Simulation of Hydrogen Auto-Ignition in a Turbulent Co-flow of Heated Air with LES and CMC Approach

Large-Eddy Simulations (LES) with the first order Conditional Moment Closure (CMC) approach of a nitrogen-diluted hydrogen jet, igniting in a turbulent co-flowing hot air stream, are discussed. A detailed mechanism (nine species, 19 reactions) is used to represent the chemistry. Our study covers the following aspects: CFD mesh resolution; CMC mesh resolution; inlet boundary conditions and conditional scalar dissipation rate modelling. The Amplitude Mapping Closure for the conditional scalar dissipation rate produces acceptable results. We also compare different options to calculate conditional quantities in CMC resolution. The trends in the experimental observations are in general well reproduced. The auto-ignition length decreases with an increase in co-flow temperature and increases with increase in co-flow velocity. The phenomena are not purely chemically controlled: the turbulence and mixing play also affect the location of auto-ignition. In order to explore the effect of turbulence, two options were applied: random noise and turbulence generator based on digital filter. It was found that stronger turbulence promotes ignition.     

Heat Transfer Study of the Hencken Burner Flame

The Hencken burner flame is often used in combustion laser diagnostics as a calibration flame because of its near adiabatic condition. For a fast burning H₂ flame, it can tolerate high flow rate and the flame is indeed near adiabatic; however, for a slow burning CH₄ flame, the flow rate is not always high enough to maintain near adiabatic conditions. The heat transfer of the H₂ and CH₄ Hencken burner flames are studied numerically and experimentally. Three heat loss mechanisms are analyzed: the burner surface radiation, the hot gas radiation, and the convection heat transfer between the main flow and the co-flow. The surface radiation produces negligible temperature drop while the gas radiation and the convection heat loss contribute significant temperature drop. Reducing the co-flow rate can decrease the convection heat loss slightly. The temperature drop caused by the heat loss is inversely proportional to the main flow rate. Increasing the burner size and running the flame premixed mode can increase the flow rate and reduce the temperature deviation from the adiabatic equilibrium value. Based on the heat loss and temperature drop analysis, suggestions are given to maintain the flame at near adiabatic conditions.     

The Structure of the Auto-Ignition Region of Turbulent Dilute Methanol Sprays Issuing in a Vitiated Co-flow

This paper presents planar imaging of laser induced fluorescence (LIF) from key reactive species in the auto-ignition region of dilute turbulent spray flames of methanol. High-speed (5?kHz) LIF-OH imaging as well as low speed (10?Hz) imaging of joint LIF-OH-CH₂O is performed. The product of the OH and CH₂O signals is used as a qualitative indicator of local heat release. The burner is kept intentionally simple to facilitate computations and the spray is formed upstream of the jet exit plane and carried with air or nitrogen into a hot co-flowing stream of vitiated combustion products. The studied flames are all lifted but differ in the shape of their leading edge and heat release zones. Similarities with auto-ignition of gaseous fuels, as well as differences, are noted here. Formaldehyde is detected earlier than OH implying that the former is a key precursor in the initiation of auto-ignition. Growing kernels of OH that are advected from upstream, close in on the jet centreline and ignite the main flame. The existence of double reaction zones in some flames may be due to ignitable mixtures formed subsequent to local evaporation of droplets and subsequent mixing. When air is used as spray carrier, reaction zones broaden with distance, possibly due to increased partial premixing and regions of intense heat release occur near the flame centreline further downstream. With nitrogen as carrier, the flame maintains a nominal diffusion-like structure with reaction zones of uniform width and substantially less concentration of heat release on the flame centreline.     

An Experimental Investigation of the Turbulence Structure of a Lifted H2/N2 Jet Flame in a Vitiated Co-Flow

Measurements of mean velocity components, turbulent intensity, and Reynolds shear stress are presented in a turbulent lifted H₂/N₂ jet flame as well as non-reacting air jet issuing into a vitiated co-flow by laser doppler velocimetry (LDV) technique. The objectives of this paper are to obtain a velocity data base missing in the previous experiment data of the Dibble burner and so provide initial and flow field data for evaluating the validity of various numerical codes describing the turbulent partially premixed flames on this burner. It is found that the potential core is shortened due to the high ratio of jet density to co-flow density in the non-reacting cases. However, the existence of flame suppressed turbulence in the upstream region of the jet dominates the length of potential core in the reacting cases. At the centreline, the normalized axial velocities in the reacting cases are higher than the non-reacting cases, and the relative turbulent intensities of the reacting flow are smaller than in the non-reacting flow, where a self-preserving behaviour for the relative turbulent intensities exists at the downstream region. The profiles of mean axial velocity in the lifted flame distribute between the non-reacting jet and non-premixed flame both in the axial and radial distributions. The radial distributions of turbulent kinetic energy in the lifted flames exhibit a change in distributions indicating the difference of stabilisation mechanisms of the two lifted flame. The experimental results presented will guide the development of an improved modelling for such flames.     

Influence of Co-flow on Flickering Diffusion Flame

The influence of air co-flow on flickering methane diffusion flame was studied experimentally using the image processing technique and the proper orthogonal decomposition (POD) analysis. The flickering of the flame is characterized by the mean height, the oscillation amplitude and the Strouhal number, which are measured by the digital image analysis of the diffusion flame. The experiments are carried out for various combinations of burner diameters, fuel velocities and co-flow velocities. With increasing the velocity ratio of the co-flow to the fuel flow, the oscillation amplitude is decreased and the Strouhal number is increased slightly in proportional to the inverse Froude number, while the frequency jump occurs in the low co-flow velocity ratio. These results are commonly observed in all the burners of different diameters, while the critical co-flow velocity ratio to suppress the flickering is found to be increased with increasing the burner diameters due to the influence of Froude number. The POD analysis of the flickering flame shows that the flickering energy is dominant in the first two POD modes and they are axisymmetric except for the zero co-flow velocity case and fully suppressed case. The correlation of POD coefficients in the first two fluctuating POD modes suggests the suppression of large-scale structure of flickering due to the influence of co-flow.     

Three-dimensional Linear Eddy Modeling of a Turbulent Lifted Hydrogen Jet Flame in a Vitiated Co-flow

A new methodology for modeling and simulation of reactive flows is reported in which a 3D formulation of the Linear Eddy Model (LEM3D) is used as a post-processing tool for an initial RANS simulation. In this hybrid approach, LEM3D complements RANS with unsteadiness and small-scale resolution in a computationally efficient manner. To demonstrate the RANS-LEM3D model, the hybrid model is applied to a lifted turbulent N₂-diluted hydrogen jet flame in a vitiated co-flow of hot products from lean H₂/air combustion. In the present modeling approach, mean-flow information from RANS provides model input to LEM3D, which returns the scalar statistics needed for more accurate mixing and reaction calculations. Flame lift-off heights and flame structure are investigated in detail, along with other characteristics not available from RANS alone, such as the instantaneous and detailed species profiles and small-scale mixing.     

高马赫数下超声速燃烧的自点火查表方法

超燃冲压发动机燃烧室工作在高马赫数工况时, 入口来流空气的总焓非常高, 自点火在高焓条件下成为维持火焰稳定的重要物理化学过程. 本文借鉴火焰面/进度变量模型的降维思路, 发展了一种基于化学动力学的自点火建表方法. 通过定义混合分数和进度变量将复杂多维的化学反应降维, 并成功将数据库方法结合到现有的大涡模拟求解器中. 经过测试和验证, 该方法初步具备对超声速自点火燃烧进行仿真描述的能力. 针对自点火诱导的超声速燃烧问题开展数值模拟, 该方法通过查表的方式有效降低了化学反应求解过程中的计算量. 在采用详细化学反应机理时能够准确地再现自点火行为和火焰结构, 并且预测的温度和重要组分分布与实验吻合较好.      

Hysteretic heat transfer study of liquid–liquid two-phase flow in a T-junction microchannel

Liquid–liquid two-phase flow in microchannels is capable of boosting the heat removal rate in cooling processes. Formation of different two-phase flow patterns which affect the heat transfer rate is numerically investigated here in a T-junction containing water-oil flow. For this purpose, the finite element method (FEM) is applied to solve the unsteady two-phase Navier–Stokes equations along with the level set (LS) equation in order to capture the interface between phases. It is shown that the two-phase flow pattern in microchannels depends on the flow initial condition which causes hysteresis effect in two-phase flow. In this study, the hysteresis is observed in flow pattern and consequently in the heat transfer rate. The effect of wall contact angle on the hydrodynamics and heat transfer in the microchannel is investigated to gain useful insight into the hysteresis phenomenon. It is observed that the hysteresis is significant in super-hydrophilic microchannels, while it disappears at the contact angle of 75^°. The effect of water to oil flow rate ratio (Q_wat/Q_oil) on the heat transfer is also studied. The flow rate ratio has a negligible effect on the Nusselt number (Nu) in the dripping regime, while the Nu decreases with an increase of Q_wat/Q_oil in the co-flow regime. The thickness of the oil film, velocity, and temperature distribution are studied in the co-flow regime. It is revealed that the normalized slip velocity reduces at higher values of Q_wat/Q_oil, which causes a reduction in the averaged Nu. In dripping regimes, higher flow rate ratios lead to a more frequent generation of droplet/slugs at a smaller size. The passage of the slugs or droplets increases the local Nu. Larger droplets generated at lower flow rate ratios cause a larger increase in the local Nu than smaller droplets. The temperature and velocity field around the droplets are also illustrated to investigate the heat transfer improvement. The generated vortex at the tip of the oil jet causes an increase in the velocity and Nu on the water side.     

Large Eddy Simulation of a Turbulent Lifted Flame using Conditional Moment Closure and Rate-Controlled Constrained Equilibrium

The objective of this paper is the simulation of a turbulent flame by employing the Rate-Controlled Constrained Equilibrium (RCCE) approach for the chemistry reduction, and Large-Eddy Simulation (LES) coupled with Conditional Moment Closure (CMC) for the turbulence-chemistry interaction modelling. RCCE is a systematic method for mechanism reduction, based on the concept that certain species characterized by faster time scales are in a constrained equilibrium state, determined by the concentration of the species controlled by the chemical kinetics. A general system of differential equations can be derived, independent on the selection of the fast and slow species (which appears as a parameter). The RCCE system is used to compute the conditional source term in the CMC equation. The flame simulated here is a methane flame issuing into a vitiated co-flow formed by hot combustion products, the ??Cabra?? flame, which is controlled by auto-ignition and is therefore sensitive to the chemical mechanism. The results show an influence of the chosen chemistry in the ignition length.     

The Problems of the Turbulent Burning Velocity

The paper reviews the practical problems in measuring a turbulent burning velocity that gives the mass rate of burning. These largely centre on identifying an appropriate flame surface to associate with the turbulent burning velocity, u _t , and the density of the unburned mixture. Such a flame surface has been identified, in terms of the mean reaction progress variable, $\bar {c}$ , for explosive flame propagation in a fan-stirred bomb. Measurement of $\bar {c}$ makes possible an estimation of the flame surface density, ??, from the relationship ${\it \Sigma} =k\bar {c}\left( {1-\bar {c}} \right)$ . It is shown that in such explosions, mass rates of burning derived from the measured total flame surface area agreed well with those found from the measured turbulent burning velocity. Flamelet considerations identify appropriate dimensionless correlating parameters for u _t . As a result, correlations of turbulent burning velocity divided by the effective rms turbulent velocity, are plotted against the turbulent Karlovitz stretch factor, K, for different values of the Markstein number for flame strain rate, Ma_sr. These plots cover a wide range of variables, including pressure and fuels, and are indicative of different regimes of turbulent combustion. At the lower values of K, there is some evidence of increases in u _t and k due to high-frequency flame surface wrinkling arising from flame instabilities. These increase as Ma_sr becomes more negative. It is found from the developed value of the mean flame surface density throughout the flame brush that, to a first approximation, an increase in u _t for a given mixture is accompanied by a proportional increase in the volume of the brush. The analysis shows that the volume fraction of the turbulent flame brush that is reacting is quite small.     

Radiative heat flux characteristics of methane flames in oxy-fuel atmospheres

Oxy-fuel combustion is a promising alternative for power generation with CO₂ capture, where the fuel is burned in an atmosphere enriched with oxygen and CO₂ is used as a diluent. This type of combustion is characterised by uncommon characteristics in terms of thermal heat transfer budget as compared to air supported systems. The study presents experimental results of radiative heat flux along the flame axis and radiant fractions of non-premixed jet methane flames developing in oxy-fuel environments with oxygen concentrations ranging from 35% to 70%, as well as in air. The flames investigated have inlet Reynolds numbers from 468 to 2340. The data collected have highlighted the effects of the flame structure and thermo-chemical properties of oxy-fuel combustion on the heat flux radiated by the flames. It was first observed that peak heat flux increases considerably with oxygen concentration. More generally the radiant fraction increases with both increasing Reynolds number in the laminar regime and oxygen concentration. It was found that despite a difference in flame temperature, the radiative characteristics of the flames (heat flux distributions and radiant fraction) in air were similar to those with 35% O₂ in CO₂. The radiative properties of flames in oxy-fuel atmosphere with CO₂ as diluents appear to be dominated by the flame temperature.     

Influence of heat and mass transfer on the ignition and NOx formation in single droplet combustion

The effect of heat and mass transfer on the ignition, and in a second step on the nitrogen oxide (NO _x ) generation, of single burning droplets is examined in a numerical study. Spherical symmetry with no gravity and no forced convection is presumed; ambient temperature is set at 500 K, below the auto-ignition point. The essentials of a forced droplet ignition by an external energy source are introduced. Two methods are applied: heat introduction at a fixed radial position r and heat introduction at a fixed local equivalence ratio ϕ _r . This study’s distinctiveness compared to previous research is its focus on and its combination of partially pre-vaporized droplets and detailed chemistry, both being technically relevant in kerosene and diesel fuel combustion. The fuel of choice is n-decane (C₁₀H₂₂), and NO _x production is studied exemplarily as a representative group of pollutant emissions. The conducted simulations show a decrease of NO _x formation with an increase of the pre-vaporization rate \Uppsi. \Uppsi. This decrease is generally valid for both methods of heat introduction. However, results on flame stabilization and NO _x production reveal a high sensitivity to parameters of the ignition model. The burning behavior during the initial stages is dominated by the ignition position. Extracting heat from the exhaust gas region of burning droplets shows no impact on the flame position nor on the relative NO _x production. As a consequence, a well-founded modeling of the investigated droplet regime needs to resort to an iterative adaptation of the heat introduction parameters based on the findings of droplet burning and exhaust gas production.     

Flame Stabilization Mechanisms in Lifted Flames

Flame stabilization and the mechanisms that govern the dynamics at the flame base of lifted flames have been subject to numerous studies in recent years. A combined Large Eddy Simulation-Conditional Moment Closure (LES-CMC) approach has been successful in predicting flame ignition and stabilization by auto-ignition, but accurate modelling of the competition between turbulent quenching and laminar and turbulent flame propagation at the anchor point had not been demonstrated. This paper will consolidate LES-CMC results by analysing a wide range of lifted flame geometries with different prevailing stabilization mechanisms. The simulations allow a clear distinction of these mechanisms. It is corroborated that LES-CMC accurately predicts the competition between turbulence and chemistry during the auto-ignition process, the dynamics of turbulent flame propagation can be captured, however, the dynamics of the extinction process are not approximated well under certain conditions. The averaging process inherent in the CMC methods does not allow for an instant response of the transported conditionally averaged reactive species to the changes in the flow conditions and any response of the scalars will therefore be delayed. The dimensionality of the CMC implementation affects the solution and higher dimensionality does no necessarily improve results. Stationary or quasi-stationary conditions, however, can be well predicted for all flame configurations.     

Observations on Upstream Flame Propagation in the Ignition of Hydrocarbon Jets

A variety of investigators have attempted to characterize the mechanisms of how reaction zones stabilize, or propagate, against incoming reactants, particularly in stable lifted jet flames both laminar and turbulent. In this paper, experiments are described that investigate the characteristics of upstream flame propagation in turbulent hydrocarbon jet flames. An axisymmetric, gaseous turbulent jet mixing in air has been selectively ignited at downstream positions to assess the upstream propagation of the bulk reaction zone. The farthest axial position that permitted the reaction zone to propagate upstream after application of the ignition source, referred to as the “upper propagation limit”, or UPL, is determined for a variety of jet and air co-flow parameters. There is an inverse relationship between the upper propagation limit position and the jet Reynolds number. Conversely, there is a direct relationship between the upper propagation limit and the co-flow velocity. Interpretation of the results is related to the velocity at the stoichiometric surface. Global discussion is made as to what these results imply about the stabilization and propagation of turbulent lifted jet flames.     

Leading-Edge Reaction Zones in Lifted-Jet Gas and Spray Flames

An investigation of the leading edge characteristics in lifted turbulent methane-air (gaseous) and ethanol-air (spray) diffusion flames is presented. Both combustion systems consist of a central nonpremixed fuel jet surrounded by low-speed air co-flow. Non-intrusive laser-based diagnostic techniques have been applied to each system to provide information regarding the behavior of the combustion structures and turbulent flow field in the regions of flame stabilization. Simultaneous sequential CH-PLIF/particle image velocimetry and CH-PLIF/Rayleigh scattering measurements are presented for the lifted gaseous flame. The CH-PLIF data for the lifted gas flame reveals the role that ``leading-edge' combustion plays as the stabilization mechanism in gaseous diffusion flames. This phenomenon, characterized by a fuel-lean premixed flame branch protruding radially outward at the flame base, permits partially premixed flame propagation against the incoming flow field. In contrast, the leading edge of the ethanol spray flame, examined using single-shot OH-PLIF imaging and smoke-based flow visualization, does not exhibit the same variety of leading-edge combustion structure, but instead develops a dual reaction zone structure as the liftoff height increases. This dual structure is a result of the partial evaporation (hence partial premixing) of the polydisperse spray and the enhanced rate of air entrainment with increased liftoff height (due to co-flow). The flame stabilizes in a region of the spray, near the edge, occupied by small fuel droplets and characterized by intense mixing due to the presence of turbulent structures. This revised version was published online in July 2006 with corrections to the Cover Date.     

The propagation mechanism of high speed turbulent deflagrations

The propagation regimes of combustion waves in a 30 cm by 30 cm square cross–sectioned tube with an obstacle array of staggered vertical cylindrical rods (with BR=0.41 and BR=0.19) are investigated. Mixtures of hydrogen, ethylene, propane, and methane with air at ambient conditions over a range of equivalence ratios are used. In contrast to the previous results obtained in circular cross–sectioned tubes, it is found that only the quasi–detonation regime and the slow turbulent deflagration regimes are observed for ethylene–air and for propane–air. The transition from the quasi–detonation regime to the slow turbulent deflagration regime occurs at (where D is the tube “diameter” and is the detonation cell size). When , the quasi–detonation velocities that are observed are similar to those in unobstructed smooth tubes. For hydrogen–air mixtures, it is found that there is a gradual transition from the quasi–detonation regime to the high speed turbulent deflagration regime. The high speed turbulent deflagration regime is also observed for methane–air mixtures near stoichiometric composition. This regime was previously interpreted as the “choking” regime in circular tubes with orifice plate obstacles. Presently, it is proposed that the propagation mechanism of these high speed turbulent deflagrations is similar to that of Chapman–Jouguet detonations and quasi-detonations. As well, it is observed that there exists unstable flame propagation at the lean limit where . The local velocity fluctuates significantly about an averaged velocity for hydrogen–air, ethylene–air, and propane–air mixtures. Unstable flame propagation is also observed for the entire range of high speed turbulent deflagrations in methane–air mixtures. It is proposed that these fluctuations are due to quenching of the combustion front due to turbulent mixing. Quenched pockets of unburned reactants are swept downstream, and the subsequent explosion serves to overdrive the combustion front. The present study indicates that the dependence on the propagation mechanisms on obstacle geometry can be exploited to elucidate the different complex mechanisms of supersonic combustion waves. Received 5 November 2001 / Accepted 12 June 2002 / Published online 4 November 2002 Correspondence to: J. Chao (e-mail: jenny.chao@mail.mcgill.ca) An abridged version of this paper was presented at the 18th Int. Colloquium on the Dynamics of Explosions and Reactive Systems at Seattle, USA, from July 29 to August 3, 2001.     

Flow Structure of Swirling Turbulent Propane Flames

Flow structure of premixed propane–air swirling jet flames at various combustion regimes was studied experimentally by stereo PIV, CH* chemiluminescence imaging, and pressure probe. For the non-swirling conditions, a nonlinear feedback mechanism of the flame front interaction with ring-like vortices, developing in the jet shear layer, was found to play important role in the stabilisation of the premixed lifted flame. For the studied swirl rates (S = 0.41, 0.7, and 1.0) the determined domain of stable combustion can be divided into three main groups of flame types: attached flames, quasi-tubular flames, and lifted flames. These regimes were studied in details for the case of S = 1.0, and the difference in the flow structure of the vortex breakdown is described. For the quasi-tubular flames an increase of flow precessing above the recirculation zone was observed when increased the stoichiometric coefficient from 0.7 to 1.4. This precessing motion was supposed to be responsible for the observed increase of acoustic noise generation and could drive the transition from the quasi-tubular to the lifted flame regime.     

Flame holding downstream from a co-flow injectorAccrochage d'une flamme en aval d'un injecteur co-courant

We present numerical results on the flame attachment in the downstream vicinity of the co-flow injector lip that separates the reactive fluids at injection. Two stability diagrams show the domains where the flame is anchored, blown off, or extinguished, in terms of separating plate thickness and injection velocities of both fluids. Different anchoring modes—stagnation point counter-flow holding or edge flame anchorage—are described, depending particularly on the plate rim thickness. To cite this article: C. Nicoli et al., C. R. Mecanique 334 (2006).     

Transient behaviors of a flame over a Tsuji burner

The present study investigated numerically the physical mechanisms underlying the transient behaviors of the flame over a porous cylindrical burner. The numerical results showed that a cold flow structure at a fixed inflow velocity of Uin = 0.6 m/s in a wind tunnel could be observed in two co-existing recirculation flows. Flow variations occur repeatedly until t = 4.71 s, and then a vortex existed steadily behind the burner and no shading occurred. The ignition of flammable mixture led to a rapid rise in gas temperature and a sudden gas expansion. When it reached the stable envelope flame condition, Uin is adjusted to an assigned value. Two blow-off mechanisms were identified. It was also found in the study flame shapes with buoyancy effects agreed with the ones observed experimentally by Tsai. Furthermore, the lift-off flame would appear briefly between the envelopes and wake ones, and was stabilized as a wake flame.     

Structures and stabilization of low calorific value gas turbulent partially premixed flames in a conical burner

Experiments are carried out on partially premixed turbulent flames stabilized in a conical burner. The investigated gaseous fuels are methane, methane diluted with nitrogen, and mixtures of CH₄, CO, CO₂, H₂ and N₂, simulating typical products from gasification of biomass, and co-firing of gasification gas with methane. The fuel and air are partially premixed in concentric tubes. Flame stabilization behavior is investigated and significantly different stabilization characteristics are observed in flames with and without the cone. Planar laser induced fluorescence (LIF) imaging of a fuel-tracer species, acetone, and OH radicals is carried out to characterize the flame structures. Large eddy simulations of the conical flames are carried out to gain further understanding of the flame/flow interaction in the cone. The data show that the flames with the cone are more stable than those without the cone. Without the cone (i.e. jet burner) the critical jet velocities for blowoff and liftoff of biomass derived gases are higher than that for methane/nitrogen mixture with the same heating values, indicating the enhanced flame stabilization by hydrogen in the mixture. With the cone the stability of flames is not sensitive to the compositions of the fuels, owing to the different flame stabilization mechanism in the conical flames than that in the jet flames. From the PLIF images it is shown that in the conical burner, the flame is stabilized by the cone at nearly the same position for different fuels. From large eddy simulations, the flames are shown to be controlled by the recirculation flows inside cone, which depends on the cone angle, but less sensitive to the fuel compositions and flow speed. The flames tend to be hold in the recirculation zones even at very high flow speed. Flame blowoff occurs when significant local extinction in the main body of the flame appears at high turbulence intensities.