京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第22770号 / 工博第4769号 / 新制||工||1746(附属図書館) / 京都大学大学院工学研究科機械理工学専攻 / (主査)教授 黒瀬 良一, 教授 中部 主敬, 教授 岩井 裕 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

Afin de contribuer à l'amélioration des simulations numériques de foyers industriels avec recycle de gaz brûlés et combustible liquide, la modélisation de la combustion turbulente non prémélangée est abordée sous deux de ses aspects : le traitement des problèmes diphasiques et la prise en compte des configurations complexes de mélange (dilution, recirculation interne).Une flamme spray éthanol/oxygène diluée par du dioxyde de carbone est d'abord calculée en LES dans un formalisme Euler-Lagrange et une chimie détaillée pour une résolution atteignant 250 µm. Les conditions d'injection du brouillard de gouttes sont déterminées à partir de mesures expérimentales de granulométrie et d'anémométrie phase Doppler.Dans une seconde partie, un nouveau formalisme de description de la combustion turbulente, basé sur l'introduction de temps caractéristiques de l'histoire du mélange est développé, puis validé en LES sur le cas de référence d'une flamme jet dans un écoulement co-courant vicié. / As a contribution to the improvement of numerical simulations of industrial furnaces with flue gas recirculation and liquid fuel, two aspects of the non-premixed turbulent combustion modeling are addressed: the handling of two-phase problems and the treatment of complex mixing configuration, such as dilution and internal recirculation. An ethanol/oxygen spray flame is first solved in an LES, following an Euler-Lagrange formalism, with detailed chemistry and at a resolution reaching 250 µm. The spray injection conditions are determined from granulometry and Doppler-phase anemometry measurements. In a second part, a novel formalism to describe turbulent combustion is developed, based on the introduction of characteristic timescales of the mixing history. It is then validated on the reference case of a jet flame in a vitiated co-flow.

No description available.

No description available.

A fast response method for detection of light-off in gaseous flames and liquid spray flames has been developed. The method used chemiluminescent signals from the 2Σ - ²π OH system centered at 309 nm and the ²Δ - ²π CH system centered at 430 nm to indicate the presence of a flame. Spectral scans (performed on gaseous methane, liquid hexane and liquid Jet-A aircraft fuel) from 280 nm to 610 nm indicated that these two species produced the strongest signals available for flame detection. As their light is emitted in the ultraviolet spectrum, using the OH and CH radicals will potentially provide a good signal-to-noise ratio since, in combustion chambers, most of the broadband background emissions are in the infrared and visible wavelengths. These scans also showed that the hexane and Jet-A gave OH and CH signals of approximately equal intensity. The transient histories of OH and CH were investigated by performing light-off ignition tests and intermittent light-off ignition tests. These various flame conditions showed that both signals were good indicators of flame presence, showing on average, a response time of better than 3 milliseconds. It was found that when the Hydrogen to Carbon ratio of the fuel was decreased, the CH signal strength increased as a percentage of OH signal intensity. Finally, the output signal intensity was found to be sensitive to both the flame image magnification and to the part of the flame observed. / Master of Science

A "leveling off" trend in NOx emissions with high amounts of steam addition has been observed in industrial gas turbine diffusion flame combustors. Experiments were performed to try to reproduce this trend in a laminar, opposed flow diffusion flame burner. Experiments were performed with Cli4, C2H4, CO, COIH2 (1:1), and COIH2 (1:2) as fuels. Both hydrocarbon fuels and non-hydrocarbon fuels were tested to study the contribution of the Fenimore mechanism to the "leveling off" trend. Probe sampling with chemiluminescent analysis was used to fmd NOx concentrations; Pt/PtRh thermocouples corrected for radiation losses were used to measure flame temperatures. The experiments reproduced the "leveling off" of NOx emissions, but a "leveling off" of temperatures also occurred. There were no significant differences in the results from the hydrocarbon and non-hydrocarbon fuels. The "leveling off" of NOx emissions is attributed to the "leveling off" of temperatures in the burner. It is not necessary to invoke the Fenimore mechanism to explain this trend. At least 55% of the NOx was eliminated from the flames using steam injection, which implies that at least 55% of the NOx was formed by the Zeldovich mechanism Evidence of Fenimore NO was provided by the fact that the existence of hydrocarbon coking on the fuel nozzle encouraged NOx production in all flames. / Master of Science

Gas turbine reheat thrust augmenters known as afterburners are used to provide additional thrust during emergencies, take off, combat, and in supersonic flight of high-performance aircrafts. During the course of reheat development, the most persistent trouble has been the onset of high frequency combustion instability, also known as screech, invariably followed by rapid mechanical failure. The coupling of acoustic pressure upstream of the flame stabilizer with in-phase heat-release downstream, results in combustion instability by which the amplitude at various resonant modes — longitudinal (buzz — low frequency), tangential or radial (screech — high frequency) – amplifies leading to deterioration of the afterburner components. Various researchers in early 1950s have performed extensive testing on straight jet afterburners, to identify screech frequencies. Theoretical and experimental work at test rig level has been reported in the case of buzz to validate the heat release combustion models. In this work, focus is given to study the high frequency tangential combustion instability by vibro-acoustic software and the tests are conducted on the scaled bypass flow afterburner for confirmation of predicted screech frequencies. The wave equation for the afterburner is solved taking the appropriate geometry of the afterburner and taking into account the factors affecting the stability. Nozzle of the afterburner is taken into account by using the nozzle admittance condition derived for a choked nozzle. Screech liner admittance boundary condition is imposed and the effect on acoustic attenuation is studied. A new combustion model has been proposed for obtaining the heat release rate response function to acoustic oscillations. Acoustic wave – flame interactions involve unsteady kinetic, fluid mechanic and acoustic processes over a large range of time scales. Three types of flow disturbances exist such as : vortical, entropy, and acoustic. In a homogeneous, uniform flow, these three disturbance modes propagate independently in the linear approximation. Unsteady heat release also generates entropy and vorticity disturbances. Since flow is not accelerated in the region of uniform area duct, vortical and entropy disturbances are treated as in significant, as these disturbances are convected out into atmosphere like an open-ended tube, but these are considered in deriving the nozzle admittance condition. Heat release fluctuations that arise due to fluctuating pressure and temperature are taken into consideration. The aim is to provide results on how flames respond to pressure disturbances of different amplitudes and characterised by different length scales. The development of the theory is based on large activation energy asymptotics. One-dimensional conservation equations are used for obtaining the response function for the heat release rate assuming the laminar flamelet model to be valid. The estimates are compared with the published data and deviations are discussed. The normalized acoustic pressure variation in the afterburner is predicted using the models discussed earlier to provide an indication of the resonant modes of the pressure oscillations and the amplification and attenuation of oscillations caused by the various processes. Similar frequency spectrum is also obtained experimentally using a test rig for a range of inlet mean pressures and temperatures with combustion and core and bypass flows simulated, for confirmation of predicted results. Without the heat source only longitudinal acoustic modes are found to be excited in the afterburner test section. With heat release, three additional tangential modes are excited. By the use of eight probes in the circumferential cross section of afterburner it was possible to identify the tangential modes by their respective phase shift in the experiments. Comparison of normalized acoustic pressure and phase with and without the incorporation of perforate liner is made to study the effectiveness of the screech liner in attenuating the amplitude of screech modes. By the analysis, conclusion is drawn about modes that get effectively attenuated with the presence of perforate liner. Parametric study of screech liner porosity factor of 1.5 % has not shown appreciable attenuation. Whereas with 2.5 % porosity significant attenuation is noticed, but with 4 % porosity, the gain is very minimal. Hence, the perforate screech liner with the porosity of 2.5 % is finalized. From the rig runs, first pure screech tangential mode and second screech coupled tangential modes are captured. The theoretical frequencies for first and second tangential modes with their phases are comparable with experimental results. Though third tangential mode is predicted, it was not excited in the experiments. There was certain level of deviation in the prediction of these frequencies, when compared to the experimentally obtained values. For this test section of length to diameter ratio of 5, no radial modes are encountered both in the analysis and experiments in the frequency range of interest. In summary, an acoustic model has been developed for the afterburner combustor, taking into account the combustion response, the screech liner and the nozzle to study the acoustic instability of the afterburner. The model has been validated experimentally for screech frequencies using a model test rig and the results have given sufficient confidence to apply the model for full scale afterburners as a predictive design tool.

Physical mechanism for high frequency instability in cryogenic engines at low hydrogen injection temperature has been a subject of debate for long time. Experimental and early developmental studies revealed no instabilities and it was only much later when liquid hydrogen at lower initial temperature (~50 to 100 K) was injected into the combustion chamber that instabilities were detected. From the compilations of the experimental data related to the instability of cryogenic engines by Hulka and Hutt, it was found that the instability was strongly connected to the temperature of hydrogen. Experiments conducted with hydrogen temperature ramping from a higher value to lower values indicated that the temperatures in excess of 90 K favor stability under most practical operating conditions. Even though this has been known for over forty years, there has been no clear and simple explanation for this. Many physical mechanisms have been hypothesized to explain how temperature ramping causes instability, but all appear to have limited range of applicability. Current understanding of cryogenic engine combustion instability has been achieved through a combination of experimental investigation and approximate analytical models as well as CFD tools. Various researchers have tried to link the low hydrogen injection temperature combustion instability phenomena with various potential mechanisms for combustion instability. They involve coupling of combustion acoustics with atomization, vaporization, mixing, chemical kinetics or any combination of these processes. Various studies related to the effect of recess, injector hydrodynamics, acoustic damping of gas liquid scheme injectors and effect of drop size distribution on the stability characteristics of cryogenic engines were compiled in the thesis. Several researchers examined fuel droplet vaporization as the rate controlling mechanism. Recently a new method for the evaluation of stability characteristics of the engine using model chamber were proposed by Russians and this is based on mixing as the rate controlling mechanism. Pros and cons of this method were discussed. Some people examined the combustion instability of rocket engines based on chemistry dynamics. A considerable amount of analytical and numerical studies were carried out by various researchers for finding out the cause of combustion instability. Because of the limitations of their analysis, they could not successfully explain the cause of combustion instability at low hydrogen injection temperature. A compilation of previous numerical studies were carried out. A number of researchers have applied CFD in the study of combustion instabilities in liquid propellant rocket engines. In the present thesis, a theoretical model has been developed based on the vaporization of droplets to predict the stability characteristics of the engine. The proposed concept focuses on three dimensional simulation of combustion instability for giving some meaningful explanations for the experimental work presented in the literature. In the present study the pressure wave corresponding to the transverse modes were superimposed on a three dimensional steady state operating conditions. Steady state parameters were obtained from the three dimensional combustion modeling. The conservation equations for mass, momentum and energy are non dimensionalized for facilitating the order of magnitude analysis. In order to do the stability analysis, variables are represented as the sum of their steady values and deviation from the steady state. A harmonic time dependence is assumed for the perturbations. For the transverse mode of oscillations independent variables of the zeroth order equations are r and θ only and the dependant variables are not functions of the axial distance. The axial dependence comes only through the first order equations. In this analysis, the wave motion in the combustion chamber is assumed to be linear, confining the nonlinearity to the vaporization process only. The reason behind making this assumption is that the vaporization process is the major mechanism driving the instability. Vaporization histories of liquid oxygen drops in a combustor with superimposed transverse oscillations were computed and stability characteristics of the engine were estimated. The stability characteristics of the engine are accessed from the solutions of first order equations. Effects of various parameters like droplet diameter, hydrogen injection temperature and hydrogen injection area on the stability characteristics of cryogenic engines are studied. A comparison of predicted and published experimental results was made which showed general agreement between experiment and computation. The present study and experimental results show clearly that hydrogen injection velocity is the critical parameter for instability rather than hydrogen injection temperature. What has happened in actual experiments when hydrogen injection temperature is varied is an effective alteration of the injection velocity that leads to the situation of instability. For higher relative velocity between hydrogen and liquid oxygen, the response of the vaporization rate in the presence of pressure wave is minimum compared to lower relative velocity. Due to this cryogenic engines will go to unstable mode at lower relative velocity.

Les progrès technologiques des dispositifs expérimentaux ainsi que les récentes avancées pour la simulation des spectres Raman à haute température rendent aujourd’hui possible la mise en oeuvre d’une technique de mesures multi-espèces de température et de concentration uniquement fondée sur l’ajustement des spectres mono-coups de diffusion Raman spontanée collectés au sein des flammes turbulentes. Dans cette étude, cette stratégie de post-traitement, associée à une chaîne de mesure spécifique, est développée selon plusieurs axes pour permettre l’extension des mesures à des cas de flammes ordinairement hostiles aux mesures classiques Rayleigh/Raman résolues par inversion matricielle ou par méthode hybride. Dans une flamme diphasique d’éthanol, une thermométrie fondée sur l’ajustement des spectres de N2 a été mise en place afin de s’affranchir des contraintes liées à la diffusion de Mie des gouttes. L’intégration d’une thermométrie Raman à partir du spectre du méthane ainsi que d’une procédure de minimisation de l’émission de C2 dans le post-traitement des spectres ont permis la réalisation de cartographies multi-scalaires (température et toutes espèces majoritaires) dans toute la zone de recirculation d’une flamme turbulente légèrement fuligineuse générée par un brûleur bluff-body. Une thermométrie fondée sur la minimisation du spectre Raman de CO2 a également été développée et éprouvée au cours d’une campagne de mesures dans une installation d’échelle semi-industrielle reproduisant les conditions d’oxycombustion des cycles de turbines à gaz dans l’objectif de la capture et de la séquestration du CO2. / Recent progress in experimental devices and simulation of high-temperature Raman spectra enable the implementation of a spectral fitting method to solve single-shot Spontaneous Raman Scattering spectra collected in turbulent flames. In this study, this post-processing method, associated to a specific experimental set-up, has been developed to extend measurements to several cases of non-Raman friendly flames where matrix inversion or hybrid methods are usually limited. In a two-phase flame, N2 Raman thermometry has been used to overcome issues from Mie scattering of droplets. The implementation of a CH4 Raman thermometry and a minimization procedure of C2 emission in the data post-processing allowed the achievement of multi-scalar cartographies (temperature and all major species) throughout the recirculation zone of a slightly sooting turbulent flame stabilized on a bluff-body burner. A thermometry based on the minimization of CO2 Raman spectrum has also been developed and tested during a measurement campaign in a semi-industrial scale installation designed for the study of oxyfuel gas turbine cycle in the aim of carbon capture and sequestration.

A research report submitted to the Faculty of Engineering and the Built Environment, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Master of Science in Engineering, 2016 / Spontaneous combustion of coal may occur when coal is mined, stored or transported and is influenced by a combination of intrinsic and /or extrinsic factors. While it is unusual for intact seams to burn in the highwall, the most common occurrence is when surface mines extract seams previously partially mined by underground bord and pillar operations. The aim of the study is to provide a predictive model (matrix) of the spontaneous combustion potential of remnant pillars at Colliery X. A number of different thermal, chemical and petrographic tests (coal factors) will be undertaken to determine their individual and collective impacts on the sponcom predictive model. The primary geology at the mine is conformable with that of the Witbank Coalfield. Battacharyya (1982) described 3 main factors in the spontaneous combustion of coal, mining factor, coal factor and geological factor which have an aggregate effect. Some of the main historical and present theories of sponcom are the pyrite theory, the bacterial theory, the oxidation theory and the humidity theory. It is important to note that no single factor is responsible for spontaneous combustion. The oxidation of coal occurs constantly. The temperature of the coal is a function of the rate of heat generation versus the rate of heat loss. Fires can start at outcrops and move through interconnected workings with heat transfer by conduction (into the overburden) or convection (between panels).The overburden can also insulate the burning coal seam. Geological factors such as depth of overburden, the degree of fracturing, and the nature of the overlying strata vary between coalfields. A coal seam fire or mine fire is the underground smouldering of a coal deposit, often in a coal mine. Such fires have economic, social and ecological impacts In order to extinguish a fire, one of three elements, fuel, oxygen, or energy, must be removed. The components of the fire triangle can be further subdivided into conventional mine control techniques and more or less unconventional or unproven mine fire control techniques. The thermal techniques discussed include the crossing point temperature, thermogravimetric analyses and oxygen absorption. Macerals, the microscopically identifiable organic constituents of coal, are one of the three basic parameters that define coal. The other two parameters are the coal rank and the mineral matter Vitrinite is the principal maceral group of the No.5 seam and inertinite dominates the No.2 and No.4 seams. The results obtained from the 22 drill-core samples and 2 ROM samples were matched to the existing borehole dataset (2296 boreholes) based on similarity of heat value (figure 3.11). A total of 24 test results (thermal, chemical and petrographic) from borehole A and borehole B were thus assigned to the borehole database which has approximately 1500 samples for each seam. By linking the laboratory datasets (borehole A and B) and the existing borehole database used for resource modelling, the sponcom variables could be modelled in a similar way to the coal resources. The overall risk matrix was calculated on a full seam basis by combining 15 variable scores, each variable having a score of 0, 1 or 2 (low-mod-high probability). The overall results from this research produced clear and unambiguous contour plans of different factors effecting sponcom of coal using single variable and combined variable datasets. In conclusion, it appears that the acceptability of a method for determining spontaneous heating characteristics of coal mainly depends upon how closely it predicts the spontaneous heating behaviour in the field conditions / CK2018