Flame Flicker Research Articles

Dynamic application of digital image and colour processing in characterizing flame radiation features

In this work, the experimental investigation of the dynamic flame properties of flame flickering and equivalence ratio sensing of a combustion process was done. In particular, the time-varied flame properties were examined using a novel digital image and colour processing methodology. This technique makes use of the observed correlation between a digital image colour signal and physical flame radiation characteristics in the visible wavelength domain. Aspects of RGB and HSV colour modelling principles were applied to show that the addition of colour identification in the image processing of high-speed flame image data could yield three useful parameters which are related to the dynamic behaviour of different flame emanating components. First, the validity of the colour identities for tracking the yellowish-red diffusion and greenish-blue premixed flame colourations were examined by comparing their respective flickering frequency profiles. Then, the usefulness of the extracted Rdiffusion, Gpremixed and Bpremixed colour signals to abstractly represent the behaviour of soot, C2* and CH* emission characteristics in a dynamic flame transition from diffusion to stoichiometric premixed condition was demonstrated. In particular, the colour signal ratio Bpremixed/Gpremixed was correlated to exemplify the approximate time-varied state of the equivalence ratio from the imaged combustion phenomenon.

Fuel mixing effect on the flickering of jet diffusion flames

An experimental study has been performed for a buoyant jet diffusion flame, which was observed to oscillate at different frequencies spatially. The flame dynamics and structure were visualized by a commercial digital camera and a high-speed camera. Mixed fuel of methane and propane at certain proportion was found to generate very different vortex shedding behaviours. As a result, the flickering frequency of a methane/propane 1:1 mixture can be half of that of typically observed for pure methane or propane flame. The distance between adjacent flame puffs or the size of vortical structures in the reacting flow field, which can be modified by the fuel composition, was identified to be the key factor that affects the flickering frequencies. Repeated tests confirmed that mixed fuel at certain proportions can have a significant effect on the flame flickering frequency through the modification of vortex structure and dynamics.

Effect of a uniform electric field on soot in laminar premixed ethylene/air flames

The effect of a nominally uniform electric field on the initially uniform distribution of soot has been assessed for laminar premixed ethylene/air flames from a McKenna burner. An electrophoretic influence on charged soot particles was measured through changes to the deposition rate of soot on the McKenna plug, using laser extinction (LE). Soot volume fraction was measured in situ using laser-induced incandescence (LII). Particle size and morphologies were assessed through ex situ transmission electron microscopy (TEM) using thermophoretic sampling particle diagnostics (TSPD). The results show that the majority of these soot particles are positively charged. The presence of a negatively charged plug was found to decrease the particle residence times in the flame and to influence the formation and oxidation progress. A positively charged plug has the opposite effect. The effect on soot volume fraction, particles size and morphology with electric field strength is also reported. Flame stability was also found to be affected by the presence of the electric field, with the balance of the electrophoretic force and drag force controlling the transition to unstable flame flicker. The presence of charged species generated by the flame was found to reduce the dielectric field strength to one seventh that of air.

Monitoring combustion systems using HMM probabilistic reasoning in dynamic flame images

Abstract In this paper, a novel method of on-line flame detection in video is proposed. Processing the data generated by an ordinary camera monitoring scene, it aims to early detect the current state of the combustion system and prevent the system from further degradation and occurrence of failure. Due to the dynamic change of the combustion system, the turbulent flame flicker produces images with different spatial and high temporal resolutions. The proposed method consists of hidden Markov model (HMM) and multiway principal component analysis (MPCA). MPCA is used to extract the cross-correlation among spatial relationships in the low dimensional space while HMM constructs the temporal behavior of the sequential observation. Although the prior process knowledge may not be available in the operation processes, the probability distribution of the normal status can be trained by the images collected from the normal operation processes. Subsequently, monitoring of a new observed image is achieved by a recursive Viterbi algorithm which can find the transition state sequence from series of observed image data. The proposed method, like the philosophy of traditional statistical process control, can generate simple probability monitoring charts to track the progress of the current transition state sequence and monitor the occurrence of the observable upsets. The advantages of the proposed method, data from the monitoring practice in the real combustion systems, are presented to help readers delve into the matter.

Multi-feature fusion based fast video flame detection

A video flame detection method based on the multi-feature fusion is presented in this paper. The temporal and spatial characteristics of flames, such as ordinary flame movement and color clues, a flame flickering detection algorithm is incorporated into the scheme to detect fires in color video sequences. An improved Gaussian mixture model method is firstly adopted to extract moving foreground objects from the still background of detection scenes; secondly, detected moving objects are then categorized into candidate and non-candidate flame regions by using a flame color filtering algorithm; finally, a flame flicker identification algorithm based on statistical frequency counting is used to distinguish true flames from fire-like objects in video images. Testing results show that the proposed algorithms are effective, robust and efficient. The processing rate of the flame detection method can achieve 24 fps with image size of 320 × 240 pixels on a PC with an AMD 2.04 GHz processor.

Low-Frequency Oscillation of a Non-Premixed Flame on a Bluff-Body Burner

The flickering characteristics of a non-premixed flame interacting with von Kármán vortex street on a bluff-body burner were investigated experimentally with varied velocities of the central fuel jet and annular air stream. Flow visualization in terms of a shadowgraph, planar laser-induced fluorescence of hydroxyl, and Mie scattering, which was obtained via PIV used to analyze the flow field, indicated that the flame flickering was caused by periodically growing, toroidal vortices outside the flame surface that were much larger than those observed in buoyant jet flames. In contrast to a weak dependence of the latter on the co-flowing air velocity and the central fuel velocity, the present case was strongly correlated with the von Kármán vortex street shed from the rim of the air stream that produced vortex pairing downstream, and was coupled with buoyancy effects rendered by the pilot flame in the recirculation zone as well as the central jet flame. The flame surface was consequently wrinkled substantially by the strong flame-vortex interaction, leading to frequent local extinction. Furthermore, due to a significant extension of the residence period, the frequency of oscillation was greatly decreased when the velocity of air flow was large enough to introduce a recirculating flame behind the bluff body. It was then increased when the air velocity increased. In contrast, a maximal frequency of oscillation was generated when the fuel velocity was increased, which occurred as the flame mode altered from a recirculation zone flame to a jet-dominated flame, for which the central jet flame was about to penetrate the recirculation zone. This frequency was caused by the intense interactions among the recirculated air, the central jet flame, and small eddies convened around the dome of the recirculation zone. The global behaviors of flame flickers could be further summarized using the Strouhal number and the Reynolds number of fuel jet in a relation of (St)(Re c ) = C, whereas the constant C was modified by the Reynolds number of air flow.

Fire Detection in Video Using LMS Based Active Learning

In this paper, a video based algorithm for fire and flame detection is developed. In addition to ordinary motion and color clues, flame flicker is distinguished from motion of flame colored moving objects using Markov models. Irregular nature of flame boundaries is detected by performing temporal wavelet analysis using Hidden Markov Models as well. Color variations in fire is detected by computing the spatial wavelet transform of moving fire-colored regions. Boundary of flames are represented in wavelet domain and irregular nature of the boundaries of fire regions is also used as an indication of the flame flicker. Decisions from sub-algorithms are linearly combined using an adaptive active fusion method. The main detection algorithm is composed of four sub-algorithms (i) detection of fire colored moving objects, (ii) temporal, and (iii) spatial wavelet analysis for flicker detection and (iv) contour analysis of fire colored region boundaries. Each algorithm yields a continuous decision value as a real number in the range [−1, 1] at every image frame of a video sequence. Decision values from sub-algorithms are fused using an adaptive algorithm in which weights are updated using the least mean square (LMS) method in the training (learning) stage.

Flame-Intrinsic Kelvin−Helmholtz Instability of Flickering Premixed Flames

The flame-intrinsic Kelvin−Helmholtz (KH) instability, a self-oscillation of laminar flame fronts extending to the streamwise direction, was experimentally investigated using inverted conical premixed flames, and the governing parameters of flame flickering were deduced. We supposed that the flame flickering is caused by the KH instability, and the feature may be determined by which is dominant between the buoyancy-driven one taking place in the shear layer between hot products and ambient air and the flame-intrinsic destabilization because of the vorticity generation of flame fronts. For the distinction between the two, the former effect was excluded using a cylindrical quartz tube preventing the interaction with the ambient air and the latter was characterized by the flickering frequencies and wavelengths obtained by chemiluminescence signals and instantaneous images, respectively. Besides, laminar burning velocities were measured from the velocity normal to the flame front by the particle image velocimetry (PIV) method. From the dimensional analysis and physical considerations on an oblique plane flame front, it was estimated that the frequency will be determined by the flame structure and described by the modified Richardson number Rif involved with the flame thickness and laminar burning velocity. From the experiments, it was revealed that the Strouhal number, dimensionless frequency, is proportional to Rif0.869 rather than the conventional Richardson number within the experimental conditions. This means that the flickering motions could be influenced by the flame-intrinsic KH instability described by the flame structure, in addition to the buoyancy-driven KH instability determined by the mixture velocity and length scale of burner geometry.

Characteristics of Flow and Flame Behavior Behind Rifled/Unrifled Nozzles

A novel rifled nozzle was installed behind a conventional combustion exhauster to improve combustion efficiency. The rifled nozzles improve the momentum transmission, turbulent strength, and mixing efficiency between the central jet and annular jet. The flow characteristics behind the nozzles (rifled and unrifled) were visualized and detected using the smoke-wire flow visualization, particle image velocimetry, and hot-wire anemometry. The cold flow structures were categorized into four modes—jet flow, single bubble, dual bubble, and turbulent flow. The topological scheme was adopted to analyze and verify these flow modes. The flame structures behind the nozzles (rifled and unrifled) are classified into three modes—jet flame, flickering flame, and turbulent flame—using the direct-photo visualization. The flame height of a 12-rifled nozzle is decreased by about 50% under that of an unrifled nozzle. The flame shedding frequency declines rapidly in the flickering flame mode and the relationship between the Strouhal number (Sr) and annular velocity (ua) is Sr=0.0238+0.13/ua.

Effects of fuel type and equivalence ratios on the flickering of triple flames

An experimental study has been conducted in axisymmetric, co-flowing triple flames with different equivalence ratios of the inner and outer reactant streams ( 2 < ϕ in < 3 and 0 ⩽ ϕ out < 0.7 ). Different fuel combinations, like propane/propane, propane/methane or methane/methane in the inner and outer streams respectively, have been used in the experiments. The structures of the triple flames have been compared for the different fuel combinations and equivalence ratios. The conditions under which triple flames exhibit oscillation have been identified. During the oscillation, the non-premixed flame and the outer lean premixed flame flicker strongly, while the inner rich premixed flame remains more or less stable. The flickering frequency has been evaluated through image processing and fast Fourier transform (FFT) of the average pixel intensity of the image frames. It is observed that, for all the fuel combinations, the frequency decreases with the increase in the outer equivalence ratio, while it is relatively invariant with the change in the inner equivalence ratio. However, an increase in the inner equivalence ratio affects the structure of the flame by increasing the heights of the inner premixed flame and non-premixed flame and also enlarges the yellow soot-laden zone at the tip of the inner flame. A scaling analysis of the oscillating flames has been performed based on the measured parameters, which show a variation of Strouhal number ( St) with Richardson number ( Ri) as St ∝ Ri 0.5. The fuel type is found to have no influence on this correlation.

Impact of soot on flame flicker

The impetus for the present study is a perceived anomaly in which a vertically mounted ethylene jet diffusion flame used to study soot is observed to be steady; whereas similar flames used in other studies have been observed to flicker or oscillate at a frequency of about 12Hz. This difference raises a question as to why the sooting flame is steady. A well-validated, Navier–Stokes-based, time-dependent numerical code is used to provide an answer to this question. A prediction is made, and later confirmed by experiments, of a phenomenon that is new to the authors and perhaps to the combustion community. That is, soot radiation can influence flame flicker resulting from a buoyancy-induced instability to such a point that the oscillation is completely suppressed. Experiments are used to validate the prediction and the result of numerical studies of the impact of soot on flame flicker. It is found that in a jet diffusion flame, the magnitude of flame oscillations (flicker) decreases with the amount of soot generated in the flame; however, the frequency of the oscillations does not change. Good agreement is obtained between measurements and computations for the dynamic behavior of flames with different soot levels. Both experiments and calculations yield a steady flame when the generated soot is sufficiently high and then produce the initial dynamic flame when the soot formation is restricted. The impact of soot on flame flicker is further studied by performing a simulation for the sooty flame after neglecting radiation from soot.

The effect of a lifted flame on the stability of round fuel jets

The stability and dynamics of an axisymmetric lifted flame are studied by means of direct numerical simulation (DNS) and linear stability analysis of the reacting low-Mach-number equations. For light fuels (such as non-premixed methane/air flames), the non-reacting premixing zone upstream of the lifted flame base contains a pocket of absolute instability supporting self-sustaining oscillations, causing flame flicker even in the absence of gravity. The liftoff heights of the unsteady flames are lower than their steady counterparts (obtained by the method of selective frequency damping (SFD)), owing to premixed flame propagation during a portion of each cycle. From local stability analysis, the lifted flame is found to have a significant stabilizing influence at and just upstream of the flame base, which can truncate the pocket of absolute instability. For sufficiently low liftoff heights, the truncated pocket of absolute instability can no longer support self-sustaining oscillations, and the flow is rendered globally stable.

Experimental study of diffusion flame oscillations and empirical correlations

A diagnostic tool, the cadmium sulfide photocell, was used to study flame oscillation. This light sensitive device proved to be relatively accurate and cheap for measuring flame oscillations. This was done by measuring the light variations associated with – and as a result of – flame oscillation or flicker. It is speculated that the flame flow conditions and mechanisms responsible for flame oscillations are highly sensitive to various burner configurations. In the study, various burner configurations were placed inside an enclosure and the resulting flame oscillation frequency was measured. Considered were: (a) size of slot burner orifice, (b) size of slot burner nominal diameter, and (c) gas supply flow rate. Each of these variables was changed to determine whether the flame oscillation frequency followed the Von Karman vortex shedding or other models. It was shown that there is a strong relation between St and Fr numbers.

Reviving the "L" Word

Few of us will forget November 2004. I remember driving myself to the point of pneumonia, having spent the previous two months making "persuasion" calls to my fellow Ohioans during the evenings and doing weekend "lit drops" in tiny rural towns, consistently chased away from houses by snarling dogs. Then the results came in. Most liberals and leftists turned morose; some thought the country had lost its mind. Friends told me they planned to move to Canada (they never did). The journalist Marc Cooper argued that it became easy for a progressive in 2004 to "fancy" oneself "a member of a persecuted minority, bravely shielding the flickering flame of enlightenment from the increasing Christo-Republican darkness."

Fire detection in infrared video using wavelet analysis

A novel method to detect flames in infrared (IR) video is proposed. Image regions containing flames appear as bright regions in IR video. In addition to ordinary motion and brightness clues, the flame flicker process is also detected by using a hidden Markov model (HMM) describing the temporal behavior. IR image frames are also analyzed spatially. Boundaries of flames are represented in wavelet domain and the high frequency nature of the boundaries of fire regions is also used as a clue to model the flame flicker. All of the temporal and spatial clues extracted from the IR video are combined to reach a final decision. False alarms due to ordinary bright moving objects are greatly reduced because of the HMM-based flicker modeling and wavelet domain boundary modeling.

Flame flickering frequency on a rotating Bunsen burner

The empirical correlation ( St * 2 / Ri = 0.00028 Re * 2 / 3 : St * is the reduced Strouhal number, Ri the Richardson number, Re * the reduced Reynolds number) proposed by Kostiuk and Cheng [1995. The coupling of conical wrinkled laminar flames with gravity. Combustion and Flame 103, 27–40] for predicting buoyancy-induced flame flickering frequency, has been experimentally investigated under the swirling flow conditions produced by a rotating Bunsen burner, focusing on how the flame flickering frequency changes with increasing swirl number S. Under low swirling conditions up to S ≈ 0.1 , the flame flickering frequency f t did not change and monotonically increased with increasing the bulk flow velocity from the burner tube U. The trend of the data is similar to those obtained by Durox et al. [1995. Some effects of gravity on the behavior of premixed flames. Combustion and Flame 82, 66–74] and Kostiuk and Cheng (1995) and could be fitted well with the empirical correlation Kostiuk and Cheng (1995). However, under high swirling conditions ( S > 1 ), f t significantly decreased up to ≈ 80 % of that with non-swirling condition, and became insensitive to U. As a result, the data under S > 1 could not be correlated by the empirical equation. From results obtained by laser doppler velocimetry (LDV), this inconsistency is due to flow divergence shown by the change in the velocity distribution between the burner exit and the flame tip with burner rotation. A minimum value of the centerline velocity u j , min between the burner exit and the flame tip fulfills an important role in controlling the flame flickering frequency. The use of this parameter allows the empirical correlation to be extended to the high swirling case by means of a modified empirical correlation St * 2 / Ri = 0.00028 ( Re * exp ( - 0.64 S 1.78 ) ) 2 / 3 .

Vortex-coupled oscillations of edge diffusion flames in coflowing air with dilution

The unsteady characteristics of oscillating methane diffusion flames in coflowing air diluted with CO 2 in earth gravity have been studied experimentally and computationally. The measured frequency of flame flickering due to buoyancy-driven large-scale vortices was bi-modal; it jumped from ≈11 Hz to ≈15 Hz with an increase in the air velocity (at ≈11 cm/s). As CO 2 was added into coflowing air gradually, the base (edge) of the flame detached from the burner rim, oscillated at half the flickering frequency, and blew off eventually. Numerical simulations with full chemistry predicted the internal flame structure and unsteady flame behavior: flame flickering, tip separation, base detachment, oscillation, and blowoff, in good agreement with the experiment. The mechanism of the edge diffusion flame oscillation was due to a cyclic series of events: (1) flame-base detachment and drifting downstream as a result of weakening due to dilution and a momentary increase in the entrainment-flow velocity associated with the vortex evolution, (2) fuel–air mixing in widened, lower-speed, wake space between the flame base and the burner rim, and (3) flame-base propagation through the flammable mixture layer back to the burner rim. A peak reactivity spot (reaction kernel) at the edge diffusion flame controlled the unsteady behavior through its dramatic changes in characteristics from the passively drifting to (premixed-type) propagating phase during a cycle. Because a mixing time of approximately 100 ms was required before propagation was enabled, a subsequent vortex evolved and passed. Thus, the flame-base oscillation was strongly coupled with the buoyancy-driven vortex evolution and the oscillation frequency was locked-in to half the flame-flickering frequency. The results have implications in turbulent flame structure; more specifically, the local extinction–mixing–reignition processes, in that the slow molecular mixing can become rate-limiting and the edge diffusion flame structure can be significantly different, depending on the phase in the process.

Quantitative measurements in an unsteady flame using high-speed rainbow schlieren deflectometry

Rainbow schlieren deflectometry was integrated with a high-speed digital imaging system to quantify the scalar flow structure of flames at a high temporal resolution. The schlieren image data were processed to determine the temperature field across the whole field assuming chemical equilibrium in the flame. First, shapes of steady hydrogen jet diffusion flames estimated from the schlieren technique were compared with those obtained by direct visualization. Next, the schlieren technique was applied to flame experiments at a jet exit Reynolds number of 300, which produced periodic oscillations or flame flicker. Rainbow schlieren images taken at acquisition rates of 30, 60, 125, 250, 500 and 1000 frames per second were analysed to construct the instantaneous temperature distributions across the whole field, the vortex convection velocity profiles, and the mean and root-mean-square temperature profiles. Results demonstrate that high-speed rainbow schlieren deflectometry is effective in quantitatively describing the transient flow structure of unsteady flames.

Audio–video fire-detection of open fires

In the scope of the development of a video-based fire detection system, investigations on the dynamic behaviour of open flames have been made [Schultze. Untersuchungen zur Audio-Video-Detektion von offenen Bränden, Diploma thesis, Duisburg, 2003]. Standardised fires were recorded with different audio and video systems and their flow and flickering analysed. The analyses show that the flickering of flames is not only viewable in the video data, but also detectable in the acoustic recordings. Using a block-based motion estimator it is possible to trace the movements of flames. Due to the high velocity of flame movements this motion estimation is however only successful in high speed recordings. Therefore a simplified flow analysis has been developed in order to trace the characteristic pumping movement of flame balls in turbulent open flames. The simplified flow analysis is applicable on usual video recordings and should be considered for implementation in an automatic flame detection system.

A lattice Boltzmann algorithm for calculation of the laminar jet diffusion flame

A new two-distribution lattice Boltzmann equation (LBE) algorithm is presented to solve the laminar diffusion flames within the context of Burke–Schumann flame sheet model. One distribution models the transport of the Schvab–Zeldovich coupling function, or the mixture fraction to combine the energy and species equations. The other distribution models the quasi-incompressible Navier–Stokes equations with the low Mach number approximation. In the quasi-incompressible flows, the thermodynamics quantities depend on the coupling function but not on the hydrodynamic pressure, and the fluid components are assumed to be compressible only in the mixing/reaction region. A systematic and consistent approach to deriving LBEs for the general advection–diffusion equation and the quasi-incompressible Navier–Stokes equations are also presented. The streaming step of the LBEs are discretized by the total variation diminishing (TVD) Lax–Wendroff scheme. Numerical simulations are carried out to reproduce the low frequency flame oscillation (or flame flicker) of buoyant jet diffusion flame. Comparison between the quasi-incompressible model and the incompressible model is presented and the role of non-solenoidal velocity is examined.