Flat Flame Research Articles

An in-situ method for time-resolved sodium release behaviour during coal combustion and its application in industrial coal-fired boilers

To mitigate the slagging, fouling and high-temperature corrosion problems caused by alkali metals during coal combustion process, measurement of time-resolved alkali metals release is very important. The paper proposed an in-situ approach for measuring sodium (Na) release in coal combustion by Flame Emission Spectroscopy (FES). Through the analysis of spontaneous emission spectra and a calibration procedure, the concentration of gas phase Na, temperature and thermal radiation can be obtained. Firstly, experimental measurement of Zhundong coal particles burning in a flat flame burner was done. Two kinds of Zhundong coal with similar proximate and ultimate analyses, but different ash composition were used. The Na-release history measured by FES was compared with that by LIBS. Results showed that the Na-release at the devolatilization, char, and ash stages can be distinguished by FES. The higher Si/Al content in ash can suppress the Na-release at the ash stage. Moreover, FES method was extended to the measurement of Na-release in four industrial boiler furnaces of two Zhundong coal-fired power plants. Results showed the Na-release measured by FES can reflect the change of fuel and load, and both temperature and thermal radiation play key roles in Na-release in coal combustion.

LES of a lifted methanol spray flame series using the sparse Lagrangian MMC approach

Two-phase multiple mapping conditioning / large eddy simulation (MMC-LES) is applied for the first time to lifted spray flames on a vitiated coflow burner. Three flames with different inlet fuel mass loading are investigated. MMC-LES uses a hybrid Eulerian-Lagrangian-Lagrangian approach for the evolution of gas-phase flow, stochastic particles, and liquid fuel droplets, respectively. Due to the benefits of localizing mixing in a reference mixture fraction space, a sparse-Lagrangian distribution of particles is able to be used, permitting computationally efficient simulations of the flame series using gas-phase chemical kinetics that includes 32 species and 167 reactions. Heat and mass transfer between the stochastic particles and evaporative fuel droplets are also local in the reference mixture fraction space. Analysis of the gas-phase temperature fields reveals two distinct flame base stabilization phenomena. The first, which is observed for the low and intermediate fuel loading cases, produces a flat flame base spreading across the central region of the jet. The second, observed for the highest fuel loading case, produces an annular flame base in the shear layer that surrounds a cold central jet region, which is characteristic of autoignition due to entrainment of hot oxidizer from the co-flow. Except for a few axial locations, predicted radial profiles of temperature and liquid-phase velocity/size statistics yield a reasonable agreement with experimental measurements for all three flames. Furthermore, the trend of decreasing lift-off height with an increase in fuel loading is captured well.

Particle temperature and potassium release during combustion of single pulverized biomass char particles

This work investigated the combustion characteristics of single pulverized biomass-derived char particles. The char particles, in the size range 224–250 µm, were prepared in a drop tube furnace at pyrolysis temperatures of 1273 or 1473 K from four types of biomass particles – wheat straw, grape pomace, kiwi branches and rice husk. Subsequently, the char particles were injected upward into a confined region of hot combustion products produced by flat flames stabilized on a McKenna burner, with mean temperatures of 1460, 1580 and 1670 K and mean O2 concentrations of 4.5, 6.5 and 8.5 vol%. The data reported include particle temperature, obtained using a two-color pyrometry technique, and potassium release rate, measured using a laser-induced photofragmentation fluorescence imaging technique. In addition, particle ignition delay time and burning time, obtained from the temporal evolution of the thermal radiation intensity of the burning char particles, are also reported. The results indicated that ignition of the char particles occurs simultaneously with the starting of the potassium release, then the particle burning intensity increases rapidly until it reaches a maximum, after which both the particle temperature and the potassium release rate remain approximately constant until the end of the char oxidation process. The char ignition process is temperature controlled, and the char oxidation process is oxygen diffusion controlled, with the total potassium release being independent of the oxygen concentration and the temperature of the combustion products. The combustion behavior of the chars studied is more affected by the char type than by the conditions used to prepare them.

Experimental study on aluminum particles combustion in a turbulent jet

In this study, experimental investigation on the aluminum particles turbulent jet flame is carried out. The fluidized aluminum particles ignite and burn in the post-flame environment supported by a premixed flat flame burner. High-speed camera and Mie scattering technique are employed to photograph the flame structure. The jet flame is divided into a preheating zone, a flame zone and a burnout zone based on the self-luminous intensity. In the upstream region of the flame, aluminum particles spread into the hot coflow and need to be heated within a certain distance to be ignited. In the flame region, typical characteristics of a single aluminum particle combustion, including condensation trail and particle rotation are observed using a high magnification zoom lens. The average image of the jet flame indicates that the condensed combustion products are mainly distributed in a thin region at the periphery of the particle clouds. Moreover, the temperature of condensed phase emitters in the flame is derived by fitting Planck’s law, and the results show that increasing the oxidizing ability of the environments can significantly improve the flame temperature and reduce the lift-off height and visible length of the jet flame, but has little influence on the flame structure.

Determination of the volume fraction of soot accounting for its composition and morphology

The soot volume fraction, fv, is essential for combustion engineering, air quality and climate modeling. It is commonly obtained from mobility or optical data assuming soot spheres of constant composition that tend to overestimate fv by up to an order of magnitude. Here, a method is presented for estimation of the fv from soot mobility or optical diagnostics data accounting for the realistic soot morphology and composition. With mobility data, the fv is determined by a relation for soot morphology (fractal-like structure) and the average primary particle diameter from microscopy, gas adsorption or power laws. Such fv and the corresponding soot volume/mass size distributions are in excellent agreement with accurate but tedious soot mass-mobility measurements in premixed flames along various heights above the burner (HAB) as well as from the exhaust of diffusion flames and diesel engines. With light extinction data, the fv is obtained by explicitly accounting for soot composition (through its optical band gap and refractive index, RI) and the light absorption enhancement (due to multiple light scattering within soot agglomerates) in the absorption function, E(RI). So as nascent soot matures by surface growth and agglomeration, its optical band gap decreases while its mobility diameter, carbon to hydrogen (C/H) ratio and fv increase, increasing the soot E(RI) from 0.22 to 0.40, in excellent agreement with laser induced incandescence data from flat premixed ethylene flames. Using this variable E(RI), the soot fv from light extinction data in these flames is up to 50% lower than that using standard constant E(RI) for soot spheres.

TDLAS-based photofragmentation spectroscopy for detection of K and KOH in flames under optically thick conditions.

Photofragmentation spectroscopy is combined with tunable diode laser absorption spectroscopy to measure the line shape of the fragment species. This provides flexibility in choosing the UV pulse location within the line shape and accurate quantification of both target species and background fragment concentrations, even under optically thick conditions. The technique is demonstrated by detection of potassium hydroxide (KOH) and atomic potassium K(g) above solid KOH converted in a premixed methane-air flat flame. Time series of KOH(g) and K(g) concentrations are recorded as a function of solid KOH mass and flame stoichiometry. The total substance released during the conversion is in good agreement with the initial solid KOH mass. Under fuel-rich conditions, increased K(g) concentrations at the expense of KOH(g) are observed compared to thermodynamic equilibrium.

Mid-infrared laser-induced thermal grating spectroscopy of hot water lines for flame thermometry

In this work we report the impact of using mid-infrared laser-induced thermal grating spectroscopy (IR-LITGS) for temperature measurements in flames by probing hot water lines. The measurements have been performed in the product zone of laminar atmospheric CH4/H2/air flat flames with equivalence ratio ranging between 0.6 and 1.05. LITGS is a technique based on thermalization through collisions of the excited molecules for generation of a laser-induced grating, which then decays through thermal diffusion. As such, it tends to have limited application in atmospheric flames compared to flame measurements at elevated pressures due to the faster decay at low gas densities. However, by using mid-IR pump laser beams, it enables the generation of laser-induced gratings with large grating spacing, resulting in strong signal intensities and long signal durations. Single-shot IR-LITGS signals were recorded in the different CH4/H2/air flames that covered a temperature range between 1500 and 1800 K. To test the accuracy, the IR-LITGS flame temperature measurements were compared with laser Rayleigh scattering measurements and the result were in good agreement with each other. The IR-LITGS flame temperature measurements show a repetitive single-shot temperature precision better than 1% and an accuracy of 2.5% of the flame temperature. An IR-LITGS excitation scan of water in the flame shows that some ro-vibrational transitions exhibit no IR-LITGS signal, probably due to less efficient collisional energy transfer mechanism. This is important when deciding the wavelength to use for IR-LITGS flame temperature measurements using water absorption.

Measurement of the adiabatic flame speed and overall activation energy of a methane enriched H2/CO/CO2/N2 low heating value mixture

The objective of this work is to investigate the effect of the methane blending in the adiabatic laminar flame speed and overall activation energy of a CO/CO2 rich, low calorific value, syngas fuel premixed combustion with air. The measurements were performed in a McKenna flat flame burner at constant equivalence ratio, unburned temperature and pressure. A flame asymptotic model curve-fitted to the measurements provided the overall activation energy and adiabatic laminar flame speed. The mixtures tested presented lower heating values from 4.5 MJ/kg to 50 MJ/kg as the CH4 fraction was changed from 0% to 100%. While the measured laminar flame speed remained approximately constant for all fuels tested, the overall activation energy evidenced the higher reactivity of the mixture blended with 20% CH4. Predictions using GRI-Mech 3.0 and San Diego mechanisms closely followed the measurements of the laminar flame speed for higher methane content, but deviations were found for the fuels with higher hydrogen content. Conversely, the mechanisms underpredicted the values of overall activation energy for all mixtures tested. A sensitivity analysis revealed the competition between the CO wet oxidation and the chain branching formation of OH and the increase in reactivity as the fraction of CH4 increased from 0 was related to a switch from a higher sensitivity of the former by the later.

Experimental flat flame study of monoterpenes: Insights into the combustion kinetics of α-pinene, β-pinene, and myrcene

Pinenes and pinene dimers have similar energy densities to petroleum-based fuels and are regarded as alternative fuels. The pyrolysis of the pinenes is well studied, but information on their combustion kinetics is limited. Three stoichiometric, flat premixed flames of the C10H16 monoterpenes α-pinene, β-pinene, and myrcene were investigated by synchrotron-based photoionization molecular-beam mass spectrometry (PI-MBMS) at the Advanced Light Source (ALS). This technique allows isomer-resolved identification and quantification of the flame species formed during the combustion process. Flame-sampling molecular-beam mass spectrometry even enables the detection of very reactive radical species. Myrcene was chosen because of its known formation during β-pinene pyrolysis. The quantitative speciation data and the discussed decomposition steps of the fuels provide important information for the development of future chemical kinetic reaction mechanisms concerning pinene combustion. The main decomposition of myrcene starts with the unimolecular cleavage of the carbon-carbon single bond between the two allylic carbon atoms. Further decompositions by β-scission to stable combustion intermediates such as isoprene (C5H8), 1,2,3-butatriene (C4H4) or allene (aC3H4) are consistent with the observed species pool. Concentrations of all detected hydrocarbons in the β-pinene flame are closer to the myrcene flame than to the α-pinene flame. These similarities and the direct identification of myrcene by its photoionization efficiency spectrum during β-pinene combustion indicate that β-pinene undergoes isomerization to myrcene under the studied flame conditions. Aromatic species such as phenylacetylene (C8H6), styrene (C8H8), xylenes (C8H10), and indene (C9H8) could be clearly identified and have higher concentrations in the α-pinene flame. Consequently, a higher sooting tendency can generally be expected for this monoterpene. The presented quantitative speciation data of flat premixed flames of the three monoterpenes α-pinene, β-pinene, and myrcene give insights into their combustion kinetics.

A new pattern of flame/flow dynamics for lean-premixed, low-swirl hydrogen turbulent jet flames under thermoacoustic instability

This paper presents a phenomenological investigation of a new pattern of lean-premixed hydrogen-flame and flowfield dynamics under thermoacoustic instabilities (TIs) observed inside a low-swirl combustor (LSC) with two combustion chamber lengths of 300 and 500 mm. In this study, the initial bulk inlet velocity was set to Vm=15 m/s. The fuel flow rate was linearly increased while the air flow rate was kept constant, causing a transition from stable operation to TI. Simultaneous time-resolved electronically excited hydroxyl radical (OH*) chemiluminescence imaging and two-dimensional particle image velocimetry were performed along with pressure-fluctuation measurements to investigate spatiotemporal relationships between fluctuating quantities, i.e., OH* chemiluminescence, velocity fields, and pressure, before and after TI onset. A typical TI pattern characterized by a lifted, inverted conical flame caused by diverging, decelerating flow near the injector exit was observed in the long combustor. In contrast, the flame/flow dynamics pattern under TI in the short combustor had strong pressure oscillations and anomalous LSC flame structures which were characterized by a radially flat flame region and counter-rotating vortex pair (CVP) flame structures. In this case, the radial profile of the axial velocity assumed an atypical top-hat-like shape. This feature resulted in thermoacoustic coupling that satisfied the Rayleigh criterion over an almost entirely flat flame region and further strengthened the TI. Thermal expansions of CVP structures served as “converging nozzles” constricting the central flow region. Periodic appearances of this Venturi-like effect caused an alternating diverging (decelerating) and converging (accelerating) flowfield, with a significant peak-to-peak velocity-oscillation amplitude of 15 m/s, equivalent to Vm. These pioneering observations regarding LSCs provide further insights into TI mechanisms in hydrogen-fueled gas turbine combustors.

Measurement of CO temperature and concentration in flat flame based on mid infrared wavelength modulation

CO in combustion products of aeroengine is a sign product of incomplete combustion. Real-time monitoring of CO can feedback combustion efficiency, which is of great significance to the design of combustion chamber and the optimization of fuel ratio of aeroengine. In this paper, based on the mid infrared absorption spectroscopy, a single optical path calibration-free wavelength modulation method is proposed to measure the temperature and concentration of CO in a flat flame. Firstly, the spectrum of light intensity signal and the amplitude frequency response of different frequencies are analysed, and the appropriate scanning frequency and modulation frequency are selected. Secondly, the equalization of Butterworth filter in DPLA is proposed to extract the first and second harmonics. The filter can effectively reduce the influence of nonlinear phase. Furthermore, calibration-free wavelength modulation method is simulated to measure CO at different temperature and concentration. Finally, CO in flat flame with different equivalence ratio at 5mm was measured for verification. The relative error between the measurement results of calibration-free wavelength modulation and that of thermocouple measurement is within 3.6% in the range of concentration 0.11% ∼ 0.24%, which verified the feasibility and effectiveness of the measurement scheme, and provided a reference for the CO concentration measurement at the outlet of aeroengine.

Research on a quantitative method for three-dimensional computed tomography of chemiluminescence.

To develop a more advanced 3D computed tomography of the chemiluminescence method, the first quantitative 3D diagnosis was realized. The nonlinearity coefficient, the nonuniformity coefficient of the camera response, and various optical fiber attenuation coefficients were obtained through correction experiments. The conversion relationship between the number of photons released by the target object per unit time and the camera gray value at a specified solid angle was also calibrated. To verify the quantitative reconstruction equation, 3D reconstructions of a methane-air flat flame and a simulated phantom were performed for comparison. The method can overcome artificial distortions caused by uncorrected reconstruction.

High-speed 1D Raman analyzer for temperature and major species measurements in a combustion environment.

In this Letter, we demonstrate a 5 kHz 1D Raman instrument for temporally and spatially resolved quantitative measurements of temperature and all the major species (N2, O2, H2, and H2O) concentration in H2-air flames. The major constituents of the system are a pulse-burst laser operated at 5 kHz and four back-illuminated CCD cameras operated in subframe burst-gating mode. The use of CCD cameras allows achieving a high sampling rate with no compromise on instrument precision, but it requires one camera for each species of interest. A cascade of dichroic mirrors and bandpass filters spectrally separates the Raman signal associated with each of the four species and directs it to a separate camera. Measurements in a well-characterized H2-air premixed flat flame show that the system has precision comparable with the low-speed Raman system. The measuring uncertainty of the species mole fraction ranges between 1% (N2) and 3∼4% (O2 in lean flames). Measurements in laminar and turbulent H2/N2 jet flames show good agreement with the theoretical prediction. By measuring all species simultaneously, important combustion quantities such as the mixture fraction are also derived.

A Compact Laser Absorption Spectroscopy Tomographic System With Short Spectral Scanning Time and Adjustable Frame Rate

A compact hardware system for laser absorption spectroscopy tomography is introduced to monitor dynamic flame evolution with a short and fixed spectral scanning time and an adjustable frame rate. The system mainly consists of two parts, a pentagon sensor and a reconfigurable laser control and data acquisition hardware. In the sensor, 12 photodetectors are installed in each of the five sides of the pentagon. Fan-shaped laser beams from each vertex illuminate the opposite side and half of its two adjacent sides. Compared with the traditional illuminating scheme from each vertex to only the opposite side, this design reduces the sensor size and improves the spatial resolution. In the hardware, the distributed feedback lasers are integrated with the controller and modulated to generate a programmed waveform and implement spectral scanning. The spectral scanning time is short and fixed, and an adjustable frame rate is achieved by changing the interval, i.e., the waiting time, between two successive scans. The sampling rate of each channel can reach up to 20 MHz and is synchronized to the laser emission. A DDR2 memory is used for temporary data storage for four channels, enabling a full frame rate of up to 2 kHz. In the experiment, water vapor in a flat flame burner was chosen as the target absorbing species. Local integral absorbance and temperature distributions in the case of a moving steel ruler across the burner were reconstructed and agreed well with the observed distributions.

Measurements of the laminar burning velocities of rich ethylene/air mixtures

Laminar burning velocities of premixed ethylene/air flames were investigated under fuel lean and rich conditions. The laminar burning velocities were measured with the heat flux method at atmospheric pressure and unburnt gas temperatures of 298 K. The measurements have been performed for the equivalence ratio range of Φ = 0.7–2.5 using stabilized and flat flames on a perforated burner plate under adiabatic conditions. This is the first time that experimental measurements with the heat flux method of the ethylene/air flames under super fuel rich conditions are performed. The experimental data were compared against predictions using three different kinetic models and published flame speed. The measured flame speeds agree with other published data within the error margin. The experimental and predicted laminar flames do agree at fuel lean conditions, but there are some notable discrepancies under fuel rich conditions.

A New Analytic pdf for Simulations of Premixed Turbulent Combustion

A new reaction rate source term omega _m(c) for modelling of premixed combustion with a single progress variable c is proposed. omega _m(c) mimics closely the Arrhenius source term omega _A(c) for a large range of activation energies and density ratios while admitting analytic evaluation of many quantities of interest. The analytic flame profile c_m(xi ) very closely approximates the numerically integrated Arrhenius flame profiles c_A(xi ). An important feature of c_m(xi ) is that it is analytically invertible into a xi _m(c). Analytic estimates of the laminar flame Eigenvalue Lambda and of the Le dependence of the laminar flame speed s_L are derived, which are more accurate than classic results based on asymptotic analyses. The flamelet pdf p(c)=1/(Delta *c*(1-c^m)) for a flat laminar flame front in a LES cell of width Delta is derived. The exact mean of the reaction rate overline{omega (c)} is compared to a beta pdf result, which is shown to be inaccurate for large ratios of filter width to flame thickness Delta /delta _f and particularly for high activation energy flames. For multidimensional flame wrinkling we derive the exact relationship p(c)=p_{1D}(c)I(c)Xi (c) between the 3D pdf p(c), the 1D flat flame pdf p_{1D}(c), a correction factor I(c) for change of inner flame structure and a geometrical wrinkling factor Xi (c). We show that the c dependence of these quantities cannot be neglected for small Delta /delta _f. A simple model of a sinusoidally wrinkled flame front qualitatively demonstrates the effect of flame wrinkling on p(c). We show that for large Delta /delta _f, a wrinkling of the reaction zone almost constantly increases p(c) in the reaction zone by a wrinkling factor Xi ^* (defined for the surface of the isosurface of maximum heat release) while reducing it near c=0,1 as required for normalisation of p(c). The 1D p(c) evaluated using a reduced filter width Delta '=Delta /Xi ^* may be a good approximation of the wrinkled flame pdf for evaluation of overline{omega (c)} for such cases.

Cellular instabilities of n-butane/air flat flames probing by PLIF-OH and PLIF-CH2O laser diagnosis

The structure and instability characteristics of non-adiabatic fuel-rich n-butane/air cellular flames on McKenna burner were experimentally investigated at atmospheric pressure. Planar Laser Induced Fluorescence (PLIF-OH and PLIF-CH2O) flame diagnosis technology was utilized to probe the flame structure under varied equivalence ratio and inflow mixture velocity respectively. Results show that, the equivalence ratio plays an important role in forming cellular flames. The flat, wrinkled and cellular flames appear in turn when increasing equivalence ratio and inflow velocity. Differed to the separated cells of cellular flames appeared in direct Digital camera images and PLIF-OH images, the PLIF-CH2O images clearly display that all cellular flames are connected. In the PLIF-OH images of cellular flames, the OH radical fluorescence signal intensity is higher in the convex region toward unburned mixture than in the concave region, but in PLIF-CH2O images, the fluorescence signal intensity is nearly unchanged in both convex and concave regions. Quantitative stand-off distance and amplitude data probed by PLIF-OH and PLIF-CH2O diagnosis together reveal that the onset of non-adiabatic n-butane/air cellular flames on flat flame burner is dominantly governed by diffusive-thermal mechanism.

Range of “complete” instability of flat flames propagating downward in the acoustic field in combustion tube: Lewis number effect

Downward propagating flames ignited at the open end of an open-closed tube exhibit thermo-acoustic instability due to interaction of combustion generated acoustic fluctuations with the flame front. At sufficiently high laminar burning velocity (SL) two regimes of thermo-acoustic instability are observed, namely, primary instability (where initial cellular flame transitions to a vibrating flat flame) and a secondary instability (where vibrating flat flame transitions to vibrating turbulent flame due to parametric instability of flame front). On further increasing SL to a particular value, “complete instability” of flat flames is observed meaning flat flame cannot be stabilized and initial cellular flame transitions directly to parametric instability. This particular SL introduced in this work is termed “critical SL”. In past experimental works, stability of flat flames in the acoustic field had only been studied in terms of acoustic velocity amplitude and a critical acoustic velocity amplitude had been measured at the onset of parametric instability. The novelty of this work is that boundary of unconditional instability of flat flame (flat flame is unstable irrespective of acoustic velocity amplitude) is determined in terms of mixture conditions, e.g., SL. Particularly for propagating flames, this critical SL can be measured more easily and accurately than the critical acoustic velocity. This work presents the effect of Le (Lewis number) on critical SL. Three different fuels, CH4, C2H4 and C3H8 are tested with two different dilution gases (N2 and CO2) for equivalence ratio of 0.8 (lean) and 1.2 (rich). Twelve different Le ranging from 0.7 to 1.9 are generated through these mixture combinations. Generally, larger Le mixtures show higher critical SL than lower Le mixtures for any fuel. Theoretical calculations are performed to predict critical SL by studying instability of planar flame fronts in presence of acoustic forcing. Theoretical calculations successfully captured the effect of Le as predicted stability region of planar flame is narrower for lower Le than that for higher Le. However, accurate quantitative predictions of critical SL couldn't be obtained from existing theory, particularly for non-unity Le. Hence, a correction (a function of Zeldovich number, β and Le) to width of stability region is proposed to obtain better quantitative agreement for critical SL between experiments and theory and performs significantly well. The correction factor acts to compensate for the inaccuracies in Markstein number obtained from an analytical relationship during calculation of stability region width.

Chemiluminescence-based characterization of tail biogas combustion stability under syngas and oxygen-enriched conditions

After the biogas is upgraded to CH4, a waste stream (tail biogas) is obtained, which contains methane in the order of few to tens vol.%. However, the composition of tail biogases can be unstable, therefore, safety should be maintained during the onsite combustion processes. Accordingly, a process control system with various sensors needs to be introduced. This study aims to analyze the experimental limits of flame stability using the chemiluminescence technique during the premixed combustion of low calorific tail biogas (mixture of CH4 with 15–30 vol% concentration in CO2) by adding oxygen or syngas. For the experimental investigation, a flat flame burner mounted in an enclosed combustion chamber was used. The stability of the flame was analyzed via the lift-off height by capturing the flame chemiluminescence from the excited OH* (282.9 nm), CH* (387.1 nm), and C2* (514.0 nm) radicals.This study reveals that the chemiluminescence spatial intensity of species such as OH*, CH*, and C2* is enhanced with an increase in the addition of syngas and oxygen. The peak position of the emission intensities in the flame shifts depending on the composition of gases, showing the existence of unstable flame that can reach its blow-off limit. When syngas is added, an excess of air leads to an increase in flame lifting; however, an O2 surplus results in more stable flame with minimized lifting. Moreover, it was determined that the lifting height of the flame is higher when syngas is added as compared to O2 enrichment. Finally, the flue gas emissions, such as CO and NOx, were examined and the most influential formation parameter was determined.

Experimental and modeling study of the low to high temperature oxidation of the linear pentanone isomers: 2-pentanone and 3-pentanone

Short-carbon-chain ketones are known for their high octane numbers, their knock resistance, and their low soot emission. However, studies on the combustion behavior of ketones, particularly in the low temperature regime, are sparse, and small ketones as acetone and butanone show limited low temperature chemistry (LTC). Therefore, saturated linear five-carbon ketones are good candidates to better understand the LTC of ketones and the effect of the functional carbonyl group. In this study, the ignition delay times of stoichiometric non-diluted 2- and 3-pentanone-oxygen-inert gas mixtures were measured in a rapid compression machine at pressures between 20 and 40 bar and temperatures ranging between 650 and 950 K. Furthermore, the mole fraction profiles of fuel, oxygen, intermediate species, and products were measured in a laminar flow reactor coupled with molecular-beam mass spectrometry at 0.97 bar, an equivalence ratio of 0.8 and temperatures ranging from 800 to 1050 K, to achieve a better understanding of the oxidation process of linear pentanones. A consistent detailed kinetic model was developed and validated against the newly measured experimental data for 2-pentanone and 3-pentanone to provide insight into the oxidation mechanism of pentanones. The model shows also good agreement with available data from the literature, including ignition delay times measured in a shock tube, speciation in a flat flame burner and laminar burning velocities.