Search

Downsized spark ignition engines running under high loads have become more and more attractive for car manufacturers because of their increased thermal efficiency and lower CO2 emissions. However, the occurrence of abnormal combustions promoted by the thermodynamic conditions encountered in such engines limits their practical operating range, especially in high efficiency and low fuel consumption regions. One of the main abnormal combustion is knock, which corresponds to an auto-ignition of end gases during the flame propagation initiated by the spark plug. Knock generates pressure waves which can have long-term damages on the engine, that is why the aim for car manufacturers is to better understand and predict knock appearance. However, an experimental study of such recurrent but non-cyclic phenomena is very complex, and these difficulties motivate the use of computational fluid dynamics for better understanding them. In the present article, large-eddy simulation (LES) is used as it is able to represent the instantaneous engine behavior and thus to quantitatively capture cyclic variability and knock. The proposed study focuses on the large-eddy simulation analysis of knock for a direct injection spark ignition engine. A spark timing sweep available in the experimental database is simulated, and 15 LES cycles were performed for each spark timing. Wall temperatures, which are a first-order parameter for knock prediction, are obtained using a conjugate heat transfer study. Present work points out that LES is able to describe the in-cylinder pressure envelope whatever the spark timing, even if the sample of LES cycles is limited compared to the 500 cycles recorded in the engine test bench. The influence of direct injection and equivalence ratio stratifications on combustion is also (MAPO) analyzed. Finally, focusing on knock, a Maximum Amplitude Pressure Oscillation analysis (MAPO) is conducted for both experimental and numerical pressure traces pointing out that LES well reproduces experimental knock tendencies.

A Large-Eddy Simulation (LES) modeling framework, dedicated to ultra-lean spark-ignition engines, is proposed and validated in the present work. A direct injection research engine is retained as benchmark configuration. The LES model is initially validated using the cold gas-exchange conditions by comparing numerical results with PIV (Particle Imaging Velocimetry) experimental data. Then, the fired configuration is investigated, combining ECFM (Extended Coherent Flame Model) turbulent combustion model with Adaptive Mesh Refinement (AMR). The capability of the model to reproduce experimental pressure envelope and cycle-to-cycle variability is assessed. Within the major scope of the work, a particular focus on the Combustion Cyclic Variability (CCV) is made correlating them with the variability encountered in the in-cylinder aerodynamic variations. R3P4. Finally two post-processing tools, Empirical Mode Decomposition (EMD) and Γ3p function, are proposed and combined to analyse for the first time the aerodynamic tumble-based in-cylinder velocity field. Both tools make it possible to get deeply into the insight and visualization of the flow field and to understand the links between its cyclic variability and the combustion cyclic variability.

International audience; Relying on two recent contributions by Massot et al. [SIAM J. Appl. Math. 70 (2010), 3203--3234] and Kah et al. [J. Comput. Phys. 231 (2012), 394--422], where a Eulerian Multi-Size Moment (EMSM) model for the simulation of polydisperse evaporating sprays has been introduced, we investigate the potential of such an approach for the robust and accurate simulation of the injection of a liquid disperse phase into a gas for automotive engine applications. The original model used a high order moment method in droplet size to resolve polydispersity, with built-in realizability preserving numerical algorithm of high order in space and time, but only dealt with one-way coupling and was restricted to fixed meshes. Extending the approach to internal combustion engine and fuel injection requires solving two major steps forward, while preserving the properties of robustness, accuracy and realizability: 1- the extension of the method and numerical strategy to two-way coupling with stable integration of potential stiff source terms, 2- the introduction of a moving geometry and meshes. We therefore present a detailed account on how we have solved these two issues, provide a series of verification of the proposed algorithm, showing its potential in simplified configurations. The method is then implemented in the IFP-C3D unstructured solver for reactive compressible flows in engines and validated through comparisons with a structured fixed mesh solver; it finally proves its potential on a free spray jet injection where it is compared to a Lagrangian approach and its reliability and robustness are assessed, thus making it a good candidate for realistic injection applications.

International audience; To assist industrial engine design, 3D simulations are increasingly used as they allow evaluation of a wide range of engine configurations and operating conditions and bring a comprehension of the underlying physics comple-mentary to experiments. While the gaseous flow description has reached a certain level of maturity, the multiphase flow description involving the liquid jet fuel injected into the chamber still faces some major challenges. There is a pressing need for a spray model that is time efficient and accurately describes the fuel-particle cloud dynamics downstream of the injector, which is an essential prerequisite for predictive combustion simulations. Due to the highly unsteady nature of the flow following the high-pressure injection process and the complexity of the flow regimes from separated/dense compressible phases to fully developed turbulent spray with evaporating droplets, Eulerian-Eulerian descriptions of two-phase flows are seen as very promising approaches towards realistic and predictive simulations of the mixing process. However they require some effort in terms of physical modeling and numerical analysis related to the more complex mathematical structure of the system of equations and to the unclosed terms appearing in space/time-average equations. Among the various challenges faced, one critical as-pect is to capture spray polydispersity in this framework. A review of recent developments that have permitted key advances in the spray modeling community is proposed in this paper. It is divided into four parts. First, an introduction to automotive spray modeling is provided. Then the formalisms for the description of the disperse region of an engine spray are presented with particular emphasis on the pros and cons of classical Lagrangian par-ticle methods versus Eulerian approaches. The third part presents the motivation for and the recent developments of Eulerian high-order moment methods for size polydispersion. Finally, the extension to fully two-way coupled interactions with the gas phase and the implementation of such methods for variable-geometry applications in CFD codes is described in the fourth part. Using realistic direct injection conditions computed with the IFP-C3D solver, the application and efficiency of Eulerian approaches is illustrated.

[EN] In this work, both experimental and numerical investigations have been carried out in order to improve the modelling of the vaporization of wall liquid-deposits in internal combustion engines. A comprehensive model is suggested for the vaporization of liquid films in the different boiling regimes, including nucleate boiling regime, the Leidenfrost boiling regime, as well as the transition boiling regime occurring between the two latter. This work extends the validity of the Liquid Film Boiling model (Habchi, Oil & Gas Science and Technology – Rev. IFP, Vol. 65, No. 2, 2010) for dispersed liquid films that may be formed when a dilute spray impinges a wall. A sub-grid liquid film is indeed considered when the wetted-area is smaller than the wall cell-face area. A sessile droplet model is used to estimate the wall area wetted by the liquid film and whether it is resolved by the grid or located in the sub-grid scale (SGS). In addition, a novel Leidenfrost vaporization model is proposed for spray droplets located near a hot wall. The above vaporisation/boiling models has been implemented in the Large-Eddy simulation (LES) AVBP code. The validation has been carried out using two different experiments. First, the experimental lifetime curve of a sessile droplet (Stanglmaier et al., SAE paper 2002-01-0838) has been used for a quantitative validation in the different boiling regimes. Second, the wall impingement of a heptane spray from a typical gasoline injector from Continental Automotive, has been simulated. The numerical results obtained under boiling conditions, are compared to the liquid film footprints and lifetime provided by the Refractive Index Matching (RIM) experiment which is described in this article.

Predictive simulation of liquid fuel injection in automotive engines has become a major challenge for science and applications. The key issue in order to properly predict various combustion regimes and pollutant formation is to accurately describe the interaction between the carrier gaseous phase and the polydisperse evaporating spray produced through atomization. For this purpose, we rely on the EMSM (Eulerian Multi-Size Moment) Eulerian polydisperse model. It is based on a high order moment method in size, with a maximization of entropy technique in order to provide a smooth reconstruction of the distribution, derived from a Williams-Boltzmann mesoscopic model under the monokinetic assumption [O. Emre (2014) PhD Thesis , Ecole Centrale Paris; O. Emre, R.O. Fox, M. Massot, S. Chaisemartin, S. Jay, F. Laurent (2014) Flow, Turbulence and Combustion 93 , 689-722; O. Emre, D. Kah, S. Jay, Q.-H. Tran, A. Velghe, S. de Chaisemartin, F. Laurent, M. Massot (2015) Atomization Sprays 25 , 189-254; D. Kah, F. Laurent, M. Massot, S. Jay (2012) J. Comput. Phys. 231 , 394-422; D. Kah, O. Emre, Q.-H. Tran, S. de Chaisemartin, S. Jay, F. Laurent, M. Massot (2015) Int. J. Multiphase Flows 71 , 38-65; A. Vie, F. Laurent, M. Massot (2013) J. Comp. Phys. 237 , 277-310]. The present contribution relies on a major extension of this model [M. Essadki, S. de Chaisemartin, F. Laurent, A. Larat, M. Massot (2016) Submitted to SIAM J. Appl. Math.], with the aim of building a unified approach and coupling with a separated phases model describing the dynamics and atomization of the interface near the injector. The novelty is to be found in terms of modeling, numerical schemes and implementation. A new high order moment approach is introduced using fractional moments in surface, which can be related to geometrical quantities of the gas-liquid interface. We also provide a novel algorithm for an accurate resolution of the evaporation. Adaptive mesh refinement properly scaling on massively parallel architectures yields a precise integration of transport in physical space limiting both numerical dissipation as well as the memory trace of the solver. A series of test-cases is presented and analyzed, thus assessing the proposed approach and its parallel computational efficiency while evaluating its potential for complex applications.

In the context of engine development for low fuel consumption and low emissions, combustion modeling is a challenging subject as the requirements for lower pollutant emissions mean that accurate temporal and spatial predictions of the heat release and species concentrations are needed. Among the various approaches developed recently to account for these processes in realistic configurations, tabulated techniques appear to be a promising approach. A good compromise can be obtained between the accuracy brought by detailed chemistry at limited computational cost. Tabulation approaches were first developed and successfully applied to stationary combustors at constant pressure. However, these approaches are severely limited for internal combustion engine applications: indeed, they are not capable of accounting for the enthalpy or energy loss due to the pressure work in the expansion stroke. To overcome this limitation a new Variable Volume Tabulated Homogeneous Chemistry (VVTHC) approach is proposed in this study. It provides the evolution of major species and radicals from the onset of auto-ignition up to the end of the expansion stroke for compression ignited and spark ignited engine applications. It is first validated for homogeneous engine cases where it compares very well to complex chemistry simulations. Implemented in a piston engine combustion model, it also succeeds in matching the burned gases volume variation behind a propagating flame at constant pressure and in reproducing the subsequent composition evolution. In these conditions, the new approach gives results similar to those obtained with classical constant pressure tabulations. Finally, a feasibility calculation of a realistic stratified Diesel engine configuration is presented. On these cases, a constant volume tabulation approach is shown to compare well with VVTHC and the experiment on the evolution of pressure but not on that of species mass fractions. This comparison demonstrates the interest of the VVTHC approach for piston engine applications.

The concerns about global warming and long-term lack of fossil fuels are strong incentives for alternative fuel research and adaptation of the internal combustion engines (ICE) to these fuels. Because it is free of any carbon compounds and can be produced from alternative sources, hydrogen is an interesting candidate for future ICE-based powertrains. However, the peculiar properties of hydrogen, among those its low density and its very high laminar flame speed, impose specific operating strategies and the adaptation of the conventional research tools. In this context, the 3D CFD models dedicated to combustion and pollutant prediction have to be modified. In the present work, the ECFM (Extended Coherent Flame Model) is adapted to hydrogen combustion through the addition of a new laminar flame speed correlation and a new laminar flame thickness expression. Furthermore, the prediction of the NOx emissions is performed with a modified version of the extended Zeldovitch model.

Concepts of surface density are exploited to model the coupled processes of atomization, vaporization, and combustion in turbulent jet flames formed by cryogenic propellants. An Eulerian framework is used to describe the two-phase flow formed by a coaxial injector fed by liquid oxygen and gaseous hydrogen. A transport equation for the density of the liquid interface represents the jet break-up, subsequent atomization, and space-time evolution of the liquid surface area which combined with a local description of the vaporization flux provides the volumetric source of gaseous oxygen. Subcritical and transcritical conditions are successively considered. Another transport equation for the flame surface density is then used to evaluate the mean reaction rate per unit volume. This reaction rate depends on an effective strain rate and on operating parameters such as the mean pressure prevailing in the combustion chamber. A simple analysis based on the fast chemistry limit indicates that the local rate of reaction varies like the square root of pressure and strain rate. A computational model combining these various elements is established. Calculated cryogenic flame structures are compared to recent experimental data corresponding to different gas to liquid momentum flux ratios and chamber pressures. The mean reaction rate spatial maps are in agreement with experimental OH light emission images which are used to locate the chemical reaction regions. Parametric studies indicate how the liquid or transcritical phase evolution modifies the flame pattern.

To assist the industrial engine design process, 3-D computational fluid dynamics simulations are widely used, bringing a comprehension of the underlying physics unattainable from experiments. However, the multiphase flow description involving a liquid fuel jet injected into the chamber is still in its early stages of development. There is a pressing need for a spray model that is time efficient and accurately describes the cloud of fuel and droplet dynamics downstream of the injector. Eulerian descriptions of the spray are well adapted to this highly unsteady configuration. The challenge is then to capture accurately both the evaporating spray polydispersity and its two-way mass, momentum and energy interactions with the surrounding gas phase in this framework. The Eulerian Multi-Size Moment (EMSM) model, a high-order (in size) moment model has proved to be well adapted for the treatment of polydispersity. Moreover, it requires less computational effort relative to competing methods, with a single section for the size phase space. Academic test cases have demonstrated its great potential for industrial applications using one-way coupling. Recent modeling and numerical development efforts have resulted in an extension to two-way coupling with an unconditionally stable and accurate numerical scheme, while preserving the size moment space. All these developments have been successfully verified through injection simulations in the context of laminar flow. In the present contribution, in preparation for comparisons with experimental data, an extension to two-way coupling to account for the droplet-gas turbulence interactions is developed and validated for academic cases using physical data for realistic internal combustion engine conditions.

A model for the auto-ignition of hydrocarbons applicable to 3D internal combustion engine calculations is proposed in this paper. The limits of classical methods using an auto-ignition delay are investigated when cool flame phenomena are present. A method based on tabulated reaction rates is presented to capture the early heat release induced by low temperature combustion. Cool flame ignition delay when present and cool flame fuel consumption are also tabulated. The reaction rate, fuel consumption, and cool flame ignition delay tables are built a priori from complex chemistry calculations. The reaction rates, which directly depend on instantaneous changes of thermodynamic conditions, are then integrated during the 3D engine calculation. The model is first validated through comparisons with complex chemistry calculations in constant and variable volume configurations where good agreement has been found. The model is then applied to both a Diesel computation with spray injection and residual gases, and a Diesel HCCI configuration. Comparisons with experimental results show that the auto-ignition essential features are well reproduced in these cases.

International audience; A major issue for the simulation of two-phase flows in engines concerns the modeling of the liquid disperse phase, either in the Lagrangian or the Eulerian approach. In the perspective of massively parallel computing, the Eulerian approach seems to be more suitable, as it uses the same algorithm as the gaseous phase solver. However taking into account the whole physics of a turbulent spray, especially in terms of polydispersity, requires an additional modeling effort. The Mesoscopic Eulerian Formalism (MEF) accounts for the effect of turbulence on the dispersed phase, and was extended to the Large Eddy Simulation framework, but is limited to monodisperse flows. In Massot et al. 2004, The influence of poly- dispersity on resolved and unresolved turbulent motions of the disperse phase was highlighted, and a first model was proposed, based on size-conditioned statistics. Starting from this idea, a coupling between the MEF and the Multifluid Approach (MA) is proposed. The MA decomposes the Eulerian phase into several fluids classes called sections, and corresponding to size intervals. Each section uses then size-conditioned closures. The original idea of this work is to use the MEF closures independently in each section, taking into account the mean droplet size of this section. This new approach, called Multifluid Mesoscopic Eulerian Formalism (MMEF), is then able to capture polydispersion with associated size-conditioned turbulent dynamics. First, the importance of polydispersity and the ability of MMEF to capture it are highlighted with a 0D evaporation case and a 2D vortex case, showing its impact on dynamics in both size and physical spaces. Then, the MMEF is applied to the MERCATO configuration of ONERA. Results are compared to monodisperse Eulerian, Lagrangian and experimental results.

International audience

Accepted for Publication; International audience; In this paper, we tackle the modeling and numerical simulation of sprays and aerosols, that is dilute gasdroplet flows for which polydispersity description is of paramount importance. Starting from a kinetic description for point particles experiencing transport either at the carrier phase velocity for aerosols or at their own velocity for sprays as well as evaporation, we focus on an Eulerian high order moment method in size and consider a system of partial differential equations (PDEs) on a vector of successive integer size moments of order 0 to N, N > 2, over a compact size interval. There exists a stumbling block for the usual approaches using high order moment methods resolved with high order finite volume methods: the transport algorithm does not preserve the moment space. Indeed, reconstruction of moments by polynomials inside computational cells coupled to the evolution algorithm can create N-dimensional vectors which fail to be moment vectors: it is impossible to find a size distribution for which there are the moments. We thus propose a new approach as well as an algorithm which is second order in space and time with very limited numerical diffusion and allows to accurately describe the advection process and naturally preserves the moment space. The algorithm also leads to a natural coupling with a recently designed algorithm for evaporation which also preserves the moment space; thus polydispersity is accounted for in the evaporation and advection process, very accurately and at a very reasonable computational cost. These modeling and algorithmic tools are referred to as the EMSM (Eulerian Multi Size Moment) model. We show that such an approach is very competitive compared to multi-fluid approaches, where the size phase space is discretized into several sections and low order moment methods are used in each section, as well as with other existing high order moment methods. An accuracy study assesses the order of the method as well as the low level of numerical diffusion on structured meshes. Whereas the extension to unstructured meshes is provided, we focus in this paper on cartesian meshes and two 2D test-cases are presented: Taylor-Green vortices and turbulent free jets, where the accuracy and efficiency of the approach are assessed.

International audience; The bivariate two dimensional empirical mode decomposition (Bivariate 2D-EMD) is extended to estimate the turbulent fluctuations and to identify cycle-to-cycle variations (CCV) of in-cylinder flow. The Bivariate 2D-EMD is an adaptive approach that is not restricted by statistical convergence criterion, hence it can be used for analyzing the nonlinear and non-stationary phenomena. The methodology is applied to a high-speed PIV dataset that measures the velocity field within the tumble symmetry plane of an optically accessible engine. The instantaneous velocity field is decomposed into a finite number of 2D spatial modes. Based on energy considerations, the in-cylinder flow large-scale organized motion is separated from turbulent fluctuations. This study is focused on the second half of the compression stroke. For most of the cycles, the maximum of turbulent fluctuations is located between 50 and 30 crank angle degrees before top dead center (TDC). In regards to the phase-averaged velocity field, the contribution of CCV to the fluctuating kinetic energy is approximately 55% near TDC.

Ultra-lean burn conditions (λ > 1.8) is seen as a way for improving efficiency and reducing emissions of spark-ignition engines. It raises fundamental issues in terms of combustion physics and its modeling, among which the significant reduction of the laminar flame speeds and increase of the laminar flame thickness, as well as an increased sensitivity to local fuel/air equivalence ratio variations are essential to be accounted for as compared to conventional stoichiometric mixture conditions. In particular, the effects of the modified laminar flame characteristics on flame stretch during the early flame development in a spark ignited gasoline engine can be expected to become of importance. In the present work, a Large-Eddy Simulation combustion approach is presented and applied to the study of the cycle-to-cycle combustion variations of a direct injection gasoline engine operating both in stoichiometric and ultra-lean burn conditions. The Coherent Flame Model approach is used and enriched via a correlation for the laminar flame velocity accounting for nonlinear stretch effects. The stretched flame calculations are validated against experimental results. Then, different engine operating points are computed in stoichiometric and ultra-lean burn conditions assessing the capacity of the approach to reproduce variations of combustion regimes. The results are analyzed in terms of cycle-to-cycle combustion variabilities and the influence of the spark-plug orientation is studied. Finally, a detailed analysis of the flame development is presented with a particular emphasis on the analysis of the initial flame kernel development accounting for stretch effects in lean conditions and the analysis of extreme cycles in lean burn. A strong reduction of the flame velocity by one third was observed for lean-burn conditions due to non-linear stretch effects occurring during the early stage of the flame development while almost no change was observed for stoichiometric conditions. Moreover, the proposed approach was capable of handling the various conditions featuring significantly different combustion regimes (one order of magnitude for the Karlovitz number) with only a minor change in the model parameterization.