Blowoff dynamics of bluff body stabilized turbulent premixed flames

This article concerns the flame dynamics of a bluff body stabilized turbulent premixed flame as it approaches lean blowoff. Time resolved chemiluminescence imaging along with simultaneous particle image velocimetry and OH planar laser-induced fluorescence were utilized in an axisymmetric bluff body stabilized, propane-air flame to determine the sequence of events leading to blowoff and provide a quantitative analysis of the experimental results. It was found that as lean blowoff is approached by reduction of equivalence ratio, flame speed decreases and the flame shape progressively changes from a conical to a columnar shape. For a stably burning conical flame away from blowoff, the flame front envelopes the shear layer vortices. Near blowoff, the columnar flame front and shear layer vortices overlap to induce high local stretch rates that exceed the extinction stretch rates instantaneously and in the mean, resulting in local flame extinction along the shear layers. Following shear layer extinction, fresh reactants can pass through the shear layers to react within the recirculation zone with all other parts of the flame extinguished. This flame kernel within the recirculation zone may survive for a few milliseconds and can reignite the shear layers such that the entire flame is reestablished for a short period. This extinction and reignition event can happen several times before final blowoff which occurs when the flame kernel fails to reignite the shear layers and ultimately leads to total flame extinguishment.     

Blowoff mechanism of two dimensional bluff-body stabilized turbulent premixed flames in a prototypical combustor

Near blowoff dynamics and characteristics of turbulent premixed flames stabilized by a triangular flame holder in the midspan of a rectangular duct were studied using high speed imaging and simultaneous particle imaging velocimetry and OH planar laser-induced fluorescence. Near blowoff dynamics manifested by the onset of asymmetric vortex shedding and local extinction were observed. It has been proposed that a partial or total extinction of the flame sheet along the shear layers is the major factor that determines the final blowoff event. Observation of flame kernels within the recirculation zone, which under stable conditions contain only combustion products, is further evidence of the shear layer extinction. For a stably burning planar V flame away from blowoff, the flame front envelopes the Kelvin Helmholtz vortices. Near blowoff, the two flame fronts become more aligned with the flow direction due to reduction in flame speed and interact with the vortices emanating from the shear layer. This overlap induces high local stretch rates that exceed the extinction stretch rates, resulting in local flame extinction along the shear layers. Following extinction, fresh reactants are entrained into and react within the recirculation zone, with all other parts of the flame extinguished. This flame kernel within the recirculation zone may survive for time scales of about one hundred milliseconds, potentially reigniting the shear layers such that the entire flame is re-established for a short period. This shear layer extinction and re-ignition event can happen several times before final blowoff, which occurs when the flame kernel fails to reignite the shear layers before being extinguished itself, thus leading to global flame extinction.     

Lean blowoff of bluff body stabilized flames: Scaling and dynamics

This paper overviews the dynamics of bluff body stabilized flames and describes the phenomenology of the blowoff process. The first section of the paper provides an overview of the fluid mechanics of the non-reacting and reacting bluff body wake flow. It highlights the key features of the flow (the boundary layer, separated shear layer, and wake), the flow instabilities that influence each of these features, and the influences of the flame on these instabilities. A key point from these studies is the large differences between the non-reacting wake (dominated by an absolutely unstable, sinuous instability associated with vortex shedding from the bluff body) and the reacting wake of high dilatation ratio flames. The latter are dominated by the lower intensity, convective instability of the shear layer. Very low dilatation ratio flames begin to approach the behavior of the non-reacting wake, as might be expected.     

Blowoff mechanism of harmonically forced bluff body stabilized turbulent premixed flames

Flame blowoff under harmonic forcing of upstream mixture velocity is analyzed for bluff-body stabilized conical premixed flames in a laboratory burner. It is found that the forced vortex shedding phenomenon within the recirculation zone, accompanies flame blowoff when the convective wavelength of the imposed oscillation is larger than the recirculation zone length in the streamwise direction. The experimental results obtained from combined particle image velocimetry and planar laser induced fluorescence of OH radical provide evidence for this behavior. The differences between forced and unforced flame blowoff are also discussed.     

Effect of high hydrogen enrichment on the outer-shear-layer flame of confined lean premixed CH4/H2/air swirl flames

In this study, we investigated the H₂-induced transition of confined swirl flames from the “V” to “M” shape. H₂-enriched lean premixed CH₄/H₂/air flames with H₂ fractions up to 80% were conducted. The flame structure was obtained with Planar Laser-Induced Fluorescence (PLIF) of the OH radical. Flow fields were measured with Particle Image Velocimetry (PIV). It was observed that the flame tip in the outer shear layer gradually propagated upstream and finally anchored to the injector with the hydrogen fractions increase, yielding the transition from the “V” to “M” flame. We examined the flame structures and the flame flow dynamics during the transition. The shape transition was directly related to the evolution of the corner flame along the outer shear layer. With H₂ addition, the outer recirculation zone first appeared downstream where the corner flame started to propagate upstream; then, the recirculation zone expanded upward to form a stable “M” flame gradually. The flow straining was observed to influence the stabilization of the outer shear layer flame significantly. This study can be useful for the understanding of recirculation-stabilized swirling flames with strong confinement. The flame structure and the flow characteristics of flames with a high H₂ content are also valuable for model validation.     

Flame speed and tangential strain measurements in widely stratified partially premixed flames interacting with grid turbulence

Methane-air partially premixed flames subjected to grid-generated turbulence are stabilized in a two-slot burner with initial fuel concentration differences leading to stratification across the stoichiometric concentration. The fuel concentration gradient at the location corresponding to the flame base is measured using planar laser induced fluorescence (PLIF) of acetone in the non-reacting mixing field. Simultaneous PLIF of the OH radical and particle image velocimetry (PIV) measurements are performed to deduce the flow velocity and the flame front. These flames exhibit a convex premixed flame front and a trailing diffusion flame, with flow divergence upstream of the flame, as indicated by the instantaneous OH–PLIF, Mie scattering images, and PIV data. The mean streamwise velocity profile attains a global minimum just upstream of the flame front due to expansion of a gases caused by heat release. The flame speed measured just upstream of the flame leading edge is normalized with respect to the turbulent stoichiometric flame speed that takes into account variations in turbulent intensity and integral length scale. The turbulent edge flame speed exceeds the corresponding stoichiometric premixed flame speed and reaches a peak at a certain concentration gradient. The mean tangential strain at the flame leading edge locally peaks at the concentration gradient corresponding to the peak flame speed. The strain varies non-monotonically with the flame curvature unlike in a non-stratified curved premixed flame. The mechanism of peak flame speed is explained as the competition between availability of hot excess reactants from the premixed flame branches to the flame stretch induced due to flame curvature. The results suggest that the stabilization of lifted turbulent partially premixed flames occurs through an edge flame even at a relatively gentle concentration gradient. The strain is also evaluated along the flame front; it peaks at the flame leading edge and decreases gradually on either side of the leading edge. The present results also show qualitatively similar trends as those of laminar triple flames.     

Investigation of local flame structures and statistics in partially premixed turbulent jet flames using simultaneous single-shot CH and OH planar laser-induced fluorescence imaging

We report on the application of simultaneous single-shot imaging of CH and OH radicals using planar laser-induced fluorescence (PLIF) to investigate partially premixed turbulent jet flames. Various flames have been stabilized on a coaxial jet flame burner consisting of an outer and an inner tube of diameter 22 and 2.2 mm, respectively. From the outer tube a rich methane/air mixture was supplied at a relatively low flow velocity, while a jet of pure air was introduced from the inner one, resulting in a turbulent jet flame on top of a laminar pilot flame. The turbulence intensity was controlled by varying the inner jet flow speed from 0 up to 120 m/s, corresponding to a maximal Reynolds number of the inner jet airflow of 13,200. The CH/OH PLIF imaging clearly revealed the local structure of the studied flames. In the proximity of the burner, a two-layer reaction zone structure was identified where an inner zone characterized by strong CH signals has a typical structure of rich premixed flames. An outer reaction zone characterized by strong OH signals has a typical structure of a diffusion flame that oxidizes the intermediate fuels formed in the inner rich premixed flame. In the moderate-turbulence flow, the CH layers were very thin closed surfaces in the entire flame, whereas the OH layers were much thicker. In the high-intensity-turbulence flame, the CH layer remained thin until it vanished in the upper part of the flame, showing local extinction and reignition behavior of the flame. The single-shot PLIF images have been utilized to determine the flame surface density (FSD). In low and moderate turbulence intensity cases the FSDs determined from CH and OH agreed with each other, while in the highly turbulent case a locally broken CH layer was observed, leading to a significant difference in the FSD results determined via the OH and CH radicals. Furthermore, the means and the standard deviations of CH and OH radicals were obtained to provide statistical information about the flames that may be used for validation of numerical calculations.     

Blowoff characteristics of bluff-body stabilized conical premixed flames with upstream spatial mixture gradients and velocity oscillations

This experimental study concerns determination of blowoff equivalence ratios for lean premixed conical flames for different mixture approach velocities ranging from 5 to 16 m/s in the presence of spatial mixture gradients and upstream velocity modulation. Conical flames were anchored on a disk-shaped bluff body that was attached to a central rod in the burner nozzle. A combustible propane-air mixture flowed through a converging axisymmetric nozzle with a concentric insert, allowing radial mixture variation by tailoring the composition in the inner and outer parts of the nozzle. The radial mixture profiles were characterized near the location of the flame holder by laser Rayleigh light scattering. Additionally, a loudspeaker at the nozzle base allowed introduction of periodic velocity oscillations with an amplitude of 9% of the mean flow velocity up to a frequency of 350 Hz. The flame blowoff equivalence ratio was experimentally determined by continuously lowering the fuel flow rates and determining the flame detachment point from the flame holder. Flame detachment was detected by a rapid reduction of CH* emission from the flame base imaged by a photomultiplier detector. It was found that the flame blowoff is preceded by progressive narrowing of the flame cone for the case of higher inner jet equivalence ratios. In this case, the fuel-lean outer flow cannot sustain combustion, and clearly this is not a good way of operating a combustor. Nevertheless, the overall blowoff equivalence ratio is reduced by inner stream fuel enrichment. A possible explanation for this behavior is given based on the radial extent of the variable-equivalence-ratio mixture burning near the flame stabilization region. Fuel enrichment in the outer flow was found to have no effect on blowoff as compared to the case of uniform mixture. The results were similar for the whole range of mean flow velocities and upstream excitation frequencies.     

Numerical study of premixed twin edge flames in a counterflow field

Characteristics of premixed edge flames established in a counterflow field with a stretch-rate gradient were numerically investigated by solving three-dimensional governing equations with detailed chemistry in the general curvilinear coordinates system. Local mole fractions of radicals, such as OH or CH, at the flame edge of a CH₄/air mixture were found to be larger than those in other parts of the flame. On the other hand, local mole fractions of radicals in the flame edge of a C₃H₈/air mixture were smaller than those in other parts. These numerical results agreed well with the experimental results of the present authors. Moreover, it was elucidated that two flame edges of twin counterflow flames did not merge at the edge even in the case of the CH₄/air mixture. The ratio of the local stretch rate at the flame edge to the extinction stretch rate for planar twin flames with the same equivalence ratio was 0.6 for the CH₄/air mixture and 0.7 for the C₃H₈/air mixture. These numerical results also agreed with results of the past experiments. Moreover, as for relatively low stretch-rate gradients, the stretch-rate gradient had no strong influence on the characteristics of the edge flames.     

Spatially resolved heat release rate measurements in turbulent premixed flames

Heat release rate is a fundamental property of great importance for the theoretical and experimental elucidation of unsteady flame behaviors such as combustion noise, combustion instabilities, and pulsed combustion. Investigations of such thermoacoustic interactions require a reliable indicator of heat release rate capable of resolving spatial structures in turbulent flames. Traditionally, heat release rate has been estimated via OH or CH radical chemiluminescence; however, chemiluminescence suffers from being a line-of-sight technique with limited capability for resolving small-scale structures. In this paper, we report spatially resolved two-dimensional measurements of a quantity closely related to heat release rate. The diagnostic technique uses simultaneous OH and CH₂O planar laser-induced fluorescence (PLIF), and the pixel-by-pixel product of the OH and CH₂O PLIF signals has previously been shown to correlate well with local heat release rates. Results from this diagnostic technique, which we refer to as heat release rate imaging (HR imaging), are compared with traditional OH chemiluminescence measurements in several flames. Studies were performed in lean premixed ethylene flames stabilized between opposed jets and with a bluff body. Correlations between bulk strain rates and local heat release rates were obtained and the effects of curvature on heat release rate were investigated. The results show that the heat release rate tends to increase with increasing negative curvature for the flames investigated for which Lewis numbers are greater than unity. This correlation becomes more pronounced as the flame gets closer to global extinction.     

Hysteresis and transition in swirling nonpremixed flames

Strongly swirling nonpremixed flames are known to exhibit a hysteresis when transiting from an attached long, sooty, yellow flame to a short lifted blue flame, and vice versa. The upward transition (by increasing the air and fuel flow rates) corresponds to a vortex breakdown, i.e. an abrupt change from an attached swirling flame (unidirectional or with a weak bluff-body recirculation), to a lifted flame with a strong toroidal vortex occupying the bulk of the flame. Despite dramatic differences in their structures, mixing intensities and combustion performance, both flame types can be realised at identical flow rates, equivalence ratio and swirl intensity. We report here on comprehensive investigations of the two flame regimes at the same conditions in a well-controlled experiment in which the swirl was generated by the rotating outer pipe of the annular burner air passage. Fluid velocity measured with PIV (particle image velocimetry), the qualitative detection of reaction zones from OH PLIF (planar laser-induced fluorescence) and the temperature measured by CARS (coherent anti-Stokes Raman spectroscopy) revealed major differences in vortical structures, turbulence, mixing and reaction intensities in the two flames. We discuss the transition mechanism and arguments for the improved mixing, compact size and a broader stability range of the blue flame in comparison to the long yellow flame.     

Experimental study of vortex-flame interaction in a gas turbine model combustor

The interaction of a helical precessing vortex core (PVC) with turbulent swirl flames in a gas turbine model combustor is studied experimentally. The combustor is operated with air and methane at atmospheric pressure and thermal powers from 10 to 35 kW. The flow field is measured using particle image velocimetry (PIV), and the dominant unsteady vortex structures are determined using proper orthogonal decomposition. For all operating conditions, a PVC is detected in the shear layer of the inner recirculation zone (IRZ). In addition, a co-rotating helical vortex in the outer shear layer (OSL) and a central vortex originating in the exhaust tube are found. OH chemiluminescence (CL) images show that the flames are mainly stabilized in the inner shear layer (ISL), where also the PVC is located. Phase-averaged images of OH-CL show that for all conditions, a major part of heat release takes place in a helical zone that is coupled to the PVC. The mechanisms of the interaction between PVC and flame are then studied for the case P = 10 kW using simultaneous PIV and OH-PLIF measurements with a repetition rate of 5 kHz. The measurements show that the PVC causes a regular sequence of flame roll-up, mixing of burned and unburned gas, and subsequent ignition of the mixture in the ISL. These effects are directly linked to the periodic vortex motions. A phase-averaged analysis of the flow field further shows that the PVC induces an unsteady lower stagnation point that is not present in the average flow field. The motion of the stagnation point is linked to the periodic precession of the PVC. Near this point burned and unburned gas collide frontally and a significant amount of heat release takes place. The flame dynamics near this point is also coupled to the PVC. In this way, a part of the reaction zone is periodically drawn from the stagnation point into the ISL, and thus serves as an ignition source for the reactions in this layer. In total, the effects in the ISL and at the stagnation point showed that the PVC plays an essential role in the stabilization mechanism of the turbulent swirl flames. In contrast to the PVC, the vortices in the OSL and near the exhaust tube have no direct effect on the flame since they are located outside the flame zone.     

Measurements in turbulent premixed bluff body flames close to blow-off

The structure of unconfined lean premixed methane–air flames stabilized on an axisymmetric bluff body has been examined for conditions increasingly closer to blow-off and during the blow-off event. Fast imaging (5 kHz) of OH¹ chemiluminescence and OH-PLIF and PIV (at 1 kHz) were used to obtain instantaneous and time-averaged images, temporal sequences, spectra of OH, 2-D estimates of flame surface density, curvature, turbulence statistics, and measurements of the duration of the blow-off transient. Blow-off was approached by slowly reducing the fuel flow rate, and the flame shape was seen to change from a cylindrical shape at stable burning conditions, with the flame brush closing across the flow at conditions close to the blow-off condition. This was followed by entrainment of fresh reactants from the downstream end of the recirculation zone (RZ), and fragmentation of the downstream flame parts. Just before the blow-off event, reaction fronts were observed inside the RZ, with progressive fragmentation occurring, leading to a shorter flame brush. Complete extinction occurred once the flame at the attachment point had been destroyed, and stabilization at the shear layers was no longer possible. Measurements showed a gradual reduction in FSD during the approach to blow-off and during the extinction event itself, and higher values of flame front curvature at conditions approaching extinction. The local Karlovitz number was estimated based on the local turbulence velocity and lengthscale characteristics and it reached a maximum value of about 10 at the location where the flame bends towards the axis. Quantification of the duration of the blow-off event showed that it was an order of magnitude longer than the characteristic timescale of the burner d/U_b. The measurements reported here are useful for model validation and for exploring the changes in turbulent premixed flame structure as extinction is approached.     

Instabilities and soot formation in high-pressure, rich, iso-octane-air explosion flames: 1. Dynamical structure

Simultaneous OH planar laser-induced fluorescence (PLIF) and Rayleigh scattering measurements have been performed on 2-bar rich iso-octane-air explosion flames obtained in the optically accessible Leeds combustion bomb. Separate shadowgraph high-speed video images have been obtained from explosion flames under similar mixture conditions. Shadowgraph images, quantitative Rayleigh images, and normalized OH concentration images have been presented for a selection of these explosion flames. Normalized experimental equilibrium OH concentrations behind the flame fronts have been compared with normalized computed equilibrium OH concentrations as a function of equivalence ratio. The ratio of superequilibrium OH concentration in the flame front to equilibrium OH concentration behind the flame front reveals the response of the flame to the thermal-diffusive instability and the resistance of the flame front to rich quenching. Burned gas temperatures have been determined from the Rayleigh scattering images in the range 1.4???1.9 and are found to be in good agreement with the corresponding predicted adiabatic flame temperatures. Soot formation was observed to occur behind deep cusps associated with large-wavelength cracks occurring in the flame front for equivalence ratio ??1.8 (C/O?0.576). The reaction time-scale for iso-octane pyrolysis to soot formation has been estimated to be approximately 7.5-10 ms.     

Asymptotic analysis of transport properties and burning velocities for premixed hydrocarbon flames

IntroductionThe fundamental meChedsm Of a p~xed flamewith the flow near the front stagnation point of a platewall has receiVed considerable attention in the field ofcombushon, which helps us to realize the behavior offlame Propagation. The CO~thew teChnique,inboduced by Law and coworkers["n, has produced theIndnar flame speed data that are ~ntiy usedextensively fof validation Of chemical ldnetics and themodeling of turbulent combustion. The laminar flamespeed is an important Property of a …     

Experimental study of premixed single edge-flame in a counterflow field

A single premixed edge-flame established in a counterflow field of a combustible mixture and an inert nitrogen was experimentally investigated by using twin rectangular burners which were misaligned by a few degrees. The stretch-rate gradient was quantitatively defined as a function of the angle between the two burners and the distance from the edge of the burner. The flame weakly curved at the edge toward the stagnation plane and the shape of the flame edge did not depend on the composition of the mixture. The response of edge-flames to changes in flow conditions, such as the equivalence ratio of the mixture or the injection velocity at the burner exit, was basically determined by the Damkohler number of the mixture. The ratio of the local stretch rate at the flame edge and the extinction stretch rate for a single planar flame with the same composition was slightly smaller than unity, though the extinction stretch rate for the planar flame was about half of that for the twin planar flame due to large heat conduction to the opposing inert gas. This ratio of stretch rates for the single edge-flame was larger than that for twin edge-flames in a previous work, and this result agreed with the previous theoretical analysis by Buckmaster et al. and the previous experimental results of Ronney et al. Moreover, the effect of the stretch-rate gradient on the characteristics of the edge-flame did not appear for the single edge-flame.     

On stretch-affected pulsating instability in rich hydrogen/air flames: asymptotic analysis and computation

Effects of stretch on the pulsating instability of rich hydrogen/air premixed flames, which are characterized by large Lewis numbers, were analytically and computationally investigated via the negatively stretched inwardly propagating spherical flame (IPF) and the positively stretched counterflow flame (CFF). Analytical results yield explicit criteria for the onset of pulsation, and show that positive stretch promotes pulsation while negative stretch retards it. Computational results for the IPF show that the flame initially propagates at the laminar flame speed when the flame radius is large. Oscillation subsequently develops, and is then amplified, damped, and eventually suppressed when the flame is still sufficiently far away from the center. Thus negative stretch tends to suppress the occurrence of pulsating instability, and thereby extend the flammable range of rich hydrogen/air flames beyond that of their unstretched, planar counterpart. Furthermore, the pulsating propagation is quasi-steady in that it is independent of the initial state. Computational results for the CFF show that oscillation is initiated at an equivalence ratio much smaller than the one-dimensional rich threshold, and that the critical strain rate leading to pulsation is smaller than the corresponding static extinction limit. Furthermore, similar to the one-dimensional, unstretched cases, with progressive increase in the strain rate for a sufficiently rich mixture, the pulsation mode changes from that of monochromatic, to period doubling, and to permanent extinction. The pulsating flames are also quasi-steady in nature in that the period of oscillation is larger than the characteristic flame time. As such, the unsteady flame cannot recover once the instantaneous flame temperature is reduced below the corresponding steady-state extinction temperature. Because pulsating extinction occurs at a smaller strain rate than the steady extinction limit, the flame extinguishes in the pulsating instead of the steadily propagating mode, and the flammable range is accordingly narrowed. Finally, the numerically calculated critical states at which the IPF and the CFF respectively lose stability are predicted well by using the analytically derived transition criteria and global flame parameters.     

Laminar propagation of lean premixed flames ignited in stratified mixture

Combustion in stratified mixtures is envisaged in practical energy systems such as direct-injection spark-ignited (DISI) car engines, gas turbines, for reducing CO₂ and pollutant emissions while protecting their efficiency. The mixture gradients change the fundamental properties of the flame, especially by a difference in temperature and composition between the burnt gases and those of a flame consuming a homogeneous mixture. This paper presents an investigation of the properties of the flame propagating in a lean homogeneous mixture after ignition in a richer mixture according to the magnitude of the stratification. Three magnitudes of stratification are investigated. The local flame burning velocity is determined by an original PIV algorithm developed previously. The local equivalence ratio in the fresh gases is measured from anisole PLIF. From the simultaneous PIV–PLIF measurements, the flame burning velocities conditioned on the local stretch rate and equivalence ratio in fresh gases are measured. The flame propagating through the homogeneous lean mixture has properties depending on the ignition conditions in the stratified layer. The flame propagating in the lean mixture is back-supported longer for ignition under the richer condition. The change of stretch sensitivity and burning velocity of the flame in the lean mixture is measured over time for the three magnitudes of mixture stratification investigated. The ignition in richer mixtures compensates for the nonequidiffusion effect of lean propane flame and sustains its robustness to stretch. The flame propagation in the lean homogeneous mixture is enhanced by ignition in a richer stratified layer, as much by their robustness to stretch as by an increase in the flame speed or the burning velocity. The decay time of this influence of the stratification, called memory effect, is determined.     

Tubular premixed and diffusion flames: Effect of stretch and curvature

Tubular flames are ideal for the study of stretch and curvature effects on flame structure, extinction, and instabilities. Tubular flames have uniform stretch and curvature and each parameter can be varied independently. Curvature strengthens or weakens preferential diffusion effects on the tubular flame and the strengthening or weakening is proportional to the ratio of the flame thickness to the flame radius. Premixed flames can be studied in the standard tubular burner where a single premixed gas stream flows radially inward to the cylindrical flame surface and products exit as opposed jets. Premixed, diffusion and partially premixed flames can be studied in the opposed tubular flame where opposed radial flows meet at a cylindrical stagnation surface and products exit as opposed jets. The tubular flame flow configurations can be mathematically reduced to a two-point boundary value solution along the single radial coordinate. Non-intrusive measurements of temperature and major species concentrations have been made with laser-induced Raman scattering in an optically accessible tubular burner for both premixed and diffusion flames. The laser measurements of the flame structure are in good agreement with numerical simulations of the tubular flame. Due to the strong enhancement of preferential diffusion effects in tubular flames, the theory-data comparison can be very sensitive to the molecular transport model and the chemical kinetic mechanism. The strengthening or weakening of the tubular flame with curvature can increase or decrease the extinction strain rate of tubular flames. For lean H₂-air mixtures, the tubular flame can have an extinction strain rate many times higher than the corresponding opposed jet flame. More complex cellular tubular flames with highly curved flame cells surrounded by local extinction can be formed under both premixed and non-premixed conditions. In the hydrogen fueled premixed tubular flames, thermal-diffusive flame instabilities result in the formation of a uniform symmetric petal flames far from extinction. In opposed-flow tubular diffusion flames, thermal-diffusive flame instabilities result in cellular flames very close to extinction. Both of these flames are candidates for further study of flame curvature and extinction.