微尺度预混合火焰结构和熄火特性研究

本文以空气中的无约束甲烷预混合火焰为对象,用实验和数值解析的方法研究了微尺度预混合火焰的火焰结构和熄火特性。实验测得不同尺寸下混合气当量比和喷出速度与熄火关系图,在不到理论当量比(φ>1)时,火焰已经熄灭,管径越小,极限混合气当量比φu越大。数值解析研究了d=0．3 mm无约束甲烷预混合火焰,在混合气当量比大于 1的富燃料燃烧条件下,空气中形成的预混合火焰结构是内层预混合火焰和外层扩散火焰,极限当量比约为1,解析结果再现了实验现象。     

喷管直径对微尺度扩散火焰特性的影响

对均匀空气流中微尺度甲烷扩散燃烧进行了数值模拟,重点考察微喷管内的流动和传热传质对微尺度燃烧特性的影响.研究结果表明,在保持燃料喷出速度一定的条件下,随着喷管直径的减小,喷管内与甲烷喷出速度相反方向上发生热量和质量的传递,燃料与空气的混合在喷管内已经发生,火焰的一部分热量回流到喷管内顶热了未燃混合气,同时也增加了火焰的热损失.当管径为0.15 mm时,甲烷在微喷管内就开始发生化学反应,在进行微尺度解析计算时,必须包含一定的喷管区域.     

H_2/N_2/O_2射流扩散火焰的直接数值模拟

直接数值模拟(DNS),大涡模拟(LES)与雷诺时均模拟(RANS)是数值模拟燃烧流动的三大主要方法,而射流扩散火焰在燃烧理论,实验研究与数值模拟中都扮演着十分重要的角色,本文采用完全可压缩的Naive-Stokes方程,对喷口直径为D=1 mm,Re=2000的射流扩散火焰进行了直接数值模拟.本文首先分析了冷态时H_2,O_2的混合,发现燃料与氧气在流向长度大于6D后的区域混合得十分充分.随后本文分析了燃烧后的统计特性,主要包括速度场,密度,温度以及主要组分与混合分数的分布,并将DNS结果与实验结果进行对比分析.     

微喷管甲烷非预混射流火焰燃烧特性实验研究

为发展微尺度燃烧器并拓展微尺度燃烧理论,对具有外部伴流空气的甲烷非预混微喷管射流火焰燃烧特性进行了实验研究。微喷管采用内径为710μm、425μm及280μm的不锈钢管,通过实验得到了微喷管非预混射流火焰的火焰形态、高度、最小熄灭流速及吹熄极限,并与常规尺度(管内径2 mm)非预混射流火焰进行了对比。研究表明微喷管射流火焰只有层流火焰一种形态;微喷管射流火焰高度主要取决于燃料流速而不受外部伴流速度影响;微喷管射流火焰的吹熄极限随伴流速度先增加后减小,而微射流火焰的最小熄灭流速受伴流空气速度影响较小,随管径减小微喷管射流火焰的可燃范围急剧减小。     

液体乙醇微尺度扩散火焰高度的影响因素研究

通过实验研究了微尺度扩散火焰高度的影响因素,结果表明:对液体乙醇微尺度扩散火焰,在稳定燃烧区域,其火焰高度与乙醇流量,燃烧器内径,燃料雷诺数,陶瓷管露出在空气中长度有关。另外,外加电场也对微尺度扩散火焰高度有影响。实验结果为掌握液体燃料微尺度扩散火焰高度的影响因素,提高火焰燃烧稳定性和燃烧效率提供了有益的参考。     

超音速氢气燃烧火焰结构特性的直接数值模拟

在课题组前期工作的基础上,对一马赫数为1.2的三维超音速氢气射流抬升火焰进行了直接数值模拟研究,其中空间离散采用波带优化的四阶WENO格式,时间积分采用带有TVD性质的三步三阶龙格库塔格式,边界条件采用了无反射特征边界条件,总的计算网格数达到9.75亿。结果表明:超音速射流氢气燃烧火焰可分为根部层流状的高温高热量释放率稳燃区、高度褶皱的湍流剧烈混合区和远场燃烧区。火焰自燃稳燃点出现在喷口附近的x/D=0.86处,对应着最易反应混合分数。在此下游,预混燃烧和扩散燃烧两种模式同时存在,其中在剧烈混合区和远场区火焰以扩散燃烧为主,但在火焰根部的局部区域预混燃烧热量释放率达到35%左右。     

甲烷/富氧扩散火焰燃烧区域的分层特性研究

本文对甲烷／富氧扩散火焰燃烧区域的分层特性进行了数值模拟和实验研究,结果表明氧化剂中氧浓度的增加加剧了火焰的分层现象(黄焰层与蓝焰层),使蓝色火焰变厚,并且使NOx生成大量增加;火焰面上的速度梯度主要影响黄色火焰厚度,蓝色火焰随着速度梯度的增加而减小, NOx生成也随之较少。对比温度及火焰结构还表明,研究中所采用的数值模拟方法可以正确地预测对向流扩散火焰特性。     

稀释剂对H2/CH4预混火焰熄灭特性影响

掺氢天然气在稀释气体作用下的熄灭特性研究对实际燃烧设备的设计和优化具有重要的指导意义。本文利用对冲火焰法测量了掺氢天然气层流火焰在N₂和CO₂作用下的熄灭拉伸率,并采用数值模拟耦合详细化学反应机理对N₂,CO₂和He的稀释剂效应展开研究。结果表明,Li、GRI Mech 3.0和FFCM-1机理均能定性反映燃料熄灭拉伸率随当量比的变化规律,且FFCM-1机理综合预测精度最高。实验和模拟发现,不同稀释剂气体对掺氢天然气熄灭拉伸率降低幅度满足:He22。进一步研究发现,CO₂由于热容大,在反应体系中会降低火焰温度,同时增强了链终止反应强度,通过热效应和化学效应两方面对火焰熄灭特性起作用。He则能显著改变燃料混合物的平均摩尔质量,从而改变体系中重要反应物和自由基的扩散特性,从扩散效应方面影响火焰的熄灭特性。     

甲烷微尺度催化燃烧的数值模拟

本文联合使用计算流体力学软件FLUENT和可以计算表面反应的化学反应动力学软件DETCHEM对有逆流换热的微尺度燃烧器进行了数值计算。计算中忽略空间反应。燃料-空气混合物的当量比为0.4,反应器壁面采用等温边界条件。计算结果表明,采用催化燃烧可以实现微尺度下通常情况下无法实现的甲烷稳定燃烧。通过适当设置催化表面,可以实现燃料低温、高效转变。甲烷的总转变率受流动状态、反应温度和催化表面的大小等因素的影响。     

湍流扩散火焰的非稳态火焰面模拟

采用稳态的和非稳态的火焰面模型同时对一个湍流甲烷射流扩散火焰进行了数值模拟,比较了两者对湍流平均火焰结构、活性自由基和污染物(氮氧化物)排放的模拟效果。速度场采用κ-ε模型计算,守恒标量混合物分数的分布通过其概率密度函数(PDF)输运方程的求解得到。稳态的火焰面结构由查询火焰面数据库得到,而非稳态的火焰面结构由火焰面方程和流场方程耦合求解来计算。采用详细的GRI—Mech 3.0机理描述甲烷的氧化和氮氧化物的形成。数值模拟结果和实验数据作了广泛的对比,验证了火焰面模型对湍流扩散燃烧的定量模拟能力。     

A numerical study on extinction behaviour of laminar micro-diffusion flames

We conducted a numerical study on the fluid dynamic, thermal and chemical structures of laminar methane–air micro flames established under quiescent atmospheric conditions. The micro flame is defined as a flame on the order of one millimetre or less established at the exit of a vertically-aligned straight tube. The numerical model consists of convective–diffusive heat and mass transport with a one-step, irreversible, exothermic reaction with selected kinetics constants validated for near-extinction analyses. Calculations conducted under the burner rim temperature 300 K and the adiabatic burner wall showed that there is the minimum burner diameter for the micro flame to exist. The Damköhler number (the ratio of the diffusive transport time to the chemical time) was used to explain why a flame with a height of less than a few hundred microns is not able to exist under the adiabatic burner wall condition. We also conducted scaling analysis to explain the difference in extinction characteristics caused by different burner wall conditions. This study also discussed the difference in governing mechanisms between micro flames and microgravity flames, both of which exhibit similar spherical flame shape.     

Experimental low-stretch gaseous diffusion flames in buoyancy-induced flowfields

The present study experimentally investigates the structure and instabilities associated with extremely low-stretch (1 s⁻¹) gaseous diffusion flames. Ultra-low-stretch flames are established in normal gravity by bottom burning of a methane/nitrogen mixture discharged from a porous spherically symmetric burner of large radius of curvature. OH-PLIF and IR imaging techniques are used to characterize the reaction zone and the burner surface temperature, respectively. A flame stability diagram mapping the response of the ultra-low-stretch diffusion flame to varying fuel injection rate and nitrogen dilution is explored. In this diagram, two main boundaries are identified. These boundaries separate the stability diagram into three regions: sooting flame, non-sooting flame, and extinction. Two distinct extinction mechanisms are noted. For low fuel injection rates, flame extinction is caused by heat loss to the burner surface. For relatively high injection rates, at which the heat loss to burner surface is negligible, flame radiative heat loss is the dominant extinction mechanism. There also exists a critical inert dilution level beyond which the flame cannot be sustained. The existence of multi-dimensional flame phenomena near the extinction limits is also identified. Various multi-dimensional flame patterns are observed, and their evolutions are studied using direct chemiluminescence and OH-PLIF imaging. The results demonstrate the usefulness of the present burner configuration for the study of low-stretch gaseous diffusion flames.     

On the global structure and asymptotic stability of low-stretch diffusion flame: Forced convection

The present work analyzes cylindrical diffusion flames (Tsuji burner) under low stretch condition, considering fuel injection also from the backward region of the burner. To highlight the fundamental aspects of this flame, some assumptions are imposed, like constant thermodynamic and transport coefficients, unitary Lewis number and no radiative heat loss. It is also considered potential flow model and incompressible Navier–Stokes model. Despite the simplicity of the former model, results from both models show good agreement. Also, an asymptotic analysis describing the problem far from the burner is able to capture the most important mechanisms controlling the flame, then the flame shape is determined and the dependence of the characteristic length scales on Peclet number (based on the burner properties), free stream velocity and stoichiometry is revealed. The results show that the flame width is proportional to the mass stoichiometric coefficient and reciprocal to the Peclet number the 1/4 power and free stream velocity the 3/4 power, and that the flame height is proportional to the square of the mass stoichiometric coefficient and to the square root of the ratio of Peclet number to free stream velocity. In addition, an asymptotic stability analysis reveals low-stretch flame extinction to be caused by reduction in fuel and oxidizer concentrations, which provides the range of the stoichiometric coefficient for stable regime, and at the same time the range of heat released.     

An experimental realization of an unstrained, planar diffusion flame

A unique burner was constructed to experimentally realize a one-dimensional unstrained planar non-premixed flame, previously considered only in idealized theoretical models. One reactant, the fuel mixture in the current experiments, is supplied through a porous plug at the bottom of the combustion chamber and flows vertically up towards the horizontal flame. The crux of the design is the introduction of the oxidizer from above in such a way that its diffusion against the upward product flow is essentially one-dimensional, i.e., uniform over the burner cross-section. This feature was implemented by introducing the oxidizer into the burner chamber from the top through an array of 625 closely spaced hypodermic needles, and allowing the hot products to escape vertically up through the space between the needles. Due to the injection of oxidizer through discrete tubes, a three-dimensional “injection layer” exists below the exit plane of the oxidizer supply tubes. Experimental evidence suggests that this layer is thin and that oxidizer is supplied to the flame by 1-D counterdiffusion, producing a nearly unstrained flame. To characterize the burner, flame position measurements were conducted for different compositions and flowrates of H₂–CO₂ and O₂–CO₂ mixtures. The measured flame locations are compared to an idealized one-dimensional model in which only diffusion of oxidizer against the product flow is considered. The potential of the new burner is demonstrated by a study of cellular structures forming near the extinction limit. Consistent with previous investigations, cellular instabilities are shown to become more prevalent as the initial mixture strength and/or the Damköhler number are decreased. As the extinction limit is approached, the number of cells was observed to decrease progressively.     

Extinction characteristics of isolated n-alkane fuel droplets during low temperature cool flame burning in air

Large carbon number n-alkanes are a notable component in all real transportation fuels, and their chemical structure fosters substantial low temperature kinetic reactivity. Normal alkanes have been studied in various canonical configurations but rarely in systems with strong coupling between low temperature chemistry and transport for pure as well as for multi-component n-alkane mixtures. The Flame Extinguishment (FLEX) experiments onboard the International Space Station provided a unique platform for investigating low temperature multi-phase n-alkane and iso-alkane combustion. Among the many interesting phenomena experimentally observed, cool flame extinction can occur, accompanied by the concurrent formation of a surrounding cloud of condensed vapor. In this work we conduct numerical simulations of high and low temperature combustion of large, initially single-component n-heptane, n-decane and n-dodecane droplets. The role of initial droplet diameter, operating pressure, and n-alkyl carbon number on the extinction of hot and low temperature flames is investigated and compared against the available experimental data. While all three fuels exhibit similar hot flame behavior, cool flame activity increases with the carbon number, resulting in an increased cool flame temperature and decreased extinction diameter. Multi-cyclic “hot/cool flame transitions” are found in air as pressure is slightly increased above one atmosphere. The cyclic behaviors correspond to continuously varying hot and cool flame transitions across the high, low, and negative temperature coefficient (NTC) kinetic regimes. Further increase in pressure results in a second stage steady “Warm flame” transition. The extinction of hot and cool flame has a strong non-linear dependence on ambient pressure but as the hot flame extinction diameter increases with pressure the extinction diameter of the cool flame decreases. The computational results are compared with a recent asymptotic analysis of FLEX n-alkane cool flames.     

Kinetics and extinction of non-premixed cool and warm flames of dimethyl ether at elevated pressure

The growing demand of clean and efficient propulsion and energy systems has sparked an interest in understanding low-temperature combustion at high pressure. Cool flame transition and extinction limits as well as oxygen concentration dependence at elevated pressures provide insights of the low-temperature and high-pressure fuel reactivity. A new experimental high-pressure counterflow burner platform was designed and developed to achieve the studies of high-pressure cool flames. Dimethyl ether (DME) was chosen to study its non-premixed cool flame in high-pressure counterflow burner at pressure up to 5 atm, perhaps for the first time. This paper investigates the effects of pressure on cool flame structure, extinction and transition limits, and oxygen concentration dependence as well as ozone assisted warm flames of DME in experiments and numerical simulations. The results show that the reignition transition from cool flame to hot flame occurs either with the decrease of the strain rate at a given fuel concentration and pressure or with the increase of fuel mole fraction or pressure at a given strain rate. Furthermore, it is shown that the higher pressure shifts the cool flame to higher strain rates and results in higher cool flame extinction strain rates. However, the existing kinetic model of DME fails in predicting the cool flame extinction limit at elevated pressures. Besides, the cool flame extinction limits are proportional to nth power of the oxygen concentration, [O₂]ⁿ, and the increase of pressure leads to stronger extinction limit dependence (larger n) on oxygen concentration. The present experiment and detailed kinetic analysis show clearly that increasing pressure promotes the low-temperature chemistry including the oxygen addition reactions. In addition, stable warm flame was first experimentally observed by using DME at elevated pressure with ozone sensitization.     

Studies of autoignition-assisted nonpremixed cool flames

Autoignition-assisted nonpremixed cool flames of diethyl ether (DEE) are investigated in both laminar counterflow and turbulent jet flame configurations. First, the ignition and extinction limits of laminar nonpremixed cool flames of diluted DEE are measured and simulated using detailed kinetic models. The laminar flame measurements are used to validate the kinetic models and guide the turbulent flame measurements. The results show that, below a critical mixture condition, for elevated temperature and dilute mixtures, the cool flame extinction limit and the low-temperature ignition limit merge, leading to autoignition-assisted cool flame stabilization without hysteresis. Based on the findings from the laminar flame experiments, autoignition-assisted turbulent lifted cool flames are established using a Co-flow Axisymmetric Reactor-Assisted Turbulent (CARAT) burner. The lift-off heights of the turbulent cool flames are quantified using formaldehyde planar laser-induced fluorescence. Based on an analogy with autoignition-assisted lifted hot flames, a correlation is proposed such that the autoignition-assisted cool flame lift-off height scales with the product of the flow velocity and the square of the first-stage ignition delay time. Using this scaling, we demonstrate that the kinetic mechanism that most accurately predicts the laminar flame ignition and extinction limits also best predicts the turbulent cool flame lift-off height.     

Experimental investigation on burner lip thickness effect on the liftoff and blowout velocities of jet diffusion flame

This experimental study addresses issue on the effect of burner lip thickness on the liftoff and blowout velocities of jet diffusion flame discharging into quiescent air. Burner tubes of two kinds of inner diameter (2 and 3 mm) with a wide range of lip thicknesses (0.25–16.5 mm) are implemented with methane or propane jet diffusion flame, respectively. The results show that the burner lip thickness has a profound effect on flame liftoff velocities, especially the blowout velocities. With the increase of the lip thickness, the blowout velocities firstly increase, then decrease and lastly remain unchanged. Specifically, the blowout velocities of 2 and 3 mm inner diameters tubes reach the maximum values when the corresponding burner lip thicknesses are 2 and 1.5 mm, respectively. In addition, compared with free (unconfined) jet diffusion flame, the jet confinement results in slight reductions of flame liftoff and blowout velocities. The existence range of lifted flame issuing from larger diameter burner tube is wider than that of smaller diameter burner tube. The existence range of propane lifted flame is wider than that of methane.     

Characteristics of microjet methane diffusion flames

Characteristics of microjet methane diffusion flames stabilized on top of the vertically oriented, stainless-steel tubes with an inner diameter ranging from 186 to 778 μ m are investigated experimentally, theoretically and numerically. Of particular interest are the flame shape, flame length and quenching limit, as they may be related to the minimum size and power of the devices in which such flames would be used for future micro-power generation. Experimental measurements of the flame shape, flame length and quenching velocity are compared with theoretical predictions as well as detailed numerical simulations. Comparisons of the theoretical predictions with measured results show that only Roper's model can satisfactorily predict the flame height and quenching velocity of microjet methane flames. Detailed numerical simulations, using skeletal chemical kinetic mechanism, of the flames stabilized at the tip of d = 186, 324 and 529 μ m tubes are performed to investigate the flame structures and the effects of burner materials on the standoff distance near extinction limit. The computed flame shape and flame length for the d = 186 μm flame are in excellent agreement with experimental results. Numerical predictions of the flame structures strongly suggest that the flame burns in a diffusion mode near the extinction limit. The calculated OH mass fraction isopleths indicate that different tube materials have a minor effect on the standoff distance, but influence the quenching gap between the flame and the tube.     

Aspects of oscillating flames

Photography and chemieluminescence from CH radicals have been used to identify the reaction zones and quantify the areas and shapes of kerosene-fuelled flames with swirl numbers of 0.7 and 0.8 and an overall equivalence ratio of 0.25. The air flow was oscillated at a frequency of 350 Hz and the results suggest that the oscillations caused a sequence of vortex rings at the burner exit and that these distorted the reaction zone and increased its area in the near burner region leading to an overall shorter flame. For the swirl number of 0.7, the flame was lifted and the oscillations led to an increase in the average lift off length whereas the higher swirl number caused an attached flame with and without oscillations. The stretch rate, evaluated from the variation of the flame area in time, was higher for the lifted flame suggesting that lift off was caused by local extinction.