Search Results

Emissions and in-cylinder pressure traces are used to compare the relative importance of soot formation and soot oxidation in a heavy-duty diesel engine. The equivalence ratio at the lift-off length is estimated with an empirical correlation and an idealized model of diesel spray. No correlation is found between the equivalence ratio at lift-off and the soot emissions. This confirms that trends in soot emissions cannot be directly understood by the soot formation process. The coupling between soot emission levels and late heat release after end of injection is also studied. A regression model describing soot emissions as function of global engine parameters influencing soot oxidation is proposed. Overall, the results of this analysis indicate that soot emissions can be understood in terms of the efficiency of the oxidation process.

Increased research is being driven by the automotive industry facing challenges, requiring to comply with both current and future emissions legislation, and to lower the fuel consumption. The reason for this legislation is to restrict the harmful pollution which every year causes 3.3 million premature deaths worldwide [1]. One factor that causes this pollution is NOx emissions. NOx emission legislation has been reduced from 8 g/kWh (Euro I) down to 0.4 g/kWh (Euro VI) and recently new legislation for ammonia slip which increase the challenge of exhaust aftertreatment with a SCR system. In order to achieve a good NOx conversion together with a low slip of ammonia, small droplets of Urea solution needs to be injected which can be rapidly evaporated and mixed into the flow of exhaust gases.

In this paper, computation fluid dynamics (CFD) simulations are employed to describe the effect of flow parameters on the formation of soot and NOx in a heavy duty engine under low load and high load. The complexity of diesel combustion, specially when soot, NOx and other emissions are of interest, requires using a detailed chemical mechanism to have a correct estimation of temperature and species distribution. In this work, Multiple Representative Interactive Flamelets (MRIF) method is employed to describe the chemical reactions, ignition, flame propagation and emissions in the engine. A phenomenological model for soot formation, including soot nucleation, coagulation and oxidation with O2 and OH is incorporated into the flamelet combustion model. Different strategies for modelling NOx are chosen to take into account the longer time scale for NOx formation. The numerical results are compared with experimental data to show the validity of the model for the cases under study.

Two competing in-cylinder processes, soot formation and soot oxidation, govern soot emissions from diesel engines. Previous studies have shown a lack of correlation between the soot formation rate and soot emissions. The current experiment focuses on the correlation between soot oxidation rates and soot emissions. Laser extinction is measured using a red (690nm) laser beam, which is sent vertically through the cylinder. This wavelength is long enough to minimize absorption interference from poly-aromatic hydrocarbons, while still in the visible regime. It is modulated at 72 kHz in order to produce 10 pulses per crank angle degree at an engine speed of 1200 rpm. The intake oxygen concentration is varied between 9% and 21%. The time resolved extinction measurements are used to estimate soot oxidation rates during expansion.

The soot distribution as function of ambient O₂ mole fraction in a heavy-duty diesel engine was investigated at low load (6 bar IMEP) with laser-induced incandescence (LII) and natural luminosity. A Multi-YAG laser system was utilized to create time-resolved LII using 8 laser pulses with a spacing of one CAD with detection on an 8-chip framing camera. It is well known that the engine-out smoke level increases with decreasing oxygen fraction up to a certain level where it starts to decrease again. For the studied case the peak occurred at an O₂ fraction of 11.4%. When the oxygen fraction was decreased successively from 21% to 9%, the initial soot formation moved downstream in the jet. At the lower oxygen fractions, below 12%, no soot was formed until after the wall interaction. At oxygen fractions below 11% the first evidence of soot is in the recirculation zone between two adjacent jets.

A heavy duty diesel engine operating case producing no engine-out smoke was studied using combined simultaneous optical diagnostics. The case was close to a typical low load modern diesel operating point without EGR. Parallels were drawn to the conceptual model by Dec and results from high-pressure combustion vessels. Optical results revealed that no soot was present in the upstream part of the jet cross-section. Soot was only observed in the recirculation zones close to the bowl perimeter. This indicated very slow soot formation and was explained by a significantly higher air entrainment rate than in Dec's study. The local fuel-air equivalence ratio, Φ, at the lift-off length was estimated to be 40% of the value in Dec's study. The lower Φ in the jet produced a different Φ -T-history, explaining the soot results. The increased air entrainment rate was mainly due to smaller nozzle holes and increased TDC density.

When applying high amount of EGR (exhaust gas recirculation) in Partially Premixed Combustion (PPC) using diesel fuel, an increase in soot emission is observed as a penalty. To better understand how EGR affects soot particles in the cylinder, a fast gas sampling technique was used to draw gas samples directly out of the combustion chamber in a Scania D13 heavy duty diesel engine. The samples were characterized on-line using a scanning mobility particle sizer for soot size distributions and an aethalometer for black carbon (soot) mass concentrations. Three EGR rates, 0%, 56% and 64% were applied in the study. It was found that EGR reduces both the soot formation rate and the soot oxidation rate, due to lower flame temperature and a lower availability of oxidizing agents. With higher EGR rates, the peak soot mass concentration decreased. However, the oxidation rate was reduced even more.

The local equivalence ratio, Φ, was measured in fuel jets using laser-induced spontaneous Raman scattering in an optical heavy duty diesel engine. The measurements were performed at 1200 rpm and quarter load (6 bar IMEP). The objective was to study factors influencing soot formation, such as gas entrainment and lift-off position, and to find correlations with engine-out particulate matter (PM) levels. The effects of nozzle hole size, injection pressure, inlet oxygen concentration, and ambient density at TDC were studied. The position of the lift–off region was determined from OH chemiluminescence images of the flame. The liquid penetration length was measured with Mie scattering to ensure that the Raman measurement was performed in the gaseous part of the spray. The local Φ value was successfully measured inside a fuel jet. A surprisingly low correlation coefficient between engine-out PM and the local Φ in the reaction zone were observed.

The major challenge of the post-processing of soot aggregates in transmission electron microscope (TEM) images is the detection of soot primary particles that have no clear boundaries, vary in size within the fractal aggregates, and often overlap with each other. In this study, we propose an automated detection code for primary particles implementing the Canny Edge Detection (CED) and Circular Hough Transform (CHT) on pre-processed TEM images for particle edge enhancement using unsharp filtering as well as image inversion and self-subtraction. The particle detection code is tested for soot TEM images obtained at various ambient and injection conditions, and from five different combustion facilities including three constant-volume combustion chambers and two diesel engines.

The exact shape of the combustion chamber will not have a major effect in case of homogeneous type of combustion. However, presently engine needs to be operated in diesel mode during start-up and outside HCCI operating range to meet stringent emission norms. Hence, the combustion chamber is therefore optimized for conventional diesel mode operation for minimum emissions as well as benefit to HCCI mode operation with better swirl. When operated in pure HCCI mode, the exact shape of the combustion chamber is of little relevance-mainly because the fuel jets do not interact with the bowl at early injection conditions. However, at some operating conditions it is necessary to operate with partially premixed conditions where fuel is injected between 20 to 30° BTDC. Then bowl geometry plays significant role to promote homogeneous mixture of air and fuel. There is no need for late cycle turbulence generation in HCCI mode, since it is not a mixing controlled process.

Laser-Induced Incandescence (LII) has traditionally been considered a straightforward and reliable optical diagnostic technique for in-cylinder soot measurements. As a result, it is nowadays even possible to buy turn-key LII measurement systems. During recent years, however, attention has been drawn to a number of unresolved challenges with LII. Many of these are relevant mostly for particle sizing using time-resolved LII, but also two-dimensional soot volume fraction measurements are affected, especially in regions with high soot concentrations typically found in combustion engines. In this work the focus is on the specific challenges involved in performing high-repetition rate measurements with LII in diesel engines. All the mentioned issues might not be possible to overcome but they should nevertheless be known and their potential impact should be considered.

The combustion process in a homogeneous charge compression ignition (HCCI) engine is mainly governed by ignition wave propagation. The in-cylinder pressure, heat release rate, and the emission characteristics are thus largely driven by the chemical kinetics of the fuel. As a result, CFD simulation of such combustion process is very sensitive to the employed reaction mechanism, which model the real chemical kinetics of the fuel. In order to perform engine simulation with a range of operating conditions and cylinder-piston geometry for the design and optimization purpose, it is essential to have a chemical kinetic mechanism that is both accurate and computational inexpensive. In this paper, we report on the evaluation of several chemical kinetic mechanisms for methanol combustion, including large mechanisms and skeletal/reduced mechanisms.

Soot emissions from diesel internal combustion engines are strictly regulated nowadays. Laser extinction measurement (LEM) and natural luminosity (NL) of sooty flames are commonly applied to study soot. LEM measures soot along the laser beam path and it can probe soot regardless of temperature. NL integrates the whole field of view and relies on soot temperature. In this work, a comparison of simultaneously recorded LEM and NL data has been performed in a heavy-duty optical engine. A 685 nm laser beam is used for LEM. The laser was modulated at 63 kHz, which facilitated subtraction of the background NL signal from the raw LEM data. By Beer-Lambert’s law, KL factor can be calculated and used as a metric to describe soot measurements. A compensation of transmitted laser intensity fluctuation and soot deposits on optical windows has been performed in this work.

The presence of OH radicals as a marker of the high temperature reaction region usually has been used to determine the lift-off length (LOL) in diesel engines. Both OH Laser Induced Fluorescence (LIF) and OH* chemiluminescence diagnostics have been widely used in optical engines for measuring the LOL. OH* chemiluminescence is radiation from OH being formed in the exited states (OH*). As a consequence OH* chemiluminescence imaging provides line-of-sight information across the imaged volume. In contrast, OH-LIF provides information on the distribution of radicals present in the energy ground state. The OH-LIF images only show OH distribution in the thin cross-section illuminated by the laser. When both these techniques have been applied in earlier work, it has often been reported that the chemiluminescence measurements result in shorter lift-off lengths than the LIF approach.

In order to separate the HC-emissions from two-stroke engines into short-circuit losses and emissions due to incomplete combustion, Laser Induced Fluorescence (LIF) measurements were performed on the exhaust gases just outside the exhaust ports of two engines of different designs. The difference between the two engines was the design of the transfer channels. One engine had “finger” transfer channels and one had “cup handle” transfer channels. Apart from that they were similar. The engine with “finger” transfer channels was earlier known to give more short-circuiting losses than the other engine, and that behavior was confirmed by these measurements. Generally, the results show that the emission of hydrocarbons has two peaks, one just after exhaust port opening and one late in the scavenging phase. The spectral information shows differences between the two peaks and it can be concluded that the latter peak is due to short-circuiting and the earlier due to incomplete combustion.

Experiments on a modern DI Diesel engine were carried out: The engine was fuelled with standard Diesel fuel, RME and a mixture of 85% standard Diesel fuel, 5% RME and 10% higher alcohols under low load conditions (4 bar IMEP). During these experiments, different external EGR levels were applied while the injection timing was chosen in a way to keep the location of 50% heat release constant. Emission analysis results were in accordance with widely known correlations: Increasing EGR rates lowered NOx emissions. This is explained by a decrease of global air-fuel ratio entailing longer ignition delay. Local gas-fuel ratio increases during ignition delay and local combustion temperature is lowered. Exhaust gas analysis indicated further a strong increase of CO, PM and unburned HC emissions at high EGR levels. This resulted in lower combustion efficiency. PM emissions however, decreased above 50% EGR which was also in accordance with previously reported results.

The development of a new combustion system for a light-duty diesel engine is presented. The soot-NOx trade-off is significantly improved with maintained or improved efficiency. This is accomplished only by altering the combustion chamber geometry, and thereby the in-cylinder flow. The bowl geometry is developed in CFD and validated in single cylinder tests. Tests and simulations align remarkably well. Under identical conditions in the engine the new combustion chamber decreases smoke by 11-27%, NOx by 2-11%, and maintains efficiency as compared to the baseline geometry. The injector nozzle is matched to the new bowl using design of experiments (DoE). By this method transfer functions are obtained that can be used to optimize the system using analytical tools. The emissions show a complex dependence on the nozzle geometry. The emission dependence on nozzle geometry varies greatly over the engine operating range.

The liquid phase penetration of diesel sprays under reacting conditions is measured in an optical heavy-duty Direct Injection (DI) diesel engine. Hot gas reservoirs along the diffusion flames have previously been shown to affect the liftoff length on multi hole nozzles. The aim of this study is to see if they also affect the liquid length. The inter-jet spacing is varied together with the Top Dead Center density and the inlet temperature. To avoid unwanted interferences from the natural flame luminosity the illumination wavelength is blue shifted from the black body radiation spectrum and set to 448 nm. Filtered Mie scattered light from the fuel droplets is recorded with a high speed camera. The liquid fuel penetration is evaluated from the start of injection to the quasi steady phase of the jets. Knowledge of jet-jet interaction effects is of interest for transferring fundamental understanding from combustion vessels to practical engine applications.

Heat transfer losses are one of the largest loss contributions in a modern internal combustion engine. The aim of this study is to evaluate the contribution of the piston bowl type and swirl ratio to heat losses and performance. A commercial CFD tool is used to carry out simulations of four different piston bowl geometries, at three engine loads with two different swirl ratios at each load point. One of the geometries is used as a reference point, where CFD results are validated with engine test data. All other bowl geometries are scaled to the same compression ratio and make use of the same fuel injection, with a variation in the spray target between cases. The results show that the baseline case, which is of a conventional diesel bowl shape, provides the best emission performance, while a more open, tapered, lip-less combustion bowl is the most thermodynamically efficient.

In computational fluid dynamic simulations of partially premixed combustion engines it is common to find simplifications of the in cylinder flow conditions in order to save computational cost. One common simplification is to start the simulation at the moment of intake valve closing with an assumed initial flow condition, rather than making a full scavenging simulation. Another common simplification is the periodic sector assumption, limiting all sector cuts of the full cylinder to be identical periodic copies of each other. This work studies how such flow simplifications affect the spray injection and in turn the fuel/air mixing at different injection timings. Focus is put on the stratification of fuel concentration and gas temperature due to interaction of the spray, turbulence and piston geometry. The investigated engine setup consists of a light duty engine with a piston bowl and a five-hole injector.