Effects of Turbulence on Self-sustained Combustion in Premixed Flame Kernels: A Direct Numerical Simulation (DNS) Study

The effects of mean flame radius and turbulence on self-sustained combustion of turbulent premixed spherical flames in decaying turbulence have been investigated using three-dimensional direct numerical simulations (DNS) with single step Arrhenius chemistry. Several flame kernels with different initial radius or initial turbulent field have been studied for identical conditions of thermo-chemistry. It has been found that for very small kernel radius the mean displacement speed may become negative leading ultimately to extinction of the flame kernel. A mean negative displacement speed is shown to signify a physical situation where heat transfer from the kernel overcomes the heat release due to combustion. This mechanism is further enhanced by turbulent transport and, based on simulations with different initial turbulent velocity fields, it has been found that self-sustained combustion is adversely affected by higher turbulent velocity fluctuation magnitude and integral length scale. A scaling analysis is performed to estimate the critical radius for self-sustained combustion in premixed flame kernels in a turbulent environment. The scaling analysis is found to be in good agreement with the results of the simulations.     

Influence of flow field scaling on flashback of swirl flames

In this paper, the effect of geometrical scaling on the onset of flashback into a cylindrical premixing zone of a swirl flame is investigated. We discriminate two types of flashback. In the first type of flashback the flame propagates upstream inside an already present axial recirculation zone. This flashback is caused by turbulent burning along the vortex axis (TBVA¹) and is controlled by flame extinction inside the recirculation zone. The second type of flashback is caused by combustion induced vortex breakdown (CIVB²). This type of flashback is characterised by the aerodynamic influence of the combustion heat release that leads to propagation of the axial recirculation zone and the flame in upstream direction.To study the effects of geometrical scaling on the flow fields and the two types of flashback, the operation of two geometrically scaled burners are compared at equal Reynolds number. By this method it is possible to observe the flashback phenomena in similar swirl flow fields but with different turbulent scales affecting the combustion process. To check flow field similarity and to indentify the flashback type, the non-reacting and reacting flow fields have been examined by planar particle imaging velocimetry and simultaneous recording of the flame luminescence.It is shown that geometrical scaling of the burner shifts the equivalence ratio at which flashback occurs and that this shift is different for the two types of flashback. Consistency and inconsistency with known scaling and stability criterions is discussed. Analysing the fluid dynamics and turbulent combustion gives a first explanation of why CIVB and TBVA are affected differently by geometrical scaling at constant Reynolds number which is in good agreement with the experimental observations.     

Dynamics of premixed confined swirling flames

Considerable effort is currently being extended to examine the fundamental mechanisms of combustion instabilities and develop methods allowing predictions of these phenomena. One central aspect of this problem is the dynamical response of the flame to incoming perturbations. This question is examined in the present article, which specifically considers the response of premixed swirling flames to perturbations imposed on the upstream side of the flame in the feeding manifold. The flame response is characterized by measuring the unsteady heat release induced by imposed velocity perturbations. A flame describing function is defined by taking the ratio of the relative heat release rate fluctuation to the relative velocity fluctuation. This quantity is determined for a range of frequencies and for different levels of incoming velocity perturbations. The flame dynamics is also documented by calculating conditional phase averages of the light emission from the flame and taking the Abel transform of these average images to obtain the flame geometry at various instants during the cycle of oscillation. These data can be useful to the determination of possible regimes of instability. To cite this article: P. Palies et al., C. R. Mecanique 337 (2009).     

Stationary modes of combustion in a fine-scale turbulent stream

The majority of models of the turbulent combustion of gases are based mainly on intuitive concepts concerning the processes occurring in the flame. The characteristics of a turbulent flame are estimated from considerations of dimensionality and similarity. A detailed review of works on turbulent combustion is given in [1]. Problems on the calculation of the combustion rate in a turbulent stream as a proper value of the equations of heat and mass transfer and of the corresponding boundary conditions have recently been raised. Here too one must rest on assumptions of a semiempirical nature, which in large measure is connected with the inadequate level of development of turbulence theory. In the present work the equation of propagation of the zone of chemical reactions in the stream is averaged statistically by analogy with studies of turbulent flows. Correct averaging is possible at scales of hydrodynamic disturbances smaller than the flame thickness (fine-scale turbulence). The temperature pulsations are related with the size of the heat flux using the theory of mixing lengths. The main influence is specific to effects arising during averaging of the heat release function. Two stationary modes, distinguished by the normal propagation velocity ₁, are isolated within the framework of the Cauchy problem with a given initial mixture temperature and zero heat flux in the burned gas. A heat conduction mode occurs with a stream velocity > ₁ and an induction mode with < ₁. An expression is found for ₁ which reflects the principal effects in the flame and which in the limit coincides with the equation of Zel'dovich and Frank-Kamenetskii for a laminar flame. In those cases when the distorting effect of the heat release function is small, the turbulence affects the combustion rate through mechanisms of intensification of transport processes.Translated from Zhurnal Prikladnoi Mekhaniki i Tekhnicheskoi Fiziki, No. 5, pp. 118–124, September–October, 1973.     

A Comprehensive Study of Improved Delayed Detached Eddy Simulation with Wall Functions

The mechanisms for nonlinear saturation of a bluff-body stabilised turbulent premixed flame are investigated using LES with the transported flame surface density (TFSD) approach to combustion modelling. The numerical simulation is based on a previous detailed experimental investigation. Results for both the unforced non-reacting and reacting flows are validated against experiment, demonstrating that the fundamental flow features and predicted flame structure are well captured. Key terms in the FSD transport equation are then used to describe the generation and destruction of flame surface area for the unforced reacting flow. In order to investigate the non-linear response of the unsteady heat release rate to acoustic forcing, four harmonically forced flames are considered having the same forcing frequency (160 Hz) but different amplitudes of 10 %, 25 %, 50 % and 64 % of the mean inlet velocity. The flame response is characterised using the Flame Describing Function (FDF). Accurate prediction of the FDF is obtained using the current approach. The computed forced flame structure matches well with the experiment, where effects of shear layer rollup and growth of the vortices on the flame can be clearly observed. Transition to nonlinearity is also observed in the computed FDF. The mechanisms leading to the saturation of the flame response in the higher amplitude case are characterised by inspecting the terms in the FSD transport equation at conditions when the integrated heat release is at its maximum and minimum, respectively. Pinch-off and flame rollup can be seen in snapshots taken at the phase angle of maximum integrated heat release. Conversely, intense vortex shedding and flame-sheet collapse around the shear-layer, as well as small-scale destruction of flame elements in the wake, can be seen in snapshots taken at the phase angle of minimum integrated heat release. The pivotal role of FSD destruction on nonlinear saturation of the flame response is confirmed through the analysis of phase-averaged terms in the FSD transport equation taken at different locations. The phase-averaged subgrid curvature term is found to concentrate in the cusps and downstream regions where flame annihilation is dominant.     

Temperature response of turbulent premixed flames to inlet velocity oscillations

Flame–turbulence interactions are at the heart of modern combustion research as they have a major influence on efficiency, stability of operation and pollutant emissions. The problem remains a formidable challenge, and predictive modelling and the implementation of active control measures both rely on further fundamental measurements. Model burners with simple geometry offer an opportunity for the isolation and detailed study of phenomena that take place in real-world combustors, in an environment conducive to the application of advanced laser diagnostic tools. Lean premixed combustion conditions are currently of greatest interest since these are able to provide low NO _x and improved increased fuel economy, which in turn leads to lower CO₂ emissions. This paper presents an experimental investigation of the response of a bluff-body-stabilised flame to periodic inlet fluctuations under lean premixed turbulent conditions. Inlet velocity fluctuations were imposed acoustically using loudspeakers. Spatially resolved heat release rate imaging measurements, using simultaneous planar laser-induced fluorescence (PLIF) of OH and CH₂O, have been performed to explore the periodic heat release rate response to various acoustic forcing amplitudes and frequencies. For the first time we use this method to evaluate flame transfer functions and we compare these results with chemiluminescence measurements. Qualitative thermometry based on two-line OH PLIF was also used to compare the periodic temperature distribution around the flame with the periodic fluctuation of local heat release rate during acoustic forcing cycles.     

Turbulent Flame Speeds of Premixed Supersonic Flame Kernels

It is unclear whether turbulent flame speed scalings established in low speed regimes are applicable to supersonic flames. To investigate this question, the canonical flame kernel is investigated in a scramjet-like channel having a one degree wall divergence. The growth, shape and internal kernel dynamics are investigated. Results are presented for three Mach numbers, four equivalence ratios, and three turbulence generators. Schlieren photography provides flame images for growth rate statistics and particle image velocimetry (PIV) provides turbulence statistics and investigation of internal kernel dynamics. Supersonic flame kernels are self-propagating and respond to the equivalence ratio in a fashion that is similar to low speed flames. However, supersonic flame kernels have features that are not present in subsonic flame kernels. Baroclinicity, resulting from pressure-density misalignment, creates a reacting vortex ring structure. Further, the mean kernel shape has a Mach number dependence and the vortex ring enhances the turbulent flame speed through entrainment of reactants and augmented flame surface growth. Hence, the previously established (low speed) flame speed scalings are inappropriate for supersonic flame kernels. Drawing motivation from vortex ring literature, the ring propagation velocity is used as the characteristic velocity and a new flame speed scaling is proposed.     

An experimental investigation of unsteady turbulent-wake/boundary-layer interaction

An experimental investigation of unsteady-wake/boundary-layer interaction, similar to that occurring in turbomachinery, has been conducted in a specially modified wind tunnel. Unsteadiness in a turbomachine is periodic in nature, due to the relative motion of rotor and stator blades, resulting in travelling-wave disturbances that affect the blade boundary layers. In the experimental rig, travelling-wave disturbances were generated by a moving airfoil apparatus installed upstream of a flat plate to provide a two-dimensional model of a turbomachine stage. The boundary layer on the flat plate was tripped near the leading edge to generate a turbulent flow prior to interaction with the wakes, and measurements of velocity throughout the boundary layer were taken with a hot-wire probe. The Reynolds number, based on distance along the plate, ranged from 0.144×10⁵ to 1.44×10⁵, and all data were reduced through a process of ensemble averaging. Due to the nonlinear interactions with the boundary layer, the travelling discrete frequency wakes were found to decrease the shape factor of the velocity profile and to increase the level of turbulent fluctuations. Unlike the phase advance found with stationary-wave external disturbances, velocity profiles subject to the travelling wake fluctuations exhibited increasingly negative phase shifts from the free-stream towards the wall.     

Experimental criteria for the determination of fractal parameters of premixed turbulent flames

The influence of spatial resolution, digitization noise, the number of records used for averaging, and the method of analysis on the determination of the fractal parameters of a high Damköhler number, methane/air, premixed, turbulent stagnation-point flame are investigated in this paper. The flow exit velocity was 5 m/s and the turbulent Reynolds number was 70 based on a integral scale of 3 mm and a turbulent intensity of 7%. The light source was a copper vapor laser which delivered 20 nsecs, 5 mJ pulses at 4 kHz and the tomographic cross-sections of the flame were recorded by a high speed movie camera. The spatial resolution of the images is 155 × 121 m/pixel with a field of view of 50 × 65 mm. The stepping caliper technique for obtaining the fractal parameters is found to give the clearest indication of the cutoffs and the effects of noise. It is necessary to ensemble average the results from more than 25 statistically independent images to reduce sufficiently the scatter in the fractal parameters. The effects of reduced spatial resolution on fractal plots are estimated by artificial degradation of the resolution of the digitized flame boundaries. The effect of pixel resolution, an apparent increase in flame length below the inner scale rolloff, appears in the fractal plots when the measurent scale is less than approximately twice the pixel resolution. Although a clearer determination of fractal parameters is obtained by local averaging of the flame boundaries which removes digitization noise, at low spatial resolution this technique can reduce the fractal dimension. The degree of fractal isotropy of the flame surface can have a significant effect on the estimation of the flame surface area and hence burning rate from two-dimensional images. To estimate this isotropy a determination of the outer cutoff is required and three-dimensional measurements are probably also necessary.     

On the Flame-generated Vorticity Dynamics of Bluff-body-stabilized Premixed Flames

This investigation considers the dynamics of flame-generated vorticity for a premixed, submerged bluff-body stabilized flame. Experimentation characterizes the far-field region in particular with a level of detail not previously afforded to this type of flow. Simultaneous particle imaging velocimetry (PIV), Mie scattering and CH ^? chemiluminescence are used to obtain velocity fields and flame location. Mean static pressure measurements at the combustion chamber wall capture the pressure field. Analysis of the flame fronts in relation to the mean velocity and vorticity fields provides useful insight into the interaction of the flame and the flow. The unique nature of the velocity and vorticity fields and their effect on downstream flame structures are explained by the baroclinic torque generation of vorticity. The coupling that exists among pressure, heat release, and baroclinic generation is acknowledged and will influence strategies for control of the baroclinic mechanism.     

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.     

Modeling droplet dispersion and interphase turbulent kinetic energy transfer using a new dual-timescale Langevin model

Dispersion of spray droplets and the modulation of turbulence in the ambient gas by the dispersing droplets are two coupled phenomena that are closely linked to the evolution of global spray characteristics, such as the spreading rate of the spray and the spray cone angle. Direct numerical simulations (DNS) of turbulent gas flows laden with sub-Kolmogorov size particles, in the absence of gravity, report that dispersion statistics and turbulent kinetic energy (TKE) evolve on different timescales. Furthermore, each timescale behaves differently with Stokes number, a non-dimensional flow parameter (defined in this context as the ratio of the particle response time to the Kolmogorov timescale of turbulence) that characterizes how quickly a particle responds to turbulent fluctuations in the carrier or gas phase. A new dual-timescale Langevin model (DLM) composed of two coupled Langevin equations for the fluctuating velocities, one for each phase, is proposed. This model possesses a unique feature that the implied TKE and velocity autocorrelation in each phase evolve on different timescales. Consequently, this model has the capability of simultaneously predicting the disparate Stokes number trends in the evolution of dispersion statistics, such as velocity autocorrelations, and TKE in each phase. Predictions of dispersion statistics and TKE from the new model show good agreement with published DNS of non-evaporating and evaporating droplet-laden turbulent flow.     

Spatial Coherence of the Heat Release Fluctuations in Turbulent Jet and Swirl Flames

The noise generation of turbulent flames is governed by temporal changes of the total flame volume due to local heat release fluctuations. On the basis of the wave equation an expression for the relationship between the acoustic power and the heat release fluctuations is derived and a correlation function is obtained which reveals that the sound pressure level of flames is governed by the spatial coherence. Noise models rely on precise coherence information in terms of characteristic length scales, which are the measure of the acoustic efficiency of the flame. Since the published length scale information is scarce and inconsistent, length scales were measured for a number of laboratory flames using two measurement techniques developed for this purpose: A planar LIF-system with a repetition rate of 1 kHz acquires the instantaneous flame front position and heat release, whereas two chemiluminescence probes with an optics confining the measurement volume to a line of sight provide further spatial correlation data. For all flames investigated the length scales are smaller than the height of the burner exit annulus and they are of the order of the local flame brush thickness. Using the measured length scales, the coherent volume and the efficiency of the noise generation are calculated, which are three orders of magnitude higher than measured. However, the proper order of magnitude is obtained, if only the measured fluctuating part of the thermal power is used in the model and if the periodic formation of local zones with heat release overshoot and deficit are properly incorporated.     

Subgrid-Scale Stress Modelling Based on the Square of the Velocity Gradient Tensor

A new subgrid scale model is proposed for Large Eddy Simulations in complex geometries. This model which is based on the square of the velocity gradient tensor accounts for the effects of both the strain and the rotation rate of the smallest resolved turbulent fluctuations. Moreover it recovers the proper y ³ near-wall scaling for the eddy viscosity without requiring dynamic procedure. It is also shown from a periodic turbulent pipe flow computation that the model can handle transition.     

Research on the effect of cylinder particles on the turbulent properties in particulate flows

ＩｎｔｒｏｄｕｃｔｉｏｎＴｈｅｔｗｏ＿ｐｈａｓｅｆｌｏｗｗｉｔｈｃｙｌｉｎｄｅｒｐａｒｔｉｃｌｅｓｉｓａｄｉｆｆｉｃｕｌｔｍａｔｔｅｒｉｎｔｈｅｏｒｅｔｉｃａｌｒｅｓｅａｒｃｈｏｆｍｕｌｔｉｐｈａｓｅｆｌｏｗａｎｄｎｏｎ＿Ｎｅｗｔｏｎｉａｎｆｌｏｗ ,ｓｏｉｔｉｓａｃａｄｅｍｉｃａｌｌｙｖａｌｕａｂｌｅｔｏｓｔｕｄｙｉｔ.Ｍｅａｎｗｈｉｌｅ ,ｓｕｃｈｆｌｏｗｓｈａｖｅａｂｒｏａｄｂａｃｋｇｒｏｕｎｄｉｎｉｎｄｕｓｔｒｙ .Ｉｎｔｈｅｍａｔｅｒｉａｌｓｃｉｅｎｃｅ,ｔｈｅｍｏｌｄｉｎｇａｎｄｐｒｏｃｅｓ…     

Statistical Analysis of Turbulent Flame-Droplet Interaction: A Direct Numerical Simulation Study

Turbulent combustion of mono-disperse droplet-mist has been analysed based on three-dimensional Direct Numerical Simulations (DNS) in canonical configuration under decaying turbulence for a range of different values of droplet equivalence ratio (?_d), droplet diameter (a_d) and root-mean-square value of turbulent velocity (u^′). The fuel is supplied in liquid phase and the evaporation of droplets gives rise to gaseous fuel for the flame propagation into the droplet-mist. It has been found that initial droplet diameter, turbulence intensity and droplet equivalence ratio can have significant influences on the volume-integrated burning rate, flame surface area and burning rate per unit area. The droplets are found to evaporate predominantly in the preheat zone, but some droplets penetrate the flame front, reaching the burned gas side where they evaporate and some of the resulting fuel vapour diffuses back towards the flame front. The combustion process in gaseous phase takes place predominantly in fuel-lean mode even for ?_d > 1. The probability of finding fuel-lean mixture increases with increasing initial droplet diameter because of slower evaporation of larger droplets and this predominantly fuel-lean mode of combustion exhibits the attributes of low Damköhler number combustion and gives rise to thickening of flame with increasing droplet diameter. The chemical reaction is found to take place under both premixed and non-premixed modes of combustion and the relative contribution of non-premixed combustion to overall heat release increases with increasing droplet size. The statistical behaviours of the flame propagation and mode of combustion have been analysed in detail and detailed physical explanations have been provided for the observed behaviour.     

Temperature Imaging in an Unsteady Propane-Air Counterflow Diffusion Flame Subjected to Low Frequency Oscillations

Current emphasis in non-premixed turbulent combustion research is focused on the effects of hydrodynamic unsteadiness and transient effects in these flames. To address these effects, measurements of the temperature field in unsteady propane-air flames were made using planar laser-induced fluorescence of the hydroxyl radical to ascertain the effect of fluctuating hydrodynamics on flame temperature. Planar temperature measurements were made at four temporal locations within the 25 Hz velocity fluctuation as a function of initial steady strain rate and forcing amplitude. Results show that temperature in the propane flame is rather insensitive to initial strain rate for these weakly strained flames due to the balance between decreased heat release rate and reduced radiative losses resulting from diminished soot production as the strain rate is increased. The temperature is significantly influenced by the imposed velocity oscillation, however, which causes large fluctuations in the instantaneous strain rate. A decoupling of peak flame temperature and maximum PAH and excited CH concentration indicates significant transient effects resulting from the unsteady flow field even at these low oscillation frequencies. This revised version was published online in July 2006 with corrections to the Cover Date.     

Flame Structure and Propagation in Turbulent Flame-Droplet Interaction: A Direct Numerical Simulation Analysis

Three-dimensional Direct Numerical Simulations (DNS) in canonical configuration have been employed to study the combustion of mono-disperse droplet-mist under turbulent flow conditions. A parametric study has been performed for a range of values of droplet equivalence ratio ?_d, droplet diameter a_d and root-mean-square value of turbulent velocity u^′. The fuel is supplied entirely in liquid phase such that the evaporation of the droplets gives rise to gaseous fuel which then facilitates flame propagation into the droplet-mist. The combustion process in gaseous phase takes place predominantly in fuel-lean mode even for ?_d>1. The probability of finding fuel-lean mixture increases with increasing initial droplet diameter because of slower evaporation of larger droplets. The chemical reaction is found to take place under both premixed and non-premixed modes of combustion: the premixed mode ocurring mainly under fuel-lean conditions and the non-premixed mode under stoichiometric or fuel-rich conditions. The prevalence of premixed combustion was seen to decrease with increasing droplet size. Furthermore, droplet-fuelled turbulent flames have been found to be thicker than the corresponding turbulent stoichiometric premixed flames and this thickening increases with increasing droplet diameter. The flame thickening in droplet cases has been explained in terms of normal strain rate induced by fluid motion and due to flame normal propagation arising from different components of displacement speed. The statistical behaviours of the effective normal strain rate and flame stretching have been analysed in detail and detailed physical explanations have been provided for the observed behaviour. It has been found that the droplet cases show higher probability of finding positive effective normal strain rate (i.e. combined contribution of fluid motion and flame propagation), and negative values of stretch rate than in the stoichiometric premixed flame under similar flow conditions, which are responsible for higher flame thickness and smaller flame area generation in droplet cases.     

Simultaneous velocity and scalar measurements in premixed recirculating flames

 The use of a laser-Doppler velocimeter has been extended to the analysis of turbulent heat transfer in a strongly sheared disc-stabilised propane-air flame through its combination with either laser Rayleigh scattering or digitally-compensated fine-wire thermocouples. The laser velocimeter was based on a conventional forward scattering system from the green light of a 5W Argon-Ion laser, while the Rayleigh signals used the blue line of the same laser. The procedure for the numeric compensation of the thermocouple signals included analysis of the effect of velocity and temperature on the time constant of the thermocouple and was optimised to allow combined velocity–temperature samples acquired by a purpose-built digital interference with a frequency up to 2000 Hz, without deterioration of the thermocouple by particle accretion. The maximum effective data rate for the combined Rayleigh/LDV system is shown to be around 130 Hz, which corresponds to a data rate of valid Doppler signals around 400 Hz and statistics based on more than 15 000 measurements is made possible. The results obtained with the two systems agree qualitatively, although the use of thermocouples attenuates the measured velocity-temperature correlations. The results are used to assess the extent to which turbulent mixing in flames is altered by the accompanying heat release and quantify the processes of non-gradient diffusion in a strongly recirculating premixed flame. Received: 15 November 1996/Accepted: 2 September 1997     

Numerical Study of Ignition Process in Turbulent Shear-less Methane-air Mixing Layer

In this work, ignition process in a turbulent shear-less methane-air mixing layer is numerically investigated. A compressible large eddy simulation method with Smagorinsky sub-grid scale model is used to solve the flow field. Also, a thickened flame combustion model and DRM-19 reduced mechanism are used to compute species distribution and the heat release. Non-reacting mean and RMS axial velocity profiles and mean mixture fraction are validated against experimental data. Instantaneous mixture fraction contours show that the large bursts penetrate from the fuel stream into that of the oxidizer and vice versa and a random behaviour in the cross-stream direction. Flame kernel initiation, growth and propagation are analysed and compared with the experimental data. The ignition results show that the flame is not stable and blow-off occurs, but a more detailed investigation shows that local and short time flame stabilization exist during blow-off. During these local stabilization, heat release increased at the upstream edge of the flame. Most_upstream flame edge scalar analysis shows that the methane mass fraction has a dominant role in the local flame stabilization. OH, HO₂, CH₂O and heat release contours demonstration reveal that HO₂ and CH₂O mass fraction as well as the heat release reach a maximum on the border of the flame, but the maximum OH concentration is located in the middle of flame kernel.