Search

With growing concern about global warming, alternatives to fossil fuels in internal combustion engines are searched. In this context, hydrogen is one of the most interesting fuels as it shows excellent combustion properties such as laminar flame speed and energy density. In this work a CFD methodology for 3D-CFD in-cylinder simulations of engine combustion is proposed and its predictive capabilities are validated against test-bench data from a direct injection spark-ignition (DISI) prototype. The original engine is a naturally aspirated, single cylinder compression ignition (Diesel fueled) unit. It is modified substituting the Diesel injector with a spark plug, adding two direct gas injectors, and lowering the compression ratio to run with hydrogen fuel. A 3D-CFD model is built, embedding in-house developed ignition and heat transfer models besides G-equation one for combustion. Three different lean-burn conditions are selected among the tested ones for the validation of the numerical framework. The investigated conditions are characterized by the same revving speed (3000 rpm) but different equivalence ratios (0.4, 0.6 and 0.8, respectively). A good agreement with the experimental dataset is observed, confirming the validity of the proposed CFD approach, and opening the possibility of further virtual optimizations of the engine.

Hybrid URANS/LES turbulence modeling is rapidly emerging as a valuable complement to standard LES for full-engine multi-cycle simulation. Among the available approaches, zonal hybrids are potentially attractive due to the possibility of clearly identify URANS and LES zones, eventually introducing further zone types with dynamically switching behavior. The present work aims at evaluating the impact of different zonal configurations on the simulated flow statistics using the well-assessed TCC-III experimental engine setup. More specifically, different methods (URANS, LES or seamless DES) are applied inside the cylinder volume, as well as into the intake/exhaust ports and plenums. For each of the five tested configurations, in-cylinder flow features are compared against the reference TCC-III experimental measurements, in terms of velocity field statistics and quality indices. In addition, a detailed analysis using Proper Orthogonal Decomposition (POD) is carried out to quantitatively compare the results from experiments and simulation sets. The study outcomes are used as a starting point for discussing the applicability of zonal hybrid turbulence modeling to realistic engine geometries, critically analyze the model assumptions (e.g. the domain zonal decomposition) and provide guidelines for general application of such method.

Recently, a growing interest in the development of more accurate phenomenological turbulence models is observed, since this is a key pre-requisite to properly describe the burn rate in quasi-dimensional combustion models. The latter are increasingly utilized to predict engine performance in very different operating conditions, also including unconventional valve control strategies, such as EIVC or LIVC. Therefore, a reliable phenomenological turbulence model should be able to physically relate the actuated valve strategy to turbulence level during the engine cycle, with particular care in the angular phase when the combustion takes place. Similarly, the capability to sense the effects of engine architecture and intake geometry would improve the turbulence model reliability. 3D-CFD codes are recognized to be able to accurately forecast the evolution of the in-cylinder turbulence field, taking into account both geometrical features (compression ratio, bore-to-stroke ratio, intake runner orientation, valve, piston and head shapes, etc.) and operating conditions (engine speed, boost level, valve strategy). Instead, more common 0D turbulence models usually synthesize geometrical effects in a number of tuning constants and “try” to be sensitive to the operating conditions as much as possible. In this two-part paper, the final goal is the refinement of a previously developed 0D turbulence model, here extended to directly predict the tumble vortex intensity and its close-to-TDC collapse into turbulence. In addition, the model is enhanced to become sensitive to engine geometrical characteristics, such as intake runner orientation, compression ratio, bore-to-stroke ratio and valve number, without requiring any preliminary estimation of the tumble coefficient on a flow bench. Part I describes a background study, where 3D analyses are performed to highlight the effects of operating conditions and main engine geometrical parameters on tumble and turbulence evolution during the engine cycle. In a preliminary stage, the averaging process influence to define representative quantities of mean flow and turbulence is discussed, in order to take into account not-uniformities inside the combustion chamber. 3D simulations are carried out under motored conditions on a VVA engine, at various engine speeds. The VVA device is controlled to simulate both standard, early and late valve closures. The results highlight substantial differences in the mean flow velocity, turbulence intensity and tumble speed among the above cases. To focus the engine geometry impact on the turbulence evolution, further analyses are performed on a different engine, by changing the angle between the intake runners and the cylinder axis. The geometrical compression ratio and the bore-to-stroke ratio are modified, as well. Finally, a two-valve version of this engine is also considered. Results of 3D analyses are discussed to widely assess the effects of valve strategy and main engine geometrical parameters on mean flow, tumble and turbulence evolution inside the combustion chamber. The presented information constitute an extended database for the development and validation of a refined quasi-dimensional model, discussed in the companion part II of the paper.

The increased limitations to both NOx and soot emissions have pushed engine researchers to rediscover gasoline engines. Among the many technologies and strategies, gasoline direct injection plays a key-role for improving fuel economy and engine performance. The paper aims to investigate an extremely complex task such as the idle operating engine condition when the engine runs at very low engine speeds and low engine loads and during the warm-up. Due to the low injection pressure and to the null contribution of the turbocharger, the engine condition is far from the standard points of investigation. Taking into account the warm-up engine condition, the analyses are performed with a temperature of the coolant of 50°C. The paper reports part of a combined numerical and experimental synergic activity aiming at the understanding of the physics of spray/wall interaction within the combustion chamber and particular care is used for air/fuel mixing and the combustion process analyses. In order to properly describe the engine condition, different injection strategies are investigated. Late and early injection strategies are deeply analyzed and compared in terms of combustion stability and pollutant emissions. UV-visible imaging and spectral measurements are carried out in real engine with wide optical accesses... Measurements are performed in the optically accessible combustion chamber realized by modifying a real engine. The cylinder head was modified in order to allow in the fourth cylinder the visualization of the fuel injection and the combustion process with high spatial and temporal resolution. The 3D-CFD engine simulations are reproduced by means the commercial code Star-CD. Due to the warm-up condition and the many physical sub-models a numerical methodology is implemented and particular care is used to boundaries conditions analyses. CFD analysis is used to find a possible explanation of the high cycle to cycle variability. The experimental and numerical comparisons, in terms fuel mixing and front flame propagation, give an explanation of the idle condition.