Flame Speed Research Articles

Numerical investigation of premixed syngas/air flame evolution in a closed duct with tulip flame formation

In this study, the dynamic behavior of the premixed syngas/air flames in enclosed ducts with obstacles at the blockage ratio of 0.5, as well as without obstacles, was studied via large eddy simulation and the turbulent flame speed closure model. The hydrogen volume fraction (hydrogen content in fuel) was set to 50 %. Results indicate that the simulation reproduced the flame shapes, flame speed and explosion overpressure observed in the experiments. In cases without obstacles, a “tulip” flame with a tulip cusp can be seen, and secondary cusps emerge on tulip lips due to the high-pressure zone ahead of “tulip” flame lips, leading to in a distorted “tulip” flame. In cases with obstacles, the flame develops hemispherical and finger shapes upstream of the obstacles and evolves into a “tulip” flame downstream of the obstacles due to vortices and adverse pressure gradients. The alteration of the flow direction within the unburned and burned zones results in intricate flame shapes. Flame speed and overpressure show a close relationship and exhibit oscillations following the generation of the “tulip” flame. Typically, the velocity, pressure and flow fields adjacent to the flame front can affect dynamic behaviors of premixed syngas/air flames.

Effect of H[formula omitted] addition on NH[formula omitted] flame structure and dynamics in a mixing layer

Hydrogen addition has been widely used to improve the combustion performance of low-reactivity fuels such as NH3. In this study, a series of two-dimensional NH3/H2 flames in mixing layers were investigated numerically. The aim is to assess and interpret the effects of H2 addition on the structure and dynamics of ammonia flames. Detailed chemistry and transport models were considered in the simulations. It was shown that for pure hydrogen or ammonia flames, a classical tribrachial structure was observed. The non-premixed branch in the tribrachial flame of NH3 combustion features high heat release rate (HRR), while the two premixed branches are weak. The mixture fraction corresponding to the maximum HRR of NH3 combustion is significantly lower than the stoichiometric mixture fraction. When blending ammonia with hydrogen, a tetrabrachial flame structure consisting of two premixed branches and two non-premixed branches was observed, where one non-premixed branch corresponds to the reaction of NH3, while the other is formed due to the reaction of H2. It was suggested that the formation of the tetrabrachial flame is related to the preferential diffusion of hydrogen. After decreasing artificially the diffusivity of hydrogen to suppress the preferential diffusion effect, the flame exhibits only three branches with the merging of the two non-premixed branches. By analyzing the flame dynamics at the leading edge, it was found that the flame speed (Sd∗) is negatively correlated with the flame stretch (κ) and scalar dissipation rate (χ), with Sd∗ decreasing almost linearly with κ or χ, consistent with previous studies of methane and hydrogen flames. As the H2 concentration is increased, the values of Sd∗ increase due to the enhanced reactivity.

Study on Explosion Mechanism of Dimethyl Ether/H2-Blended Gas Based on Chemical Kinetics Method

In order to reveal the deflagration mechanism of DME/H2-blended gasses, the micro-mechanism was studied based on the constructed UC San Diego 2016 pyrolysis oxidation mechanism model. The results show that adiabatic flame temperature and laminar flame speed increase with the increase in the equivalence ratio (Φ); they first increase and then decrease with the increase in the hydrogen (H2)-blended ratio (λ), and with the increase in λ, the Φ corresponding to the peak laminar flame speed of the blended gas increases. The addition of H2 increases the consumption of O2, and H2 reacts with CO to form H2O and CO2, promoting complete combustion. When Φ = 1.0–1.2, the equilibrium mole fraction of H and OH-activated radicals reach the maximum, and with the addition of H2, the concentration of activating radicals gradually increases, while the number of promoted elementary reactions increases by two, and the number of inhibited elementary reactions does not increase. Meanwhile, the addition of H2 increases the reaction rate of most reactions on the main chemical reaction path CH3OCH3→CH3OCH2→CH2O→HCO→CO→CO2 of DME and increases the risk of the deflagration of DME/H2-blended gas.

Turbulent Burning Velocity of High Hydrogen Flames

Abstract The turbulent burning velocity (ST) is one of the most important combustion properties controlling combustor operability limits, directly influencing blowoff, flashback, and combustion instabilities. Hydrogen has particularly significant influences on the turbulent flame speed. This paper presents new H2/CH4 data of high pressure, high hydrogen turbulent burning velocities. The datasets were designed to address fundamental questions as well as provide engineering/design relevant insights. This paper presents new scaling analysis of fuel composition, pressure, and preheat temperature effects on turbulent burning velocity. We also discuss the importance of considering what is being held constant (temperature, flame speed, Reynolds number, etc.) when one is analyzing these sensitivities. Data show that hydrogen fraction and pressure cause an increase in turbulent flame speed; whether quantified as raw ST,GC or normalized as ST,GC/SL,0 or ST,GC/SL,max (where SL,0 and SL,max are the unstretched and stretched laminar flame speed, respectively). We also propose that observed increases in ST,GC with pressure are due to increases in Reynolds number and not a kinetics/stretch sensitivity effect. With increasing preheat temperature, ST,GC increases while its normalized value (ST,GC/SL,0 and ST,GC/SL,max) can either increase or decrease, depending upon fuel composition. We also show how these sensitivities vary, depending on whether these comparisons are made at constant raw turbulence intensity urms or at constant normalized urms/SL,0 or urms/SL,max.

Large Eddy Simulation of Ammonia-Hydrogen Non-Premixed Flames Stabilized on a Bluff Body Burner

Abstract The demand for carbon-free fuels such as NH3 and H2 has been growing due to stricter environmental policies on carbon emissions. Since ammonia is easier to store and transport, it is considered an alternative fuel for marine, aviation, and gas turbine industries. However, the narrow flammability, low reactivity, and potential emissions of NOx are important challenges that need to be overcome if ammonia is to be used as a practical fuel for industrial use. Recent works have provided experimental data on the characteristics of NH3 flames under premixed and non-premixed combustion, focusing on flame speed, turbulence-chemistry interaction, and especially the dissociation/cracking of NH3 into H2 and N2. These measurements have encouraged the use of CFD for the simulation and scaling of practical ammonia systems, evaluating the performance of different numerical models for NH3 combustion. In this study, LES is utilized to predict a series of NH3/H2/N2 non-premixed flames stabilized on a bluff body burner. The accuracy of the FGM combustion model combined with LES has been examined, using a commercial CFD solver. The different cracking levels of NH3 into H2 and N2 lead to simulation cases ranging from an H2/N2 flame (i.e., “fully cracked”), to a flame with 50% NH3 (i.e., “half cracked”). The different flame shapes, temperature distributions, and NOx emissions have been simulated, generally matching well experimental data. The developed numerical settings and workflows may be applied to simulating more complex combustors which use NH3 as a fuel.

Impact of Diluents on Flame Stability With Blends of Natural Gas and Hydrogen

Abstract Two potential decarbonization pathways for natural gas (NG)-fueled gas turbine engines include blending hydrogen (H2) into NG and postcombustion carbon capture. H2 blending changes several combustion properties, including flame speed and stretch sensitivity. The use of post-combustion carbon capture systems is typically facilitated by the implementation of exhaust gas recirculation (EGR), where exhaust gases are injected into the inlet of the engine, increasing carbon dioxide (CO2) concentration at the outlet and, hence, increasing the efficiency of carbon capture technologies. In this work, we explore the impact of H2 blending and EGR on the stability of a swirl-stabilized, central-piloted flame. Mixtures of NG and H2 are tested at a range of different diluent compositions, with oxygen varied from 21% to 15% by volume in the oxidizer. In all cases, a constant adiabatic flame temperature is maintained to mimic the operation of a gas turbine at a given turbine inlet temperature. A variable-length combustor is used for testing, where combustor length is varied to understand the dynamic stability characteristics of the system. Results show that EGR and H2 work in opposition to each other, where higher levels of EGR result in poor flame holding and higher levels of H2 result in better flame holding. Increasing H2 generally increases the amplitude of thermoacoustic instability at each condition, a result of the change in flame position in this particular combustor. Importantly, H2 can be added to NG to improve flame holding without significantly decreasing CO2 levels in the products, showing that H2 blending can be a method for counteracting combustor operability issues that arise from high levels of EGR necessary to improve the efficiency of typical carbon capture systems.

Experimental and chemical kinetic study on effects of H2-DME fusion addition on laminar premixed flame speed and flame instability for ammonia composite combustion

Improvement of the laminar flame speed and stability for ammonia combustion by using H2 and DME has received significant attention. In this study, the characteristics of NH3/H2, NH3/DME and NH3/H2/DME were investigated in constant-volume bomb with high-speed schlieren technique and chemical kinetics. The results show that DME or H2 addition mainly acts as “start-up" or “acceleration" for ammonia laminar flame speed increasing, respectively. In NH3/H2/DME laminar flame, the increase of DME addition ratio not only increases the intensity but also advances the onset of self-acceleration. At ammonia substitution ratio of 10 %–70 %, the Markstein length of NH3/H2/DME flame are all significantly improved compared with pure ammonia flame. The H2-DME fusion-addition can mitigate the reduction of flame thickness and keep a relatively minor thermal expansion ratio. In NH3/H2/DME laminar flame, DME addition can increase the Lewis number greater than 1.0 at ammonia substitution ratio of 10 %–40 %, which significantly reduce the thermo-mass diffusion instability in comparison with H2 addition. Based on mixture design method and experimental validation, it was found that the flame speed at 473 K, 5 bar and φ = 1 for NH3/H2/DME blending ratios of 53.94 %/26.61 %/19.45 % are greater than those of NH3/H2 and NH3/DME with the same ammonia substitution ratios.

Turbulent flame propagation of C10 hydrocarbons/air expanding flames: Possible unified correlation based on the Markstein number

In this study, we experimentally analyzed the influence of turbulence and thermal-diffusional effect on the morphology and turbulent flame speed of premixed expanding flames. The turbulent flames of the n-decane/air and decalin/air mixtures were measured using a fan-stirred constant volume combustion bomb generating near-isotropic turbulence at elevated pressures (1, 2, and 5 bar), temperature (443 K), and a wide range of equivalence ratios (0.8–1.6). The results found that the wrinkling degree of the flame front was greater for Ma<0 compared to Ma>0. However, the distribution patterns of the curvature radius of the flame contours exhibited similarities across different equivalence ratios. Furthermore, an increase in turbulence intensity would aggravate the randomness of flame contour distribution. The turbulent expanding flames of n-decane/air and decalin/air mixtures were self-similar under different turbulence intensities and pressures, and this self-similar propagation followed a correlation between the normalized turbulent flame speed and the turbulent flame Reynolds number (ReT,f) to the one-half power. The similarity in normalized turbulent flame speeds was extended from normal alkanes (C4-C8) and isomeric alkanes (C8, C16) to longer straight-chain alkane (n-decane) and cycloalkane (decalin). As the increase of equivalence ratio, their normalized turbulent flame speeds increased nonlinearly due to the thermal-diffusional effect. Additionally, it was found that the classical, effective, and experimental Lewis numbers were confirmed as unsuitable parameters for characterizing the role of thermal-diffusional effect on the turbulent flame speed for hydrocarbon fuels with large molecular weight. Finally, two possible unified correlations were proposed based on the Markstein number, (d〈r〉/dt)/(σSL)=0.178e−0.231MaReT,f1/2 and (d〈r〉/dt)/(σSL)=(0.170−0.0447Ma+0.0067Ma2)ReT,f1/2, which could predict the turbulent flame speeds of hydrocarbon fuels with large molecular weight such as n-alkanes, iso-alkanes, and decalin.

FGM modeling considering preferential diffusion, flame stretch, and non-adiabatic effects for hydrogen-air premixed flame wall flashback

Preferential diffusion plays an important role especially in hydrogen flames. Flame stretch significantly affects the flame structure and induces preferential diffusion. A problematic phenomenon occurring in real combustion devices is flashback, which is influenced by non-adiabatic effects, such as wall heat loss. In this paper, an extended flamelet-generated manifold (FGM) method that explicitly considers the preferential diffusion, flame stretch, and non-adiabatic effects is proposed. In this method, the diffusion terms in the transport equations of scalars, viz. the progress variable, mixture fraction, and enthalpy, are formulated employing non-unity Lewis numbers that are variable in space and different for each chemical species. The applicability of the extended FGM method to hydrogen flames is investigated using two- and three-dimensional numerical simulations of hydrogen-air flame flashback in channel flows. The results of the extended FGM method are compared with those of detailed calculations and other FGM methods. The two-dimensional numerical simulations show that considering both preferential diffusion and flame stretch improves the prediction accuracy of the mixture fraction distribution and flashback speed. The three-dimensional numerical simulations show that the prediction accuracy of the flashback speed, backflow region, and distributions of physical quantities near the flame front is improved by employing the extended FGM method, compared with the FGM method that considers only the heat loss effect. In particular, the extended FGM method successfully reproduced the relationship between the reaction rate and curvature. These results demonstrate the effectiveness of the extended FGM method.Novelty and Significance Statement The novelty of this research is the development of a flamelet-generated mani-fold (FGM) method that explicitly considers preferential diffusion, flame stretch, and non-adiabatic effects. To the best of the authors’ knowledge, no studies have performed numerical simulations of pure hydrogen flames using such an FGM method. The developed FGM method was applied to numerical simula- tions of hydrogen-air premixed flame flashback at an equivalence ratio of 0.5 and reproduced the flashback speed of lean hydrogen-air premixed flame. The applicability of the FGM method to the numerical simulation of hydrogen-air flashback is reported first. This research is significant because the FGM method is one of the most widely used combustion models for premixed combustion, and the development of an accurate FGM method will contribute to the engineering field. The accurate prediction of the flame flashback attempted in this study is particularly important for the development of hydrogen-fueled combustion devices.

Reassessment of the High-Temperature Oxidation of Di Ethyl Ether through Ab Initio Calculations.

Recent investigations of diethyl ether (DEE) high-temperature pyrolysis and fuel-rich oxidation have highlighted the failure of existing kinetic models to describe experimental CO production. The DEE high-temperature pyrolysis and oxidation chemistry is thus investigated through ab initio calculations. Geometries, frequencies, and hindered-rotor potentials of reactants, products, and transition states of key reactions (fuel decomposition radical decomposition and H-abstraction reactions) are calculated with the B2PLYP-D3/def2-TZVPD method, whereas final energies are refined using CCSD(T)/aug-cc-pV(D,T)Z. Temperature- and pressure-dependent rate constants are then derived from either canonical transition state theory (CTST) or ME/RRKM analysis with the inclusion of tunneling effect and hindered-rotor corrections and compared to experimental measurements when available as well as to previously suggested values. This new information is then merged with a C0-C3 core chemistry model and a low-temperature chemistry DEE subset from the literature to propose a new kinetic model for the combustion of DEE. This model is tested successfully against a large database related to the high-temperature oxidation and pyrolysis chemistry of DEE, including ignition delay times, shock tube speciation data, time-resolved CO profiles, laminar flame speeds, flame structures, and jet-stirred reactor data.

Effects of Vent Size and Pressure on Hydrogen Explosion Dynamic Characteristics.

Explosion venting is an effective method to reduce the explosion damage; in order to study the mechanism of an explosion venting process in internal and external space, this paper investigates the influence of vent parameters on hydrogen-air explosion in a rectangular duct through numerical simulation. The model including the internal and external space is first constructed, and then the explosion dynamic behaviors of the full flow field are analyzed under different vent pressures and sizes. The study aims to reveal the coupling effect of the flame, pressure, and flow field on hydrogen explosion venting. The results indicate that the explosion intensity increases with the growth of the vent pressure and the reduction of the vent size. The maximum external overpressure increases to 2.6 and 2.3 times as the vent pressure increased to 10 times or vent size reduced by 90%. The flame and combustible gas mixture evolve from a mushroom cloud into a jet form as the vent size decreases, and vortexes formed at the flame front suppress flame propagation. However, the flame speed increases significantly as the flame passes the vent under the impact of larger pressure gradient, which results in a more violent turbulence intensity and secondary external explosion.

Reduced Kinetic Mechanism for Modeling High-Temperature Ignition and Flames for n-Pentanol.

n-Pentanol has emerged as a promising substitute for fossil fuels due to its potential in reducing greenhouse gas emissions. This work presents a reduced combustion mechanism for modeling high-temperature ignition (T > 1000 K) and flames for n-pentanol/air mixtures. The model comprises only 64 species and 282 chemical reactions. As we know, this is the first comprehensive study to include all available experimental data for these phenomena and this specific fuel/oxidizer mixture, reported in the literature. Our approach involves coupling a detailed submechanism for n-pentanol to the San Diego mechanism, a reduced scheme for C1-C4 hydrocarbons, followed by a systematic reduction process, mainly using sensitivity analysis and steady-state approximation. We quantitatively assessed the agreement between experimental data and simulation results using deviation or error indicators along with graphical comparisons. Remarkably, the ignition delay times and flame speeds calculated using the reduced kinetic model exhibit a high agreement with the experimental data reported in the literature.

Turbulent flame acceleration and deflagration-to-detonation transitions in ethane–air mixture

The deflagration-to-detonation transition (DDT) poses significant risks in the oil, gas, and nuclear industries, capable of causing catastrophic explosions and extensive damage. This study addresses a critical knowledge gap in understanding the DDT of ethane–air mixtures on a large scale, amid increasing industrial utilization and production of ethane. A novel computational framework is introduced, utilizing the finite-volume code named Morris Garages, which incorporates reactive compressible Navier–Stokes equations, adaptive mesh refinement, and correlations of turbulent burning velocities. This model integrates the most recent data on laminar and turbulent burning velocities for premixed ethane–air mixtures, simulating flame acceleration and DDT within a two-dimensional large-scale setting, measuring 21 m in length and 3 m in height, with obstacles mimicking pipe congestion. Two mixture scenarios, lean and near-stoichiometric, are analyzed to evaluate the effects of equivalence ratios on flame propagation and DDT. The simulations, validated against large-scale experimental data from Shell, show reasonable agreement and provide critical insights into the onset conditions of DDT, such as temperature, pressure, flame speed, and turbulent kinetic energy. Furthermore, the ξ–ε detonation peninsula diagram is utilized to explore autoignition and detonation behaviors in ethane–air mixtures.

Conversion of a small size passenger car to hydrogen fueling: simulation of full load performance

Abstract Hydrogen offers a valid choice when considering zero emissions transport. This alternative fuel is well suited for spark ignition (SI) engines and it can be implemented with relatively minor changes in terms of added components. Initial evaluations of feasibility with respect to available space for the fuel tank revealed that it is possible to convert a small size passenger car to hydrogen fueling and maintain a range comparable to the fully electric alternative. Peak power assessment showed that additional boosting was required compared to gasoline operation, close to the limit of compressor surge. As a result, the present study extended the engine speed range so as to evaluate full load performance levels that can be obtained when using hydrogen. 0D/1D simulation was applied for simulating power unit output as well as related control parameters. A dedicated laminar flame speed sub-model was implemented for including fuel chemistry effects. Predicted combustion phasing showed the required changes in ignition timing when switching fuels and revealed that the engine could be operated close to stoichiometry when using hydrogen. A reduction of torque was calculated at low engine speed, even if peak power ratings were practically the same as with gasoline fueling.

Effects of partial fuel cracking on the forced ignition and spherical flame propagation in ammonia/air mixtures

As a zero-carbon fuel, ammonia has promising applications and thereby has received great attention recently. However, compared to traditional hydrocarbon fuels, ammonia has very low laminar flame speed and extremely high minimum ignition energy (MIE). A feasible method to promote ammonia combustion is partial fuel cracking, during which ammonia undergoes decomposition into hydrogen and nitrogen. This work reports a computational study on forced ignition and spherical flame propagation in partially cracked ammonia/air mixtures. The objective is to assess the effects of fuel cracking on the MIE and flame propagation. It is found that even a slight amount of ammonia cracking can significantly reduce the MIE and promote ignition. In addition, the influence of positive stretch rate on the spherical flame propagation speed, which is quantified by the Markstein length, is shown to be greatly affected by ammonia cracking. The mechanism of ignition promotion and change in spherical flame propagation by ammonia cracking is interpreted and discussed.

A comparative analysis of the performances of a syngas port injection plus gasoline direct injection combined-injection spark-ignition engine under lean-homogeneous charge and Lean-Stratified Charge modes

This study conducts a comparative analysis of the performances of a syngas port injection plus gasoline direct injection combined-injection spark-ignition engine operating in Lean-Homogeneous Charge (LHC) and Lean-Stratified Charge (LSC) modes. Results demonstrate that adjusting direct injection timing achieves LHC with early intake stroke injection and LSC with late compression stroke injection. Regarding mixture charge, increasing excess air ratio (λ) or syngas injection proportion (SIP) in LHC mode improves mixture homogeneity, while a higher SIP in LSC mode enhances mixture distribution for ignition. Concerning combustion characteristics, both modes benefit from lower λ and higher SIP with increased peak cylinder pressure, heat release rate, flame speed, and in-cylinder temperature. A higher λ extends CA0-10 and CA10-90, while a higher SIP shortens these durations. Regarding emissions, a higher λ positively affect HC, CO, and NOx emissions, whereas a higher SIP reduces HC, CO and soot but increases NOx emissions. The comparison between LHC and LSC highlights the superior combustion efficiency of LSC, whereas LHC excels in emission control.

On the Influence of H2 Addition on NH3 Laminar Flame Speed under Engine-like Conditions

On the Influence of H2 Addition on NH3 Laminar Flame Speed under Engine-like Conditions

Dynamics of slowly propagating flames: Role of the Rayleigh-Taylor instability

Environmental concerns have motivated the development of alternate fuels and refrigerant working fluids that have low global warming potential. Ammonia (NH3) is a candidate zero-carbon fuel and hydrofluorocarbons (HFCs) such as R-32 and R-1234yf are being adopted as refrigerants. Upon mixing with air, these compounds can sustain flames that are slowly propagating i.e., with laminar flame speeds less than 10 cm/s. These flames are referred to as slow as they are affected by buoyancy-induced flow and radiation heat loss, in contrast to typical hydrocarbon-fueled flames. In this study, we investigate the flame dynamics of slowly propagating refrigerant/air flames, with a focus on the effect of the Rayleigh-Taylor (RT) instability. Dispersion relations, that characterize the range of unstable wavelengths and their growth rates during early stages of instability growth, are derived from two-dimensional direct numerical simulations (DNS, including detailed chemistry and transport) of RT-unstable R-32/air flames. These results are compared to predictions from reduced-order theoretical models. DNS are also performed to study the long-term evolution of such flames and to quantify flame speed enhancements. Results showed that RT-unstable flames can propagate up to 6–7 times the laminar flame speed. Although slow from a flamelet point of view, these flames propagate much faster at larger scales. These results have implications on developing metrics to assess explosion risk, and minimize it while storing, transporting, and utilizing these compounds. (225 words)

Development of Detailed and Reduced Chemical Kinetic Models for Ammonia Gas Turbines

Abstract The power generation sector has been recently moving towards decarbonization and there is an increased interest in replacing conventional fossil fuels with fuels that produce reduced/zero carbon emissions. One such fuel is ammonia (NH3). However, ammonia is hard to ignite, has a low flame speed, and produces a significantly large amount of nitrogen oxide (NOx) emissions. Hence, using 100% ammonia as fuel in gas turbines requires significant modifications and the development of novel combustors. Blending hydrogen with ammonia, however, helps in having better control over the combustion properties. Before utilizing hydrogen-blended ammonia in an actual gas turbine combustor, thorough simulation studies are required to evaluate its performance, possible hazards, and emissions. The literature lacks well-validated chemical kinetic models for the combustion of hydrogen-blended ammonia for undiluted mixtures at gas turbine-relevant conditions (~20 bar). Hence, in this work, we develop a detailed chemical kinetic model for hydrogen blended ammonia combustion and validate it with a wide range of experimental data for both dilute and undiluted mixtures relevant to gas turbine operating conditions. The detailed chemical kinetic mechanism was reduced to a smaller version (32 species mechanism) without significant loss in accuracy using the direct relation graph with error propagation (DRGEP) and full species sensitivity analysis. The resultant mechanism can predict a wide range of experimental results with the least cumulative error and will be a valuable tool in CFD simulations that will enable the development of gas turbines for zero-carbon power generation.

Double flame dynamics in hotspot ignition

Hotspot ignition is a key problem in fundamental flame initiation, combustion control and prevention of abnormal combustion phenomena. Depending on the characteristics of the hotspot and the thermochemical condition of the reactive mixture, subsequent reaction front initiated from a hotspot can vary substantially. The current paper focus on the dynamics and chemistry for one of the most complicated scenarios - hotspot induced double flames, which include a cool and a hot flame segment. This work adopts a recently developed compressible reacting flow simulation code COGNAC to investigate the initiation and propagation of double flame dynamics from hotspot ignition in a one-dimensional spherical coordinate. The results have shown that under atmospheric pressure, double flame initiation is not feasible for sufficiently small hotspot. Under elevated pressures, double flame can be initiated within a much wider hotspot window, exhibiting complex flame dynamics, such as variations in cool and hot flame bifurcation behaviors and their initiation locations. Analysis on thermochemical structure and dynamics of double flame has shown that the interaction of cool and hot flames in a double flame structure is inherently intermutual, in that the leading cool flame enhances trailing hot flame speed through reform of the unburnt mixture, while the trailing hot flame initiation affects the leading cool flame via thermal expansion effects.