Christopher R. Shaddix

EFFECT OF SYNGAS COMPOSITION AND CO2-DILUTED OXYGEN ON PERFORMANCE OF A PREMIXED SWIRL-STABILIZED COMBUSTOR

Future energy systems based on gasification of coal or biomass for co-production of electrical power and fuels may require gas turbine operation on unusual gaseous fuel mixtures. In addition, global climate change concerns may dictate the generation of a CO2 product stream for end-use or sequestration, with potential impacts on the oxidizer used in the gas turbine. In this study the operation at atmospheric pressure of a small, optically accessible swirl-stabilized premixed combustor, burning fuels ranging from pure methane to conventional and H2-rich and H2-lean syngas mixtures is investigated. Both air and CO2-diluted oxygen are used as oxidizers. CO and NOx emissions for these flames have been determined from the lean blowout limit to slightly rich conditions (ϕ ∼ 1.03). In practice, CO2-diluted oxygen systems will likely be operated close to stoichiometric conditions to minimize oxygen consumption while achieving acceptable NOx performance. The presence of hydrogen in the syngas fuel mixtures results in more compact, higher temperature flames, resulting in increased flame stability and higher NOx emissions. Consistent with previous experience, the stoichiometry of lean blowout decreases with increasing H2 content in the syngas. Similarly, the lean stoichiometry at which CO emissions become significant decreases with increasing H2 content. For the mixtures investigated, CO emissions near the stoichiometric point do not become significant until ϕ > 0.95. At this stoichiometric limit, CO emissions rise more rapidly for combustion in O2–CO2 mixtures than for combustion in air.

Time-Resolved Laser-Induced Incandescence and Laser Elastic Scattering Measurements in a Propane Diffusion Flame

Laser-induced incandescence (LII) and laser elastic-scattering measurements have been obtained with subnanosecond time resolution from a propane diffusion flame. Results show that the peak and time-integrated values of the LII signal increase with increasing laser fluence to maxima at the time of the onset of significant vaporization, beyond which they both decrease rapidly with further increases in fluence. This latter behavior for the time-integrated value is known to be characteristic for a laser beam with a rectangular spatial profile and is attributed to soot mass loss from vaporization. However, there is no apparent explanation for the corresponding large decrease in the peak value. Analysis shows that the peak value occurs at the time in the laser pulse when the time-integrated fluence reaches approximately 0.2 J/cm(2) and that the magnitude of the peak value is strongly dependent on the rate of energy deposition. One possible explanation for this behavior is that, at high laser fluences, a cascade ionization phenomenon leads to the formation of an absorptive plasma that strongly perturbs the LII process.

Idealized gas turbine combustor for performance research and validation of large eddy simulations

This paper details the design of a premixed, swirl-stabilized combustor that was designed and built for the express purpose of obtaining validation-quality data for the development of large eddy simulations (LES) of gas turbine combustors. The combustor features nonambiguous boundary conditions, a geometrically simple design that retains the essential fluid dynamics and thermochemical processes that occur in actual gas turbine combustors, and unrestrictive access for laser and optical diagnostic measurements. After discussing the design detail, a preliminary investigation of the performance and operating envelope of the combustor is presented. With the combustor operating on premixed methane/air, both the equivalence ratio and the inlet velocity were systematically varied and the flame structure was recorded via digital photography. Interesting lean flame blowout and resonance characteristics were observed. In addition, the combustor exhibited a large region of stable, acoustically clean combustion that is suitable for preliminary validation of LES models.

Simultaneous correction of flat field and nonlinearity response of intensified charge-coupled devices

Intensified charge-coupled devices (ICCDs) are used extensively in many scientific and engineering environments to image weak or temporally short optical events. Care has to be taken in interpreting the images from ICCDs if quantitative results are required. In particular, nonuniform gain (flat field) and nonlinear response effects must be properly accounted for. Traditional flat-field corrections can only be applied in the linear regime of the ICCD camera, which limits the usable dynamic range. This paper reports a more general approach to image correction whereby the nonlinear gain response of each pixel of the ICCD is characterized over the full dynamic range of the camera. Image data can then be corrected for the combined effects of nonuniform gain and nonlinearity. The results from a two-color pyrometry measurement of soot field temperature are used to illustrate the capabilities of the new correction approach.

Soot property measurements in a two-meter diameter JP-8 pool fire

ABSTRACT An in situ adsorption/emission diagnostic was used to measure soot properties in 2 m diameter JP-8 pool fires. Twelve tests were performed at the Lurance Canyon Burn Site operated by Sandia in Albuquerque, New Mexico. Seven of the tests were conducted with the probe positioned close to the centerline at heights above the pool surface ranging from 0.5 m to 2.0 m in 0.25 m increments. For the remaining five tests, the probe was positioned at two heights 0.3 m from the centerline and at three heights 0.5 m from the centerline. Soot concentration was determined using a soot absorption measurement based on the transmission of a solid-state red laser (635 nm) through the 3.7 cm long probe volume. Soot temperature and a second estimate of soot concentration were measured using two-color optical pyrometry at 850 nm and 1000 nm. The effective data rate for these measurements was 10 kHz. The results presented include the statistics, probability density functions, and spectral density functions of soot concentration and soot temperatures at the different measurement locations throughout the fire.

Design of “model-friendly” turbulent non-premixed jet burners for C2+ hydrocarbon fuels

Experimental measurements in laboratory-scale turbulent burners with well-controlled boundary and flow configurations can provide valuable data for validating models of turbulence-chemistry interactions applicable to the design and analysis of practical combustors. This paper reports on the design of two canonical nonpremixed turbulent jet burners for use with undiluted gaseous and liquid hydrocarbon fuels, respectively. Previous burners of this type have only been developed for fuels composed of H(2), CO, and/or methane, often with substantial dilution. While both new burners are composed of concentric tubes with annular pilot flames, the liquid-fuel burner has an additional fuel vaporization step and an electrically heated fuel vapor delivery system. The performance of these burners is demonstrated by interrogating four ethylene flames and one flame fueled by a simple JP-8 surrogate. Through visual observation, it is found that the visible flame lengths show good agreement with standard empirical correlations. Rayleigh line imaging demonstrates that the pilot flame provides a spatially homogeneous flow of hot products along the edge of the fuel jet. Planar imaging of OH laser-induced fluorescence reveals a lack of local flame extinction in the high-strain near-burner region for fuel jet Reynolds numbers (Re) less than 20,000, and increasingly common extinction events for higher jet velocities. Planar imaging of soot laser-induced incandescence shows that the soot layers in these flames are relatively thin and are entrained into vortical flow structures in fuel-rich regions inside of the flame sheet.

CONTAMINATION OF CARBON MONOXIDE WITH METAL CARBONYLS: IMPLICATIONS FOR COMBUSTION RESEARCH

Abstract In a study on syngas combustion in a swirl-stabilized combustor, red deposits quickly formed on the quartz combustor walls, preventing laser diagnostic measurements. Analysis of the deposit showed that is was composed of iron and nickel. A literature review revealed that typical pressurized cylinders of CO, even at the highest CO purity level, contain up to 100 ppmv of iron pentacarbonyl and somewhat smaller levels of nickel tetracarbonyl. Only semiconductor grade CO stored in aluminum cylinders has sub-ppm levels of carbonyls. The demonstrated strong flame inhibition effect of carbonyls suggests that the use of CO from steel cylinders in combustion experiments may affect studies of flame ignition or extinction, in addition to the effect of particle formation and deposition observed here.

Effects of char content and simple additives on biomass pyrolysis oil droplet combustion

Over the past few years, a laminar-flow, isolated-droplet combustion facility has been used to evaluate the fundamental combustion properties of biomass pyrolysis oils. This work has focused on characterizing oils produced from various feedstocks with a small-scale, fast-ablative vortex reactor coupled with various hot-gas filtration techniques. Recently, combustion of an oak oil produced in a large-scale, entrained fluidized bed has been characterized. This oil exhibits microexplosions very early in the combustion process, presumably as a consequence of the significant loading of small char particles in the oil. Backlit stroboscopic imaging of the burning droplets reveals that the oak oil microexplosions are characterized by the formation of a cellular network of vapor bubbles that ultimately contracts and results in the formation of coke particles. In addition, combustion experiments have been performed on a low-char poplar oil with small amounts of water and methanol additives. These simple additives have been suggested as an inexpensive means of reducing the viscosity and rate of chemical aging of biomass oils. Stroboscopic imaging of the droplets of these mixtures reveals a dramatic effect of both additives on the microexplosion process. However, long-exposure photographs of the complete combustion history suggest that the addition of methanol does not improve the shattering effect of the microexplosion and hence the total droplet burnout time. The addition of water delays the occurrence of microexplosion but improves its effectiveness, resulting in a small reduction in the droplet burnout time. Addition of both methanol and water yields the shortest burnout time.

Mineral-char interactions during char combustion of a high-volatile coal

We report on recent investigations of the role of inorganic mineral matter on the evolution of char structure during carbon burnout. Char samples collected in a carefully controlled, laminar flame-supported entrained flow reactor have been characterized using a number of microscopy tools. Observations of the inorganic structure of chars produced at a variety of combustion conditions are coupled with in situ particlesizing pyrometry measurements of the char particle population with an eye toward identifying the mechanism of mineral interaction and its effects on carbon burnout kinetics during pulverized coal char combustion. No evidence of a macroscopic ash film which has been hypothesized to retard char oxidation kinetics, was found on the chars. High-resolution electron microscopy, however, shows a surprising amount of inorganic mineral in solid solution within the carbonaceous matrix. This intimate mixing of organic and inorganic constituents may affect reactivity by both blocking oxygen access to active carbon sites and influencing the microscopic carbon structure that evolves during combustion.

Laser-induced breakdown spectroscopy of alkali metals in high-temperature gas

Laser-induced breakdown spectroscopy (LIBS) measurements of alkali in the high-temperature exhaust of a glass furnace show an attenuation of the Na and K LIBS signals that correlates with the stoichiometry of the bath gas surrounding the spark. The results are explained as being due to (1) a strong increase in the concentration of atomic Na and K, resulting in neutral line signal absorption by these atoms, and to (2) a change of phase of the major Na- and K-containing species from an aerosol to a gaseous phase when the gas mixture becomes fuel rich, resulting in a reduced LIBS emission intensity. LIBS sampling at lower temperatures, or in a consistently oxidizing environment, or both are suggested strategies for circumventing these difficulties.