Secondary Flame Research Articles

Turbulent flame dynamics of homogeneous solid propellant in a rocket motor

A comprehensive numerical analysis has been conducted to study the combustion of a double-base homogeneous propellant in a rocket motor. Emphasis was placed on the motor internal flow development and its influence on propellant combustion. The formulation is based on the Favre-averaged, filtered equations for the conservation laws and takes into account finite-rate chemical kinetics and variable thermophysical properties. Turbulence closure is obtained using the large-eddy-simulation technique. The contribution of large energy-carrying structures to mass, momentum, and energy transfer is computed explicitly, and the effect of small scales of turbulence is modeled. The governing equations and associated boundary conditions are solved using a time-accurate, semi-implicit Runge-Kutta scheme coupled with a fourth-order central difference algorithm for spatial discretization. The motor internal flowfield is basically determined by the balance between the inertia force and the pressure gradient arising from the mass injection at the propellant surface. The temporal evolution of the vorticity field shows a laminar upstream region, a transition zone in the midsection of the chamber, and a fully developed turbulent regime further downstream. The turbulent mixing proceeds at a rate faster than chemical reactions, and the flame stretch is strong enough to regard propellant combustion as a well-stirred reactor. The combustion wave in the laminar region exhibits a two-stage structure consisting of a primary flame, a dark zone, and a secondary luminous flame. The enhanced energy and mass transport in the turbulent region partially merges the primary and secondary flames, thereby raising the temperatures in the dark zone. In the present study, the smoother axial velocity gradient and vertical flow convection prevent turbulence from deeply penetrating into the primary flame zone. The turbulence energy spectra indicate dominant harmonics in a frequency range capable of triggering combustion instabilities.

Flame structure and burning rate of ammonium perchlorate/hydroxyl-terminated polybutadiene propellant sandwiches

The gas-phase flame structure that forms over a lamina of hydroxyl-terminated polybutadiene (HTPB) binder (50 to 450 μm) sandwiched between layers of ammonium perchlorate (AP) has been investigated through measurements of the surface regression rate of the interface region, ultraviolet emission and transmission imaging of the gas-phase flame zone, and numerical simulations using simplified kinetics. The results show that over the observed pressure range (0.2 to 3.2 MPa), for binder widths from 100 to 450 μm, the regression rate is primarily a function of pressure ( r b ≈p 0.4 ) and is relatively independent of binder width. At binder widths below 100 μm, the regression rate is reduced, presumably due to locally lean conditions created by diffusive mixing. The imaging results suggest the existence of relatively UV-inactive leading-edge flames that form on the AP side of the interfaces adjacent to the surface and probably dominate the regression rate. These leading-edge flames are followed by a relatively UV-active secondary flame region downstream. Both the leading-edge and secondary flames show a tendency to merge into a single flame for thin binder and at lower pressures (low Peclet and Damkohler numbers), and to split into two distinct flames for thick binder ( Pe> 2) and at higher pressures (high Da ). The modeling results show that simplified kinetics representing the AP monopropellant flame and the AP-binder non-premixed flame, in the framework of a species and energy transport analysis, can predict surface heat feedback distributions that correlate reasonably well with observed surface profiles.

An eigenvalue method for computing the burning rates of HMX propellants

A mathematical model for a three-tiered system consisting of solid, liquid, and gas is derived for studying the combustion of HMX propellants. The resulting nonlinear two-point boundary value problem is solved by Newton’s method with adaptive gridding techniques. In this study the burning rate is computed as an eigenvalue, which removes the uncertainty associated with employing evaporation and condensation rate laws in its evaluation. Results are presented for laser-assisted and self-deflagration of HMX monopropellants and are compared with experimental results. The burning rates are computed over a wide range of ambient pressures and compare well with experimental results from 1 to 90 atmospheres. The burning rate is found to be proportional to the pressure raised to the 0.82 power. Sensitivity of the burning rate to initial propellant temperature is calculated and found to be extremely low, in agreement with past theoretical predictions and experimental data. Results for laser-assisted combustion show a distinct primary and secondary flame separated by a dark zone, the length of which is dependent upon the incident laser flux intensity.

Numerical Modeling of NO Formation in Laminar Bunsen Flames—A Flamelet Approach

Based on the flamelet concept, a numerical model has been developed for fast predictions of NO X and CO emissions from laminar flames. The model is applied to studying NO formation in the secondary nonpremixed flame zone of fuel-rich methane Bunsen flames. By solving the steady-state flamelet equations with the detailed GRI2.1 methane–air mechanism, a flamelet library is generated containing thermochemical information for a range of scalar dissipation rates at the ambient pressure condition. Modeling of NO formation is made by solving its conservation equation with chemical source term evaluated based on flamelet library using the extended Zeldovich mechanism and “NO reburning reactions.” The optically-thin radiation heat transfer model is used to explore the potential effect of heat loss on thermal NO formation. The numerical scheme solves the two-dimensional Navier–Stokes equations as well as three additional equations: the mixture fraction, the NO mass fraction, and the enthalpy deficit due to radiative heat loss. With an established flamelet library, typical computing times are about 5 hours per calculation on a DEC-3000 300LX workstation. The predicted mixing field, radial temperature profiles, and NO distributions compare favorably with recent experimental data obtained by Nguyen et al. [2]. The dependence of NO X emission on equivalence ratio is studied numerically and the predictions are found to agree reasonably well with the measurements by Muss [3]. The computed results show a decreasing trend of NO X emission with the equivalence ratio but an increasing trend in the CO emission index. By examining this trade-off between NO X and CO, an optimal equivalence ratio of 1.4 is found to yield the lowest combined emission.

Effect of potassium hydroxide on PAH formation during toluene incineration

A laboratory-scale liquid-injection incinerator was used to incinerate toluene at 750°C and PAH emission was analyzed. The results indicate that the emission of solid-phase PAHs considerably increased when potassium hydroxide was added into a fuel-lean toluene flame; more potassium hydroxide resulted in more solid-phase PAHs. But this phenomenon was not observed for the fuel-rich toluene flame. The enhanced PAH formation is proposed as by-product of the partial oxidation or incomplete burning of soot in the secondary flame zone of the fuel-lean flame. It is suggested that reducing soot formation via potassium hydroxide addition may increase PAHs.

An Eigenvalue Method for Computing the Burning Rates of RDX Propellants

A mathematical model for a three-tiered system consisting of solid, liquid and gas is derived for studying the combustion of RDX propellants. The resulting nonlinear two-point boundary value problem is solved by Newton's method with adaptive gridding techniques. In this study the burning rate is computed as an eigenvalue, which can remove the uncertainty associated with employing evaporation and condensation rate laws in its evaluation. Results are presented for laser-assisted and self-deflagration of RDX monopropellants and are compared with experimental results. The burning rates are computed over a wide range of ambient pressures and compare well with experimental results from one to ninety atmospheres. The burning rate is found to be proportional to the pressure raised to the 0.76 power. Sensitivity of the burning rate to initial propellant temperture is calculated and found to be extremely low, in agreement with past theoretical predictions and experimental data. Results for laser-assisted combustion show a distinct primary and secondary flame separated by a dark zone, the length of which is dependent upon the incident laser flux intensity.

Nitramine deflagration - Reduced chemical mechanism for primary flame facilitating simplified asymptotic analysis

It is proposed that nitramines burn by staged combustion in which most of the energy is released in a secondary flame that is convective-re active in character and that has a negligible influence on the deflagration velocity. A nearly adiabatic condition is then achieved at the end of the primary gas-phase flame that precedes this secondary flame, the flame temperature 7} at the adiabatic point lying between about 1200-1500 K. The detailed gas-phase chemical kinetics for this primary flame are addressed here. The chemistry is reasoned to involve initially a set of elementary steps that result in HONO, HCN, H2CO, and N2O as decomposition products obtained hi an overall endothermic manner at temperatures in the vicinity of 1000 K, followed by steps that exothermically produce NO, N2, CO2, H2O, and HOCN at a rate found through systematic reduction of the chemistry to be that of bimolecular HONO decomposition. Systematic reduction of the chemistry of the initial process enables the ratio of HCN to H 2CO to be determined as a function of pressure, thereby predicting the pressure dependence of 7} and showing that this primary flame temperature decreases gradually with increasing pressure. Because of the activation energy of HONO decomposition, this decrease causes the effective pressure exponent of the exothermic reduced step to be less than two, and thereby reproduces the experimentally observed pressure exponent of about 0.8 for the deflagration velocity. Excellent agreement is obtained with measured deflagration velocities of cyclotrimethylenetrinitramine.

Solid propellant diffusion flame structure

Laser-supported deflagration of ADN, AP, HNF, TNA, RDX, or HMX in sandwich configurations with various energetic and nonenergetic binders was studied. Planar laser-induced fluorescence (PLIF) was used to measure the two-dimensional NH, OH, and CN species profiles for these sandwiches. For the HNF sandwiches, obvious diffusion flames were present at low pressure, but they were weak, further from the surface than the HNF monopropellant flame, and not expected to be burn-rate controlling. ADN was found to have diffusion flames with energetic binders at low pressures, but they were very far from the surface. The high burn rate of ADN at pressures above 3 atm left the binder behind, and the diffusion flames became weak. AP showed clear evidence for strong close diffusion flames with binders and even with other oxidizers. The combination of HMX or TNA with AP leads to obvious diffusion flame interactions: the highly reactive AP-decomposition products accelerated the nitramine dark zone kinetics and caused the secondary flame to hold on the interface region. This leads to higher heat feedback in the interface region and a somewhat faster nitramine regression rate. TNA, RDX, and HMX were not found to have any diffusion flames with energetic or nonenergetic binders: the monopropellant flame was lifted in the region around the binder. These results imply that ADN, HNF, TNA, RDX, and HMX propellants will not have the clear particle size ballistic tailorability seen with AP propellants, because diffusion flames appear to be strong and close to the surface (and therefore burn-rate controlling) only in AP propellants.

Laser recoil, plume emission, and flame height combustion response of HMX and RDX at atmospheric pressure

The laser recoil technique was used to measure the combustion response of neat HMX and RDX atatmospheric pressure in air. Heat flux was modulated approximately sinusoidally at frequencies from 2.5 to 1000 Hz and heat fluxes of 8–52 W/cm 2 . Thrust, light emission, and flame position were measured simultaneously so that amplitude and phase relationships could be determined as a function of frequency. Thrust was measured using a microforce transducer, light emission was measured by photo diode, and the flame height was photographed using a gated intensified charge coupled device (CCD) camera. The CCD camera recorded CN radical chemiluminescence, which is a measure of flame structure for the final (secondary) flame sheet. HMX had a broad flat thrust response, suggesting that the recoil technique needs further improvement.HMX had a light emission resonance at 10 Hz. RDX thrust and light emission oscillations were about half the amplitude of HMX. Monopropellant burn rates were 0.080 cm/s and 0.054 cm/s for HMX and RDX, respectively. HMX flame height was found to vary sinusoidally with phase angle, having a response peak between 8 and 50 Hz and crossing from phase lead to phase lag at about 20 Hz. The flame height response function of HMX was similar to light emission response. The resonant peak occurred at approximately the same frequency, the phase reacted the same as a function of frequency, and no unusual effects were seen at high frequency. Based on this data, the flame is reasonably quasi-steady below 40 Hz, in that the difference in phase lag between thrust and flame oscillations is less that 30° with almost no difference at the response peak of 10–20 Hz.

Diamond coating using multiple flame torches in an atmospheric chamber

Using an oxyacetylene flame, diamond coatings were applied to cutting inserts made of cemented carbides with a chamber which contained four torches. By equipping torches in the chamber and shutting off outside air, secondary and detrimental combustion flames could be eliminated and deposition of the diamond coating using only the primary combustion flame (acetylene feather) was possible even at normal pressures. Application to cutting inserts was attempted, leading to a success in coating four corners of a cutting insert with diamond in as short a time as 40 min. Evaluation of the cutting performance of the insert was carried out on an 18% Si−Al alloy, and the results showed that the performance was nearly equivalent to that of diamond-coated inserts prepared by the hot filament chemical vapor deposition process.

Evaluation method for the effect of salts on the elimination of secondary flames of energetic materials by differential scanning calorimetry

The effect of LiF, KC1 and KHCO 3 on the thermal decomposition of HTPB/RDX energetic composite (where HTPB is hydroxyl-terminated polybutadiene and RDX is cyclotrimethylene trinitramine) was studied and a method suitable for the evaluation of the inhibitory behavior of alkali metal salts on the secondary flame combustion reaction of HTPB/RDX composite was developed, utilizing thermogravimetry-derivative thermogravi- metry differential scanning calorimetry (DSC) and a multi-channel recorder. It was found that the addition of salts not only limited the extent, but also changed the mechanism, of the decomposition reaction of HTPB/RDX composite. The original one-stage exothermic reac- tion was found to be split into two, three or even more stages by the addition of salts. DSC results proved that the decomposition reaction species are different for the various stages. The mixture of LiF and KHCO 3 had a stronger effect as a retardant of the secondary flame than pure compounds such as LiF, KCl or KHCO 3 alone. The maximum peak temperature ( T m) decreased as KHCO 3 concentration increased, at a concentration of KHCO 3 lower than 35%. A relationship between the DSC data and secondary flame intensity was also found.

The structure of hydrogen cyanide-nitrogen dioxide premixed flames

The gas-phase chemistry of HCN/NO 2 flames has been studied both experimentally and theoretically. FTIR (Fourier Transform Infrared Spectroscopy) has been used to analyze the reaction zones of burner stabilized, premixed HCN/NO 2 flames having HCN to NO 2 stoichiometries of 1:1 and 1.4:1 (equivalence ratios of 1.25 and 1.74, respectively). The FTIR spectra have been used to determine both species profiles and temperature profiles as a function of axial distance above the burner. These adiabatic flames were modeled using Sandia's steady laminar one-dimensional premixed flame code. The detailed chemistry included 28 species and 130 reactions. The flame structure model was used to develop reaction mechanisms for the carbon-and nitrogen-containing species in the flames. The model results indicate that the reaction pathways are similar to those identified previously for the C 2 N 2 /NO 2 flames. These mechanisms are also similar to those presented previously for the secondary flame chemistry of the RDX model flame. The primary differences among these various flames occur due to the relative availability of hydrogen containing species.

An investigation of diamond film deposition in a premixed oxyacetylene flame

Diamond film has been deposited in a single-nozzle premixed oxyacetylene flame. Results of runs of varying duration suggest that diamond is deposited via the transport of hydrocarbon fragments produced at the secondary flame front. Planar laser-induced photodissociation fluorescence suggests that this region is rich in C2H species. Emission studies also suggest that the post-primary flame zone presents a source of C2 radicals that may account for the observed graphite and diamondlike carbon deposited on the substrate exposed to this region of the flame. The results on the premixed flame suggest that it would be possible and more convenient to attempt large area deposition using a multinozzle diffusion flame.

Flame Structure and Diamond Growth Mechanism of Acetylene Torch

To understand the diamond growth mechanism in C2H2/O2 flame, we carried out gas analyses for various mixture ratios (R=C2H2/O2) with laser-induced fluorescence and mass spectrometric techniques. These measurements show that CO and H2 are the dominant gases in the feather, and that C2H2 and C-containing radicals (e.g., C2H, CR, Cn, n=1-3) are minor species. Their concentrations are found to be consistent with the equilibrium values estimated by the adiabatic calculation. The feather is formed by the inter-diffusion and reactions between these C-radicals and O-radicals (e.g., O and OH), which are produced in the intermediate zone by the oxidation with O2 supplied through the secondary flame. It is also found that R-dependences of the diamond growth rate is in good agreement with those of CH and C2 concentrations in the feather.

A model for double base propellant combustion

A model has been developed to describe the combustion of double base propellants with a primary objective to develop a better understanding of the combustion mechanisms of double base propellants in both condensed and gas phases. The gas phase in double base propellants sustains two flames which are separated by a dark zone at low pressure. At high pressure, the secondary flame merges into the primary flame, eliminating the dark zone. The proposed model calculates the temperature profile in both the condensed phase and through both flames of the gas phase. the model solves the energy equations and species equations for both the condensed and gas phases separately, and then matches the boundary conditions between the two phases. The model assumes that there are three parallel reactions occurring in the condensed phase; a zero order degradation reaction, a first order reaction with NO 2 , and a second order reaction of complex aldehyde species. The gas phase kinetics are summarized with four reactions; a first order NO 2 reaction, a second order reaction of aldehydes, a first order NO-carbon reaction, and a second order NO reduction reaction. p In the past, several models have been developed to predict the burning rate of double base propellant as a function of pressure and initial temperature. Increasingly numerous experimental results (most from Kubota and Zenin) provide a well established data base which, along with increased computing capability, now allow one to verify the reliability of numerical models using a much broader variety of experimental data. A complete model must determine the burning rate as a function of pressure and initial temperature. In this study, the behavior of each reaction zone through the propellant combustion wave has been examined.

Mechanisms of formation and destruction of soot particles in a laminar methane-air diffusion flame

A laser light scattering/extinction method has been used to examine the evolution of soot particles in a cylindrical laminar diffusion flame of methane and air at atmospheric pressure. The soot particle characteristics (volume fraction, mean diameter, number density) were plotted versus time along the streamline trajectories, assuming a monodisperse distribution of particle-sizes. Different steps of soot evolution have been deduced: (1) The soot formation phase consists of two steps, identical to those described by several authors in rich premixed flames: nucleation, then coagulation or aggregation; surface growth, which occurs simultaneously, is responsible for the main part of the total mass of soot. (2) Three steps have been observed in the destruction phase: -In the first step, the mean diameter increases while the particles number decreses; this phenomenon may be considered as an aggregation process or as an effect of the change of particle size-distribution. -In the second step, the mean diameter decreases but the number increases. According to Neoh's study in premixed secondary flames, this phenomenon has been attributed to a “break-up” of the aggregates; it occurs when about 75 percent of the soot mass has disappeared. -The last step is characterized by the decreases of diameter and number of particles. In the diffusion flame we have considered, the soot destruction phase takes about five milliseconds. The specific burnout rate has a maximal value of 10−3 g cm−2 s−1 which agrees with the order of magnitude previously measured in secondary premixed flames.

Formation of acute pulmonary toxicants following thermal degradation of perfluorinated polymers: Evidence for a critical atmospheric reaction

In some small-scale tests of combustion-product toxicity, perfluorinated polymers (FP) have been shown to be much more toxic than other common materials. These studies were conducted to determine the conditions under which highly toxic decomposition products of FP are formed. A modification of the National Bureau of Standards exposure system (Levin et al. National Bureau of Standards 1982, NBS IR 82-2532) was used, in which materials could be heated either in a cup furnace or above a small methane flame. At 700°C, the approximate lethal concentrations (ALCs) of the products formed from polytetrafluoroethylene (pTFE) and from hexafluoropropylene/tetrafluoroethylene copolymer (FEP) in the CH 4 flame (applied for 1 min) are approximately 850 times higher than the ALCs of the products formed in the cup furnace. Analytically, the major products formed initially from pTFE at 700°C under either condition (flame or cup furnace) are similar but they disappear rapidly in the presence of continuous heat. When the cup furnace is removed 1 min after pTFE is added (a procedure temporally similar to the use of the flame) the toxicity of the products is again low. However, when heat from either the cup furnace or from a small secondary flame is applied continuously (for up to 30 min) to the initial products formed from pTFE in the flame, high toxicity is observed. These observations are supported by pathological evaluation of the respiratory tracts of exposed rats. Thus, the inhalation toxicity of FP thermal decomposition products is related to a heat- and time-dependent atmospheric reaction. Such mechanisms should be considered in estimating the fire hazard of these materials in use.

Soot formation in an opposed flow polymer diffusion flame

This study describes an investigation of the mechanisms of soot formation in an opposed flow diffusion flame on a polyethylene rod. A specially designed burner which maintained the flame in a steady position in the laboratory frame of reference was built. Temperature distributions were measured using thin thermocouples while combined laser light scattering and absorption measurements were used to determine distributions of soot volume fractions, soot aggregate diameters and their number densities. The gaseous flow field as well as that of the soot particles were determined from a model developed for this flame. These results predicted the existence of two separate stagnation planes for the flame gases and the soot particles. Similar to what has been established for jet-like gaseous diffusion flames, nucleation in this flame was also found to take place in the pyrolysis region near the flame front. After nucleation, the particle grow by coalescence, agglomeration and surface growth as they move towards the burning polymer surface, with thermophoretic velocity first in concert with the convective velocity and then against the convective velocity of the gases. A secondary flame front surrounding the sides of the sooty zone cause a circumferential annulus of heavy soot loadings. The local rates of soot formation were determined from the soot mass conservation equation using measured axial distributions of the temperature and the soot volume fraction and calculated gas and soot particle velocities.

New thermionic detector utilizing orthogonal nebulization for capillary column liquid chromatography

A new version of the previously studied phosphorus-sensitive, dual-flame thermionic detector has been developed for microcolumn liquid chromatography. The total column effluent is orthogonally nebulized and aspirated directly into a primary air—hydrogen diffusion flame. Phosphorus compounds with molecular weights extending beyond 500 a.m.u. are then selectively detected by measuring the conductivity of the secondary flame in the presence of a rubidium silicate bead. The sensitivity was found to be 1.36 · 10 −11 g phosphorus/sec at the maximum of a Gaussian peak, and the signal increased linearly with concentration over 2 orders of magnitude for dilute samples. Possible mechanisms accounting for the negative orientation of the signal and for the limited dynamic range are discussed.

Flame propagation through an air-fuel spray mixture with transient droplet vaporization

Flame initiation and propagation through an air/fuel vapor/fuel drop system is numerically modeled in a cylindrical one-dimensional closed combustor. An unsteady formulation of the flow problem eliminates the cold-boundary difficulty and gas-phase ignition problem. A velocity lag between the gas and the liquid phase is allowed and unsteady heat transfer to the droplets is taken into account. The surface temperature of the droplet is evaluated by using an unsteady spherically symmetric formulation of the droplet heat conduction problem with no internal motion and with a time-varying heat flux specified at the surface as a boundary condition. Results have been obtained for two commercially important fuels, namely, n-octane and n-decane. The activation energy and the preexponential factor in the Arrhenius-type expression for chemical rate, along with initial temperature, initial droplet size, stoichiometric ratio, and diffusivity are parametrically varied and flame speed and flame temperatures are observed. Flame speed is seen to increase with increasing preexponential factor, decreasing activation energy, increasing ambient temperature, decreasing initial droplet radii, and increasing diffusivities. It is also observed that unlike premixed combustion, heterogeneous combustion gives rise to local variations of equivalence in the axial direction. This phenomenon could give rise to a secondary diffusion flame in the wake of a propagating flame and produce local variations in the flame temperature.