Large-eddy simulations of a stratified-charge direct-injection spark-ignition engine: Comparison with experiment and analysis of cycle-to-cycle variations

Multiple-cycle large-eddy simulations (LES) have been performed for an optically accessible, single-cylinder, four-stroke-cycle, spray-guided direct-injection spark-ignition (SG-DISI) engine operating in a stratified globally fuel-lean mode. The simulations combine a standard Smagorinsky turbulence model, a stochastic Lagrangian parcel method for liquid fuel injection and fuel spray modeling, a simple energy-deposition spark-ignition model, and a modified thickened flame model for turbulent flame propagation through highly stratified reactant mixtures. Comparisons between simulations and experiments include individual-cycle and ensemble-average pressure and apparent-heat-release-rate traces, individual-cycle and ensemble-average indicated mean effective pressures (IMEP), and instantaneous two-dimensional vapor-equivalence-ratio contours. Although the number of LES cycles is small (35), the results show that the simulations are able to capture the global combustion behavior that is observed in the experiments, including cycle-to-cycle variations. The simulation results are then analyzed further to provide insight into the conditions that lead to misfire versus robust combustion. As has been reported in earlier experimental and LES studies for homogeneous-charge SI engines, local conditions in the vicinity of the spark gap at the time of ignition largely determine the subsequent flame development. However, in contrast to homogeneous-charge engines, no single local or global quantity correlates as strongly with the eventual peak pressure or IMEP for each cycle. Rather, it is the interplay among the early flame kernel, the velocity field that it experiences, and the fuel distribution that it encounters that ultimately determines the fate of each combustion event. Deeper analysis and quantitative statistical comparisons between experiments and simulations will require the simulation of larger numbers of engine cycles.     

Studies of the flow and turbulence fields in a turbulent pulsed jet flame using LES/PDF

A turbulent piloted jet flame subject to a rapid velocity pulse in its fuel jet inflow is proposed as a new benchmark case for the study of turbulent combustion models. In this work, we perform modelling studies of this turbulent pulsed jet flame and focus on the predictions of its flow and turbulence fields. An advanced modelling strategy combining the large eddy simulation (LES) and the probability density function (PDF) methods is employed to model the turbulent pulsed jet flame. Characteristics of the velocity measurements are analysed to produce a time-dependent inflow condition that can be fed into the simulations. The effect of the uncertainty in the inflow turbulence intensity is investigated and is found to be very small. A method of specifying the inflow turbulence boundary condition for the simulations of the pulsed jet flame is assessed. The strategies for validating LES of statistically transient flames are discussed, and a new framework is developed consisting of different averaging strategies and a bootstrap method for constructing confidence intervals. Parametric studies are performed to examine the sensitivity of the predictions of the flow and turbulence fields to model and numerical parameters. A direct comparison of the predicted and measured time series of the axial velocity demonstrates a satisfactory prediction of the flow and turbulence fields of the pulsed jet flame by the employed modelling methods.     

Identification and analysis of instability in non-premixed swirling flames using LES

Large eddy simulations (LES) of turbulent non-premixed swirling flames based on the Sydney swirl burner experiments under different flame characteristics are used to uncover the underlying instability modes responsible for the centre jet precession and large scale recirculation zone. The selected flame series known as SMH flames have a fuel mixture of methane-hydrogen (50:50 by volume). The LES solves the governing equations on a structured Cartesian grid using a finite volume method, with turbulence and combustion modelling based on the localised dynamic Smagorinsky model and the steady laminar flamelet model respectively. The LES results are validated against experimental measurements and overall the LES yields good qualitative and quantitative agreement with the experimental observations. Analysis showed that the LES predicted two types of instability modes near fuel jet region and bluff body stabilised recirculation zone region. The mode I instability defined as cyclic precession of a centre jet is identified using the time periodicity of the centre jet in flames SMH1 and SMH2 and the mode II instability defined as cyclic expansion and collapse of the recirculation zone is identified using the time periodicity of the recirculation zone in flame SMH3. Finally frequency spectra obtained from the LES are found to be in good agreement with the experimentally observed precession frequencies.     

Experimental and numerical study of a conical turbulent partially premixed flame

The structure and dynamics of a turbulent partially premixed methane/air flame in a conical burner were investigated using laser diagnostics and large-eddy simulations (LES). The flame structure inside the cone was characterized in detail using LES based on a two-scalar flamelet model, with the mixture fraction for the mixing field and level-set G-function for the partially premixed flame front propagation. In addition, planar laser induced florescence (PLIF) of CH and chemiluminescence imaging with high speed video were performed through a glass cone. CH and CH₂O PLIF were also used to examine the flame structures above the cone. It is shown that in the entire flame the CH layer remains very thin, whereas the CH₂O layer is rather thick. The flame is stabilized inside the cone a short distance above the nozzle. The stabilization of the flame can be simulated by the triple-flame model but not the flamelet-quenching model. The results show that flame stabilization in the cone is a result of premixed flame front propagation and flow reversal near the wall of the cone which is deemed to be dependent on the cone angle. Flamelet based LES is shown to capture the measured CH structures whereas the predicted CH₂O structure is somewhat thinner than the experiments.     

Experimental and numerical investigation of a stagnation pulverised coal flame

A multi-stream Flamelet Progress Variable (FPV) model, specifically developed for coal combustion, is proposed. The model accounts for the different fuel streams associated with the volatile and char burnout products. The applicability of the new FPV model is investigated in a laminar stagnation pulverized coal flame. The flame considered is a premixed mixture of CH₄, O₂ and N₂, carrying pulverized coal particles, stabilized in an impinging wall. Spontaneous emissions of OH*, CH* and C${}_2^{*}$ are measured to identify the flame. The 1D numerical simulations of the experimental conditions are able to reproduce the main features of the flame. The applicability of the multi-stream FPV model to coal combustion is further evaluated with the a posteriori analysis of the FPV results, comparing the results with a reference model, where the species are fully transported and the chemistry directly evaluated. Then, with the budget analysis, the influence of the control variables used to build the look-up table is assessed by examining the conditional contributions to the overall transport terms of scalar quantities (e.g. species, temperature). The results of both analyses show that the proposed multi-stream FPV model can accurately predict the main features of coal combustion, with only minor issues related to the manifold used to build the look-up table.     

LES studies of the flow in a swirl gas combustor

Environmental and other practical concerns have led to the development of compact gas turbine combustors burning lean mixtures leading to potentially low CO and NO_x emissions. The compact design requires efficient atomization and mixing together with a compact premixed flame. Associated with these requirements are higher temperatures, increased heat transfer, and thermal load, thus increasing the danger of combustion instabilities (causing performance deterioration and excessive mechanical loads), and possible off-design operation. Numerical simulations of reacting flows are well suited to address these issues. To this end, large eddy simulation (LES) is particularly promising. The philosophy behind LES is to explicitly simulate the large scales of the flow and the thermochemistry, affected by boundary conditions whilst modeling only the small scales, including the interaction between the flow and the combustion processes. Here, we examine the flow and the flame in a model gas turbine combustor (General Electric’s lean premixed dry low NO_x LM6000) to evaluate the potential of LES for design studies of engineering applications and to study the effects of the combustor confinement geometry on the flow and on the flame dynamics. Two LES models, a Monotone Integrated LES model with 1 and 2 step Ahrrenius chemistry, and a fractal flame-wrinkling LES model coupled to a conventional one-equation eddy-viscosity subgrid model, are used. Reasonable agreement is found when comparing predictions with experimental data and with other LES computations of the same case. Furthermore, the combustor confinement geometry is found to strongly affect the vortical flow, and hence also the flame and its dynamics.     

Evaluation of a flamelet/progress variable approach for pulverized coal combustion in a turbulent mixing layer

A steady flamelet/progress variable (FPV) approach for pulverized coal flames is employed to simulate coal particle burning in a turbulent shear and mixing layer. The configuration consists of a carrier-gas stream of air laden with coal particles that mixes with an oxidizer stream of hot products from lean combustion. Carrier-phase DNS (CP-DNS) are performed, where the turbulent flow field is fully resolved, whereas the coal is represented by Lagrangian point particles. CP-DNS with direct chemistry integration is performed first and provides state-of-the-art validation data for FPV modeling. In a second step the control variables for FPV are extracted from the CP-DNS and used to test if the tabulated manifold can correctly describe the reacting flow (a priorianalysis). Finally a fully coupled a posteriori FPV simulation is performed, where only the FPV control variables are transported, and the chemical state is retrieved from the table and fed back to the flow solver. The a priori results show that the FPV approach is suitable for modeling the complex reacting multiphase flow considered here. The a posteriori data is similarly in good agreement with the reference CP-DNS, although stronger deviations than a priori can be observed. These discrepancies mainly appear in the upper flame (of the present DNS), where premixing and highly unsteady extinction and re-ignition effects play a role, which are difficult to capture by steady non-premixed FPV modeling. However, the present FPV model accurately captures the lower, more stable flame that burns in non-premixed mode.     

Large eddy simulation of bluff-body stabilized swirling non-premixed flames

Large eddy simulations (LES) using a subgrid mixing and combustion model are carried out to study two bluff-body stabilized swirling non-premixed flames (SM1 and SMA2). The similarities and differences between the two flames are highlighted and discussed. Flow features, such as, the recirculation zone (RZ) size and the flame structure are captured accurately in both cases. The SM1 flame shows a toroidal RZ just behind the bluff body and a vortex breakdown bubble (VBB) downstream. In addition, a highly rotational non-recirculating region in-between the RZ and VBB is observed as well. On the other hand, the SMA2 shows a single elongated recirculation zone downstream the bluff body. Flame necking is observed downstream the bluff body for the SM1 flame but not for the SMA2 flame. The time-averaged velocity and temperature comparison also shows reasonable agreement. The study shows that the sensitivity of the flame structure to inflow conditions can be captured in the present LES without requiring any model changes.     

Generation of Turbulent Inflow Data for Spatially-Developing Boundary Layer Simulations

A method for generating three-dimensional, time-dependent turbulent inflow data for simulations of complex spatially developing boundary layers is described. The approach is to extract instantaneous planes of velocity data from an auxiliary simulation of a zero pressure gradient boundary layer. The auxiliary simulation is also spatially developing, but generates its own inflow conditions through a sequence of operations where the velocity field at a downstream station is rescaled and re-introduced at the inlet. This procedure is essentially a variant of the Spalart method, optimized so that an existing inflow–outflow code can be converted to an inflow-generation device through the addition of one simple subroutine. The proposed method is shown to produce a realistic turbulent boundary layer which yields statistics that are in good agreement with both experimental data and results from direct simulations. The method is used to provide inflow conditions for a large eddy simulation (LES) of a spatially evolving boundary layer spanning a momentum thickness Reynolds number interval of 1530–2150. The results from the LES calculation are compared with those from other simulations that make use of more approximate inflow conditions. When compared with the approximate inflow generation techniques, the proposed method is shown to be highly accurate, with little or no adjustment of the solution near the inlet boundary. In contrast, the other methods surveyed produce a transient near the inlet that persists several boundary layer thicknesses downstream. Lack of a transient when using the proposed method is significant since the adverse effects of inflow errors are typically minimized through a costly upstream elongation of the mesh. Extension of the method for non-zero pressure gradients is also discussed.     

Large-eddy simulation of MILD combustion using partially stirred reactor approach

Subgrid-scale (SGS) parameterization and method for calculating filtered reaction rate are critical components of an accurate large-eddy simulation (LES) of turbulent flames. In this study, we integrate gradient-type structural SGS models with a partially stirred reactor approach by using detailed chemical kinetics to simulate a turbulent methane/hydrogen jet flame under moderate or intense low-oxygen dilution (MILD) conditions. The study examines two oxygen dilution levels. The framework is assessed through a systematic and comprehensive comparison of temperature, and mass fractions of major and minor species with experimental data and other reference simulation results. Overall, the statistics of the combustion field show excellent agreement with measurements at different axial locations, and a significant improvement compared to some previous simulations. It suggests that the proposed nonlinear LES framework is able to accurately model MILD combustion with reasonable computational cost.     

Effects of buoyancy on turbulent premixed V-flames by large-eddy simulation

Buoyancy effects on turbulent premixed V-flames are investigated under normal gravity (+g) and reversed gravity (–g). Numerical simulations employ large eddy simulation (LES) with a dynamic model for sub-grid scale stress. With the assumption of fast chemistry combustion, a progress variable c-equation is applied to describe the flame front propagation. The equations are solved using a projection-based fractional step method in two dimensions for low-Mach number flows. Computed LES results of buoyancy effects on flame angle and flame brush thickness are consistent with those obtained from experiments. In both +g and –g conditions, the effects of buoyancy become important with increase in Richardson number (Ri). Buoyancy force tends to close up the flame under +g, but has the opposite effect under –g. Buoyancy force also suppresses flame wrinkling in +g and enhances wrinkling in –g. While there is a lack of experimental data available, computed axial velocity is shown to be significantly affected by buoyancy downstream from the flame holder under moderate Reynolds number.     

Computational investigations of the coupling between transient flame dynamics and thermo-acoustic instability in a self-excited resonance combustor

Combustion instability due to thermo-acoustic interactions is a critical combustion problem that requires a thorough understanding because of its adverse impact on stable and reliable operation of combustors in high-speed propulsion devices like gas turbines and rockets. This work conducts computational investigations of the coupling between the transient flame dynamics such as the ignition delay and local extinction and the thermo-acoustic instability developed in a self-excited resonance combustor to gain deep insights into the mechanisms of thermo-acoustic instability. A 2D modelling framework that employs different flamelet models (the steady flamelet model and the flamelet/progress variable approach) is developed to enable the examination of the effect of the transient flame dynamics caused by the strong coupling of the turbulent mixing and finite-rate chemical kinetics on the occurrence of thermo-acoustic instability. The models are validated by using the available experimental data for the pressure signal. Parametric studies are performed to examine the effect of the occurrence of the transient flame dynamics, the effect of artificial amplification of the Damköhler number, and the effect of neglecting mixture fraction fluctuations on the predictions of the thermo-acoustic instability. The parametric studies reveal that the occurrence of transient flame dynamics has a strong influence on the onset of the thermo-acoustic instability. Further analysis is then conducted to localise the effect of a particular flame dynamic event, the ignition delay, on the thermo-acoustic instability. The reverse effect of the occurrence of the thermo-acoustic instability on the transient flame dynamics in the combustor is also investigated by examining the temporal evolution of the local flame events in conjunction with the pressure wave propagation. The above observed two-way coupling between the transient flame dynamics (the ignition delay) and the thermo-acoustic instability provides a plausible mechanism of the self-excited and sustained thermo-acoustic instability observed in the combustor despite the fact that the results are obtained from 2D simulations. The same analysis is expected to be extensible to fully 3D simulations.     

Modelling effects of subgrid-scale mixture fraction variance in LES of a piloted diffusion flame

A posteriori analysis of the statistics of two large-eddy simulation (LES) solutions describing a piloted methane–air (Sandia D) flame is performed on a series of grids with progressively increased resolution reaching about 10.5 million cells. Chemical compositions, density and temperature fields are modelled with a steady flamelet approach and parametrised by the mixture fraction. The difference between the LES solutions arises from a different numerical treatment of the subgrid scale (SGS) mixture fraction variance – an important quantity of interest in non-premixed combustion modelling. In the first case (model I), the variance transport equation is solved directly, while in the second (model II), an equation for the square of the mixture fraction is solved, and the variance is computed from its definition. The comparison of the LES solutions is based on the convergence properties of their statistics with respect to the turbulence resolution length scale. The dependence of the LES statistics is analysed for velocity and the mixture fraction fields, and tested for convergence. For the most part, the statistics converge for the finest grids, but the variance of the mixture fraction shows some residual grid dependence in the high-gradient regions of the jet near field. The SGS variance given by model I exhibits realisability everywhere, whereas in regions of the flame model II is non-realisable, predicting negative variances. Furthermore, the LES statistics of model I exhibit superior convergence behaviour.     

甲烷平面射流扩散火焰的大涡模拟

本文对甲烷-空气平面自由射流扩散火焰进行了大涡模拟,采用分步投影法求解动量方程,湍流亚格子项采用动态模式模拟,化学反应速率亚格子项采用动态相似模式模拟,压力泊松方程采用修正的循环消去法快速求解,空间方向采用二阶精度的差分格式,在时间方向上采用二阶精度的显式差分格式。模拟结果给出了湍流扩散火焰的瞬态发展变化过程,表明射流扩散火焰的发展过程存在着“湍流控制”和“化学反应控制”两个不同阶段。 “湍流控制”阶段仅存在于火焰发展初期的极短时间内。     

Visualisation of results from LES of combustion

We examine the role of visualisation in the context of LES simulations of premixed turbulent combustion. The physical processes involved in premixed turbulent combustion are extremely complex, and the modelling of both the turbulence (via LES) and the combustion (via flame-wrinkling models) is difficult. Appropriate visualisation is required to understand the behaviour of the models, and ultimately to understand better the flow processes which are important in many industrial applications. We examine visualisations of two specific cases; simple flame kernel growth in a box of turbulence, and combustion behind a backward-facing step. A number of visualisation techniques are used to produce results that are similar to experimentally determined Schlieren and Mie photography for the flame kernel. In addition, isosurfaces of the reaction regress variable coloured by the laminar flame speed and sub-grid wrinkling are also plotted in an attempt to gain deeper insight into the physics of turbulent combustion in the context of these particular cases. Finally we discuss the role of the WWW in the continuing development of scientific visualisation techniques.     

DNS and LES of spark ignition with an automotive coil

The cycle to cycle combustion variability which is observed in spark-ignition engines is often caused by fluctuations of the early flame development. LES can be exploited for a better understanding and mastering of their origins. For that purpose appropriate models taking into account energy deposition, mixture ignition and transition to propagation are necessary requirements. This paper presents first DNS and LES of spark ignition with a real automotive coil and simplified pin-pin electrodes. The electrical circuit characteristics are provided by ISSIM while the energy deposition is modelled by Lagrangian particles. The ignition model is first evaluated in terms of initial spark radius on a pin-pin ignition experiment in pure air performed at CORIA and EM2C laboratories, showing that it pilots the radius of the torus formed by the initial shock wave. DNS of a quiescent lean propane/air mixture are then performed with this ignition system and a two-step mechanism. The impact of the modelled transferred energy during glow phase as well as the initial arc radius on the minimum ignition energy (MIE) are examined and compared to experimental values. Replacing the two-step chemistry by an analytically reduced mechanism leads to similar MIE but shows a different ignition kernel shape. Finally, LES of turbulent ignition using a Lagrangian arc model show a realistic prediction of the arc shape and its important role on the energy transfer location and thus on the flame kernel shape.     

Flame structure analysis and flamelet progress variable modelling of strained coal flames

Strained two-phase pulverised coal flames in a counterflow configuration are investigated numerically. Three operating conditions with different coal-to-primary-air ratios and inlet velocities were evaluated in order to establish different flame regimes. At first, the two-phase flow of the fully resolved reference cases is calculated solving the transport equation for the species and directly evaluating the reaction rates. Different flame structures are identified using the heat release rate and the chemical explosive mode as markers, showing that complex structures with a combination of lean premixed and non-premixed flames can be observed in strained counterflow coal flames. In addition to the fully resolved simulation, the suitability of the Flamelet-Progress Variable (FPV) model is investigated. Both premixed and non-premixed tables are employed. At first, the suitability of the look-up tables is evaluated by means of an a priori analysis, using the fully resolved simulations as reference solutions, showing that the non-premixed flamelet table correctly predicts the structure of the strained coal flames, while the premixed table shows sensible deviations in terms of temperature and species, especially at rich conditions. Finally, the a posteriori analysis shows that the fully coupled FPV model with a non-premixed flamelet look-up table can accurately predict strained coal flames.     

Eulerian transported probability density function sub-filter model for large-eddy simulations of turbulent combustion

Reactive flow simulations using large-eddy simulations (LES) require modelling of sub-filter fluctuations. Although conserved scalars like mixture fraction can be represented using a beta-function, the reactive scalar probability density function (PDF) does not follow an universal shape. A one-point one-time joint composition PDF transport equation can be used to describe the evolution of the scalar PDF. The high-dimensional nature of this PDF transport equation requires the use of a statistical ensemble of notional particles and is directly coupled to the LES flow solver. However, the large grid sizes used in LES simulations will make such Lagrangian simulations computationally intractable. Here we propose the use of a Eulerian version of the transported-PDF scheme for simulating turbulent reactive flows. The direct quadrature method of moments (DQMOM) uses scalar-type equations with appropriate source terms to evolve the sub-filter PDF in terms of a finite number of delta-functions. Each delta-peak is characterized by a location and weight that are obtained from individual transport equations. To illustrate the feasibility of the scheme, we compare the model against a particle-based Lagrangian scheme and a presumed PDF model for the evolution of the mixture fraction PDF. All these models are applied to an experimental bluff-body flame and the simulated scalar and flow fields are compared with experimental data. The DQMOM model results show good agreement with the experimental data as well as the other sub-filter models used.     

Flame characterisation of gas-assisted pulverised coal combustion using FPV-LES

A multiphase flamelet/progress variable (FPV) model for the large eddy simulation (LES) of gas-assisted pulverised coal combustion (PCC) is developed. The target of the simulation is the Darmstadt turbulent gas-assisted swirling solid fuel combustion chamber. The coal particles are treated as Lagrangian point particles, the position, momentum and energy of which are tracked. The gas phase is described by the low-Mach Navier-Stokes equations alongside the Eulerian transport equations of the governing variables for the FPV model. The set of chemical states of the PCC flame is pre-tabulated in a six-dimensional flamelet table and determined by the mixing of the primary fuel stream, volatiles and char off-gases with the oxidising air, the progress of chemical reactions, the interphase heat transfer, as well as sub-grid scale variations. A presumed $\beta$-PDF approach for the total mixture fraction is applied to capture sub-grid scale effects. The discrete ordinate method (DOM) with the weighted sum of grey gases model (WSGGM) is employed to model radiation. The FPV-LES results are validated against the experimental evidence and a good agreement of the predicted mean and RMS velocities, as well as the mean gas temperature between experiments and simulations is obtained. The contributions of the pilot, volatile and char off-gas fuel streams to the coal flame are analysed. It is found that most regions of the furnace are dominated by either pilot or volatile combustion, while char conversion only occurs in the far downstream and outer furnace regions. The pilot gas dominates the near-wall region inside the quarl, whereas the volatile gas mainly released from small particles dominates a first volatile combustion zone in the interior of the internal recirculation zone. Larger particles heat up more slowly and release their volatile content further downstream, leading to a secondary volatile combustion zone.