No abstract

Low-speed pre-ignition has become of great concern since it represents a strong limit to any further downsizing of gasoline engines. The increase of pressure and temperature inside the combustion chamber at high loads can indeed lead to a premature auto-ignition of the mixture and to severe engine damages. Several hypotheses have been formulated but there is no consensus to explain and demonstrate the cause of low-speed pre-ignition. This article provides an overview of the mechanisms that might lead to low-speed pre-ignition. A general discussion and a synthesis of the favourable conditions for auto-ignition are first introduced. Fundamentals of auto-ignition in homogeneous gaseous phase are discussed and confronted to concrete experimental observations of low-speed pre-ignition. The real operating conditions of modern gasoline engines complicate the analysis of auto-ignition because spatial and cycle-to-cycle fluctuations are linked to various physicochemical phenomena involved in the mixture preparation process. At the same time, the increase in temperature and pressure with load enhances the mixture reactivity and any perturbation at the end of the compression can unbalance the mixture and lead to uncontrolled low-speed pre-ignition. A classification of these mechanisms into three groups is proposed to categorize pre-ignitions resulting from auto-ignitions in the gaseous phase, or from auto-ignitions resulting from interactions between either a liquid or a solid phase and the gaseous mixture. Finally, the impact of mixture dilution with burned gases, mixture temperature, liquid films and deposits is examined for different engine settings, fuels and charge motions to illustrate each of these three groups.

Lean burn gasoline spark-ignition engines can support the reduction of CO 2 emissions for future hybrid passenger cars. Very high efficiencies and very low NOx raw emissions can be achieved, if relative air/fuel ratios λ of 2 and above can be reached. The biggest challenge here is to assure a reliable ignition process and to enhance the fuel oxidation in order to achieve a short burn duration and a good combustion stability.This article aims at introducing an innovative combustion system fully optimized for ultra-lean operation and very high efficiency. Thereto, a new cylinder head concept has been realized with high peak firing pressure capability and with a low surface-to-volume ratio at high compression ratios. 1D and 3D simulations have been performed to optimize the compression ratio, charge motion and intake valve lift. Numerical calculations also supported the development of the ignition system. Stable ignition and fast flame propagation were achieved thanks to a centrally located active prechamber which allows to control the air/fuel ratio independently of the air/fuel ratio in the main combustion chamber. Experimental investigations have then been performed with a single cylinder engine to demonstrate the capabilities of this new combustion system in a sweet spot operating point. A maximal indicated thermal efficiency of 47% was achieved at λ = 2 with optimized injection settings in the pre and main combustion chambers. The fuel efficiency could be maximized thanks to a fast and knock-free combustion process. Compared to the reference operation with stoichiometric air/fuel ratio, only a seventieth of the NOx raw emissions were measured (i. e. 50 ppm), and the particulate mass emissions were halved. The energy balance analysis points out that these promising results could be further improved by working on the reduction of the unburnt hydrocarbon emissions and by jointly optimizing the scavenging process.

Recent developments on highly downsized spark ignition engines have been focused on knocking behaviour improvement and the most advanced technologies combination can face up to 2.5 MPa IMEP while maintaining acceptable fuel consumption. Unfortunately, knocking is not the only limit that strongly downsized engines have to confront. The improvement of low-end torque is limited by another abnormal combustion which appears as a random pre-ignition. This violent phenomenon which emits a sharp metallic noise is unacceptable even on modern supercharged gasoline engines because of the great pressure rise that it causes in the cylinder (up to 20 MPa).The phases of this abnormal combustion have been analysed and a global mechanism has been identified consisting of a local ignition before the spark, followed by a propagating phase and ended by a massive autoignition. This last step finally causes a steep pressure rise and pressure oscillations.One of our objectives was to evaluate the sensitivity of an engine to pre-ignition regarding its design and settings. Therefore, in addition to our comprehension work, we have developed a first methodology based upon robust statistics to define new reliable and repeatable criteria to quantify this stochastic pre-ignition but also to detect each of its occurrence, suggesting the possibility of an on-line detection during steady state and transient operation as well. The statistical approach also showed that the distributions of well chosen combustion indicators are strongly altered by pre-ignitions. A second methodology was then defined to evaluate the influence of different parameters on pre-ignition by quantifying this alteration. This analysis notably gives the opportunity to achieve a deeper analysis of pre-ignition during engine development on test bench.

Heat losses through combustion chamber walls are a well-known limiting factor for the overall efficiency of internal combustion engines. Thermal insulation of the walls has the potential to decrease substantially these heat losses. However, evaluating numerically the effect of coating and of its location in the combustion chamber and then design an optimized combustion system require the use of high-fidelity engine models. The objective of this article is to present the whole workflow implying the use of three-dimensional computational fluid dynamics techniques with conjugate heat transfer (CHT) models to investigate the potential benefits of a coating on a passenger car Diesel engine. First, the baseline combustion system is modeled, using CHT models to solve in a coupled simulation the heat transfers between the fluid in the intake and exhaust lines and in the combustion chamber, on one hand, and the solid piston, head and valves, on the other hand. Based on this setup, a second simulation is performed, modeling a thermo-swing insulation on all combustion chamber walls by a contact resistance, neglecting its thermal inertia to keep a manageable computational cost. Results show a decrease of 3.3% in fuel consumption with an increase in volumetric efficiency. However, decoupled one-dimensional/three-dimensional simulations highlight the inaccuracy of these results and the necessity to model the coating thermal inertia, as they show an overestimation of the heat insulation rate and, consequently, of the gain in fuel consumption (−2.1% instead of −1.6%), for a coating on the piston with no thermal inertia.

Since the homologation drive cycles are not all entirely representative of real everyday driving conditions, additional cycles must be used to properly quantify the energy consumption of electric vehicles. Three methodologies have been developed to objectively determine a limited number of additional drive cycles. The first methodology is based upon a simple comparison of velocity profiles expressed on the same time scale. The second is based on a correlation analysis of various statistical criteria defined for each cycle. Finally, the third methodology is based on an automatic clustering technique ( K-means algorithm). These three methodologies have been used to analyse 18 drive cycles including official homologation cycles used in Europe (NEDC), Japan (JC08) and in the US (FTP72), and transient non-official cycles representing real world driving patterns (Artemis, Hyzem and Eurev cycles). When applied to these cycles, all these approaches allow the identification of a reduced number of five representative cycles enabling an efficient characterization in terms of the performance of electric vehicles in urban, extra-urban and highway conditions.

Performance of lean-burn gasoline spark-ignition engines can be enhanced through hydrogen supplementation. Thanks to its physicochemical properties, hydrogen supports the flame propagation and extends the dilution limits with improved combustion stability. These interesting features usually result in decreased emissions and improved efficiencies. This article aims at demonstrating how hydrogen can support the combustion process with a modern combustion system optimized for high dilution resistance and efficiency. To achieve this, chemical kinetics calculations are first performed in order to quantify the impacts of hydrogen addition on the laminar flame speed and on the auto-ignition delay times of air/gasoline mixtures. These data are then implemented in the extended coherent flame model and tabulated kinetics of ignition combustion models in a specifically updated version of the CONVERGE code. Three-dimensional computational fluid dynamics engine calculations are performed at λ = 2 with 3% v/v of hydrogen for two operating points. At low load, numerical investigations show that hydrogen enhances the maximal combustion speed and the flame growth just after the spark which is a critical aspect of combustion with diluted mixtures. The flame front propagation is also more isotropic when supported with hydrogen. At mid load, hydrogen improves the combustion speed and also extends the auto-ignition delay times resulting in a better knocking resistance. A maximal indicated efficiency of 48.5% can thus be reached at λ = 2 thanks to an optimal combustion timing.

Two main abnormal combustions are observed in spark-ignition engines: knock and lowspeed pre-ignition. Controlling these abnormal processes requires understanding how auto-ignition is triggered at the "hot spot" but also how it propagates inside the combustion chamber. The original theory regarding the auto-ignition propagation modes was defined by Zeldovich and developed by Bradley who highlighted different modes by considering various hot spot characteristics and thermodynamic conditions around the hot spot. Two dimensionless parameters (ε, ξ) were then defined to classify these modes and a so-called detonation peninsula was obtained for H 2 -CO-air mixtures.Similar simulations as those performed by Bradley et al. are undertaken to check the relevancy of the original detonation peninsula when considering realistic fuels used in modern gasoline engines. First, chemical kinetics calculations in homogeneous reactor are performed to determine the auto-ignition delay time τ i , and the excitation time τ e of E10-air mixtures in various conditions (calculations for a RON 95 TRF surrogate with 42.8% isooctane, 13.7% n-heptane, 43.5% toluene, and using the LLNL kinetic mechanism considering 1388 species and 5935 reactions). Results point out that H 2 -COair mixtures are much more reactive than E10-air mixtures featuring much lower excitation times τ e . The resulting maximal hot spot reactivity ε is thus limited which also restrains the use of the detonation peninsula for the analysis of practical occurrences of auto-ignition in gasoline engines.The tabulated (τ i , τ e ) values are then used to perform 1D LES of auto-ignition propagation considering different hot spots and thermodynamic conditions around them. The detailed analysis of the coupling conditions between the reaction and pressure waves shows thus that the different propagation modes can appear with gasoline and that the original detonation peninsula can be reproduced, confirming for the first time that the propagation mode can be well defined by the two non-dimensional parameters for more realistic fuels. KeywordsAuto-ignition, Deflagration, Detonation, Fuel Reactivity, LES first observations of low-speed pre-ignition (LSPI) date back to the beginning of years 2000 [2]. However, in both cases, the triggering and the development of the abnormal combustion process rely on the auto-ignition characteristics of the air/fuel mixture. In order to control these abnormal phenomena, it is necessary not only to better understand how and when an auto-ignition can be triggered by "hot spots", but also how it will propagate inside the combustion chamber since the auto-ignition intensity and the potential resulting engine damages are linked to both aspects. Different approaches such as the Livengood-Wu integral can be used to predict the autoignition temporal onset. However, advanced tools and methodologies are still being developed to better understand and predict the auto-ignition propagation modes. The original theory regarding the auto-ignition propagation mode was p...