Flame Images Research Articles

Premixed CO/air combustion in a closed duct with inhibition

Syngas is a promising alternative energy carrier with low pollutant emissions. Carbon monoxide is the main component of syngas. Once accidentally exposed to air, CO/air mixture could be easily ignited starting fire or explosion, which has been scarcely investigated. Therefore, experiments were conducted in a closed duct to study the premixed stoichiometric CO/air combustion characteristics with/without CO2 addition. Flame images and overpressure were captured by the high-speed schlieren photography and pressure sensor, respectively. It shows that the inhibitor had remarkable negative effect on flame and pressure dynamics. The competitive effect between flame instabilities and buoyancy on flame deformation was established. The classic tulip flame was more robust when the flame instabilities were strengthened along with attenuated buoyancy effect. When the buoyancy effect was significant with higher inhibitor content, an inclined flame front would form without tulip inversion and then propagate downwards. At specific conditions, tulip structure did occur at early stages when the flame instabilities were dominant, while at later stages, with decreasing flame speed, the buoyancy effect became more prominent leading to an inclined front. The thermal and kinetic effects of CO2 on premixed CO/air combustion were scrutinized for a comprehensive view of inhibiting mechanism.

Experimental study on the combustion characteristics of premixed methane-hydrogen-air mixtures in a spherical closed chamber

In this study, we conduct premixed combustion experiments on methane-hydrogen-air mixtures based on an experimental platform of a closed spherical chamber. The flame images, combustion pressure, and ion current signal of premixed combustion under different hydrogen fractions and different initial conditions are analyzed. The flame front is affected by turbulence to form cellular structures. As the hydrogen content increases, the crack and cellular structures of the flame front increase, and the scale decreases. The Markstein length decreases, and the flame stability decreases because of the addition of hydrogen. For dilute hydrogen conditions, the stretch flame speed increases linearly with the addition of hydrogen, and the concentration of hydrogen has a significant influence on the moment when the ion current signal appears. Under rich hydrogen conditions, the stretch flame speed increases exponentially with the addition of hydrogen, and the ion current signal curve is hardly affected by the hydrogen content. The peak combustion pressure in a turbulent environment is slightly higher (less than 0.5 bar) than that in a laminar environment. Turbulence accelerates the combustion velocity, and the peak combustion pressure occurs significantly earlier. However, with the addition of hydrogen, the difference between the time of peak combustion pressure in the laminar and turbulent environments decreases gradually.

Identification of CO$$_{\mathrm{2}}$$ and O$$_{\mathrm{2}}$$ emissions dynamics in a natural gas furnace through flame images, ARMAX models, and Kalman filtering

Identification of CO$$_{\mathrm{2}}$$ and O$$_{\mathrm{2}}$$ emissions dynamics in a natural gas furnace through flame images, ARMAX models, and Kalman filtering

Experimental characterization of ozone-enhanced n-decane cool flames and numerical investigation of equivalence ratio dependence

Low temperature combustion behavior is critical to understand for advanced combustor design and safety, but cool flames are challenging to study under chemistry-transport coupled conditions because of the generally transient nature of these flames. A series of studies have utilized a laminar flat flame Hencken burner platform and ozone enhancement at sub-atmospheric pressures to stabilize dimethyl ether, propane and heptane cool flames and measure flame characteristics such as propagation speed, flame temperature and product species. This work builds upon those prior studies to investigate n-decane, first through experimental characterization of ozone-enhanced n-decane cool flames, including measurement of propagation speeds and flame temperatures and imaging of formaldehyde planar laser induced fluorescence (CH2O PLIF) over a range of equivalence ratios and, second, through comparison of experimental results with one-dimensional cool flame simulations. Experimental findings showed moderate equivalence ratio dependence of cool flame propagation speeds and temperatures at lean equivalence ratios, and no significant equivalence ratio dependence for richer mixtures. Experimental CH2O profiles showed equivalence ratio dependence and no downstream consumption, as is characteristic of cool flames. Simulated n-decane cool flames showed significant dependence of propagation speeds and temperatures on equivalence ratio, as well as downstream consumption of CH2O, at the experimental conditions. Experimental and numerical discrepancies were first investigated by examining the influence of different ozone enhancement strategies on equivalence ratio dependence and the results indicate a minimal influence of ozone enhancement strategy for n-decane, while also providing insights to into equivalence ratio dependencies observed in previous ozone-enhanced and ozoneless cool flame studies. The effect of ozone concentration on simulated n-decane cool flames was then investigated, revealing a strong influence of ozone enhancement on equivalence ratio dependence. Simulated propagation speed and temperature dependence on equivalence ratio were shown to develop with increase in ozone enhancement levels. A kinetics analysis revealed enhancement of low temperature peroxy chemistry and other cool flame heat releasing reactions with ozone addition, increasing simulated flame temperatures and promoting the onset of early intermediate temperature chemistry (ITC), inducing equivalence ratio dependence which was not seen in experimental results. Finally, insights from this analysis were used to investigate reconciliation of differences in experimental and simulated cool flame temperatures and identify potential kinetics model discrepancies from experimental findings.

The effects of particle size on flameless combustion characteristics and NOx emissions of semi-coke with coal preheating technology

The experiments of the effect of particle size on flameless combustion characteristics and NOx emissions of Shenmu semi-coke were launched on a bench-scale coal preheating combustion test rig. The results revealed that the difference of fuel particle size greatly influenced the yield of each component in high-temperature coal gas and the reactive activity of coal char, thus subsequently affecting the final combustion process and NOx emissions. The specific surface area and pore volume of Shenmu semi-coke were both significantly increased after preheated, leading to a better reactivity and the improved combustion performance, especially when d50 (the mean particle size) = 163.84 μm, where the temperature of the down-fired combustor reached the highest. The different particle size of Shenmu semi-coke resulted in different preheating temperature, which further affected the content of each component in the high-temperature coal gas. During the self-preheating process, most volatile matter and fuel bound N were separated out and the released N was mainly converted to N2 and NH3, besides the conversion rate of each component (C, H, N and V) was the highest when d50 = 100.17 µm. Additionally, all experimental conditions achieved flameless combustion, with a transparent combustion area, but no flame front visible to the naked eye, and the highest combustion efficiency was 98.23% when d50 = 163.84 µm with the brightest flame image. It was discovered in this experimental that there was a critical value (d50 = 198.54 μm) of Shenmu semi-coke particles. When the particle size was lower than or higher than the critical value, NOx generated by combustion would decrease. The lowest NOx emission was 89.46 mg/m3 (@6% O2) when d50 was 35.34 μm.

Effects of oxygen on soot formation in laminar co-flow flames of binary mixtures of ethane, DME, and oxygen

Soot volume fractions of binary mixtures of ethane, DME and oxygen were evaluated experimentally in a co-flow burner. Soot volume fractions were measured by diffuse-light line-of-sight attenuation technique. The carbon mass flow rate and the C/H ratio were kept constant for all mixtures, while oxygen was introduced into the fuel stream either as an additive or as intramolecular oxygen in DME or both. Visible flame heights were decreased with increasing oxygen in the fuel stream. Binary mixtures of ethane and DME did not show any synergistic effects on soot formation. The maximum soot volume fraction was decreased from 1.41 ppm in the neat ethane flame to 0.08 ppm in the neat DME flame. On the other hand, oxygen addition to either ethane or DME resulted in an increase in soot concentrations. The maximum soot volume fraction reached to 2.1 ppm and 0.29 ppm in ethane/oxygen and DME/oxygen mixtures, respectively. The peak soot yield initially decreased for all mixtures studied as oxygen was introduced into the fuel tube. A unique transition was observed in DME/oxygen mixtures at an equivalence ratio of 6.8. When more oxygen was added beyond this ratio, the effect of oxygen on soot formation was suddenly reversed, and the peak soot volume fraction and yield were significantly decreased. This sudden reversal was explained as a result of a transition into a more premixed mode of combustion, supported by the flame images as well as chemiluminescence measurements. A distinct two-zone flame structure, that is, an inner rich zone and an outer stoichiometric zone separated by a dark region, was observed for the neat DME flame and DME/oxygen mixture flames. It is argued that as oxygen becomes available from DME decomposition, it produces a premixed region, and the flame chemistry is significantly altered beyond a critical oxygen concentration. Current radially resolved soot volume fraction measurements suggest that the effects of oxygen on soot formation can be quite complex with nonmonotonic behaviour, sudden effect reversals and shifts in soot formation and oxidation zones.

LES investigation of cycle-to-cycle variation in a SI optical access engine using TFM-AMR combustion model

Multi-cycle large-eddy simulations (LES) are performed to investigate combustion cycle-to-cycle variability (CCV) in a gasoline spark ignited optical access engine operating under homogeneous stoichiometric conditions. Combustion is addressed with the Thickened Flame Model (TFM) and finite rate chemistry is accounted for through a reduced oxidation reaction mechanism. In view of the fact that computational costs of LES engine simulations are still very high today, this work investigates the use of adaptive mesh refinement (AMR) in the flame zone in conjunction with the artificial flame thickening applied by the TFM model. The paper discusses how the resulting coupled TFM-AMR combustion model allows good resolution of the flame, maintaining accuracy at acceptable costs. First, the details of the coupled model are presented and the effects of the parameters are explored, highlighting their impact on the combustion prediction. Then, computational fluid dynamics (CFD) simulation results are validated against experimental data collected in a low-speed low-load engine point, by comparing 20 LES cycles and 100 measured cycles, for mass fraction burned, combustion phasing, flame images and CCV indices. Lastly, a detailed investigation on the fastest and slowest numerical cycles is presented, analyzing instantaneous flame structures, ignition behaviors, propagation speeds, and probability density function (PDF) of the instantaneous velocity fluctuation around the spark region. The results show that combustion variability is highly correlated to the resolved velocity field and the resolved turbulence intensity, which is found to be the main cause of CCV and affects the early flame kernel growth. This work is an early attempt to use TFM-AMR combustion model for LES simulations of internal combustion engines.

Influence of Wind Speed and Direction on the Combustion Parameters of the Exposed XPS Insulation Wall

Most of the fire accidents of exterior insulation in recent years started with the ignition of unprotected and exposed insulation materials. The fire spread on the thermal insulation layer will be affected by outdoor wind. In this paper, the influence of wind direction and wind speed on wall fire propagation characteristics was studied by carrying out medium sized experiments. The temperatures on the wall surface, radiation heat flux nearby, the infrared image and flame image were recorded during the tests with variation of wind direction and wind speed. It is found that with the increase of the angle between the wind direction and the wall, the surface temperature of the material, the radiation heat flux released decreases.This study results can provide valuable references for optimizing the fire safety design of exterior insulation layer.

Lean blow-off characteristics of a tangential entry type dual swirling free and impinging flame

An experimental study has been carried out to investigate the lean blow-off (LBO) characteristics of a tangential entry type dual swirling flame under different operating and geometrical parameters. The amount of swirl is controlled by changing the ratio of tangential to axial in-flows (T:A). The qualitative flame structure is accessed by taking the direct flame images and the radial distribution of static impingement pressure. The flame structure has also been illustrated using commercial CFD code ANSYS Fluent®. The results showed that the flame behavior and the mixing between the reactants are strongly influenced by the outer flame Reynolds number (Re(o)) and the intensity of swirl (S). Significant transitions in flame shapes were observed owing to these flow alterations. These transitions in flame shapes strongly govern the stability near blow-off conditions. It has been seen that LBO limits are considerably improved with the presence of an inner supporting flame and also in presence of the impingement surface. The inner conical flame significantly supports the outer swirling flames for stability at high flow rates and under fuel-lean operating conditions. The outer flame could be sustained stably up to ϕ(o) of 0.1 with inner stoichiometric flame and both inner and outer flames could be stabilized for simultaneous variation of equivalence ratio up to ϕ of 0.6 (S = 0.86). Geometrical design parameters of the burner showed a major impact on improving the LBO limits. Thus, the dual flame presents significant improvements to lean blow-off limits as compared to the single outer swirling flame.

The Reattachment Process of a Lifted Jet Diffusion Flame by Repetitive DC Pulse Discharges

On research of plasma assisted combustion, effects of electric and plasma discharges in DC, AC and pulse forms on reattachment of a lifted flame have attracted extensive attention. However, the detailed plasma assisted reattachment process and mechanism and roles of induced corona discharge and corona-induced ozone on the reattachment process are still unclear and undocumented. The forced reattachment process of a lifted diffusion jet flame by repetitive DC electric pulse discharges was experimentally investigated in this study using high-speed flame imaging, conditioned particle image velocimetry (PIV), and planar ozone concentration imaging. The forced reattachment process can be divided into three stages in sequence: ionic wind prior to corona initiation, corona initiation, and corona enhancement propagation. The conditioned PIV results showed that the instantaneous flame base propagation velocity is sufficiently enhanced beyond the laminar burning velocity for high pulse-repetition-frequency (PRF) cases at the instant of pulse discharge (on pulse) due to the enhanced oxidation of the corona induced ozone. By observing the dynamic flame-base behavior and evolution characteristics of the short-lived corona induced ozone for various PRFs, the novel forced reattachment process and mechanism of a lifted jet flame induced by repetitive DC electric pulse discharges is proposed.

Determination and evolution of fractal property of n-heptane pool fires caused by depressurization process in an aircraft cargo compartment

Fractality is one of the inherent characteristics to present the irregularity and self-similarity of an object. In this study, the fractal theory was applied to the fire scene under decreasing pressure in an aircraft cargo compartment, considering that the compartment is usually inaccessible in flight and flame images are a kind of convenient monitoring information compared to other detection methods. A series of n-heptane pool fires were performed under decreasing pressure with four depressurized rates. The flame image was captured by a video camera with fixed aperture and exposure time. The fractal dimension (FD) of flame images was computed via two approaches, i.e., the box-counting (BC) and differential box-counting (DBC) methods. The results show that n-heptane pool fires have a good fractal property for both binary and grayscale images. The FD via BC method shows little sensitivity with decreasing pressure. However, the FD via DBC method presents a significant linear relationship with a slope of 0.0016 in the range from 2.85 to 2.75 during the depressurization. The quantitative correlation was brought into the classical radiation modeling, and the exponential function between FD and MLR of n-heptane was established.

Image Flame Detection Method Based on Improved YOLOv3

The general flame detection method is sensitive to environmental changes and has low robustness, so the detection rate is not high. In this paper, a deep learning flame detection algorithm based on the improved YOLOv3 algorithm is proposed to automatically extract flame characteristics; with the help of MobileNet’s deep separable convolution to compress the deep network model, replace the YOLOv3 backbone network with MobileNet, decompose the standard convolution into deep convolution and point-wise convolution, which can effectively reduce the amount of model parameters, Use K-means clustering to generate Anchor Box at different scales, further enhance the detection effect of the model. This paper conducts data analysis experiments on the self-built flame data set. The results show that the optimized algorithm (M_YOLOv3) has achieved 97.34% average precision (AP) on the data set. The flame target detection algorithm proposed in this paper can achieve high-precision and rapid detection of multi-scale flame targets in flame images with different shapes and sizes.

A Two-Stream CNN Model with Adaptive Adjustment of Receptive Field Dedicated to Flame Region Detection

Convolutional neural networks (CNN) have yielded state-of-the-art performance in image segmentation. Their application in video surveillance systems can provide very useful information for extinguishing fire in time. The current studies mostly focused on CNN-based flame image classification and have achieved good accuracy. However, the research of CNN-based flame region detection is extremely scarce due to the bulky network structures and high hardware configuration requirements of the state-of-the-art CNN models. Therefore, this paper presents a two-stream convolutional neural network for flame region detection (TSCNNFlame). TSCNNFlame is a lightweight CNN architecture including a spatial stream and temporal stream for detecting flame pixels in video sequences captured by fixed cameras. The static features from the spatial stream and dynamic features from the temporal stream are fused by three convolutional layers to reduce the false positives. We replace the convolutional layer of CNN with the selective kernel (SK)-Shuffle block constructed by integrating the SK convolution into the deep convolutional layer of ShuffleNet V2. The SKnet blocks can adaptively adjust the size of one receptive field with the proportion of one region of interest (ROI) in it. The grouped convolution used in Shufflenet solves the problem in which the multi-branch structure of SKnet causes the network parameters to double with the number of branches. Therefore, the CNN network dedicated to flame region detection balances the efficiency and accuracy by the lightweight architecture, the temporal–spatial features fusion, and the advantages of the SK-Shuffle block. The experimental results, which are evaluated by multiple metrics and are analyzed from many angles, show that this method can achieve significant performance while reducing the running time.

Study of the effect of bluff body size on stability limits of a premixed natural gas swirl burner

Flame stabilization is highly important in lean premixed combustion which is considered a desirable technology for low-emission gas turbines. Swirl stabilization is one of the most common methods used for this purpose. Simple generation of swirl, however, is not enough to reach an acceptable operating range. A method of improving stability is using a bluff body in order to resist flame flashback caused by combustion-induced vortex breakdown at moderate to high swirl numbers. The present study aims to investigate the effect of bluff body size on the stability boundaries of flashback and blowout and also on flame shape. A premixed swirl burner is designed and built to operate with natural gas at atmospheric condition. In addition to gas and air flow rates at stability limits, normal and chemiluminescence flame images are documented. The non-reacting flow is also simulated to investigate the flow pattern out of the burner nozzle. It is found that bluff body size has a significant influence on flashback prevention and to a lesser extent on lean blowout limit while having minor effect on the shape of the flame. Just before blowout, low-frequency fluctuations are observed in the chemiluminescence radiations that are associated with the periodic changes in flame zone. Nonetheless, no warning signs are observed before flashback. Looking into the velocity profiles at the nozzle exit also shows an increase in axial velocity near the center of the nozzle with larger bluff bodies, revealing the mechanism of improvement of burner resistance to flame flashback.

A novel diagnostic method based on filter bank theory for fast and accurate detection of thermoacoustic instability

This study proposes and analyzes a novel methodology that can effectively detect multi-mode combustion instability (CI) in a gas turbine combustor. The experiment is conducted in a model gas turbine combustor, and dynamic pressure (DP) and flame images are examined during the transition from stable to unstable flame, which is driven by changing fuel compositions. As a powerful technique for early detection of CI in multi-mode as well as in single mode, a new filter bank (FB) method based on spectral analysis of DP is proposed. Sequential processing using a triangular filter with Mel-scaling and a Hamming window is applied to increase the accuracy of the FB method, and the instability criterion is determined by calculating the magnitude of FB components. The performance of the FB method is compared with that of two conventional methods that are based on the root-mean-squared DP and temporal kurtosis. From the results, the FB method shows comparable performance in detection speed, sensitivity, and accuracy with other parameters. In addition, the FB components enable the analysis of various frequencies and multi-mode frequencies. Therefore, the FB method can be considered as an additional prognosis tool to determine the multi-mode CI in a monitoring system for gas turbine combustors.

Compression ignition and spark assisted ignition of direct injected PRF65 spray

In this study the spark assisted compression ignition combustion (SACI) developments were investigated using PRF65 (low octane fuel), a mixture of 65% isooctane (by volume) and 35% n-heptane (by volume) with a RON of 65. Characteristics like the cumulative heat release (CHR) and the peak heat release rates (HRR) were studied pressure data from experiments conducted in a constant volume combustion chamber (CVCC) for more precise control of the tested conditions. Spray flame images were also studied using high speed imaging systems to understand the effect of the conditions tested in the luminosity of the flame. Experiments were performed to understand the effects of oxygen concentration and ambient temperatures. Results show that the heat release rate increases initially and then decreases with the increase in the ambient temperature and the peak heat release rate appears around 650 K to 700 K. The peak heat release rate timing is advanced with the increase of the ambient temperature or oxygen level. Flame luminosity was also found to increases with the increase in ambient temperature. Under a low ambient temperature, the oxygen level plays a major role in affecting the peak heat release rate. Under lower oxygen levels, the flame becomes darker, the ignition delay becomes longer, and the combustion process takes more time to complete. A well timed spark timing was found to advance the peak HRR and shorten ignition delay, but this effect becomes minor when the temperature increases.

Experimental investigation of unconfined turbulent premixed bluff-body stabilized flames operated with vapourised liquid fuels

The structure of turbulent unconfined bluff-body flames of vapourised liquid fuels was investigated at conditions far from and close to blow-off with high-speed (5 kHz) OH-PLIF imaging and 10 Hz CH2O-PLIF imaging. Four different fuels were considered: ethanol, heptane, and two different kerosene blends (a conventional Jet-A and an alcohol-to-jet kerosene, respectively denoted as A2 and C1 following the USA National Jet Fuels Combustion Programme. OH-PLIF images of ethanol flames far from blow-off display a high intensity of OH-LIF signal along the shear layer. In contrast, the OH-LIF signal was evenly distributed throughout the recirculation zone (RZ) of the heptane and kerosene flames. Regardless of the fuel used, close to blow-off the flame becomes shorter with peak OH-LIF signal intensities lying inside the RZ. All four fuels showed a decrease in flame surface density (Σ2D) and broadening of the 2-D curvature PDFs as their blow-off limits were approached. An increase in local turbulent consumption speed was observed in the downstream region as the flames approached blow-off. No significant variation in Σ2D, curvature PDF, and local turbulent consumption speed was observed between the different fuel types. The average CH2O-layer thickness was larger than the computed laminar flame value by a factor of two and six for conditions far from and close to blow-off, respectively. Moreover, heptane and kerosene flames showed more pockets of CH2O-LIF signal within the RZ as compared to ethanol, suggesting that considerably more partially-combusted fluid enters the RZ of the former than the latter. High-speed particle image velocimetry was performed to measure the local velocity fields and place various regions of the flame on the turbulent premixed regime diagram. It was observed that, regardless of fuel type, conditions close to blow-off occupy the same region on the regime diagram. However, the fact that the fuel type results in differences in some structural features near blow-off suggests that flames produced with heavy hydrocarbon fuels involve chemistry effects at blow-off that are not fully characterized by laminar flame properties.

Effect of Al2O3 particle size reduction on aluminium dust explosion

To investigate the effect of Al2O3 particle size on an aluminum explosion, the overpressure and flame velocity in a vertical duct were evaluated. The results show that the inhibitory effect of submicron Al2O3 is best, while the inhibitory effect increases with increasing inerting ratio. However, the inhibitory effect of micron Al2O3 does not increase significantly after the inerting ratio exceeds 40%. For high-concentration aluminum powder, 0.8 μm Al2O3 with an inerting ratio less than 20% promotes aluminum explosion. As the inerting ratio increases beyond 20%, however, the overpressure decreases. Furthermore, Al2O3 inhibits the formation of the intermediate product AlO and decreases the flame brightness. As the inerting ratio of 0.8 μm Al2O3 reaches 50%, the white patches in the flame image disappear. The results of scanning electron microscopy showed that the explosion products agglomerate and some dot-like protrusions appear on the surface of the unburned aluminum particles. The inhibition mechanism was qualitatively investigated. Physical heat absorption is proven to play a limited role. Thermal radiation and chemical inhibition play a key role. The chemical effect mainly influences the surface reaction energy source.

Flame Image Processing and Classification Using a Pre-Trained VGG16 Model in Combustion Diagnosis.

Nowadays, despite a negative impact on the natural environment, coal combustion is still a significant energy source. One way to minimize the adverse side effects is sophisticated combustion technologies, such as, e.g., staged combustion, co-combustion with biomass, and oxy-combustion. Maintaining the combustion process at its optimal state, considering the emission of harmful substances, safe operation, and costs requires immediate information about the process. Flame image is a primary source of data which proper processing make keeping the combustion at desired conditions, possible. The paper presents a method combining flame image processing with a deep convolutional neural network (DCNN) that ensures high accuracy of identifying undesired combustion states. The method is based on the adaptive selection of the gamma correction coefficient (G) in the flame segmentation process. It uses the empirically determined relationship between the G coefficient and the average intensity of the R image component. The pre-trained VGG16 model for classification was used. It provided accuracy in detecting particular combustion states on the ranging from 82 to 98%. High accuracy and fast processing time make the proposed method possible to apply in the real systems.

Furnace Flame Tracking for Combustion Quality Estimation in Coal Fired Boilers

This research work focus is on optimization of the combustion quality as the combustion quality plays a vital role in minimizing NOx and CO emissions in the flue gas. Power generated is also another important factor under consideration so as to meet the demand. Flame monitoring in boilers using digital imaging provides replacement for the conventional technique of temperature measurement. The indigenous method can be replaced by incorporating an online PC based system by analyzing the flame images.