2 + 46% CO2 + 7.8% Kr) was prepared prior to experiment; CO2 was added to enhance the quality of filming by decreasing the flame propagation velocity; Kr was added to diminish the discharge threshold. Chemically pure gases, 99.9% Pt and 98.5% Pd were used. To evaluate the temperature dependence of the dark reaction (the branched chain catalytic reaction of H2 oxidation below the ignition limit), the dependencies of the total pressure of the 2H2 + O2 mix on time were experimentally determined in the order given above (Fig. 1.20a). Notice that under the conditions no dark reaction in the presence of Pt wire instead of Pd was observed. As shown in Fig. 1.20a, the dependencies are straight lines. The dependence of the tangent of the slope of the lines in Arrhenius coordinates is presented in Fig. 1.20b. As shown the figure, the dependence can be approximated by a straight line (the correlation coefficient is 0.982). The data were processed with the use of the program package Statistica 9 (Statsoft). From Fig. 1.20b, one can estimate the value of effective activation energy of the gross reaction E = 4.1 ± 1 kcal/mole that is characteristic of a surface process [74]. It should be noted that the value of activation energy is close to one determined in [76] for the dependence of the H2 fraction at the ignition limit over Pd surface in mixtures with O2 on temperature in Arrhenius coordinates: 3.5 ± 1 kcal/mole. The activated (E = 16.7 kcal/mole [15, 75]) homogeneous branching step H + O2 → O + OH is the slowest elementary reaction of the branching (an increase in the number of active centers) cycle of reactions. Since the value of effective activation energy in the presence of Pd is considerably less, the branching cycle must change [75], e.g., at the expense of the additional branching step. The step can be H + HO2 → 2OH, in which a relatively inactive HO2 turns into active OH; i.e., extra branching occurs. As shown in [75], accounting for that reaction allows explaining an extension of the ignition area in the presence of external H atoms. This assumption is tested numerically in Sect. 6 of Chap. 2.
1.4 Catalytic Activity of Platinum and Palladium in Gaseous …
31
In Fig. 1.21, characteristic experiments on the high-speed filming of FF propagation in the combustible mixture at an initial pressure 170 Torr through the Pt and Pd cylinders of 3 cm in length inserted into a planar discus 14 cm in diameter are shown. As is seen, after ignition, laminar combustion occurs. The ignition after obstacle does not occur in the immediate vicinity of the obstacle; the first spot of ignition is observed at a certain distance from the obstacle surface (Fig. 1.21Ia, frame 24 and Fig. 1.21Ib–d, frame 22), in agreement with [44]. When the flame passes the obstacle, one can observe flame penetration through the obstacle in case of Pt and
Fig. 1.21 I. High-speed filming of FF propagation through a Pt cylinder 25 mm in diameter, b Pt cylinder 20 mm in diameter, c Pt cylinder 20 mm in diameter, filming through the interference filter 430 nm, d Pd cylinder 25 mm in diameter and e Pd cylinder 20 mm in diameter. Initial pressure is 170 Torr. The figure on a frame corresponds to a frame number after discharge; II. Results of the calculation of flame propagation through a single opening. Change in a dimensionless temperature, b active intermediate n for nx = 0 (boundary conditions of type I) in the opening and c active intermediate n for n = 0 (boundary conditions of type II). The scales of temperature T and dimensionless active intermediate n are presented on the right
32
1 Gas-Dynamic Factors in Combustion Processes
Pd cylinders of 25 mm in diameter as well as of Pt cylinder of 20 mm in diameter and a quenching effect at the smaller diameter (20 mm) of Pd cylinder, resulting in the extinction of a flame behind the orifice. It means that the critical diameter of the Pd cylinder is between 20 and 25 mm. The influence of obstacles can be expressed in the double way. On the one hand, the flame interaction with an obstacle can cause the development of flame instability, promoting its acceleration. On the other hand, the contact of a flame with the obstacle surface can lead to an increase in the contribution of heterogeneous reactions, in particular chain termination [75] as well as to an increase in heat losses. In our experiments, the obstacles differ in material. The rate of dark reaction, even of fast H2 oxidation on the Pd surface (see Fig. 1.20b) is too small to attain a noticeable degree of conversion for the time interval of flame penetration through the cylinder (1/600 s, see Fig. 1.21Ia– d). It means that the rate of chain termination determines the occurrence of the critical diameter. Thus, the efficiency of Pd surface in chain termination reaction is much greater than that of Pt. It should be noted that under conditions of counter flame the initiated ignition is carried out simultaneously from both sides of the obstacle, in this experiment the obstacle is not in the middle of the reactor (Fig. 1.22a), the critical diameter gets markedly smaller: the flame penetrates through a Pd cylinder 20 mm in diameter. As is shown in the figure, the boundaries of counter flame fronts interpenetrate each other. This means that gas-dynamic factors are also important, and these should be taken into consideration. We approximately estimated the contribution of chemical factors (chain termination on the noble metal surface) by numerical modeling on the basis of compressible dimensionless reactive Navier–Stokes equations in low Mach number approximation [30] which describe the flame propagation in a two-dimensional channel and showed a qualitative consent with experiments [1, 2, 44]. The problem was solved by a finite element analysis (FlexPDE 6.08 package [37]). The simple chain mechanism (see Abstract of this chapter) was used. Initiation condition was taken as T = 10 on the right boundary of the channel; there was an orifice in the channel. Boundary
Fig. 1.22 a High-speed filming of counter flame propagation through the Pd cylinder 20 mm in diameter and b results of the numerical modeling of the change in dimensionless temperature for n = 0 (boundary conditions of type II). Initial pressure is 180 Torr. The figure in each frame corresponds to a frame number after discharge
1.4 Catalytic Activity of Platinum and Palladium in Gaseous …
33
conditions (including the orifice) were C ξ = 0, u = 0, v = 0, ρ ξ = 0, nξ = 0 and n = 0 only on the inner obstacle surface (this modeled the inner surface of the cylinder made of noble metal), as well as a convective heat exchange T t = T − T 0 , where ξ is a dimensionless coordinate. The results of calculations showed that the termination of active intermediate on the inner obstacle surface (n = 0, Figs. 1.21IIc and 1.22b) markedly influences the flame penetration, namely, prevents flame penetration through the opening in comparison with the case of nξ = 0 (Fig. 1.21IIa, b). Hence, regardless of the qualitative consideration, we managed to take into account the efficient action of the active surface on the features of the flame penetration. In conclusion, a Pd catalyst may under certain conditions suppress combustion, as compared with Pt, and thereby show the effect opposite to catalytic one due to high efficiency of Pd surface in the reaction of chain termination. Therefore, kinetic factors can be determining ones even under conditions of turbulence [77]. We shortly summarize the results. The value of effective activation energy of the dark reaction over Pd is evaluated as E = 4.1 ± 1 kcal/mole that is characteristic of a surface process. The value is close to one determined for the dependence of the H2 fraction at the ignition limit over Pd surface in mixtures with O2 on temperature: 3.5 ± 1 kcal/mole. Under our conditions, no dark reaction on Pt wire was observed. It was shown that the rate of chain termination determines the value of the critical diameter for flame penetration through Pt or Pd cylinders; the efficiency of Pd surface in chain termination reaction is much greater than that of Pt. The action of noble metals on the processes of hydrocarbons oxidation is an effective tool to identify important reaction sets in their kinetic mechanism.
1.5 Thermoacoustic Regimes of Combustion of N-Pentane–Air Mixtures in the Region of Negative Temperature Coefficient Studying the ignition of hydrocarbons is of obvious importance, but there has been no complete clarity about the puzzling phenomena intrinsic to that process. These are stepwise ignition and negative temperature coefficients (NTC), observed at considerably low temperatures. NTC is the increase in the delay time of self-ignition with temperature growth in a certain interval of temperatures. It causes undesirable phenomena in internal combustion engines [78, 79]. There is no consensus on the detailed hydrocarbon oxidation mechanisms over this temperature range as well as on an understanding of NTC phenomenon. As is known, a platinum layer on the reactor surface exhibits a promoting action on the hydrogen and hydrocarbon oxidation reactions [80] caused by heterogeneous development of reaction chains [81]. The occurrence of these heterogeneous reactions enhances the possibility of spontaneous ignition of gas mixture at a surface and influences markedly on delay
34
1 Gas-Dynamic Factors in Combustion Processes
times of ignition; the state of the reactor surface is another factor that determines the spatial pattern of ignition. We emphasize once again that the design and operation of advanced reactors such as fuel reformers require reliable kinetic models that capture the dynamics of the reaction. The negative temperature coefficient phenomenon causes a reduction in mixture temperature for increasing inlet temperatures [82–84]. This behavior was previously known for higher hydrocarbons [83, 84], and limited literature exists for methane oxidation. Under practical engine conditions, the ignition characteristics of hydrocarbon fuels can be divided into two classes: those with single-stage autoignition such as aromatics and alcohols, and those with two-stage autoignition such as n-paraffins, isoparaffins and cycloparaffins, with NTC usually observed at temperatures below 850 K. When two-stage ignition occurs, the first-stage ignition assumes an essential role because the second-stage ignition depends on the heat release and intermediate species generated from the first stage. Furthermore, the negative temperature coefficient regime of the total ignition delay spans exactly around this temperature range, which is relevant to engine knock and other related combustion phenomena [85, 86]. Since hydrocarbons with two-stage ignition typically constitute more than 50% of practical fuels [87], engine processes controlled by combustion kinetics, such as homogeneous charge compression ignition would occur in two stages as well; a low-temperature heat release stage is followed by a high-temperature heat release stage. It is then significant to note that fuels with two-stage ignition have been found to offer significant advantages in controlling combustion phasing and extending the homogeneous charge compression ignition operation range [88, 89]. Consequently, it is essential to better understand the NTC phenomenon to develop novel control strategies for optimal fuel economy and lower pollutant emissions. However, the literature data are limited to experiments carried out at relatively low temperatures and pressures [90, 91]. The ability to predict system behavior for particular application situations, such as in the presence of certain diluents, or at high pressures, is largely missing. Sun et al. [92] numerically investigated the transitions from ignition to the flames as well as the combustion dynamics in stratified n-heptane/air mixture, which showed that the rich mixtures with fuel stratification can show knocking and acoustic formation. One-dimensional simulations were performed to study the autoignition and flame propagation of n-heptane/air mixture in a broad temperature range including NTC regime under elevated pressure conditions. According to one-dimensional simulations, steady premixed flame propagation affected by NTC chemistry shows a two-stage behavior, including both hot and cool flame segments [93]. It turns out that, despite all the variety of reacting systems and the conditions of development of chemical reactions in them, it is extremely difficult to determine the mechanism of ignition of the system. Really, according to [94, 95], in the shock tube and rapid compression machine, self-ignition is of kernel nature. We have shown [96] that the ignition of the n-pentane–air mixtures in a rapid mixture injection static reactor at atmospheric pressure begins with the occurrence of a primary center at the most chemically active sites on the surface. The center initiates the propagation of hemispherical flame front with the normal velocity corresponding to the temperature
1.5 Thermoacoustic Regimes of Combustion …
35
Fig. 1.23 Left frame: location of the platinum wire in the reactor; (1–4) consecutive frames of ignition of pentane–air stoichiometric mixture recorded at a frame speed of 600 frames/s. Initial conditions: a pentane–air stoichiometric mixture, T 0 = 638 K, P0 = 2 atm
of the walls and the gas composition; thus, the thermal ignition includes the stages of warming up, local ignition and flame propagation. Therefore, the reaction never starts in the entire volume, if it does not contain active surface, e.g., aerosol. We earlier [82] performed the following experiments to determine whether the reactor surface produces a catalytic effect on the ignition of hydrocarbons. To do this, we introduced into the reactor a platinum wire 0.5 mm in diameter and 0.5 m in length. Figure 1.23 displays a photograph showing the location of the wire in the reactor (the left frame), as well as frames showing the first moments of ignition of the pentane–air mixture in the presence of the catalytic surface. As can be seen from these filming frames, multiple ignition kernels occur along the wire. Without the wire, only one kernel occurs over the temperatures in the NTC interval. The high-speed filming frames shown in Fig. 1.23 correspond to the beginning of the NTC region. Figure 1.24a, b displays pressure oscillograms for ignition under the same conditions in the absence and the presence of the catalytic surface, respectively. Figure 1.24b shows the stages in the process of ignition in the presence of a catalytic surface in the reactor, whereas the oscillogram in Fig. 1.24a shows that
Fig. 1.24 Typical oscillogram of the change in pressure for the ignition of a pentane–air stoichiometric mixture near the NTC region a in the absence of the catalytic surface at initial pressure and temperature of 2.9 bar and 687 K; τ = 0.585 s; b in the presence of the catalytic surface at initial pressure and temperature of 2.9 bar and 687 K; τ = 0.275 s
36
1 Gas-Dynamic Factors in Combustion Processes
in the absence of the catalyst cool flame ignition accompanied with considerably smaller warming-up (if any arises at all) immediately transforms into a hot one. Ignition kernels develop very rapidly, signifying in fact the beginning of hot ignition, while stepwise ignition and cool flames precede the stage of hot ignition. However, the pressure time history of the process is indicative of a cool flame ignition of the mixture, which manifests itself as a small step (Fig. 1.23) that precedes the main rise in pressure and appears much earlier than the ignition kernel recorded by the camera; filming frames show no detectable cool flames. A specific effect on the ignition produced by the introduction into the reactor of a catalytic surface in the form of a platinum wire (Figs. 1.23 and 1.24) is not limited to speed filming and pressure recording. In addition, in different temperature regions, this catalytic surface influences the ignition differently; namely, at low temperature, the catalytic surface has no appreciable effect on the ignition delay time, but over the temperature region, in which NTC is usually observed, the presence of the same surface eliminates this phenomenon. In this case, the centers of catalyzed ignition are located along the surface of the wire; i.e., it serves as an ignition source. Since the ignition delay in this case behaves as if there is no difference between cool flame and hot ignition in the temperature region corresponding to NTC, we can conclude that the catalytic surface eliminates a certain stage of inhibition after the emergence of the cool flame. This fact seems to be very significant for elucidating the nature of the intermediate products of the reaction probably causing the existence of NTC phenomenon. However, it remains unclear how platinum wire localized in a small volume of the reactor has such a noticeable effect on the volume process. This paragraph is aimed at the establishment of the features of the impact of noble metals (Pt, Pd) on specific combustion modes in the area of NTC. The ignition of n-pentane–air mixtures was studied in a rapid mixture injection static reactor (Fig. 1.25). A premixed fuel–air mixture passed from a storage vessel through a solenoid valve into the reactor, which was preliminary evacuated and heated to the desired temperature. The stainless steel reactor, 12 cm in inner diameter and 25 cm in length, consisted of two hemispherical parts and narrow cylindrical one. The design of the heater provided a uniform temperature distribution in the reactor volume [97], which was controlled with a movable thermocouple placed at the reactor surface. In a number of experiments, Pt or Pd wires (0.3 mm in diameter and 80 cm long) were placed along the axis of the reactor in its central part. Experiments were carried out with stoichiometric n-pentane–air mixtures over the pressure range of 2–3 atm. The pressure time history was recorded with a Karat-CI piezoelectric transducer (4 kHz), the signal from which was fed through an ADC into the computer. Before each experiment, the reactor was evacuated with a 2NVR 5D vacuum pump to 10−2 Torr. The pressure in the reactor was measured by a vacuum meter and a standard vacuum gauge. An electromagnetic valve was used to open and close gas lines. N-pentane “Merck” of chemical pure grade, 99.9% Pt and 98.5% Pd were used. First, by means of the direct measurements of temperature [97] in the center of the reactor (10 cm in diameter and 10 cm in length) with thin 25 mm thermocouples
1.5 Thermoacoustic Regimes of Combustion …
37
Fig. 1.25 Schematic diagram of the experimental setup: (1) reactor, (2) electric heater, (3) thermal insulation, (4) valves, (5) mixer, (6) pressure-reducer safety valve, (7) removable cover, (8) semispherical inset, (9) pressure transducer, (10) ADC–computer based data acquisition system, (11) digital voltmeter and (12) spark ignition circuit
at atmospheric pressure and 800–980 K, it was shown that the time of warming up of gas mixture does not exceed 0.2 s, that is much less than the time obtained by the equation considering the only conductive heat exchange. Before the experiments, 0.5-mm-thick inner surface layer of the reactor was scraped off mechanically. In Fig. 1.26a, the experimental data on ignition delay times for a stoichiometric pentane–air mixture in the reactor in the absence (first series of experiments, rectangles) and presence (second series of experiments, triangles) of the catalytic Pt surface are compared. As shown in Fig. 1.26a, in the region of a positive temperature coefficient at lower temperatures, the catalytic Pt surface has almost no effect on the ignition delay time, i.e., on the process of ignition. However, in the region of a negative temperature coefficient, the role of the catalytic Pt surface becomes very significant in accordance with [82]. In the presence of Pt wire, the ignition delay time only decreases with increasing temperature; the NTC phenomenon is missing. Figure 1.27a, b displays pressure oscillograms at ignition under the same conditions in the absence and the presence of Pt. Pt-catalyzed ignition is located along the surface of the wire; i.e., it serves as an ignition source. Since the ignition delay in this case behaves as if there is no difference between a cool flame and hot ignition in the temperature region corresponding to NTC, we can conclude that the catalytic Pt surface eliminates a certain stage of inhibition after the occurrence of the cool flame. As shown in Fig. 1.27a, hot ignition in the absence of catalytic Pt surface within NTC region is accompanied by thermoacoustic oscillations; their maximum amplitude is attained around the middle of the NTC region by temperature. The frequency of the oscillations makes about 500 Hz; this value roughly corresponds to the first mode of oscillations of the hollow vessel with dimensions close
38
1 Gas-Dynamic Factors in Combustion Processes
Fig. 1.26 Temperature dependence of the ignition delay time τ for pentane–air stoichiometric mixture at an initial pressure of 3 atm. The squares and triangles represent the ignition delay times in the absence and presence of a catalytic surface, respectively. a Pt catalytic surface; to the right; b Pd catalytic surface, the inner surface of the reactor is scraped off before experiments. In the box: The frame of combustion of 40% H2 + 60% air mixture at the reactor walls temperature of 316 °C; Pt wire is placed in the stainless steel reactor [1, 98]
to the reactor [99]. However, in the presence of catalytic Pt surface, the oscillations are no longer observed (Fig. 1.27b). According to Lord Rayleigh principle [100] for heat-driven pressure oscillations, thermoacoustic instability is encouraged when the heat release fluctuates in phase with the pressure perturbation. It means that in the presence of Pt, heat release and pressure perturbation during the combustion occur out of phase; it is in agreement with the statement cited above that the catalytic Pt surface eliminates a certain stage of inhibition after the occurrence of the cool flame. It results in changing the rate of heat release during the hot ignition. Seemingly, thin Pt wire occupies a low volume and cannot influence the combustion process in a gas. However, it is known [101] that at the temperature over 500 °C the molecules or clusters of both platinum oxide and platinum metal occur in gaseous phase. We showed earlier [98, 102] that these particles extend into the reactor volume by diffusion and convection (see the box in Fig. 1.26) and act as catalytic centers, on which ignition takes place in the course of the flame front propagation. These are the centers, which strongly influence the combustion mechanism. In the following series of experiments, the Pt wire was extracted out of the reactor and the dependence of the logarithm of ignition delay time on reciprocal temperature for the stoichiometric pentane–air mixture within NTC region was measured again. It turned out that NTC phenomenon was missing. It is evidently due to the action of catalytic Pt-containing particles deposited on the reactor walls after ignition. To restore initial material of the surface (stainless steel), 0.5-mm-thick inner surface layer of the reactor was scraped off mechanically. In Fig. 1.25b, the experimental data on ignition delay times for a stoichiometric pentane–air mixture in the
1.5 Thermoacoustic Regimes of Combustion …
39
Fig. 1.27 Oscillograms of change in pressure for the ignition of pentane–air stoichiometric mixtures within the NTC region a in the absence of the catalytic surface, b in the presence of Pt wire, c in the absence of the catalytic surface (inner surface layer is previously scraped off), d in the presence of Pd wire
reactor in the absence (rectangles) and presence (triangles) of the catalytic Pd surface are compared. As shown in Fig. 1.25b, both in the region of positive temperature coefficient and in the region of negative temperature coefficient, the catalytic Pd surface has almost no effect on the ignition delay time, i.e., on the process of ignition. The obtained result is in agreement with the data [103]. It was shown in the work that in the presence of the Pd foil, the cellular structure of the flame front of the H2 –CH4 –air mixtures is not observed, as compared with the results obtained on the Pt surface. This is due to the greater stability of PdO in comparison with PtO2 , which is very unstable and decomposes over 500 °C (see above). Figure 1.27c, d
40
1 Gas-Dynamic Factors in Combustion Processes
displays pressure oscillograms for ignition under the same conditions in the absence (c) and the presence (d) of the catalytic Pd surface. As shown in Fig. 1.27c, the reactor inner surface was recovered from the mechanical treatment; hot ignition both in the absence and in the presence of catalytic Pd surface within NTC region is accompanied by thermoacoustic oscillations; their maximum amplitude is achieved at the middle of NTC region. It is an additional indication that in the presence of the catalytic surface, which does not noticeably react at flame temperature and does not generate catalytic centers diffusing into a volume, NTC phenomenon occurs. We took an attempt to illustrate qualitatively the influence of the chemical mechanism and heat release by a simple example of a sequence of two chain chemical reactions by means of numerical modeling using the system of compressible dimensionless reactive Navier–Stokes equations in low Mach number approximation presented in Abstract of this chapter, which describes the flame propagation in a two-dimensional area. The following reaction set was analyzed. The reaction velocity was presented by an elementary chain mechanism: C → 2n (w0) and n + C → 2n + n1 + products (β 0 , Q = β 1 ) and n1 + C → n + products (β, Q = β 2 ). We will specify that by consideration of the process of stationary flame propagation, a reaction of chain origination w0 can be neglected [104]. In this case, the equations (f, g) were replaced with the following ones: (initial condition for concentration changes to C 0 = 1). ρ Tt + vTy + uTx − (γ − 1)/γ Pt − (γ − 1)M 2 Pt + u Px + v Py = ∇ 2 T + β1 W + β2 W1 ρ Ct + vC y + uC x = 2 C + w0 − β0 nW −βn1W1 ρ n t + vn y + un x = 2 n + w0 + 2β0 nW + βn1W1 ρ n1t + vn1 y + un1x = 2 n + β0 nW − βn1W1 W = C exp(ς − ς/T ) W1 = C exp(ς1 − ς1 /T )
(5.1)
Equation (h) of the system I (see Abstract of this chapter) can be converted into the form representing Lighthill analogy [36]: Ptt − 1/M 2 ∇ 2 P = q(C P − 1) [β1 Wt + β2 (W1 )t ] The parameters were assumed to be ζ = 4, ζ 1 = 7.5, β 0 = 0.1, β = 0.15, β 1 = 0.22, β 2 = 0.3. Diffusivities Dn = Dn1 = 0.3, T t = T − T 0 . The solution of the problem was carried out by finite element analysis with the package (FlexPDE 6.08, 1996-2008 PDE Solutions Inc. [37]). The results of calculation of thermoacoustic oscillations in the reaction set (5.1) are shown in Fig. 1.28. On the top, the calculation of temperature field is shown. At the bottom, kinetic pressure curves calculated at the point at the top of the reactor (indicated by a square), corresponding to the conversion degree specified on the top of the figure, are shown. Each column corresponds to the initial dimensionless wall temperature specified at the
1.5 Thermoacoustic Regimes of Combustion …
41
Fig. 1.28 Results of numerical calculation of the set of compressible dimensionless reactive Navier–Stokes equations in low Mach number approximation. A detailed explanation of the figure is given in the text
bottom of the figure. It is seen that at T = 6, the oscillations during combustion are most intense, and at T = 3 and T = 9, the oscillations are less intense; it qualitatively agrees with the experiment. As is shown in the figure, the interchange of activation energies of two radical reactions leads to marked changes in the modes of thermoacoustic oscillations. This is most likely because the heat release and pressure perturbation at ζ = 4, ζ 1 = 7.5 during the combustion occur to a greater extent out of phase than at ζ = 7.5, ζ 1 = 4 at the expense of change in time dependencies of heat release during combustion. Therefore, it is clear that the reliable kinetic model must take into account the occurrence of thermoacoustic oscillations; the exclusion of a certain reaction should cause the NTC reaction mode to disappear. We shortly summarize the results of the paragraph. The peculiarities of ignition of premixed stoichiometric n-pentane–air mixtures were studied in a rapid mixture injection static reactor in the presence of metallic Pt and Pd in the region of negative temperature coefficient (NTC). It is shown that in the absence of noble metals thermoacoustic regimes occur within NTC region. However, in the presence of Pt catalyst surface, which reacts with oxygen at the flame temperature and generates catalytic centers diffusing into volume, thermoacoustic regimes of thermal ignition disappear; in other words, the catalytic Pt surface eliminates a certain stage of inhibition after the occurrence of the cool flame and NTC phenomenon vanishes. In the presence of the catalytic surface (Pd), which does not react at the flame temperature and does not generate catalytic centers diffusing into a volume, NTC phenomenon occurs. The detected regularities must be taken into account in numerical simulations of NTC phenomenon. Thus, the detected regularities must be taken into account in numerical
42
1 Gas-Dynamic Factors in Combustion Processes
simulations of NTC phenomenon; namely, the oscillations and NTC phenomenon must both disappear in calculations after excluding a certain reaction or a series of reaction steps from the mechanism. The step must include the superficial reaction of an active center of combustion on Pt surface.
1.6 The Features of Combustion of Hydrogen and Methane in Oxygen and Air in the Presence of Difluorodichloromethane Additives The method of inhibition of flammable gas mixes is widely used in practice to provide fire and explosion safety of the equipment [105–107]. This problem is complicated by the Montreal Protocol on the restriction of the use of allegedly ozonedepleting substances, to which effective inhibitors of combustion—fluorochlorocarbons (named halons in applications of safety, or freons, if they are considered as coolants) were assigned. Despite the ban, the ozone gap over Antarctica found in 1985 remains now almost as big as it was when the Montreal Protocol was signed in 1987. It is excessive to note that the global industrial factories and, especially refrigeration units, are missing in Antarctica. Later, it became clear that ozone gaps arise, as a rule, over zones of tectonic breaks. However, the production of widely used chlorinesubstituted halons such as promising combustion inhibitor and a coolant difluorodichloromethane CF2 Cl2 was stopped [108]. It probably means that the transition to the new series of freons, which do not contain chlorine, is caused only by the interests of the American business. In particular, for consumers, it led to a significant increase in the cost of the equipment and prices for installation and service works [109]. Not only halons are promising fire inhibitors. It is known that some organophosphorus compounds [110] and organometallic compounds [108] inhibit combustion of hydrocarbons more efficiently than, e.g., halon CF3 Br by a factor of about 100. However, they ignite in air and they are toxic, and therefore, they can be used only in places where human personnel are absent. As compared with organophosphorus and organometallic compounds, chlorinesubstituted halons do not ignite in the air under normal conditions, and they are safe for human personnel; these are only unbreathable. Among them, difluorodichloromethane is the cheapest and safe one [111]. Unfortunately, due to the ban, the data on the inhibitor efficiency of CF2 Cl2 are limited and contradictory. The mechanism of flame inhibition by means of halons has not been fully understood. It is known that combustion and explosion are branched chain processes, in which a cycle of reactions (so-called a reaction chain) results in the reproduction of the active centers of the combustion [75]. If the additive, which reacts with the active centers, is injected into a zone of combustion then reaction chains are terminated; it leads to the inhibition of combustion. Notice that the specific reaction of
1.6 The Features of Combustion of Hydrogen and Methane …
43
branching (or the set of the reactions) in the course of methane combustion has not yet been established [112]. Thus, the termination with a molecule of inhibitor can be rather effective if the molecule reacts with the active center, which participates in a branching [113] or propagation [114] step of a reaction chain. It is possible to find out the reaction only by means of kinetic or spectroscopic methods on the basis of direct measurement of the reaction rate of this active center with an inhibitor. Another experimental possibility of establishment of the chemical nature of the elementary steps, which are responsible for branching and termination of reaction chains, is the analysis of the limit phenomena in inhibition of gas flames, as these phenomena are caused by the competition of branching and termination of active centers via reaction with the inhibitor. It is shown in [115] that the most effective “ozone-safe” inhibitor is C4 F10 ; its minimum effective concentration makes 6 vol%. However, the concentration of methane in the stoichiometric mix makes about 9 vol%; i.e., the concentration of fuel is comparable with that of inhibitor. In that case, the additive may cause the change in fuel composition and withdraw the mixture out of the ignition area. Thus, the term “inhibitor” is inapplicable and it is necessary to use the terms “diluent,” “retardant” or “suppressant.” It should be noted that the vast majority of the accidents involving methane explosion take place at atmospheric pressure. Hence, it is important to have an effective (less than 10 vol% to fuel) and a safe inhibitor to reduce the risk of accidents. However, as can be seen from above, the effective amount of the perfluorinated halon additive is ≥6 vol% being comparable with fuel concentration. The establishment of the mechanism of halon inhibition in the case of the reaction of hydrogen oxidation or CO oxidation in the presence of hydrogen is a relatively clear problem because a reaction of chain branching in these processes is known: H. + O2 → O.. + OH. (the comma indicates a free valence) [75]. Atoms H participate in this reaction; therefore, the reactions of H atoms with halon or the products of its decomposition can be the competing reactions of chain termination, because reactions O.. or OH. with halon do not lead to chain termination; when only OH. is terminated, the branching is even maintained. In addition, there is no reason to consider that atoms H play the leading role in methane oxidation. Really, in the works [24, 116], it is assumed that the role of H atoms in hydrocarbon oxidation is not determining one and consists, at least, in participation in longer chains than in hydrogen oxidation. Notice that the estimate of a rate constant at room temperature of termination via CF2 Cl2 is given in [117]: k (H + CF2 Cl2 ) = 5 × 1012 exp(−9500/RT ) cm3 mole−1 s−1 . It means that hydrogen atoms react with halon molecules in the primary center of ignition at 1000 K (experimental value of gas temperature in a spark ignition zone [24]); it can provide inhibition. The paragraph is aimed at the establishment of the inhibition efficiency of CF2 Cl2 additives in the reactions of hydrogen and methane combustion in the air and oxygen as well as at identification of influence of gas dynamic factors on the efficiency of inhibition. The latter is important in practical applications; e.g., in mines, the premixed methane-inhibitor mix cannot be created anyway, because methane can leak out of the place, unknown in advance. A traditional optical emission spectroscopy
44
1 Gas-Dynamic Factors in Combustion Processes
and hyperspectrometers of visible and near-infrared spectral range were used for the establishment of the nature of the light-emitting particles in the presence of halon additive in order to obtain new data on the chemical mechanism of the action of the additive. In experiments, two installations were used. In the first installation (installation 1), both hyperspectrometers and a high-speed color digital camera were used for registration of light emission. The analysis of optical and NIR spectra of combustion of hydrogen and methane in oxygen and air at atmospheric and reduced pressures was carried out. Hyperspectrometers allowed to perform 4D measurements, namely, the 1st dimension is time, 2nd one is wavelength, 3rd one is spectral intensity at this wavelength, 4th one is coordinate of the fragment of an emitting source. In the second installation (installation 2), the spectrograph with crossed dispersion equipped with a digital video camera was used for emission recording. The spectrograph had higher resolution than an optical hyperspectrometer; it allowed to carry out the exact assignment of spectra. The stainless steel reactor 25 cm long and 12 cm in diameter was supplied with a tangential gas input (noted by a blue circle in Fig. 1.29a, b), removable covers and an optical quartz window. Experiments were performed at initial room temperature. The pumped-out reactor was filled with a gas mix from the high-pressure buffer volume up to the necessary pressure; flame initiation was carried out with a spark discharge (1.5 J). If a gas mixture was prepared immediately in the reactor, the components of the mix were injected via the open valve of the tangential input. Because of a sharp pressure difference in the buffer volume and the reactor, a gas whirl occurs in the reactor leading to the reduction of mixing time [97, 118]. The pressure in the process was recorded by means of the tensoresistive Carat DI sensor, from which the signal was transmitted to the computer via A/D converter. The value of the extent of expansion of combustion products εT was determined by
Fig. 1.29 Installation 1. a Scheme of experimental installation; b scheme of the reactor. The warmed-up reactor 1, electromagnetic valve 2, buffer volume 3, cylinder with a gas mix 4, hyper spectrometers 5, digital video camera 6, rotary mirror 7, internal asbestos isolation 8, heater 9, external asbestos isolation 10, optical window 11, pressure sensor 12, A/D converter and computer for data storage and analysis 13, millivolt meter and thermocouple 14 and spark ignition 15. The red line on which 4D-spectral shooting is carried out is shown in Fig. 1.29a. The width of this line is about 1 mm. The blue circle notes the tangential input of gas into the reactor
1.6 The Features of Combustion of Hydrogen and Methane …
45
the value of the maximum pressure Pb developed in the course of combustion of the mix [1, 3]: Pb /P0 = 1 + γ (εT − 1), the value of normal velocity of flame propagation U n was determined from a ratio [1, 3]: U n = V v /εT . P0 is initial pressure, γ is an isentropic exponent of an initial mix, and V v is a visible flame velocity. Before each experiment, the reactor was pumped out to 10−1 Torr. Pressure in the reactor was controlled with a standard vacuum gage, and in the buffer volume with a standard pressure gage. Gases (H2 , O2 , CH4 , CF2 Cl2 ) were of chemically pure grade. Fuel mixtures were mainly stoichiometric ones (see below), and the inhibitor was added in an amount of X vol%. The inhibitor concentration limit was considered as a mean inhibitor concentration X vol%, further X%; at (X − 0.1X)% the initiated ignition occurs, at (X + 0.1X)%, it does not occur, all other things being equal. The pump-down time between experiments was 2 h. Recording of the light emission during combustion was carried out by means of the spectrograph with crossed dispersion STE-1 supplied with the color Sony DCR_SR200E video camera (installation 2); or a hyperspectrometer and a color high-speed movie camera Casio Exilim F1 Pro through an optical window in a removable cover (Fig. 1.29). Hyperspectrometers sensitive over 400–970 nm range and over the near-infrared range 970–1700 nm [102] were used. The experiments on high-speed filming were performed with gas mixes (H2 + air)stoich + 10% CF2 Cl2 , (CH4 + air)stoich + 0–10% CF2 Cl2 , (H2 + O2 ) stoich + 10–15% CF2 Cl2 , (CH4 + O2 )stoich + 10–15% CF2 Cl2 , 7% CH4 + air + 0–1% CF2 Cl2 , 11% CH4 + air + 0–1% CF2 Cl2. The obtained data were processed and stored in computer memory for the further analysis. It was shown that 10% CF2 Cl2 additive does not have a noticeable inhibiting effect on the combustion of the previously prepared mix H2 –air in consent with [119]. If an estimate of the rate constant of the reaction H + CF2 Cl2 5 × 1012 exp (−9500/RT ) cm3 mole−1 s−1 [117] is true one, H atoms are terminated via halon molecules in the primary center of ignition at 1000 K (experimental value of gas temperature in a spark ignition zone [24, p. 93]). The observed weak inhibition means that CF2 Cl2 molecules participate also in some competing reactions of the development of reaction chains. However, just 2% CF2 Cl2 additive in the previously prepared stoichiometric mix of methane with air completely inhibits the initiated ignition. The experiments on the preparation of a gas mix immediately in the reactor were performed. It was shown that the gas mixture prepared by the first injection of 2% CF2 Cl2 and then stoichiometric mix CH4 + air up to 1 atm cannot be ignited. However, if the mix is prepared in reverse order (first stoichiometric mix CH4 + air is injected, and then CF2 Cl2 to 1 atm), then ignition occurs already at 4% CF2 Cl2 (Fig. 1.30a, b). Notice that in the presence of halon, the pressure jump is higher than in its absence (Fig. 1.30 right at the top). As this pressure jump is caused by heat release in the combustion [42], it means that at methane oxidation in the presence of halon, additional heat is released; i.e., halon does not act as a diluent, and it participates in the combustion. Thus, the concentration limits of the initiated combustion known from the literature are meaningful only for previously prepared mixes. If the mixes are prepared
46
1 Gas-Dynamic Factors in Combustion Processes
Fig. 1.30 High-speed filming of the flame front propagation initiated by a spark discharge in the gas mix: a (CH4 + air) stoich , initial pressure is 1 atm, b (CH4 + air) stoich + 4% CF2 Cl2 . First methane mix with air is injected, then 4% CF2 Cl2 to 1 atm. The mix, prepared in reverse order, does not burn. Oscillograms of pressure changes at the initiated ignition of mixes a and b are shown on the left. c Ignition of previously prepared mix (CH4 + air) stoich + 1% CF2 Cl2 , initial pressure 1 atm. d Ignition of the same mix at initial pressure 1.5 atm. Oscillograms of pressure change at the initiated ignition are shown on the right. e Premixed mixture 40% H2 + air in the presence of 10% CF2 Cl2 . Initial pressure is 1.5 atm. The video camera is located on the side of the reactor; in front of the reactor, the hyperspectrometers are placed. Pressure change dependency on time during ignition is shown on the right. The figure on a shot corresponds to a number of a shot after initiation, 600 frames/s
just before the experiment, then the existence or the lack of ignition is determined by both an order of injection of mixture components and the geometry of an installation. Therefore, the inhibition with halons can hardly be used to solve the problems of safety in mines: there a premixed methane-inhibitor mix cannot be created anyway, because methane can leak out of the place, unknown in advance. It was shown that in the reactor, which is not treated by ignitions, it is possible to initiate the ignition of a previously prepared mix (CH4 + air) stoich + 1% CF2 Cl2 (Fig. 1.30c); it is possible to ignite this mix at 1.5 atm only after 30 min pumping (Fig. 1.30d). Thus, even small amounts of reaction products, e.g., of water vapor, make considerable impact on the value of the limit of the initiated ignition. The visible flame velocity at the limit is lower than that of the mix without inhibitor; the flame front in the presence of inhibitor has a cellular shape. As is shown in Fig. 1.30 (down on the right), the flame propagation is followed by flame pressure oscillations. It means that the flame has unstable character.
1.6 The Features of Combustion of Hydrogen and Methane …
47
Fig. 1.31 Influence of various inhibitors on the flammability limits of methane in air: 1— C2 F4 Br2 , 2—C2 F5 H, 3—NAFS-III (mixture of fluorinated hydrocarbons), 4—CHF3 , 5—CH2 F2 . 1–4 experimental data from [119], 5 this work
Pressure oscillations and the cellular flame shape indicate the proximity of the mix containing 1% CF2 Cl2 to the limit of the initiated ignition [24, 42]. In this case, CF2 Cl2 is not a retardant, but an effective inhibitor, whose operating concentration is almost 10 times less than fuel concentration. Indeed, in that case, the additive cannot cause the change in fuel composition and withdraw the mixture out of the ignition area. The results of the experiment on ignition of leaner and richer non-stoichiometric mixtures showed that inhibitor limits make 0.5% for 7% methane–air and 0.5% for 11% methane–air mixes; these values in comparison with the data [42] indicate the high effectiveness of CF2 Cl2 additives (Fig. 1.31). Thus, under conditions of the same installation, the inhibitor limit of the mix (H2 + air)stoich in the presence of CF2 Cl2 exceeds 10%; at the same time, the inhibitor limit of the mix (CH4 + air)stoich makes 1% CF2 Cl2 . It means in consent with [115, 120], that in the work, the evidence is obtained that the active centers of methane and hydrogen combustion, which determine flame propagation, have different chemical nature. In Fig. 1.32, visible and near-IR emission spectra of combustion of the mix (H2 + air)stoich in the presence of 10% CF2 Cl2 are shown. It is shown in Fig. 1.32a that the most intensive visible bands are observed over the range 850–1000 nm; these can be assigned to HF (ν = 3) [121]. Notice that as the flame propagates with a visible velocity 210 cm/s, and ε = 12.5 from Fig. 2e, the normal flame velocity is 16.8 cm/s (see experimental). It is about 15 times less than the normal velocity of flame propagation without halon additive [107]. Thus, the warming-up in the flame front is comparably small; therefore, the bands of alkali metals (Na and K) characteristic of hot flames have low intensity in the spectrum.
48
1 Gas-Dynamic Factors in Combustion Processes
Fig. 1.32 Emission spectra of combustion of the mix (H2 + air)stoich in the presence of 10% CF2 Cl2 at 1.5 atm in visible a and near-IR area, b recorded with hyper spectrometers, c near-IR emitting bands of HF (ν = 2) and H2 O according to [38], d a near-IR emission spectrum of the mix (H2 + air)stoich in the presence of 10% CF2 Cl2 (red), imposed on a near-IR emission spectrum of the mix (H2 + air)stoich
In Fig. 1.32b, a near-IR emission spectrum of combustion of the mix (H2 + air)stoich in the presence of 10% CF2 Cl2 is shown. We carried out the line assignment, using the data [122] presented in Fig. 1.32c. It follows from the comparison of Fig. 1.32b, c that the bands at 1.25 and 1.32 µm relate to vibrationally excited molecules HF (ν = 2). The wide band at 1.4 µm is caused by H2 O radiation [122]. This assignment is illustrated in Fig. 1.32d, in which a near-IR emission spectrum of the mix (H2 + air)stoich in the presence of 10% CF2 Cl2 (red curve), imposed on a near-IR emission spectrum of the mix (H2 + air)stoich is presented. As shown, Fig. 1.32d adequately approximates the spectrum shown in Fig. 1.32b. Thus, in combustion products of the mix (H2 + air)stoich in the presence of 10% CF2 Cl2 , the molecules HF (ν = 2) and (ν = 3) are detected. Notice that molecules HF in the zero vibrational state were observed in [123] in this reaction. It should be noted that IR data [123] of the products of H2 + O2 + CF2 Cl2 combustion allow identifying HF (v = 0), CF4 , COF2 and CF3 Cl. HCl was not detected. FTIR data [124] of the products of CH4 + O2 + CF3 Br allow identifying HF (v = 0), CF4 , COF2 , H2 O and CO2 . HBr bands are very weak; i.e., the amount of halogen hydrides (except HF) in the products is probably small; it must be taken into account in considering the mechanism of inhibition with halons.
1.6 The Features of Combustion of Hydrogen and Methane …
49
We also detected molecules HF (ν = 3) at inhibition of natural gas combustion with octadecafluorodecahydronaphthalene (“artificial blood,” perfluorodecaline C10 F18 ) [125]. The reaction, in which enough energy can be emitted (11,100 cm−1 = 1.38 eV = 37 kcal/mole) to provide vibrational excitation of HF molecules (ν = 3), is almost thermally neutral one [24, 125, 126]: H + CF2 Cl2 → HF + CFCl2
(1)
Since the inhibitor concentration limit of the stoichiometric methane–air mixture is about 1% CF2 Cl2 , the intensity of HF emission is indeed rather low as compared with HF intensity at hydrogen combustion in the presence of the same amount of halon. To increase the sensitivity of the technique by means of addition of greater amounts of halon, oxygen mixes instead of air mixes were used. We draw the attention of the reader to the fact that the measurement of the inhibitor concentration limit is especially relevant for fuel–air combustion at atmospheric pressure, that is, for conditions where the issue of explosion safety arises. In the experiments, when oxygen was used instead of air, the formation of excited particles in the combustion processes studied was investigated; explosion safety issues and, accordingly, inhibitor concentration limits upon combustion in oxygen under reduced pressure were not considered. In addition, we showed that the inhibitor limit of the mix (CH4 + air)stoich makes 1% CF2 Cl2 , while, as is shown below, the mix (CH4 + O2 )stoich + 15% CF2 Cl2 can be ignited at 100 Torr. It provides a rough estimate of the inhibitor concentration limits of methane–oxygen mixtures, which are significantly greater than those of methane–air mixtures. In the experiments described below, gas mixes 2H2 + O2 + 10–15% CF2 Cl2 and (CH4 + O2 )stoich + 10–15% CF2 Cl2 were used. In Fig. 1.33, the results of experiments in the installation 2 are presented; the assignment of spectral bands of HF (ν = 3) on the basis of a spectrum with higher resolution is carried out. In the following series of experiments, the dependence of the change of a signal intensity of HF (ν = 2) and (ν = 3) on the concentration of CF2 Cl2 was investigated. In Fig. 1.34, the emission spectra of the mixes 2H2 + O2 + 10% and 15% CF2 Cl2 and (CH4 + O2 )stoich + 10% and 15% CF2 Cl2 recorded over intervals 400– 970 nm (Fig. 1.34a, b) and 970–1700 nm at an initial pressure of 100 Torr are shown (Fig. 1.34c, d). The spectrum with maximum intensity was chosen from the time sequence of the spectra for each experiment. As shown in Fig. 1.34, the emission of HF (ν = 2, 3) is also observed in the combustion of the stoichiometric methane– oxygen mix in the presence of 10% CF2 Cl2 ; in addition, the maximum intensity of HF bands (ν = 2, 3) in methane combustion is higher than in 2H2 + O2 combustion in the presence of the same amount of halon. It appears unlikely that the maximum concentration of H atoms in CH4 combustion is larger than in H2 combustion. In addition, the active centers of the combustion of hydrogen and methane determining the development of the combustion process differ from each other. It means that the reaction with CF2 Cl2 leading to formation of HF (ν = 2, 3) in methane combustion includes the active center of methane combustion
50
1 Gas-Dynamic Factors in Combustion Processes
Fig. 1.33 Emission flame spectra at initial pressure 100 Torr: a (CH4 + O2 )stoich + 10% CF2 Cl2 , b (H2 + O2 )stoich + 10% CF2 Cl2 , a the region of the spectrum (a), corresponding to the emission of HF (ν = 3) 0.87–0.91 µm processed by means of the Hesperus 3.0 program, b the region of the spectrum (b), corresponding to the emission of HF (ν = 3) 0.87–0.91 µm processed by means of the Hesperus 3.0 program. The emission spectrum of HF (ν = 3) [31] is shown on top. Three lines on each spectrum from top to bottom belong to spectral intervals 910–670 nm, 680–500 nm, 550–420 nm, respectively
rather than hydrogen combustion. As CF2 Cl2 is an effective inhibitor of methane combustion, this active center has to participate in the elementary act of branching or propagation of a reaction chain as well as to contain a hydrogen atom to form HF molecule. The data on rate constants of elementary reactions, which are accessible in the literature, are limited and contradictory; i.e., the available material is not nearly enough to make reasonable assumptions about the mechanism of halon action. According to the latest data, the rate constant of the reaction of atoms of oxygen with halon [127]. O1 D + CF2 Cl2 → 2Cl + products k = 1.2 × 10−10 cm3 mole−1 s−1
(2)
i.e., (2), is a rather fast reaction; however, in this reaction, HF is not formed. In [128], an estimate of the reaction rate constant is given (in [129]; however, it is claimed that this reaction practically does not occur) OH + CF2 Cl2 → products(CClF2 + HOCl) k = 1.0 × 10−12 cm3 mole−1 s−1 (3)
1.6 The Features of Combustion of Hydrogen and Methane …
51
Fig. 1.34 Emission spectra of ignition. a (H2 + O2 )stoich + 10% CF2 Cl2 and (CH4 + O2 ) stoich + 10% CF2 Cl2 , obtained for a spectral interval 400–970 nm. b (H2 + O2 )stoich + 15% CF2 Cl2 and (CH4 + O2 ) stoich + 15% CF2 Cl2 , obtained for a spectral interval 400–970 nm. c (H2 + O2 )stoich + 10% CF2 Cl2 and (CH4 + O2 ) stoich + 10% CF2 Cl2 , obtained for a spectral interval 970–1700 nm, d (H2 + O2 )stoich + 15% CF2 Cl2 and (CH4 + O2 ) stoich + 15% CF2 Cl2 , obtained for a spectral interval 970–1700 nm. Initial pressure is 100 Torr. The spectrum with the maximum intensity is chosen from the time sequence of spectra for each experiment
In this fast reaction, formation of HF is not also considered. However, if the molecules HF are formed in this reaction (e.g., in almost thermoneutral reaction OH + CF2 Cl2 → COFCl + HF + Cl [126]), it could not explain weak inhibition of hydrogen oxidation and strong inhibition of methane oxidation with halon. Notice that in the above reactions, the formation of experimentally detected CF2 O is not discussed at all. Let us consider the possible reactions of CH3 radical. In the reaction of a chlorine atom separation from CCl4 , CFCl3 and CF2 Cl2 with CH3 radicals at 128 °C [130], the reaction rate values obtained for these three reactions are in the relation 1:24:126. The data of [131] allow estimating a rate constant of the reaction (if the results [31, 32] are true ones) CH3 + CF2 Cl2 → products k(4)
(4)
In [131], the rate constants of the following reactions are measured CH3 + CCl4 → CH3 Cl + CCl3 k(5) = 109 exp(−10,000/RT ) cm3 mole−1 s−1 (5)
52
1 Gas-Dynamic Factors in Combustion Processes
CH3 + CCl3 Br → CH3 Br + CCl3 k(6) = 108 exp(−3500/RT ) cm3 mole−1 s−1 (6) The estimate of the rate of reaction (4) makes k(4) ≈ 126 * k(5) in the assumption that the activation energy of a rate constant of the reaction (4) makes some average of activation energies of reactions (5) and (6), i.e., k(4) ≈ 109 exp (−5000/RT ) cm3 mole−1 s−1 . The obtained value of k(4) corresponds to a rather slow reaction, which cannot be responsible for effective inhibition of gas-phase methane oxidation. It should be noted also that completely fluorinated hydrocarbons are worse inhibitors than CF2 Cl2 . For example, as is stated above, the most effective “ozonesafe” inhibitor is C4 F10 ; its minimum effective concentration makes 6% [114]. In addition, we showed earlier that 0.9% perfluorodecaline C10 F18 does not completely inhibit oxidation of stoichiometric methane–air mix; it is required to dilute this mix with carbon dioxide (15%) for complete inhibition [132]. It means that, as well as in case of hydrogen oxidation, inhibition of methane oxidation with CF2 Cl2 is caused by some cycle of reactions, which nature as well as the values of their rate constants, requires a substantial clarification [133]. We shortly summarize the results of the paragraph. It is established that the concentration limit of ignition of a premixed H2 –air mix in the presence of CF2 Cl2 at 1 atm exceeds 10%, whereas the inhibitor limit of ignition of the premixed methane–air mix makes 1% CF2 Cl2 . It means that CF2 Cl2 is an effective inhibitor to prevent undesirable ignition of methane–air mixes at atmospheric pressure. It is experimentally shown that the concentration limits of the initiated combustion known from literature are meaningful only for previously prepared mixes. If the mixes are prepared just before the experiment, then the existence or the lack of ignition is determined by both an order of injection of mixture components and the geometry of an installation. Therefore, the inhibition with halons can hardly be used to address the challenges of safety in mines. Thus, the influence of gas-dynamic factors on the efficiency of inhibition is revealed. It is shown that the active centers of hydrogen and methane combustion determining the development of combustion process have a different chemical nature. Vibrationally excited molecules HF (ν = 3) and HF (ν = 2) are for the first time detected in the products of combustion of hydrogen and methane in the presence of CF2 Cl2 . It can be promising phenomenon in laser chemical applications. It is experimentally shown that the intensity of HF (ν = 3) and HF (ν = 2) bands in methane combustion in oxygen is higher than in 2H2 + O2 combustion in the presence of the same amount of halon. It means that the reaction with CF2 Cl2 leading to HF (ν = 2, 3) formation in methane combustion has to include the active center of methane combustion. Besides, as CF2 Cl2 is an effective inhibitor of methane combustion, this active center has to participate in the elementary act of chain branching or chain propagation as well as to contain a hydrogen atom to form HF molecule.
1.7 Conclusions
53
1.7 Conclusions It has been experimentally found that a flame of dilute natural gas–oxygen mixtures does not penetrate through the central opening of a confuser, but it penetrates only through the central opening of a diffuser, even if there are additional openings on the cone elements. The numerical modeling performed using compressible dimensionless reactive Navier–Stokes equations in the low Mach number approximation made it possible to qualitatively interpret the results. The results obtained by the visualization of flame penetration through orifices of different shape are important for the solution of explosion safety problems for volumes of complex geometry such as production floor areas, combustion chambers and chemical reactors. It is established that both minimum diameter of a central opening, through which the flame of the diluted methane–oxygen mix can penetrate, and minimum pressure of flame penetration decrease with an increase in the number of openings. It was experimentally shown that the penetration limit by the diameter of the orifice for a single asymmetrical obstacle (