Abstract

This work presents a newly developed model for the oxidation of methanol and ethanol and their fuel interaction with NOx (NO and NO2) chemistry in jet-stirred and flow reactors, freely propagating and burner-stabilized premixed flames, as well as shock tubes. This work takes into account a very recent study on NO formation in pure methanol and ethanol burner-stabilized flames (Combust. Flame, 194, 363–375, 2018), augmented with so far unpublished experimental data. The paper mainly focuses on fuel interaction with nitrogen chemistry and NO formation in laminar premixed flames but also considers the formation and reduction of nitrogen oxides depending on the conditions of the surrounding gas phase. In agreement with previous experimental work, we find that doping of fuel blends with NO shifts the onset of fuel oxidation to lower temperatures depending on the gas conditions. The model suggests that the reactivity promoting effect of NO is mainly due to the net increase of OH radical concentrations, which causes increased fuel oxidation via the NO/NO2 interconversion reaction channel, NO+HO2⇋NO2+OH, NO2+H⇋NO+OH, NO2+HO2⇋HONO+O2, followed by the thermal decomposition of HONO. In burner-stabilized premixed ethanol flames, NO is mainly formed via a NCN pathway for all equivalence ratios, while for methanol flames the NCN pathway is only favored at rich conditions and the N2O pathway is favored at lean conditions.

Abstract

The prediction of local heat transfer and thermal stratification in the zero-dimensional stochastic reactor model is compared to direct numerical simulation published by Schmitt et al. in 2015. Direct numerical simulation solves the Navier–Stokes equations without incorporating model assumptions for turbulence and wall heat transfer. Therefore, it can be considered as numerical experiment and is suitable to validate approximations in low-dimensional models. The stochastic reactor model incorporates a modified version of the Euclidean Minimum Spanning Tree mixing model proposed by Subramaniam et al. in 1998. To capture the thermal stratification of the direct numerical simulation, the total enthalpy (H) is used as the only mixing limiting scalar within the newly proposed H-Euclidean-Minimum-Spanning-Tree. Furthermore, a stochastic heat transfer model is incorporated to mimic turbulence effects on local heat transfer distribution to the walls. By adjusting the Cϕ mixing time and Ch stochastic heat transfer parameter, the stochastic reactor model predicts accurately the thermal stratification of the direct numerical simulation. Comparing the Woschni, Hohenberg and Heinle heat transfer model shows that the modified Heinle model matches accurately the direct numerical simulation results. Thereby, the Heinle model accounts for the influence of turbulent kinetic energy on the characteristic velocity in the heat transfer coefficient calculation. This highlights the importance of incorporating turbulence effects in low-dimensional heat transfer models. Overall, the zero-dimensional stochastic reactor model with the H-Euclidean-Minimum-Spanning-Tree mixing model, the stochastic heat transfer model and the modified Heinle correlation have proven successfully the prediction of mean quantities like temperature and heat transfer and thermal stratification of the direct numerical simulation.

Abstract

This work introduces a newly developed reaction mechanism for the oxidation of ammonia in freely propagating and burner stabilized premixed flames as well as in shock tubes, jet stirred reactors and plug flow reactors experiments. The paper mainly focuses on pure ammonia and ammonia-hydrogen fuel blends. The reaction mechanism also considers the formation of nitrogen oxides, as well as the reduction of nitrogen oxides depending on the conditions of the surrounding gas phase. Doping of the fuel blend with NO2 can result in acceleration of H2 autoignition via the reaction NO2+HO2⇋HONO+O2 followed by the thermal decomposition of HONO, or in deceleration of H2 oxidation via NO2+OH⇋NO+HO2. The concentration of HO2 is decisive for the active reaction pathway. The formation of NO in burner stabilized premixed flames is shown to demonstrate the capability of the mechanism to be integrated into mechanisms for hydrocarbon oxidation.

Abstract

The ability of a mechanism describing the oxidation kinetics of toluene reference fuel (TRF)/n-butanol mixtures to predict the impact of n-butanol blending at 20% by volume on the autoignition and knock properties of gasoline has been investigated under conditions of a strongly supercharged spark ignition (SI) engine. Simulations were performed using the LOGEengine code for stoichiometric fuel/air mixtures at intake temperature and pressure conditions of 320 K and 1.6 bar, respectively, for a range of spark timings. At the later spark timing of 6 °CA bTDC, the predicted knock onsets for a gasoline surrogate (toluene reference fuel, TRF) and the TRF/n-butanol blend are higher compared to the measurements, which is consistent with an earlier study of ignition delay times predicted in a rapid compression machine (RCM, Agbro et al., Fuel, 2017, 187:211-219). The discrepancy between the predicted and measured knock onsets is however quite small at higher pressure and temperature conditions (spark timing of 8 °CA bTDC) and can be improved by updating a key reaction related to the toluene chemistry. The ability of the scheme to predict the influence of n-butanol blending on knock onsets requires improvement at later spark timings. The simulations highlighted that the low-intermediate temperature chemistry within the SI engine end gas, represented by the presence of a cool flame and negative temperature coefficient (NTC) phase, plays an important role in influencing the high temperature heat release and consequently the overall knock onset. This is due to its sensitisation effect (increasing of temperature and pressure) on the end gas and reduction of the time required for the high temperature heat release to occur. Therefore, accurate representation of the low-intermediate temperature chemistry is crucial for predicting knock. The engine simulations provide temperature, heat release and species profiles that link conditions in practical devices and ignition delay times predicted in an RCM. This facilitates a better understanding of the chemical processes affecting knock onsets predicted within the engine and the main reactions governing them.

Abstract

We have developed a new model utilizing our existing kinetic gas phase models to simulate experimental particle size distributions emerging in dry supersaturated H2SO4 vapor homogeneously produced by rapid oxidation of SO2 through stabilized Criegee-Intermediates from 2-butene ozonolysis. We use a sectional method for simulating the particle dynamics. The particle treatment in the model is based on first principles and takes into account the transition from the kinetic to the diffusion-limited regime. It captures the temporal evolution of size distributions at the end of the ozonolysis experiment well, noting a slight underrepresentation of coagulation effects for larger particle sizes. The model correctly predicts the shape and the modes of the experimentally observed particle size distributions. The predicted modes show an extremely high sensitivity to the H2SO4 evaporation rates of the initially formed H2SO4 clusters (dimer to pentamer), which were arbitrarily restricted to decrease exponentially with increasing cluster size. In future, the analysis presented in this work can be extended to allow a direct validation of quantum chemically predicted stabilities of small H2SO4 clusters, which are believed to initiate a significant fraction of atmospheric new particle formation events. We discuss the prospects and possible limitations of the here presented approach.

Abstract

Double-imaging photoelectron/photoion coincidence (i2PEPICO) spectroscopy using a multiplexing, time-efficient, fixed-photon-energy approach offers important opportunities of gas-phase analysis. Building on successful applications in combustion systems that have demonstrated the discriminative power of this technique, we attempt here to push the limits of its application further to more chemically complex combustion examples. The present investigation is devoted to identifying and potentially quantifying compounds featuring five heavy atoms in laminar, premixed low-pressure flames of hydrocarbon and oxygenated fuels and their mixtures. In these combustion examples from flames of cyclopentene, iso-pentane, iso-pentane blended with dimethyl ether (DME), and diethyl ether (DEE), we focus on the unambiguous assignment and quantitative detection of species with the sum formulae C5H6, C5H7, C5H8, C5H10, and C4H8O in the respective isomer mixtures, attempting to provide answers to specific chemical questions for each of these examples. To analyze the obtained i2PEPICO results from these combustion situations, photoelectron spectra (PES) from pure reference compounds, including several examples previously unavailable in the literature, were recorded with the same experimental setup as used in the flame measurements. In addition, PES of two species where reference spectra have not been obtained, namely 2-methyl-1-butene (C5H10) and the 2-cyclopentenyl radical (C5H7), were calculated on the basis of high-level ab initio calculations and Franck-Condon (FC) simulations. These reference measurements and quantum chemical calculations support the early fuel decomposition scheme in the cyclopentene flame towards 2-cyclopentenyl as the dominant fuel radical as well as the prevalence of branched intermediates in the early fuel destruction reactions in the iso-pentane flame, with only minor influences from DME addition. Furthermore, the presence of ethyl vinyl ether (EVE) in DEE flames that was predicted by a recent DEE combustion mechanism could be confirmed unambiguously. While combustion measurements using i2PEPICO can be readily obtained in isomer-rich situations, we wish to highlight the crucial need for high-quality reference information to assign and evaluate the obtained spectra.
This article offers supplementary material which is provided at the end of the article.

Abstract

A numerical platform is presented for diesel engine performance mapping. The platform employs a zero-dimensional stochastic reactor model for the simulation of engine in-cylinder processes. n-Heptane is used as diesel surrogate for the modeling of fuel oxidation and emission formation. The overall simulation process is carried out in an automated manner using a genetic algorithm. The probability density function formulation of the stochastic reactor model enables an insight into the locality of turbulence–chemistry interactions that characterize the combustion process in diesel engines. The interactions are accounted for by the modeling of representative mixing time. The mixing time is parametrized with known engine operating parameters such as load, speed and fuel injection strategy. The detailed chemistry consideration and mixing time parametrization enable the extrapolation of engine performance parameters beyond the operating points used for model training. The results show that the model responds correctly to the changes of engine control parameters such as fuel injection timing and exhaust gas recirculation rate. It is demonstrated that the method developed can be applied to the prediction of engine load–speed maps for exhaust NOx, indicated mean effective pressure and fuel consumption. The maps can be derived from the limited experimental data available for model calibration. Significant speedup of the simulations process can be achieved using tabulated chemistry. Overall, the method presented can be considered as a bridge between the experimental works and the development of mean value engine models for engine control applications.

Abstract

In this work, we apply a sequence of concepts for mechanism reduction on one reaction mechanism including novel quality control. We introduce a moment-based accuracy rating method for species profiles. The concept is used for a necessity-based mechanism reduction utilizing 0D reactors. Thereafter a stochastic reactor model for internal combustion engines is applied to control the quality of the reduced reaction mechanism during the expansion phase of the engine. This phase is sensitive on engine out emissions, and is often not considered in mechanism reduction work. The proposed process allows to compile highly reduced reaction schemes for computational fluid dynamics application for internal combustion engine simulations. It is demonstrated that the resulting reduced mechanisms predict combustion and emission formation in engines with accuracies comparable to the original detailed scheme.

Abstract

Engine knock is an important phenomenon that needs consideration in the development of gasoline-fueled engines. In our days, this development is supported using numerical simulation tools to further understand and predict in-cylinder processes. In this work, a model tool chain which uses a detailed chemical reaction scheme is proposed to predict the auto-ignition behavior of fuels with different octane ratings and to evaluate the transition from harmless auto-ignitive deflagration to knocking combustion. In our method, the auto-ignition characteristics and the emissions are calculated using a gasoline surrogate reaction scheme containing pathways for oxidation of ethanol, toluene, n-heptane, iso-octane and their mixtures. The combustion is predicted using a combination of the G-equation based flame propagation model utilizing tabulated laminar flame speeds and well-stirred reactors in the burned and unburned zone in three-dimensional Reynolds-averaged Navier–Stokes. Based on the detonation theory by Bradley et al., the character and the severity of the auto-ignition event are evaluated. Using the suggested tool chain, the impact of fuel properties can be efficiently studied, the transition from harmless deflagration to knocking combustion can be illustrated and the severity of the auto-ignition event can be quantified.Engine knock is an important phenomenon that needs consideration in the development of gasoline-fueled engines. In our days, this development is supported using numerical simulation tools to further understand and predict in-cylinder processes. In this work, a model tool chain which uses a detailed chemical reaction scheme is proposed to predict the auto-ignition behavior of fuels with different octane ratings and to evaluate the transition from harmless auto-ignitive deflagration to knocking combustion. In our method, the auto-ignition characteristics and the emissions are calculated using a gasoline surrogate reaction scheme containing pathways for oxidation of ethanol, toluene, n-heptane, iso-octane and their mixtures. The combustion is predicted using a combination of the G-equation based flame propagation model utilizing tabulated laminar flame speeds and well-stirred reactors in the burned and unburned zone in three-dimensional Reynolds-averaged Navier–Stokes. Based on the detonation theory by Bradley et al., the character and the severity of the auto-ignition event are evaluated. Using the suggested tool chain, the impact of fuel properties can be efficiently studied, the transition from harmless deflagration to knocking combustion can be illustrated and the severity of the auto-ignition event can be quantified.

Abstract

In view of a desired transition from fossil fuels to sustainably produced biofuels that should contribute to a net reduction of CO2 emissions, promising fuel candidates have been identified including 2-butanone (methyl ethyl ketone, MEK) that is qualified for use in spark-ignition (SI) engines. To support a potential, rapid integration of such biofuels into the existing infrastructure, fundamental studies of their combustion and emission behavior are highly important. In the case of 2-butanone specifically, only very few fundamental combustion experiments have been performed to date, with a notable shortage of detailed speciation data. For predictive model development, accurate and reliable species measurements are needed to describe the oxidation and combustion of 2-butanone and to elucidate the kinetic mechanism.
The present study relies on three different experiments: a laminar flow reactor coupled with molecular-beam mass spectrometry (MBMS, Bielefeld), a rapid compression machine (RCM, Aachen), and a shock tube using advanced laser absorption techniques (Stanford). This combination ensured coverage of a wide regime in temperature, pressure, and mixture composition while providing numerous species profiles. The species measurements in the flow reactor were performed at stoichiometric (Φ = 1.0) and fuel-rich (Φ = 2.0) equivalence ratios at temperatures between about 800–1100 K with an argon dilution of 95%. Ignition delay times in the RCM were measured in air for equivalence ratios of 0.5, 2.0 in the temperature range of 840–945 K to reflect application-relevant conditions. Shock tube measurements were performed at stoichiometric conditions at 1303–1509 K in argon.
To provide insight into the oxidation mechanism of 2-butanone, the newly measured experimental data was used to develop and validate a detailed chemical kinetic model. To this end, the reactions for the low-temperature regime, especially concerning early fuel consumption, radical formation, and subsequent low-temperature oxidation, were examined in detail, and used to extend and adapt an existing reaction mechanism (Burke et al., 2016 [23]) to more accurately predict the new low-temperature conditions studied. The resulting model that incorporates the latest theoretical kinetic calculations available in the literature was compared to the measurements presented here as well as validated using literature data. To the authors' knowledge, the present model represents the most robustly validated mechanism available for the prediction of 2-butanone combustion targets to date.

Abstract

A simulation process for spark ignition gasoline engines is proposed. The process is based on a zero-dimensional spark ignition stochastic reactor model and three-dimensional computational fluid dynamics of the cold in-cylinder flow. The cold flow simulations are carried out to analyse changes in the turbulent kinetic energy and its dissipation. From this analysis, the volume-averaged turbulent mixing time can be estimated that is a main input parameter for the spark ignition stochastic reactor model. The spark ignition stochastic reactor model is used to simulate combustion progress and to analyse auto-ignition tendency in the end-gas zone based on the detailed reaction kinetics. The presented engineering process bridges the gap between three-dimensional and zero-dimensional models and is applicable to various engine concepts, such as, port-injected and direct injection engines, with single and multiple spark plug technology. The modelling enables predicting combustion effects and estimating the risk of knock occurrence at different operating points or new engine concepts for which limited experimental data are available.

Abstract

Three LES models devoted to the NO prediction in under-adiabatic furnaces are evaluated in this paper: the NORA (NO relaxation Approach) model, based on the NO relaxation towards equilibrium, the linear model (LM) which employs a linear relation to rescale the NO consumption rate, and a new model, DF-NORA, in which the linear approximation of the LM is replaced by a tabulation of the reaction rate as a function of a NO progress variable. To generate this table, NO relaxation complex chemistry calculations are used like in NORA, but the homogeneous reactor is replaced by a steady laminar diffusion flame. These models are validated on Sandia Flame D and on the flameless case of Verissimo et al. (Ener. Fuel. 25, 2469-2480 ([32])). For both cases, NORA underpredicts the NO production due to its insensitivity to strain, while LM overpredicts NO by a factor 2 on Flame D and a factor 13 on the flameless case. DF-NORA presents the best prediction with a maximal underprediction of 30% on Flame D and an over-prediction of 30% on the final NO yield of the flameless case. The impact of a radiative source term is also assessed on Flame D, showing a local decrease of NO by less than 7% compared to the adiabatic calculation for the DF-NORA model.

Abstract

By applying a novel approach to evaluate photofragmentation laser-induced fluorescence (PFLIF) imaging,
experimental quantitative information on the temporal in-cylinder distribution of hydrogen peroxide
(H2O2) in a homogeneous charge compression ignition (HCCI) engine was extracted. The results from
PFLIF were then compared to those obtained from chemical kinetics simulations using computational
fluid dynamics (CFD) and a stochastic reactor model (SRM). For the CFD simulations, a sector mesh
was applied using Reynolds-averaged Navier–Stokes (RANS) equations together with a reduced chemical
kinetic model. These simulations provided detailed information on the spatial distribution of H2O2, HO2
as well as other important species and temperature. The SRM, which offers substantially reduced simulation
times but no spatial information, was used with the same reduced kinetic model. Two-dimensional
images from PFLIF and CFD show a fair temporal agreement, while details of the spatial distributions disagree.
The CFD images show that the combustion chemistry is affected by the interaction with the cylinder
walls with, for instance, a local delay of the formation and consumption of H2O2. By using probability
density functions (PDFs) of H2O2 and HO2 mass fractions, comparisons could be made between experimental
data and both the CFD and SRM simulations. In general the range of mass fractions show good
agreement but the experimental distributions are wider. Possible reasons for this discrepancy are actual
heterogeneities in the H2O2/HO2 concentration distributions not predicted by the model, spatial temperature
variations, which will influence the strength of the PFLIF signal, spatial variations in the laser
profiles, not accounted for in the data processing, and photon noise. The good agreement between the
CFD and SRM shows the relevance of fast PDF based simulation tools.
 2015 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

Abstract

An existing comprehensive kinetic hydrocarbon oxidation model has been augmented and revised for a detailed analysis of n-heptane flame chemistry. The analysis was enabled by experiments in which the detailed species composition in a fuel-rich flat premixed (ϕ = 1.69) n-heptane flame at 40 mbar has been studied by flame-sampling molecular-beam mass spectrometry using electron impact ionization. Mole fraction profiles of more than 80 different species have been measured and compared against the new detailed kinetic model consisting of 349 species and 3686 elementary reactions. For all major products and most of the minor intermediates, a good agreement of the modeling results with the experimentally-observed mole fraction profiles has been found. The presence of low- and intermediate-temperature chemistry close to the burner surface was consistently observed in the experiment and the simulation. With the same kinetic model, n-heptane auto-ignition timing, flame speeds and species composition in a jet-stirred reactor have been successfully simulated for a broad range of temperatures (500–2000 K) and pressures (1–40 bar). The comprehensive nature and wide applicability of the new model were further demonstrated by the examination of various target experiments for other C1 to C7 fuels.

Abstract

An existing detailed and broadly validated kinetic scheme is augmented to capture the flame chemistry of 1-hexene under stoichiometric and fuel rich conditions including benzene formation pathways. In addition, the speciation in a premixed stoichiometric 1-hexene flame (flat-flame McKenna-type burner) has been studied under a reduced pressure of 20–30 mbar applying flame-sampling molecular-beam time-of-flight mass spectrometry and photoionization by tunable vacuum-ultraviolet synchrotron radiation. Mole fraction profiles of 40 different species have been measured and validated against the new detailed chemical reaction model consisting of 275 species and 3047 reversible elementary reactions. A good agreement of modelling results with the experimentally observed mole fraction profiles has been found under both stoichiometric and fuel rich conditions providing a sound basis for analyzing benzene formation pathways during 1-hexene combustion. The analysis clearly shows that benzene formation via the fulvene intermediate is a very important pathway for 1-hexene.

Abstract

Aspects of zero-dimensional (0D) and three-dimensional (3D) modelling of combustion and soot formation for Diesel engines are presented. In this work 0D and 3D models were applied for the same engine experiment. The 0D simulations were carried out using a direct injection stochastic reactor model (DI-SRM). This model is built on a PDF approach, which allows for the use of detailed chemistry for calculation of combustion and emission formation and interaction between chemistry and turbulent flow. A 3D computational fluid dynamics (CFD) model in which combustion is calculated using using a PDF-time scale model and in which soot was modelled using a flamelet library model was employed. The DI-SRM results demonstrate the applicability of the flamelet model for the combustion process and also elucidate the limitations of the interactive flamelet model when calculating emission formation. The emission results, if plotted in mixture fraction space, show a wide spread for species such as NO and CO but the low spread obtained for species such as C2H2 and OH make them applicable for calculation of the source terms of soot formation in mixture fraction space. The CFD calculations were used to control assumptions made in the DI-SRM and the DI-SRM results were used to control the assumptions inferred by using tabulated chemistry. It is demonstrated that the DI-SRM can be used for soot modelling under Diesel engine conditions and that the flamelet library approach for modelling of soot formation in CFD is sound.

Abstract

Understanding the combustion chemistry of the butene isomers is a prerequisite for a comprehensive description of the chemistry of C1 to C4 hydrocarbon and oxygenated fuels such as butanol. For the development and validation of combustion models, it is thus crucial to improve the knowledge about the C4 combustion chemistry in detail.

Premixed low-pressure (40 mbar) flat argon-diluted (25%) flames of the three butene isomers (1-butene, trans-2-butene and i-butene) were studied under fuel-rich (ϕ = 1.7) conditions using a newly developed analytical combination of high-resolution in situ molecular-beam mass spectrometry (MBMS) and in situ gas chromatography (GC). The time-of-flight MBMS with its high mass resolution enables the detection of both stable and reactive species, while the gas chromatograph permits the separation of isomers from the same sampling volume. The isomer-specific species information and the quantitative mole fraction profiles of more than 30 stable and radical species measured for each fuel were used to extend and validate the C4 subset of a comprehensive flame simulation model. The experimental data shows different destruction pathways for the butene isomers, as expected, and the model is well capable to predict the different combustion behavior of the isomeric flames. The detailed analysis of the reaction pathways in the flame and the respective model predictions are discussed.

Abstract

Photochemically driven reactions involving unsaturated radicals produce a thick global layer of organic haze on Titan, Saturn’s largest moon. The allyl radical self-reaction is an example for this type of chemistry and was examined at room temperature from an experimental and kinetic modelling perspective. The experiments were performed in a static reactor with a volume of 5 L under wall free conditions. The allyl radicals were produced from laser flash photolysis of three different precursors allyl bromide (C3H5Br), allyl chloride (C3H5Cl), and 1,5-hexadiene (CH2CH(CH2)2CHCH2) at 193 nm. Stable products were identified by their characteristic vibrational modes and quantified using FTIR spectroscopy. In addition to the (re-)combination pathway C3H5+C3H5→C6H10 we found at low pressures around 1 mbar the highest final product yields for allene and propene for the precursor C3H5Br. A kinetic analysis indicates that the end product formation is influenced by specific reaction kinetics of photochemically activated allyl radicals. Above 10 mbar the (re-)combination pathway becomes dominant. These findings exemplify the specificities of reaction kinetics involving chemically activated species, which for certain conditions cannot be simply deduced from combustion kinetics or atmospheric chemistry on Earth.

Abstract

The influence of the route via the NCN radical on NO formation in flames was examined from a thermochemistry and reaction kinetics perspective. A detailed analysis of available experimental and theoretical thermochemical data combined with an Active Thermochemical Tables analysis suggests a heat of formation of 457.8 ± 2.0 kJ/mol for NCN, consistent with carefully executed theoretical work of Harding et al. (2008) [5]. This value is significantly different from other previously reported experimental and theoretical values. A combination of an extensively validated comprehensive hydrocarbon oxidation model extended by the GDFkin3.0_NCN-NOx sub-mechanism reproduced NCN and NO mole fraction profiles in a recently characterized fuel-rich methane flame only when heat of formation values in the range of 445–453 kJ/mol are applied. Sensitivity analysis revealed that the sensitivities of contributing steps to NO and NCN formation are strongly dependent on the absolute value of the heat of formation of NCN being used. In all flames under study the applied NCN thermochemistry highly influences simulated NO and NCN mole fractions. The results of this work illustrate the thermochemistry constraints in the context of NCN chemistry which have to be taken into account for improving model predictions of NO concentrations in flames.

Abstract

The influence of the route via the NCN radical on NO formation in flames was examined from a thermochemistry and reaction kinetics perspective. A detailed analysis of available experimental and theoretical thermochemical data combined with an Active Thermochemical Tables analysis suggests a heat of formation of 457.8 ± 2.0 kJ/mol for NCN, consistent with carefully executed theoretical work of Harding et al. (2008) [5]. This value is significantly different from other previously reported experimental and theoretical values. A combination of an extensively validated comprehensive hydrocarbon oxidation model extended by the GDFkin3.0_NCN-NOx sub-mechanism reproduced NCN and NO mole fraction profiles in a recently characterized fuel-rich methane flame only when heat of formation values in the range of 445–453 kJ/mol are applied. Sensitivity analysis revealed that the sensitivities of contributing steps to NO and NCN formation are strongly dependent on the absolute value of the heat of formation of NCN being used. In all flames under study the applied NCN thermochemistry highly influences simulated NO and NCN mole fractions. The results of this work illustrate the thermochemistry constraints in the context of NCN chemistry which have to be taken into account for improving model predictions of NO concentrations in flames.

Abstract

This paper reports on a turbulent flame propagation model combined with a zero-dimensional two-zone stochastic reactor model (SRM) for efficient predictive SI in-cylinder combustion calculations. The SRM is a probability density function based model utilizing detailed chemistry, which allows for accurate knock prediction. The new model makes it possible to - in addition - study the effects of fuel chemistry on flame propagation, yielding a predictive tool for efficient SI in-cylinder calculations with all benefits of detailed kinetics. The turbulent flame propagation model is based on a recent analytically derived formula by Kolla et al. It was simplified to better suit SI engine modelling, while retaining the features allowing for general application. Parameters which could be assumed constant for a large spectrum of situations were replaced with a small number of user parameters, for which assumed default values were found to provide a good fit to a range of cases. Only one parameter, the turbulence intensity, needed tuning to obtain excellent agreement for various cases. The laminar flame speed is obtained from a laminar flame speed library generated using detailed chemistry. The flame development was calculated from the turbulent flame speed under the assumption of a spherical flame. A Monte Carlo geometry calculation was applied to cater for arbitrary cylinder geometries and spark plug positions, modelling the geometrical properties of the flame with high precision. In later stages of the project, a polygon based description of the flame surface was used, to achieve faster computational times than those of the Monte Carlo model.

Abstract

3D Kinetic Monte Carlo (KMC) simulations have been carried out on the epitaxial growth of the silicon (100)2×1 surface as a function of surface temperature (570-870 °C).

The KMC model explicitly takes into account the anisotropy of the silicon (100)2×1 surface and the interaction of neighboring sites as a reaction event at a given surface site not only depends on the chemical nature of the site itself but also on steric factors and the local environment. Thus the model includes data about the local structure of the surface, the nature of the surface adatoms and their neighbors, the kinetic reaction parameters, and the incident precursor atoms. Reaction probabilities are calculated with the Arrhenius equation, kinetic parameters are taken from experimental and calculated data from the literature. First estimations are given for missing values. Silane is assumed to be the only gas-phase reactant on the surface, coupling with the gas phase is carried out by silane partial pressure.

For the first time a really complex algorithm comprehending 12 different surface sites and more than 100 reactions such as silane adsorption, SiHx decomposition and diffusion of adsorbed species is presented. The model provides a good fit to experimental observations and theoretical knowledge. Experimental data of growth rate and hydrogen coverage can be reproduced.

Abstract

Interactions between in-cylinder combustion and emission aftertreatment need to be understood for optimizing the overall powertrain system. Numerical investigations can aid this process. For this purpose, simple and numerically fast, but still accurate models are needed for in-cylinder combustion and exhaust aftertreatment. The chemical processes must be represented in sufficient detail to predict engine power, fuel consumption, and tailpipe emission levels of NOx, soot, CO and unburned hydrocarbons. This paper reports on a new transient one-dimensional catalyst model. This model makes use of a detailed kinetic mechanism to describe the catalytic reactions.

A single-channel or a set of representative channels are used in the presented approach. Each channel is discretized into a number of cells. Each cell is treated as a perfectly stirred reactor (PSR) with a thin film layer for washcoat treatment. Heat and mass transport coefficients are calculated from Nusselt and Sherwood laws. Either detailed or global surface chemistry is applied in the thin film layer. Three global parameters are used to align the detailed chemistry model with a given catalyst topology and composition; one parameter for heat transfer, one for mass transfer and one for overall reaction efficiency. This allows considering detailed surface chemistry, molecular diffusion and heat conductivity while maintaining affordable CPU time. Detailed, usually unknown, specifications of the catalyst material are insignificant for the presented approach.

The models' applicability is demonstrated for a single-channel of a NOx-storage catalyst (NSC). The detailed surface chemistry by Koop and Deutschmann is utilized. Good agreement between experimental data and model results is achieved. The investigation of surface site fractions shows, that CO and C₃H₆ from exhaust gases inhibit NO oxidation by the same process; in both cases surface bound CO blocks the sites for NO oxidation. The inhibition effect is mainly determined by the total concentration of carbon atoms contained in CO and HC in the exhaust stream. Oxidation by surface bound oxygen was further found to be the major pathway for HC conversion. The lasting inhibition effect of unburned hydrocarbons on NO oxidation was studied by a transient calculation. In this test a sudden cutoff of unburned hydrocarbons in the exhaust stream was assumed. The response time for NO oxidation was found to be 5.5 seconds. The high response time proofs the necessity of using a transient model of sufficient detail to simulate catalytic oxidation during transient engine processes or under cyclic variations.

Abstract

The combustion chemistry of the two butane isomers represents a subset in a comprehensive description of C1–C4 hydrocarbon and oxygenated fuels. A critical examination of combustion models and their capability to predict emissions from this class of fuels must rely on high-quality experimental data that address the respective chemical decomposition and oxidation pathways, including quantitative intermediate species mole fractions. Premixed flat low-pressure (40 mbar) flames of the two butane isomers were thus studied under identical, fuel-rich (φ=1.71) conditions. Two independent molecular-beam mass spectrometer (MBMS) set-ups were used to provide quantitative species profiles. Both data sets, one from electron ionization (EI)-MBMS with high mass resolution and one from photoionization (PI)-MBMS with high energy resolution, are in overall good agreement. Simulations with a flame model were used to analyze the respective reaction pathways, and differences in the combustion behavior of the two isomers are discussed.

Abstract

The influence of differential diffusion of chemical species in the soot initiation process in turbulent flows is investigated through Direct Numerical Simulations coupled to a compact global chemical mechanisms for ethylene (C2H4) flame combustion (Løvås et al., Combust Sci Tech 182(11):1945–1960, 2010) featuring the important reaction steps for acetylene production. Our focus is on the formation of acetylene (C2H2) which is one of the most important species indicative of soot formation layers, especially in relation to the location of the H and H2 layers. The effect of preferential diffusion is assessed by comparison of results from unity and non-unity Lewis number simulations. The results indicate that under moderate turbulent conditions, where preferential diffusion effects become prominent, and with the global scheme used preferential diffusion greatly enhances the spread of the radical H whose peak value in mass fraction is reduced by a factor of about two; the spread of H2 is also enhanced though to a lesser extent. Importantly, the H and H2 spread into a range of mixture fraction Z between 0.2 and 0.3 which contains the soot formation range, supporting the hypothesis that soot formation is enhanced by preferential diffusion. Nevertheless, the acetylene formation layers themselves show little adjustment in the presence of non-unity Lewis numbers suggesting that the acetylene formation is dominated under the current conditions by the direct thermal decomposition of ethylene to acetylene in the global chemistry used. The specific Fi factors that appear in flamelet models are explicitly computed; only FH, FH2 and FCO show appreciable differences on the fuel lean range of mixture fraction due to non-unity Lewis numbers, suggesting that the effects of non-unity Lewis numbers could be incorporated by a selective inclusion of only a few of the Fi factors in order to save computational time.

Abstract

Today’s car manufactures inevitably have to focus on the reduction of fuel consumption while maintaining high performance standards. In this respect, the downsized turbocharged DISI (Direct Injection Spark Ignition) engine represents an appealing solution.

However, downsizing is limited because of knocking phenomena occurring at high- and full-load conditions due to autoignition of the unburned mixture ahead the flame front. A common way of reducing knock tendencies is provided by Exhaust Gas Recirculation (EGR). However, EGR modifies the chemical composition of the cylinder charge and recirculated species like nitric oxide (NO) or unburned Hydrocarbons (HC) particularly increase the reactivity of the unburned mixture. In other words, the EGR influences the Octane Number (ON) of the in-cylinder gases.

This paper proposes a new EGR concept in which a catalytic converter is introduced into the EGR system in order to minimize the NO and HC concentrations and thereby to reduce the reactivity of EGR.

Pressure trace and infrared-spectral measurements are carried out on an engine test bed including a standard three-way-catalyst in the high-pressure EGR pipe of a turbocharged DISI engine. Pressure trace measurements allow thermodynamic analysis, while IR-spectral analysis identifies the concentrations of EGR species in the EGR pipe. Furthermore, a zero-dimensional reactor-network modeling the EGR pipe and the combustion chamber is used to numerically investigate the influence of EGR species on the autoignition processes.

Both experimental and numerical results show a potential benefit introduced by catalytic reformation of exhaust gases for turbocharged DISI engines in terms of reduced fuel consumption.

Abstract

Two compact global mechanisms for ethylene (C2H4) diffusion flame combustion have been tailored to include important reaction steps for acetylene and benzene production. One mechanism (G11) contains 11 species with 10 reaction steps including acetylene (C2H2), and the other mechanism (G12) contains 12 species with a total of 11 reactions steps to include also the formation of benzene (C6H6). The reaction steps have been carefully selected to minimize the mechanism size for the use in large-scale computational fluid dynamics (CFD) simulations. Hence, the reaction constants have been optimized for the correct prediction of important radical concentrations. Particular focus has been on the mechanisms ability to reproduce important preferential diffusion effects and on the formation of H and C2H2 due to its importance to soot formation. The two global chemical models have been validated for a transient 1-dimensional diffusion flame configuration and show very good agreement with various detailed chemical schemes. The mechanisms are found to be nonstiff reducing typical computing time for a transient flamelet calculation (F. Mauss, 1998) from a few hours (171 species mechanism) to only a few minutes (G11).

Abstract

Adaptive chemistry is based on the principle that instead of having one comprehensive model describing the entire range of chemical source term space (typically parameters related to temperature, pressure and species concentrations), a set of computationally simpler models are used, each describing a local region (in multidimensional space) or phases (in zero-dimensional space). In this work, an adaptive chemistry method based on phase optimized skeletal mechanisms (POSM) is applied to a 96 species n-heptane–isooctane mechanism within a two-zone zero-dimensional stochastic reactor model (SRM) for an spark-ignition (SI) Engine. Two models differing only in the extent of reduction in the phase mechanism, gave speed-up factors of 2.7 and 10. The novelty and emphasis of this study is the use of machine learning techniques to decide where the phases are and to produce a usable phase recognition. The combustion process is automatically divided up into an ‘optimal’ set of phases through machine learning clustering based on fuzzy logic predicates involving a necessity parameter (a measure giving an indication whether a species should be included in the mechanism or not). The mechanism of each phase is reduced from the full mechanism based on this necessity parameter with respect to the conditions of that phase. The algorithm to decide which phase the process is in is automatically determined by another machine learning method that produces decision trees. The decision tree is made up of asking whether the mass fraction values were above or below given values. Two POSM studies were done, a conservative POSM where the species in each phase are eliminated based on a necessity parameter threshold (speed-up 2.7) and a further reduced POSM where each phase was further reduced by hand (speed-up 10). The automated techniques of determining the phases and for creating the decision tree are very general and are not limited to the parameter choices of this paper. There is also no fundamental limit as to the size of the original detailed mechanism. The interfacing to include POSM in an application does not differ significantly from using the original detailed mechanism.

Abstract

There is a need for prediction models of soot particles and polycyclic aromatic hydrocarbons (PAHs) formation in parametric conditions prevailing in automotive engines: large fuel molecules and high pressure. A detailed kinetic mechanism able to predict the formation of benzene and PAHs up to four rings from C2 fuels, recently complemented by consumption reactions of decane, was extended in this work to heptane and iso-octane oxidation. Species concentrations measured in rich, premixed flat flames and in a jet stirred reactor (JSR) were used to check the ability of the mechanism to accurately predict the formation of C2 and C3 intermediates and benzene at pressures ranging from 0.1 to 2.0 MPa. Pathways analyses show that propargyl recombination is the only significant route to benzene in rich heptane and iso-octane flames. When included as the first step of a soot particle formation model, the gas-phase kinetic mechanism predicts very accurately the final soot volume fraction measured in a rich decane flame at 0.1 MPa and in rich ethylene flames at 1.0 and 2.0 MPa.

Abstract

The present study describes the compilation and validation of a compact reaction mechanism for the oxidation of n-heptane, toluene and its mixtures using the Chemistry Guided Reduction (CGR) approach. By the module-wise composition of validated reaction schemes and the successive application of chemical lumping and redundant species retrieval for the n-heptane oxidation model, a compact mechanism is generated for reference fuel blends of n-heptane and toluene. The new mechanism is validated for recently published OH-concentrations histories and ignition times front shock tube studies, HCCl engine experiments and flame speed measurements. The good agreement between experiment and prediction demonstrates the general applicability of the CGR method.

Abstract

Applied to the primary reference fuel n-heptane, we present the chemistry-guided reduction (CGR) formalism for generating kinetic hydrocarbon oxidation models. The approach is based on chemical lumping and species removal with the necessity analysis method, a combined reaction flow and sensitivity analysis. Independent of the fuel size, the CGR formalism generates very compact submodels for the alkane low-temperature oxidation and provides a general concept for the development of compact oxidation models for large model fuel components such as n-decane and n-tetradecane. A defined sequence of simplification steps, consisting of the compilation of a compact detailed chemical model, the application of linear chemical lumping, and finally species removal based on species necessity values, allows a significantly increased degree of reduction compared to the simple application of the necessity analysis, previously published species, or reaction removal methods. The skeletal model derived by this procedure consists of 110 species and 1170 forward and backward reactions and is validated against the full range of combustion conditions including low and high temperatures, fuel-lean and fuel-rich mixtures, pressures between 1 and 40 bar, and local (species concentration profiles in flames, plug flow and jet-stirred reactors, and reaction sensitivity coefficients) and global parameters (ignition delay times in shock tube experiments, ignition timing in a HCCI engine, and flame speeds). The species removal is based on calculations using a minimum number of parameter configurations, but complemented by a very broad parameter variation in the process of compiling the kinetic input data. We further demonstrate that the inclusion of sensitivity coefficients in the validation process allows efficient control of the reduction process. Additionally, a compact high-temperature n-heptane oxidation model of 47 species and 468 reactions was generated by the application of necessity analysis to the skeletal mechanism.

Abstract

A detailed reaction mechanism for n-heptane oxidation has been compiled and subsequently simplified. The model is based on a kinetic model for C1-C4 fuel oxidation of Hoyermann et al. [Phys. Chem. Chem. Phys., 2004, 6, 3824] and a detailed mechanism for n-heptane oxidation developed by Curran et al. [Combust. Flame, 1998, 114, 149]. The generated mechanism is kept compact by limiting the application of the low temperature oxidation pathways to the fuel molecule. The first reaction steps and the complex low temperature paths in the oxidation process have been simplified and reorganized by linear chemical lumping. The reported procedure allows a decrease in number of species and reactions with only a minor loss of model accuracy. The simplified model is of very compact size and gives an advantageous starting point for further model reduction. By this chemically lumped general mechanism without further adjustments the large set of experimental data for the high and low temperature oxidation ( ignition delay times, species concentration profiles, heat release and engine pressure profiles, flame speeds and flame structure data) for conditions ranging from very low to high temperatures (550-2500 K), very lean to extremely fuel rich (0.22 < phi < 3) mixtures and pressures between 1 and 42 bar is consistently described providing a basis for reliable predictions for future applications, (i) building reaction mechanisms for similar but chemically more complex fuels (e.g. iso-octane, n-decane,...) and (ii) calculating complex flow fields ("fluid dynamics'') after further simplification with advanced reduction tools.

Abstract

We present a new probability density function (PDF)-based computational model to simulate a homogeneous charge compression ignition (HCCI) engine with direct injection (DI) during gas exchange. This stochastic reactor model (SRM) accounts for the engine breathing process in addition to the closed-volume HCCI engine operation. A weighted-particle Monte Carlo method is used to solve the resulting PDF transport equation. While simulating the gas exchange, it is necessary to add a large number of stochastic particles to the ensemble due to the intake air and EGR streams as well as fuel injection, resulting in increased computational expense. Therefore, in this work we apply a down-sampling technique to reduce the number of stochastic particles, while conserving the statistical properties of the ensemble. In this method some of the most important statistical moments (e.g., concentration of the main chemical species and enthalpy) are conserved exactly, while other moments are conserved in a statistical sense. Detailed analysis demonstrates that the statistical error associated with the down-sampling algorithm is more sensitive to the number of particles than to the number of conserved species for the given operating conditions. For a full-cycle simulation this down-sampling procedure was observed to reduce the computational time by a factor of 8 as compared to the simulation without this strategy, while still maintaining the error within an acceptable limit. Following the detailed numerical investigation, the model, intended for volatile fuels only, is applied to simulate a two-stroke, naturally aspirated HCCI engine fueled with isooctane. The in-cylinder pressure and CO emissions predicted by the model agree reasonably well with the measured profiles. In addition, the new model is applied to estimate the influence of engine operating parameters such as the relative air-fuel ratio and early direct injection timing on HCCI combustion and emissions. The qualitative trends observed in the parametric variation study match well with experimental data in literature. (c) 2006 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

Abstract

Soot formation in a turbulent jet diffusion flame is modeled using an unsteady flamelet approach. In the present work, we study the effects of the choice of the dependence of scalar dissipation rate on mixture fraction and agglomeration processes on the predicted soot volume fraction. It is found that good predictions of soot volume fraction can be obtained without considering preferential diffusion effects.

Abstract

A single detailed kinetic mechanism for the oxidation and combustion of n-decane and n-heptane has been written by means of an automatic mechanism generator ( REACTION) developed in our laboratory. It shows a good prediction of the ignition delay time versus temperature for the oxidation of n-decane at 13 and 50 bar and n-heptane at 13 and 40 bar for different equivalence ratios. The n-decane/n-heptane mechanism consists of a validated and recently published O-2=H-2=C-1-C-4 mechanism produced manually and a generated C-5-C-10 set of sub-mechanisms. The mechanism includes a complete description of both n-decane and n-heptane chemistry. This mechanism has a reasonable size, 506 species and 3684 reactions, but nevertheless it has an extensive range of chemistry. This paper represents not only the validation of a specific mechanism but also a validation of the rate constants of the reaction classes used to model the oxidation of alkanes at low and high temperature.

Abstract

In this work, a progress variable approach is used to model diesel spray ignition with detailed chemistry. The flow field and the detailed chemistry are coupled using the flamelet assumption. A flamelet progress variable is transported by the computational fluid dynamics (CFD) code. The progress variable source term is obtained from an unsteady flamelet library that is evaluated in each grid cell. The progress variable chosen is based on sensible enthalpy. By using an unsteady flamelet library for the progress variable, the impact of local effects, for example variations in the turbulence field, effects of wall heat transfer etc. on the autoignition chemistry can be considered on a cell level. The coupling between the unsteady flamelet library and the transport equation for total enthalpy follows the ideas of the representative interactive flamelet (RIF) approach. The method can be compared to having an interactive flamelet in each computational cell in the CFD grid. The results obtained using the proposed model are compared to results obtained using the RIF model. Differences are exhibited during the autoignition process. After ignition, the results obtained using the proposed model and RIF are virtually identical. The model was used to study lift-off lengths in sprays as function of nozzle diameter and injection pressure. A good agreement between model predictions and experimental trends was found.

Abstract

On the basis of existing detailed kinetic schemes a general and consistent mechanism of the oxidation of hydrocarbons and the formation of higher hydrocarbons was compiled for computational studies covering the characteristic properties of a wide range of combustion processes. Computed ignition delay times of hydrocarbon–oxygen mixtures (CH4-, C2H6-, C3H8-, n-C4H10-, CH4 + C2H6-, C2H4, C3H6-O2) match the experimental values. The calculated absolute flame velocities of laminar premixed flames (CH4-, C2H6-, C3H8-, n-C4H10-, C2H4-, C3H6-, and C2H2-air) and the dependence on mixture strength agree with the latest experimental investigations reported in the literature. With the same model concentration profiles for major and intermediate species in fuel-rich, non-sooting, premixed C2H2-, C3H6- air flames and a mixed C2H2/C3H6 (1:1)-air flame at 50 mbar are predicted in good agreement with experimental data. An analysis of reaction pathways shows for all three flames that benzene formation can be described by propargyl combination.

Abstract

Homogeneous charge compression ignition (HCCI) is a potentially attractive operating mode for stationary natural gas engines. Increasing demand for efficient, clean burning engines for electrical power generation provides an opportunity to utilize HCCI combustion if several inherent difficulties can be overcome. Fuel composition, particularly the higher hydrocarbon content (ethane, propane, and butane) of the fuel is of primary concern. Fuel composition influences HCCI operation both in terms of design, via compression ratio and initial charge temperature, and in terms of engine control. It has been demonstrated that greater concentrations of higher hydrocarbons tend to lower the ignition temperature of the mixture significantly. The purpose of this paper is to demonstrate, through simulation, the effect of fuel composition on combustion in HCCI engines. Engine performance over a range of fuels from pure methane to more typical natural gas blends is investigated. This includes both the impact of various fuels and the sensitivity of engine operation for any given fuel. Results are presented at a fixed equivalence ratio, compression ratio, and engine speed to isolate the effect of fuel composition. Conclusions are drawn as to how the difficulties arising from gas composition variations may affect the future marketability of these engines.

Abstract

A stochastic model for the HCCI engine is presented. The model is based on the PaSPFR-IEM model and accounts for inhomogeneities in the combustion chamber while including a detailed chemical model for natural gas combustion, consisting of 53 chemical species and 590 elementary chemical reactions. The model is able to take any type of inhomogeneities in the initial gas composition into account, such as inhomogeneities in the temperature field, in the air-fuel ratio or in the concentration of the recirculated exhaust gas. With this model the effect of temperature differences caused by the thermal boundary layer and crevices in the cylinder for a particular engine speed and fuel to air ratio is studied. The boundary layer is divided into a viscous sublayer and a turbulent buffer zone. There are also colder zones due to crevices. All zones are modeled by a characteristic temperature distribution. The simulation results are compared with experiments and a previous numerical study employing a PFR model. In all cases the PaSPFR-IEM model leads to a better agreement between simulations and experiment for temperature and pressure. In addition a sensitivity study on the effect of different intensities of turbulent mixing on the combustion is performed. This study reveals that the ignition delay is a function of turbulent mixing of the hot bulk and the colder boundary layer.

Abstract

A PSR-PFR series reactor model has been used with different detailed C/H/N/O reaction mechanisms for the calculation of NO formation during rich-lean staged combustion of ethylene (C2H4) with monomethylamine (CH3NH2) addition as a model mixture for nitrogen-doped fuels. The PSR model and a PSR-PFR combination have been validated by comparison with measurements on a pilot-scale reactor system. Good agreement with NO measurements in the primary reactor and the flue gas can be obtained with some of the mechanisms reported in literature. For slightly fuel-rich conditions the differences in the results obtained with different mechanisms are reflected in the dissimilar reaction paths leading to NO formation. The N atom flow rates via the major formation and destruction channels in the PSR yield a net minimum production rate at Ø = .3, which corresponds well with the measured and predicted Ø for minimum NO emission. It is also demonstrated that the flue gas NO emission is largely determined already in the PSR, making the sensitivity to PFR conditions quite low.

Abstract

The effect of thermal ionization on the growth of soot particles has been analyzed by detailed kinetic modeling of a low-pressure premixed acetylene flame. The detailed kinetic model considers the oxidation of fuel, the formation and growth of polycyclic aromatic hydrocarbons, and particle inception, coagulation, as well as mass growth via surface reactions. A numerical method has been developed, which considers the production of charged particles by thermal ionization as well as coagulation and surface reactions of these particles. The enhancement of coagulation by collisions between charged-charged and charged-neutral particles is rigorously accounted for in the numerical model. The particle size distribution functions for both neutral and charged particles were solved using the method of moments. The computed relative soot volume fractions for neutral and charged soot particles were compared to measurements and found to be in good agreement with them. The results show also that omitting of thermal ionization of soot particles does not lead to significant errors in the simulation of soot formation in the acetylene flame, as long as the nature of the surface reactions between charged particles and gaseous molecules remains the same as that for neutral particles. This result can be generalized to most laboratory laminar premixed and counterflow diffusion flames with flame temperatures not exceeding 2100 K.

Abstract

A general method for automatically reducing detailed kinetic mechanisms for complex fuels is applied. The method is based on the simultaneous use of sensitivity, reaction-flow, and extended lifetime analyses. The sensitivity analysis detects species to which the overall combustion process is sensitive. Small inaccuracies in calculating these species result in large errors in the characteristic behavior of the chemical scheme. Redundant species are detected by applying a simultaneous reaction-flow and sensitivity analyses. The sensitivities are transported through the reacting flow and each species is assigned an importance according to the importance of the species itself and the flow of atoms to and from the important species. The redundant species are removed from the detailed mechanism (74 species and 510 reactions) resulting in a skeleton mechanism (63 species and 386 reactions). The skeleton mechanism is in turn the object for a further reduction by applying extended species sensitivity and lifetime analyses. These analyses are the basis for a reduction by means of a quasi-steady-state assumption. By introducing the quasi-steady-state assumption, the skeleton mechanism is reduced further to 19 species and 16 global reactions. The skeleton and reduced mechanisms generated using different cutoff levels of "relative species importance" and "lifetime," respectively, are validated against the detailed mechanism to find the final skeleton and reduced mechanisms. Gradually increasing cutoff levels results in a correspondingly gradual increase in the difference between reduced, skeleton, and detailed mechanisms. The skeleton and reduced mechanisms are valid for the predetermined parameter ranges of initial and boundary conditions, depending on experimental conditions to be modeled.

Abstract

In this paper, we present a comparison between the reduced mechanisms obtained through a computational singular perturbation method (CSP) and the reduced mechanisms obtained through a lifetime analysis based only on the diagonal elements of the Jacobian matrix and a species sensitivity The two methods are used for the analysis of autoignition, which is an interesting test situation because of the sensitivity of ignition to the radical pool and the smaller range of timescales expected. It is found that the steady-state species selected by the two methods are in good agreement. The mechanisms are reduced to a 10-step mechanism when CSP is applied and an 11-step mechanism in the case of the simpler lifetime analysis. Both mechanisms are compared with the detailed mechanism and experimental data and are found to reproduce the physical and chemical parameters very well. This shows that for a large part of the timescale range, the system is close to linear. The comparison shows the advantage of the CSP method as being somewhat more accurate. However, the simpler lifetime analysis is of sufficient accuracy and of more convenience when applied to a system requiring a considerable reduction in computational time, as is the case when applying online reduction.

Abstract

The purpose of this work is to propose a detailed model for the formation of soot in turbulent reacting flow and to use this model to study a carbon black furnace. The model is based on a combination of a detailed reaction mechanism to calculate the gas phase chemistry, a detailed kinetic soot model based on the method of moments, and the joint composition probability density function (PDF) of these scalar quantities.

Two problems, which arise when modeling the formation of soot in turbulent flows using a PDF approach, are studied. A consistency study of the combined scalar-soot moment approach reveals that the molecular diffusion term in the PDF-equation can be closed by the IEM and Curl-type mixing models. An investigation of different kernels for the collision frequency of soot particles shows that the influence of turbulence on particle coagulation is negligible for typical flame conditions and the particle size range considered.

The model is used as a simple tool to simulate a furnace black process, which is the most important industrial process for the production of carbon blacks. Despite the simplifications in the modeling of the turbulent flow reasonable agreement between the calculated soot yield and data measured in an industrial furnace black reactor is achieved although no adjustments were made to the kinetic parameters of the soot model. The effect of the mixing intensity on soot yield and different soot formation rates is investigated. In addition the influence of different operating conditions such as temperature and equivalence ratio in the primary zone of the reactor is studied.

Abstract

In this paper, an automatic method for reducing chemical mechanisms during run time based on the quasi-steady-state assumption (QSSA) is presented. The method uses a lifetime analysis of the chemical species which can be set to steady state according to a ranking procedure. Steady-state species concentrations are computed by algebraic rather than differential equations, thus yielding a significant reduction in the computational effort. In contrast to previous reduction schemes in which chemical species were selected only when they were in steady state throughout the whole process, the present method allows for species to be selected at each operating point separately generating an adaptive chemical kinetics scheme. The mechanism can change during the simulation run. This ensures that the optimal reduced mechanism is used at each time step leading to a very efficient and accurate procedure. The method is used for calculations of a natural gas fueled engine operating under homogeneous charge compression ignition (HCCI) conditions. We discuss criteria for selecting steady-state species and the influence of these criteria. on the results, such as concentration profiles and temperature. A full mechanism with 53 species can be reduced to a minimum of 14 non-steady-state species while still reproducing the physical behavior of the detailed mechanism with good agreement.

Abstract

The modeling of soot formation and oxidation under industrially relevant conditions has made significant progress in recent years. Simplified models introducing a small number of transport equations into a CFD code have been used with some success in research configurations simulating a reciprocating diesel engine. Soot formation and oxidation in the turbulent flow is calculated on the basis of a laminar flamelet library model. The gas phase reactions are modeled with a detailed mechanism for the combustion of heptane containing 89 species and 855 reactions developed by Frenklach and Warnatz and revised by Mauss. The soot model is divided into gas phase reactions, the growth of polycyclic aromatic hydrocarbons (PAH) and the processes of particle inception, heterogeneous surface growth, oxidation, and condensation. The first two are modeled within the laminar flamelet chemistry, while the soot model deals with the soot particle processes. The time scales of soot formation are assumed to be much larger than the turbulent time scales. Therefore rates of soot formation are tabulated in the flamelet libraries rather than the soot volume fraction itself. The different rates of soot formation, e.g., particle inception, surface growth, fragmentation, and oxidation, computed on the basis of a detailed soot model, are calculated in the mixture fraction/scalar dissipation rate space and further simplified by fitting them to simple analytical functions. A transport equation for the mean soot mass fraction is solved in the CFD code. The mean rate in this transport equation is closed with the help of presumed probability density functions for the mixture fraction and the scalar dissipation rate. Heat loss due to radiation can be taken into account by including a heat loss parameter in the flamelet calculations describing the change of enthalpy due to radiation, but was not used for the results reported here. The soot model was integrated into an existing commercial CFD code as a post-processing module to existing combustion CFD flow fields and is very robust with high convergence rates. The model is validated with laboratory flame data and using a realistic three-dimensional BMW Rolls-Royce combustor configuration, where test data at high pressure are available. Good agreement between experiment and simulation is achieved for laboratory flames, whereas soot is overpredicted for the aeroengine combustor configuration by 1–2 orders of magnitude.

Abstract

An existing skeletal mechanism for laminar premixed methane/air flames has been used as a starting point for further automatic reduction qv quasi-steady-state approximation (QSSA) for species with short chemical lifetimes and/or minor influence on the chemical system. Individual species are ranked with respect to static and dynamic characteristics according to a level of importance (LOI) measure obtained from their chemical lifetimes. diffusion velocities, and flame-zone residence times in combination with a species sensitivity measure. The maximum element mass fraction and the maximum enthalpy occupied hv a certain molecular species are constrained in order to limit the mass and energy deficiency caused Lv QSSA. Maximum values of lifetime and LOI are accumulated over the entire flame length for a I range of fuel/air equivalence ratios phi. Species with low LOI are selected for QSSA, and their concentrations are calculated iteratively by solving the coupled algebraic system. Kinetic models with a varying degree of reduction are then automatically generated and implemented as FORTRAN source code by setting different lower LOI and element mass fraction limits.

It is found that the lifetime and LOI measure differ due to the inclusion of sensitivity counteracting the rise in lifetime at low temperatures. The species ranking by the LOI disfavors reasonably stable species, which are removed from the system. The laminar burning velocities as predicted by the most strongly) reduced mechanism with five global reaction steps show very good agreement with detailed calculations. The profiles of steady-state species also agree well if the corresponding species lifetime is short.

Abstract

The appearance of exothermic centers caused by inhomogeneities within the end-gas of spark ignition engines is investigated. A detailed chemical mechanism is adapted to calculate the autoignition of the primary reference fuels, n-heptane and iso-octane. The pressure history in the cylinder of the engine is obtained from a homogeneous two-zone model consisting of a burned gas zone and an unburned end-gas zone. The fraction of gases burned by the spark-ignited, propagating flame front is calculated from the Wiebe function. Within the end-gas zone, we introduce a one-dimensional coordinate, the distance from cylinder wall. Conservation equations for mass, momentum, energy, and species concentrations are solved instationary along this coordinate. The pressure is assumed to be homogeneously distributed. Gas inhomogeneities are modeled as sine waves in the initial temperature field. The development of the exothermic centers is investigated for amplitudes of the sine wave between 5 and 20 K. It is found that the gas near the exothermic center is prereacted. Products and intermediate products from low-and high-temperature reactions can be found. Thus, an apparent reaction front can propagate from the exothermic center with a velocity of several meters per second. The velocity increases with decreasing temperature gradients in the inhomogeneous mixture.

Abstract

The CO and hydrocarbon emissions of a homogeneous charge compression injection engine have been explained by inhomogeneities in temperature induced by the boundary layer and crevices according to a stochastic reactor model. The boundary layer is assumed to consist of a thin film (laminar sublayer) and a turbulent buffer layer. The heat loss through the cylinder wall leads to a significant temperature gradient in the boundary layer. The partially stirred plug flow reactor (PaSPFR) model, a stochastic reactor model (SRM), has been used to model turbulent mixing between the boundary layer, crevices, and the turbulent core and to account for the chemical reactions within the combustion chamber. The combustion of natural gas in the engine is described by a detailed chemical mechanism that is incorporated in the SRM. Molecular diffusion induced by turbulent mixing is described by the simple interaction by exchange with the mean (IEM) mixing model. The turbulent mixing intensity that describes the decay of the species and temperature fluctuations is estimated from measurements of the velocity fluctuations and the integral length scale of the turbulent flow in the engine. Pressure, CO emissions, and unburned hydrocarbons are also measured. Comparison between the mean quantities obtained from the SRM and these measurements show very good agreement. It is demonstrated that the SRM clearly outperforms a previous PFR-based one-zone model. The PaSPFR-IEM model captures the pressure rise that could not be described exactly using a simple one-zone model. The emissions of CO and hydrocarbons are also predicted well. Scatter plots of the marginal probability density function of CO2 and temperature reveal that the emissions of hydrocarbons and CO can be explained by stochastic particles that undergo incomplete combustion because they are trapped in the colder boundary layer or in the crevices.