The aim of the present work is to analyze the effect of the injection pressure on the flow pattern and droplets size in a gasoline hollow-cone spray generated by a swirled injector, applying Particle Image Velocimetry (PIV) and Phase Doppler Anemometry (PDA). Experiments were carried out in the range of injection pressures between 6 and 10 Mpa using a common rail injection system, a commercial swirled type injector with a nozzle diameter of 0.50 mm and a cone angle of 67°. The investigation was carried out injecting the fuel in a chamber at ambient temperature and pressure. A 2-D imaging technique was used to follow the global evolution of the spray as function of the injection time in order to estimate the jet development, the morphology of the spray, and the instantaneous velocity field of fuel droplets. PIV images were captured, firstly, aligning the light sheet to the vertical axis of the spray; then, experiments were also taken with the light sheet placed through the cross section of the spray in order to explore the structure and velocity field, at different distances from the nozzle. A PDA system was used to acquire, simultaneously, the droplets velocity as well the droplets size. The system, equipped with an argon-ion laser, was set in forward scattering mode at an off-axis of 30°. Measurements were performed at the same operative conditions as for the visualization ones, choosing different distances from the nozzle. Droplets velocity and size data set provided a minimum of 40,000 in-cycle resolved valid data that were analyzed using the ensemble averaging technique. Results provided detailed information on the spatial velocity distribution and size of droplets close to the nozzle region. Ensemble averaged velocity field of fuel droplets obtained on the vertical plane through the injector axis showed high velocity values up to 100 m/s in the early stage of injection, whereas during the transition to completely developed main spray, droplets travel at lower velocities. The velocity field obtained on the cross section of the spray, at locations close to the nozzle (5 and 10 mm) and in the early stage of injection (up to 0.6 ms), showed a symmetric radial velocity field around the spray axis. Size measurements performed near the nozzle showed a D_<10> profile constant along the stationary period of injection and not affected by the injection pressure. Results also gave a global increasing trend of the temporally averaged mean diameter along the spray axis.

View full abstract

This paper presents simulations of Diesel spray atomisation, obtained by a quasi-direct CFD method which employs a combination of Large-Eddy-Simulation (LES) to calculate the turbulence and a Volume of Fluid (VOF) technique to track the gas-liquid interface. The simulations are for an idealised injector nozzle with both short (LID = 5) and long (LID > 40) holes, with the results for the latter taken, for comparison purposes, from a previous study. It is found that the short nozzle produces more rapid atomisation, leading to shorter jet breakup length, larger spray angle and smaller droplet sizes, all of which are in reasonable correspondence with experimental observations. The difference in the performance of the two geometries is explained in terms of the nozzle-generated turbulence and its initial perturbation of the jet surface, which are more intense for the short-hole case.

View full abstract

Experimental studies of the Homogeneous Charge Compression Ignition (HCCI) combustion have been conducted by using a compact rapid compression machine. The chemiluminescence spectra were obtained sequentially during the HCCI combustion. At the same time, the combustion processes were photographed using a high-speed camera with image intensifier or a color high-speed camera. The image of the cool flame was captured by the high-speed camera with image intensifier, and during the cool flame appearance, the Emeleus cool flame band which originated from the formaldehyde chemiluminescence was observed in the obtained emission spectrum. Then blue flame became evident, its emission spectrum being similar to that of the cool flame with overlapping by HCO bands. During the main heat release duration, the hot flame emission spectrum was observed where the distinct OH emission at 306.4 nm appeared, and it was superimposed on the carbon monoxide continuum. After the main heat release, the broad peak of the spectrum was shifted to the longer wavelength side, and the images of the flames change to red, especially at higher equivalence ratios. This red coloration must be originated from the H_2O vibration-rotation bands (580.7 - 966.9 nm). On the other hand, it has been revealed that Pulsed Flame Jet (PFJ) has a potential for controlling the start of HCCI combustion directly. The chemiluminescence spectra and the color high-speed images of combustion initiated by PFJ were also obtained. In the jet of PFJ, the emission spectrum obtained was composed of the distinct OH emission at 306.4 nm and CH emission at 431.5 nm superimposed on the carbon monoxide continuum.

View full abstract

To run a Diesel engine in Homogeneous Charge Compression Ignition (HCCI) mode has proved to be a highly promising approach towards reduced engine-out emissions of NO_X and particulate matter. A crucial issue when utilizing HCCI is the degree of charge homogeneity that is required to achieve the desired low temperature combustion. A very well mixed charge can be created through the use of traditional port injection. This approach would most often result in low emissions of NO_X and soot. However, this strategy might also see a penalty in the form of high levels of unburned hydrocarbons due to incomplete combustion, especially under low load conditions. A proposed solution to this is to utilize stratified charge in the lower load range. The creation of a stratified charge imposes no major problems in modem DI engines. The important parameter is the degree of stratification that can be tolerated. If the charge gets too highly stratified, the combustion will become more diesel-like with dramatically increased levels of NO_X and soot as a result. This paper presents simultaneous laser based measurements of formaldehyde and OH-radical distributions in an HCCI engine. Formaldehyde is formed as an intermediate species when combusting hydrocarbons. The formation occurs through low temperature reactions in an early phase of the combustion process. Later in the process formaldehyde is being consumed. Formaldehyde is, therefore, used as indicator of the first stage of combustion and a marker of zones with low-temperature reactions. The OH radical is formed as an intermediate during the high temperature reactions, and is used as a marker of zones where the combustion is ongoing. The purpose of the investigation was to study how the combustion process is affected by the change in homogeneity that arises from early and late injection, respectively. A 0.5 liter single-cylinder optical engine equipped with a DI common rail fuel system was operated with a number of different injection timings, resulting in various levels of charge stratification. A blend of iso-octane and n-heptane was used as fuel. The measurement technique used was planar laser-induced fluorescence where formaldehyde was excited at 355nm and OH at 283nm. Two separate ICCD units were used to detect the resulting fluorescence from formaldehyde and OH. Measurement series covering the process from the start of injection until late in the expansion stroke is presented for different injection timings as well as pressure traces and emission analysis. A homogeneity index is calculated and used to compare the level of homogeneity resulting from injection timings. From early injection until about 50CAD BTDC the time, between onset of low temperature reactions and start of the high temperature reactions, is long enough for the formaldehyde to form an almost homogeneous distribution before it is being consumed. For later injection timings the high temperature reactions starts before this mixing is completed and therefore the formaldehyde distribution is not longer homogeneous and the combustion is more diesel like.

View full abstract

In order to investigate the relationship between ion current and combustion characteristics, an ion current probe utilizing conventional spark plug was applied to gasoline HCCI (Homogeneous Charge Compression Ignition) combustion with hot residual gas. The ion current signals from an ion current probe and in-cylinder pressure were measured simultaneously. The ion current signal of the single cycle of the HCCI combustion had almost one peak. There were slightly differences of the start and peak timing of ion currents at two probe locations in the early stage of reaction, and the differences in both probes was almost zero in the late stage of reaction. Under the homogeneous gas composition, it seems that similar signals of the ion current will be detected even if the ion current signals were detected locally. The characteristics of ion current in whole reaction may be able to conjecture with considering the characteristics of local ion current under homogeneous gas composition. It was found that there is high correlation between the timing of the ion current and the timing of the rate of heat release in spite of various engine conditions.

View full abstract

Temporal measurements of mass concentration and mean volume equivalent diameter of soot particles in diesel exhaust were made simultaneously by a two-wavelength extinction method and a scattering method. It is shown that the soot mass concentration changes dynamically depending on the variation in engine load, while the mean particle size varies within a narrow range centering 150 nm. A theoretical approach to quantify the scattering to absorption ratio of diesel soot particles is proposed whereby the transmittance values measured at two wavelengths and electron micrographs of sampled soot particles are incorporated into the soot aggregate scattering theory.

View full abstract

Modern PM measurement techniques for current and future engine concepts with ultra-low emission levels have to fulfill various challenging requirements. Sensitivity is one of the most important issues as current standard techniques are more and more getting insufficient when used together with exhaust aftertreatment systems like particulate traps and new combustion concepts. Secondly, a temporally high resolving in-situ technique is required to enable development of further emission minimization, particularly under transient conditions. Further, because of strong evidence of health risks of carbonaceous particles a mass based technique alone will no longer be sufficient. According to several medical studies other quantities like particle composition, size, number, and surface might be important to consider, as well. Laser-induced incandescence (LIT) has been proved to be a very sensitive and flexible tool for soot characterization in different laboratory applications for nearly twenty years. Recent developments have also indicated its excellent suitability for exhaust measurements. The technique relies on a rapid heating of soot particles within the gas flow to temperatures well above 4000 K by means of a highly energetic laser pulse and a subsequent optical detection of the strongly enhanced thermal radiation. By this, concentrations as low as 10 μg/m^3 and even below can be measured. Additionally, the temporal signal decay can be evaluated giving the surface area, or equivalently the primary particle size of the soot particles. Further extensions even allow the determination of the primary particle size distribution function. If LII is combined with other techniques, like elastic laser scattering, other measurement quantities like the agglomerate size are available, too. In this contribution the basic features of the LII technique are described and the LPSA system is introduced as a system relying on this technique. Various experimental results for diesel and gasoline direct injection (GDI) engines and some specific application examples obtained by this system are shown in some detail underlining the features of the LI^2SA sensor. These are including measurements downstream of a SCR catalyst and a particulate trap applied to a heavy duty diesel engine. By this, new insights in processes during internal combustion and exhaust aftertreatment are possible. As especially rapid changes of engine speed and load are the major sources for soot emissions during transient test cycles, the high temporal resolution of the system is of particular importance for research and development applications.

View full abstract

PM emission originated in combustion process of diesel engine should be reduced to keep the clean air environment. However, the detail characteristics about the primary and aggregate diameter of the PM were still unknown and they are needed for more physically accurate measurement and more effective improvement of exhaust PM emission. In this study, the size distributions of Dry-Soot in PM emission were reported. PMs in the tail-pipe emission were sampled from three type diesel engines. Sampled PM was chemically treated to separate the solid carbon fraction from other fractions such as SOF. The electron microscopic and optical-manual size measurement procedure was used to determine the primary size distribution. The centrifugal sedimentation method was applied to measure the Stokes equivalent diameter of the aggregate size of carbon PM cluster. They were discussed with aero-dynamic PM diameter measured by SMPS. Also, the effects of fuel properties and engine types on primary and aggregate PM diameters were discussed.

View full abstract

Laser-induced incandescence (LII) is a promising new diagnostic for measuring the volume fraction of elemental carbon in engine exhaust. The technique is considerably more precise and sensitive than conventional measurement procedures, and can be applied either with or without dilution. However, LII has been slow to gain acceptance because of presumed complexity of use and high initial cost. In this paper we demonstrate a prototype LII system that offers turn-key operation and long-term cost that is highly competitive with other techniques because of very low labor costs. The LII system ran unattended for 7.5 weeks, logging 1078 heavy-duty diesel engine tests during 24/7 operation of a dilution tunnel facility. Among the tests logged were 363 FTP steady-state mode tests and 250 FTP transient tests for which gravimetric measurements of total particulate matter (PM) were obtained. Of these tests, removal of the filter- based volatile matter using supercritical fluid extraction was performed on 142 and 147 of the tests, respectively. The correlation between the time-integrated LII signals and the dry gravimetric measurements for the steady-state mode tests is used to calibrate the LII measurements in mass units. This calibration is then used to evaluate the correlation between the LII and dry gravimetric measurements for the transient tests. Finally, time-resolved LII measurements for the steady-state mode tests are presented to illustrate three forms of unsteadiness that would seem undesirable.

View full abstract

Particulate emission from diesel engines is one of the biggest environmental issues, especially in terms of the influence on human's health. At present, the gravimetric method, which is specified in the relevant regulations, is commonly used to measure particulate mass emitted from engines. Although the gravimetric method is the standard in particulate mass measurement, this method has some disadvantages. For example, by the gravimetric method, it is difficult to know when and how the engine emits the particulate matter, such as soot and soluble organic fraction (SOF), during emission tests. Additionally, this method consumes a lot of time in getting the particulate mass, and is not suitable for the measurement of low-level emissions from clean engines. In this paper, a new PM analyzing system which can measure soot and SOF in low mass level continuously has been developed. It consists of a diffusion charging (DC) detector with a dilution device for soot measurement, and two flame ionization detectors (FID) for SOF measurement. This paper describes the principles of these analyzing techniques which allow soot and SOF to be measured continuously. It is confirmed that there are good correlations in the concentration of the PM constitutes (soot and SOF) between the new system and the conventional gravimetric method with soxhlet extraction method. The correlation factor (R2) obtained for soot in the study is 0.9820 and that for SOF is 0.9507. The total PM mass concentration from the new analyzer shows a good correlation with the gravimetric method as well. Measurement results of the new analyzer in real-time test are also described in the paper.

View full abstract

In order to investigate early soot formation process in a diesel spray flame, 2-D imaging and spectral measurements of laser-induced emission from soot precursors and soot particles in a transient spray flame achieved in a rapid compression machine (2.8MPa, 710K) were conducted. 3rd harmonic (355nm) and 4th harmonic (266nm) Nd:YAG laser pulses were used as the light source for laser-induced fluorescence (LIF) from soot precursors and laser-induced incandescence (LII) from soot particles in the spray flame. The 2-D imaging covered an area between 30mm and 55mm downstream from the nozzle orifice. The results of 2-D imaging showed that strong laser-induced emission excited at 266nm appears only on the laser incident side of the spray flame, in contrast to an entire cross-sectional distribution of the emission excited at 355nm, indicating that 266nm-excited emitters are more abundant than 355nm-excited emitters in the spray flame. The spectral measurements were conducted at three different positions, 35, 45 and 55mm downstream from the nozzle orifice, along the central axis of the spray, where LIF from soot precursors was observed in a previous 2-D imaging study. The spectra measured in upstream positions showed broad emissionpeaked around 400 to 500nm, which is attributable to LIF from polycyclic aromatic hydrocarbons (PAHs). The spectra measured in downstream positions appeared very much like grey-body emission from soot particles.

View full abstract

The reduction of fuel consumption and pollutant emissions is a key point in the development of modern spark ignition engines. One of the most promising approaches is the gasoline direct injection (GDI) with fuel stratification. An important part in the frame of GDI engine development is to solve the soot problem at the combustion of heterogeneous charges. The soot formation and oxidation of a GDI engine working at stratified charge condition was investigated inside the combustion chamber by means of the laser-induced incandescence (LII) technique in a two-dimensional configuration. To investigate the phenomena of soot formation, the mixture distribution as well as the subsequent combustion have been visualized within the same engine cycle. Additionally, a quantitative online analysis of the exhaust gas was conducted by a LII sensor system (LI SA) to show the correlation of in-cylinder soot formation and emission. The collection of the results of all measurement techniques provides a deep insight into the interacting chain of combustion, soot production and emission. The subject of investigation was a single cylinder research engine with gasoline direct injection and a jet guided fuel stratification concept. For the 2D laser-induced fluorescence (LIF) measurements a KrF excimer laser operating at a wavelength of 248 nm has been used for exciting an isooctane/TEA mixture as a substitute fuel. In the same plane, the fundamental of a Nd:YAG laser (1064 nm) heated up the soot particles and the 2D image of the black body radiation has been detected (LII). For the execution of the measurements of fuel distribution (fuel/vapour), combustion characteristics (visible & UV Flame) and soot appearance within the very same cycle, an ICCD camera system has been applied which is capable for taking four images simultaneously. To analyze the soot particles in the exhaust gas, the laser-induced incandescence soot analyzer (LI2SA) which has been developed by ESYTEC (Energie- und Systemtechnik GmbH, Germany), was applied to the exhaust manifold. This LI2SA sensor is able to measure the soot volume fraction as well as the specific soot surface area quantitatively with high temporal resolution.

View full abstract

The methods to investigate the soot formation are studied experimentally with a co-flow burner. In this study, the soot volume fraction, soot particle diameter and number density in a laminar diffusion flame are measured with LII(Laser Induced Incandescence)/LIS(Laser Induced Scattering)techniques. For the purpose of quantification, validation test is needed to be conducted before LII/LIS technique is applied to measure soot formation, and the development of algorithm is also required to analyze the raw data. In the present study, the extinction and scattering tests with a co-flow burner were performed to acquire calibration data. From this result, we decided the optimal laser beam intensity and obtained the LII signal from the diffusion flame. In addition, the algorithm for LII/LIS simultaneous measurement is developed and applied to this algorithm to measure the soot volume fraction in the diffusion flame.

View full abstract

The effect of wall interaction on the soot processes of a diesel fuel jet was investigated in an optically-accessible constant-volume combustion vessel at experimental conditions typical of a DI diesel engine. Soot processes were studied in both free jets and plane wall jets at identical operating conditions. Soot distributions were visualized using planar laser-induced incandescence and quantitative soot measurements were performed using a laser extinction technique. The investigation showed that soot levels were significantly lower in a plane wall jet compared to a free jet. Possible mechanisms for the decrease hi soot upon wall interaction include an increased fuel/air mixing rate and a wall jet cooling effect, which would tend to reduce or delay soot formation.

View full abstract

A numerical study of the spark ignition engine combustion system with direct methane injection to combustion chamber is presented. Poor penetration of gaseous fuel jet and poor mixing with air in the engine cylinder together with very short time available for mixing create serious difficulties in choosing proper injection parameters and in arranging the combustion chamber geometry. The results of preliminary analysis showed that mixing process of methane with air is very complex. Early beginning of injection extends the time available for mixing and strong vortex in combustion chamber makes mixing process more efficient. Direction of injection has also influence on mixing process and thanks to this a larger amount of fuel can be involved in mixing process. When the methane jet breaks up at the surface of combustion chamber wall gaseous fuel penetrates longer distance but mixing is not so efficient and layers with well-stirred mixture are surrounded with non-flammable mixture. The two-dimensional numerical simulations of methane direct injection system were performed with the use of KIVA-3V computer code. The results of calculations allow for optimisation of this system including combustion chamber geometry, location of the spark plug and injector together with the injection and ignition timing. The results of this study create the base for the further numerical investigation of the engine combustion system with methane direct injection, which will be performed together with related experimental research.

View full abstract

A cylindrical constant volume combustion bomb is used to investigate the combustion characteristics and to analyze the simplified heat balance of stratified charge methane-air mixture by using 2-stage injection. To analyze the heatbalance, some terminologies including CHR (cumulative heat release) ratio, UHC (unburned hydrocarbon) ratio and TL (total loss) ratio are defined. The result shows that the effect of the stratification on combustion characteristics is not significant in the case of the overall excess air ratio (λ_<overall>) of 1.1. In the case of the overall excess air ratio of 1.4, as the initial charge pressure (P_<ini>) increases, the CHR ratio has been decreased, while the TL ratio has been increased, as shown in Fig. 1 (here, H1 means homogeneous condition of t_<ig> = 10,000 ms, and H3 means of t_<ig> = 300 ms, respectively). And, the effect of stratification on combustion characteristics of overall excess air ratio of 1.4 is larger than those of λ_<overall> = 1.1 compared to the homogeneous condition of H1. Generally, as the initial charge pressure increases, the amount of injection mixture has been decreased and has resulted in lower mean velocity and turbulent intensity. Also, as shown in Table 1, the excess air ratio of injection mixture (λ_<inj>) is too rich to result in mixing deficiency around the spark electrode gap in the combustion chamber. From these results, it could be possible to acquire some optimal conditions which have higher thermal efficiency and lower total loss than those of homogeneous conditions through the 2-stage injection in CNG direct injection engine.

View full abstract

This paper presents the feasibility to apply scavenging-port injection of a compressed natural gas (CNG) to a two-stroke spark-ignition engine. A two-cylinder production engine for boat racing is used as a test engine. Through a preliminary experiment with several methods of CNG flow-rate measurement, a reliable procedure could be established. The engine was run at a constant speed of 3000 rpm and WOT condition. Comparing with the operation with a gas-mixer, by applying scavenging-port injection BSFC was decreased by 20 % and HC emission by 47 % at the best condition. Three methods of injector installation and three types of fuel pipe attached to the injector tip were included in the optimization process. A mechanism to explain the optimizing was also proposed on the basis of the consideration on the interaction between the flow of CNG jet and the scavenging flow.

View full abstract

Over the last several years there has been much interest in dimethyl ether (DME) as an alternative fuel for diesel engines. DME combines advantages of a high cetane number with no soot combustion, which makes it highly suitable for compression ignition engines. On the other hand, a homogeneous charge compression ignition (HCCI) engine has been known to have high thermal efficiency, no soot and low NO_X emissions. DME for the HCCI engine has also good adaptability. But its ignition and combustion characteristics are not completely understood. The objective of this study is to investigate the ignition delay of DME as homogeneous mixtures with air by using a rapid compression machine. The investigated fuels are ethane and propane in addition to DME. When the air is used as the working gas, the maximum compressed gas temperature reaches 730K. This temperature can't realize the ignition of ethane or propane. Therefore, a mixture gas composition of 53% argon, 26% helium and 21% oxygen is used for realizing temperatures above 730K. Equivalence ratios are 0.2, 0.3 and 0.4. Ignition delays of each fuel are measured at the constant gas density with changing the compressed mixture temperature at the TDC. So far, the representative temperature at the ignition of fuel tended to use the mean gas temperature calculated by the perfect gas law using both of the measured cylinder pressure and the charged gas amount. But the temperature at the center of the combustion chamber measured by a thin Pt resistance wire thermometer having a diameter of 15 micron is used in this study. The experimental results under the constant piston displacement show that the ignition timings of ethane and propane are advanced with increasing in an equivalence ratio, on the contrary, the ignition timing of DME is late. Arrhenius plots of ignition delay on each fuel were drawn up from a series of combustion history data. The results show that ignition delay times of ethane and propane depend on both of the equivalence ratio and the gas temperature, whereas that of DME dose not depends on equivalence ratio, but only temperature. Also, the combustion histories of DME indicate two-stage combustion. The first combustion is known as the heat release by the low temperature oxidation reaction. The heat release amount during this reaction period is also independent of equivalence ratios. Furthermore, the numerical simulation of the DME combustion by means of CHEMKIN II was carried out. In the computation, detailed reaction scheme of DME proposed by Curran et al. was used in this study. Also, the computation results were compared with those from a rapid compression machine. Generally, a comparison of experiment results with computation results indicates a good agreement.

View full abstract

A diesohol fuel (10% bioethanol, 89% Thailand diesel fuel and 1% additive) may be used as an alternative fuel in a compression ignition engine. However, higher compressibility, lower energy content, and lower cetane number of the diesohol tend to increase the ignition delay and reduce engine efficiency. In this study, to get more knowledge on the diesohol combustion characteristics, an experimental study of luminous combustion in a swirl chamber of an IDI engine was performed to compare the combustion phenomena between diesel and diesohol. Two-color method was applied to obtain spatially and temporally resolved two dimentional distributions of flame temperatures and soot in flame. The imaging system used for this study was based on a wide angle endoscope that is mounted in the cylinder head of the combustion chamber. The experiments were carried out on a commercial 2.5 litre IDI engine. Through systematic experiments, it was explored the effect of ethanol in diesohol blend. Observed diesohol fuel sprays have shown either longer spray tip penetration length or wider spray angle than the reference diesel. Images of spray combustion shown in figure 1 indicate that the period of diesohol combustion phenomena occurred more retard with respect to TDC than diesel. As its consequence, together with the lower heat of combustion, the predicted combustion flame temperature and soot in flame density distribution, shown in figure 2 and 3, are lower than the reference diesel. Observation in swirl chamber shows detail of the complex inflammation and combustion processes.

View full abstract

The characteristics of spray, auto-ignition and combustion process of DME (Dimetyl ether) were investigated by 3-D simulation in a combustion vessel under high temperature and pressure conditions and engine conditions. Spray impingement and non-premixed combustion model have been developed and incorporated into the computational fluid dynamics code, STAR-CD. Laminar flamelet concept was used to simulate non-premixed combustion. In order to reduce the enormous computational time, fitted function of the integration of pre-assumed PDF (Probability density function) was introduced. For the chemical mechanism of DME, a skeletal chemical kinetics mechanism which consists of 28 species and 45 reactions was deduced by approximation of detailed mechanisms by means of sensitivity analysis and steady-state with addition of generic species or global reactions. To verify this combustion model, autoignition and combustion processes were investigated in a pre-combusted vessel. The simulation results were well matched with experimental results of the spray tip penetration of liquid and vapor phase, and ignition delay time of DME spray. In comparison with the trends observed experimentally, the flamelet-concepted combustion model predicted the essential feature of combustion processes and the autoignition characteristics of DME spray reasonably well on the various initial conditions. The combustion process of DI engine fueled by DME was simulated with moving grid of a commercial diesel engine.

View full abstract

A technique for multi-species mole fraction measurement in internal combustion engines is described. The technique is based on the spontaneous Raman scattering. It can simultaneously provide the mole fractions of several species of N_2, C_2, H_2O, CO_2 and fuel. Using the system, simultaneous measurement of air/fuel ratio and burnt residual gas are carried out during the mixture process in a Controlled Auto Ignition (CAI) combustion engine. The accuracy and consistency of the measured results were confirmed by the measured air fuel ratio using an exhaust gas analyzer and independently calculated mole fraction values. Measurement of species mole fractions during combustion process has also been demonstrated. It shows that the SRS can provide valuable data on this process in a CAI combustion engine.

View full abstract

Optical measurements have been performed to visualize the fuel distribution from an air-assisted GDI injector in a constant volume spray chamber. In the air-assisted injector, fuel droplet breakup occurs as the droplets are accelerated by the expanded air flow passing out through the injector nozzle. The injector design allows for two distinct modes of operation: stratified (low load) and homogenous (full load). Therefore, measurements were taken at back-pressure and temperature settings corresponding to both low and full loads, to investigate the influence of these variables on fuel distribution, droplet sizes and velocities in the two modes. Results from PDA measurements show that most of the droplet break-up occurs inside the nozzle. Downstream of the nozzle, rates of break-up and evaporation are low. The cause of the low breakup and evaporation rates downstream of the nozzle is believed to be the low relative velocity between the liquid fuel and the surrounding cold air jet. The fuel distribution has been visualized qualitatively by simultaneous MIE and LIF measurements with two intensified digital cameras in combination with a 266 nm YAG-laser.

View full abstract

A small flame sensor was developed to measure the local chemiluminescence intensity of OH*, CH* and 2* in a four-stroke spark-ignition engine in order to understand flame front characteristics, flame propagation speed and local air/fuel (A/F) ratios. This sensor was installed where a pressure transducer was located. The dimensions of the sensor were those of the M5 type; the specifications of the developed small Cassegrain optics were the same as those used in previous engine measurements. The measurements were carried out at engine speeds of 600 and 1200 rpm using propane and gasoline fuels with no hardware modification in a 374.7 cc. model engine. Two types of Cassegrain optics were examined to measure local chemiluminescence of the flame in the engine. The small optics developed can provide very stable data for an experiment lasting several hours, which is sufficient to conduct practical combustion diagnostics. The measurements results with propane and gasoline demonstrated that this sensor is little influenced by fuel type, and the flame propagation speed was similar under both conditions. Although the sensor must be calibrated for a particular engine initially, it can measure local A/F and flame parameters in each cycle in practical engines. On-board diagnostics can be performed using this sensor and the flame speed and local equivalence ratio at the flame front can be calculated without modifying the engine hardware. This sensor should prove useful for measuring flame structure and stoichiometry in order to reduce cyclic variation and give the optimum control of each cylinder in practical engines.

View full abstract

The cycle-to-cycle variation in the flow field inside a four stroke ignition engine has been studied. The particle image velocimetry (PIV) technique has been used to measure the flow field inside the engine. The measurements have been carried out in a motored engine equipped with suitable access. The measurements have been carried out in both the sectional and the planar directions such that the necessary data for calculating the swirl ratio and tumble motions becomes available. The standard deviations from the ensemble averaged measurements for these two motions have been used as indicators for the cyclic variations in the flow field. The effect of the engine speed, inlet valve shrouding and piston geometry on the cyclic variations in the flow has been assessed. According to the analysis of the measurements, it has been found that increasing the engine speed increases both the swirl ratio as well as the cycle-to-cycle variation. The cyclic variability in the swirl ratio decreases when 120° shrouded inlet valve is used. This trend is reverted when the shroud is increased to 180°. Also the cyclic variability in the tumble ratio increases when using the 120° and 180° shrouded inlet valves. The cyclic variability based on the swirl ratio decreases with wider and less deep piston bowel. On the contrary, the cyclic variability in the tumble motion increases with shallow and wider piston bowel. Also the cyclic variability in the swirl ratio increases as the compression ratio increases.

View full abstract

The growing employment of alternative fuels and the necessity for further reduction of in-cylinder NO formation require additional knowledge about the fuel, design and operating variable effects on internal combustion (IC) engine performance and emissions. To this end, with the present work, an enhanced combustion diagnostic tool is proposed for the analysis of both heat release and flame propagation parameters in homogeneous-charge spark-ignition (SI) engines. It includes specifically developed sub-models for evaluating thermal and prompt NO so as to rank the effects of NO formation routes and thermodynamic parameters on nitric oxide emissions. The new diagnostic method is based on a quasi-dimensional multizone combustion model that takes the non-uniform spatial distribution of in-cylinder burned gas thermodynamic and chemical properties into account. As far as temperature calculation is concerned, the energy conservation law is applied to each zone, which can be considered to be either homogeneous or composed of two distinct parts: an adiabatic core and a thermal boundary layer. Each zone is generated at user-defined crank-angle increments, so that virtually all the cases between 'fully mixed' and 'unmixed' limiting models can be considered for investigation. The multizone thermodynamic approach for mass burning rate analysis is coupled with a CAD procedure for the computation of burned-gas mean expansion speed and burning speed. A new procedure for determining the start of combustion was embedded in the model. The calculation uses detailed thermodynamic gas properties that are capable of taking fuel composition and intake air humidity into account. The diagnostic method also includes previously proposed refinements for: modelling the surface-averaged heat flux in order to take the unsteadiness of gas-wall temperature difference into account; evaluating the unburned-gas zone temperature; calibrating the mass-fraction burned at the end of the flame propagation process. The NO calculation is performed on the basis of a six-reaction thermal formation mechanism and includes the contribution from the flame-formed NO. The quasi-dimensional multizone heat-release, flame-propagation and NO-formation integrated model was shown to be an effective and reliable tool of combustion diagnostics for SI engines through its application to the analysis of accurately measured pressure time-histories in the cylinder of an upgraded multivalve engine fuelled by either natural gas or gasoline under a significant sample of operating conditions. The results obtained from two-zone and one-zone heat-release models are also reported and discussed. The predicted average NO concentrations were in good agreement with the measured values when 10 to 15 burned zones were used in the model. Finally, heat release, flame propagation parameters as well as NO concentrations obtained from mean cycle and cycle-by-cycle calculations were compared, generally showing good physical consistency. However, cyclic pressure variations led to significant differences in the heat-release rate, combustion speeds and levels of nitrogen oxide emissions. The results indicated that cycle-by-cycle calculation should be considered for NO evaluation.

View full abstract

In spark ignition engines, knock can be defined as self-ignition of a certain portion of the unbumed gas beyond the flame front. This abnormal combustion releases a chemical energy that excites within the cylinder volume pressure oscillations that are responsible of engine noise and sometimes damage. Therefore, knock is undesirable process that has to be avoided. The present paper aims to evaluate the knocking combustion by an ion current system and optical diagnostics of radical species. The ion signal was also compared with the in-cylinder combustion pressure and the engine vibration. The optical measurements were based on 2D UV-Visible digital imaging of combustion process, UV chemiluminescence and spectrally resolved emissions from radical species. The optical techniques were applied for their potential to provide accurate measurements over a wide range. In particular, in this paper the spatial and temporal evolution of flame front and radical species such as OH and HCO were correlated to pressure and ion-current sensing during knocking combustion. The engine used during experiments was an optically accessible single cylinder, ported fuel injection, four-stroke spark-ignition engine with a four-valve production head. All the measurements were realized in the transparent combustion chamber equipped with a wide quartz window (diameter=57 mm) in the bottom of the chamber. The engine operated at 1000 rpm with stoichiometric mixture and wide-open throttle.

View full abstract

During the injection and mixture formation processes of a DI gasoline engine fuel may exist in either the liquid or vapour phase and it is therefore advantageous to be able to observe both phases simultaneously. However, traditional LIF techniques are not suitable for this purpose since the large differences in fluorescence intensity from the two phases make the selection of a suitable light collection level virtually impossible. In an attempt to overcome this problem the Laser Induced Exiplex Fluorescence (LIEF) technique was applied to a single cylinder DI gasoline engine. The engine was equipped with an air assisted fuel injection system and it had full optical access through a quatz liner and transparent piston crown. Isooctane was used as the test fuel with dopants Naphthalene (7%) and N,N-dimethlylaniline (DMA) (7%). The expanded laser beam from a XeCl Excimer laser was used to illuminate the whole combustion chamber and act as the excitation source for the LIEF system. In order to collect simultaneously fluorescence signals from both liquid and vapour phases emissions an image doubling and filtering system was designed and constructed. The fluorescence images were recorded from the front of the cylinder, perpendicular to the axis of the crankshaft. In order to investigate the effect of absorption on the fluorescence signal the beam was introduced to the cylinder from 3 directions. Initially the beam entered from the left of the image. The experiment was then repeated with the beam entering from the right and finally entering from below through the transparent piston crown. The crank angle resolved results, from the start of injection to the end of the compression stroke are presented with an image for each illumination direction.

View full abstract

A new charge stratification concept - fuel stratification is being researched and developed in a three-valve twin-spark SI engine. This concept requires that two different fuels or fuel components in a specially formulated composite be introduced into the cylinder separately through two independent inlet ports. The fuels will be stratified laterally by means of strong tumble in the cylinder. Different fuel properties may be employed in this way to increase fuel economy and reduce toxic emissions. This paper reports the development of the fuel stratification technique. The intake system of the engine was firstly modified to obtain a strong in-cylinder tumbling flow, which was measured by a digital Particle Image Velocimetry (PIV) system. Then a two-tracer Planar Laser Induced Fluorescence (PLIF) system was developed to visualise the in-cylinder fuel stratification. The charge stratification was further verified by examining the lean-burn limit at part load of engine operation. These research results show that the present strong tumble field was characterised by a symmetrically-distributed mean velocity in the intake stroke and a very small velocity component along the direction of the tumble rotational axis in the compression stroke. This bulk flow characteristic resulted in a clear fuel stratification laterally and therefore the lean-burn limit was considerably extended. The cyclic variation of the bulk flow affected the charge stratification.

View full abstract

Visualization plays an effective role in the establishment of a new combustion concept by helping to find the optimal results quickly among many different parameters and contributing to a shorter development period. Laser-induced fluorescence, Raman scattering and infrared absorption were used to measure the air/fuel ratio quantitatively in a three generation direct-injection gasoline (DIG) engine including a spray-guided mixture formation process and comparisons were made with the mixture formation concepts of the first- and second-generation DIG engines. The optimum combination of fuel spray, gas flow and combustion chamber configuration was found to be different for the three generations of DIG engines. The characteristics of the stable combustion region for obtaining higher thermal efficiency and cleaner exhaust emissions differed among the three mixture formation concepts.

View full abstract

The effect of split injection on the mixture characteristics of DISI (Direct Injection Spark Ignition) engines was investigated firstly by the Laser Absorption Scattering (LAS) technique. Through splitting the fuel injection process, two possible benefits were found: 1) High density liquid phase spray piling up at the leading edge of the spray can be circumvented, subsequently the reduction of the spray tip penetration; 2) The quantity of "over lean" (φ_v<0.7) in the spray can be significantly reduced. These are believed to contribute to the reduction of the engine-out smoke and HC emissions. In order to clarify the mechanism behind the effect of the split injection, the spray-induced ambient air motion was investigated by the LIF-PIV technique. The strong ambient air entrainment into the tail region of the spray and a counter-vortex structure were found in both the single and split injections. In the case of the single injection, the spray developed in extending its length, subsequently a larger volume resulted and thus it was diluted to "over lean" by the air entrainment. In contrast, in the case of split injection. The second spray was injected into the tail region of the first spray and it distributed some distance from the leading edge of the first spray and the evaporation of the second spray was promoted by the ambient air motion. Hence the replenishment of the liquid phase second spray on the leading edge of the first spray would be reduced. As a consequence, the high density liquid phase spray piling up at the leading edge was avoided. In addition, the ambient air motion played a positive role on evaporating the spray into an "more combustible" (0.7<φv<l.3) mixture. This was especially true in the tail region of the spray and the region where the counter-vortex motion was occurring.

View full abstract

As the new emissions standards lead to a reduction of the exhaust pollutants and fuel consumption, the engine efficiency must be improved. The knock phenomenon represents the main limitation for the performance of gasoline engines, especially in the downsizing turbo-charged direct injection approach. The aim of this paper is to study the influence of the gasoline representation (single or multicomponent) on the knock timing and position in the combustion chamber. Using a gasoline lumping model, in which the octane number is one of the main parameters, two different fuels were computed : Rl with only one pseudo-component and R3 with three pseudo-components. The average octane number in the cylinder for the R3 fuel was calculated with a non-linear method. In parallel, a multicomponent evaporation model has been developed in the multidimensional IFP-C3D code to be able to simulate the vapor distribution of each pseudo-component in a gasoline direct injection (GDI) engine. Comparing the results between the Rl and R3 fuels, it has been shown that the stratification of the pseudo-components has an impact on the flame propagation and on the octane number cartography in the combustion chamber. Moreover, with the R3 fuel, position and timing of the knock are very close to experimental results.

View full abstract

The computational code KIVA-3V has been used as the modeling platform for Gasoline Direct Injection engine simulations. Improved models for fuel injection, wall impingement and stratified combustion have been implemented. To complement the modeling effort, experimental data of an optical engine at the University of Michigan has provided extensive data for validating the new models. The experimental data include Planar Laser Induced Fluorescence measurements, with toluene as a tracer, of the injection, mixing and combustion events in an optical engine. In this work, focus is given on the combustion model development and validation with experiments.

View full abstract

The computational code KIVA-3V has been used as the modeling platform for Gasoline Direct Injection engine simulations. Improved models for fuel injection, wall impingement and stratified combustion have been implemented. To complement the modeling effort, experimental data of an optical engine has provided extensive data for validating the new models. The experimental data include Planar Laser Induced Fluorescence measurements, of the injection, mixing and combustion events in an optical engine. In this work, focus is given on fuel injection, wall impingement and air-fuel mixing.

View full abstract

The structure of combusting Diesel jets is studied using Laser Induced Fluorescence (LIF) of radicals. A single hole common rail Diesel injector is used which allows high injection pressures up to 120MPa. Visualization studies have been conducted in a high pressure, high temperature cell that is designed to reproduce the typical thermodynamic conditions which exist in the combustion chamber of a Diesel engine. Planar LIF of the OH radicals (OH LIF) with excitation near 280nm and LIF with excitation at 355nm (355 LIF) have been applied. OH LIF enables the localization of ground state OH which exists as the equilibrium product in high temperature regions and acts as a tracer to both the zones of reaction and the burned gas regions. The objective of 355 LIF is to excite the fluorescence of formaldehyde in order to observe and localize this intermediate species which is also present in the reaction zone. Since no determination of the origin of the fluorescence signal collected with excitation at 355nm was made in order to distinguish formaldehyde fluorescence from other molecules which might fluoresce at the same wavelength, the technique is referred to as 355 LIF, and special care has been taken when interpreting the results. Analysis of the pressure rise in the chamber was also monitored allowing the calculation of heat release rates and auto-ignition delays. Simultaneous visualizations of 355 LIF and chemiluminescence during the early stages of combustion showed that 355 LIF is useful in identifying the precursors of auto-ignition. Furthermore, simultaneous visualizations of 355 LIF and Mie scattering of the soot during the diffusion-limited combustion phase showed that 355 LIF is useful in identifying soot precursors during the latter stages of combustion. At this corresponding stage, OH was observed in the periphery of the jet. A comparison of the results obtained with these different techniques was performed in order to analyze the combustion structure. The presence of intermediate species in the upstream part of the jet was identified and allowed the localization of the main reaction zone in the upstream periphery of the jet, in the mixing zone. The ensemble of results obtained have been used to present a conceptual model of the combustion process.

View full abstract

Spray evaporation process in a direct injection diesel engine is greatly affected by the in-cylinder temperature but also it is affected by heat transfer process from the combustion cavity wall to the impingement spray. There are lots of literatures concerning with the fundamental studies of evaporation process of diesel spray. However, there are a few investigations about the effect of wall temperature on evaporation process. In this report, the fuel vapor distribution of a diesel spray impinging on a hot wall was measured using the laser induced fluorescence method. To control the wall and the ambient temperatures independently, the surroundings around the nozzle and the wall were heated separately using electric heaters. The result indicated that the hot surroundings entrained into the impingement spray had greater effect than that of the hot wall, even if the wall temperature was higher than the boiling temperature of the fuel.

View full abstract

Non-evaporating spray penetration characteristics for multiple diesel oil injections at atmospheric conditions are studied by comparing numerical simulations, effected by means of two different numerical codes, and experimental results collected in a constant volume vessel of controlled thermodynamic conditions. Different injection strategies are implemented via a Programmable Electronic Control Unit on a Common Rail system with dwell time between injections up to the electro-mechanical limits of a Multijet injector. The instantaneous and total amount of delivered fuel is measured by an AVL Injection Rate Meter, operating according to the Bosch principle. Images of jets emerging from the nozzle in different operating conditions are captured by using an AVL Engine Videoscope with a CCD camera and a properly synchronized flash-like lightning system. The jets dynamics is extracted through a digital image processing software in order to analyze temporal and spatial behavior. The non reacting flow simulation is effected by means of the FIRE code, whereas the KTVA3 V code is used to analyze both the non reacting spray dynamics and the process of auto-ignition, combustion and pollutants formations. Reacting flow calculations are performed using a detailed chemistry mechanism for a diesel fuel surrogate. To get consistent data, the KIVA3V and FIRE predictions are preliminary compared for the same operating conditions. Reliability of both the code in predicting the spray behaviour is proven, by properly customizing the employed versions. The different multiple-injection schedules are found to have considerable effects on droplet and fuel vapor distribution, hence also on auto-ignition, combustion and emissions (soot, NO_X) formation. Among the injection schedules studied, the most optimal one is singled out as the one leading to complete fuel consumption and reduced amount of generated pollutants.

View full abstract

Spherically propagating premixed laminar flames in a large volume bomb were studied at some initial mixture pressures in order to investigate the effects of pressure on the premixed flame, especially on the burning velocity and its variation by the flame stretch. They are quite important to the modeling of combustion in internal combustion engines because the burning velocity is one of the most essential parameters. The effects of the initial mixture pressure on the unstretched laminar burning velocity and the Markstein length were investigated using methane and propane-air mixtures in the wide range of the equivalence ratio from 0.8 to 1.4 varying the initial pressure from 0.10 to 0.50MPa. The Markstein length and its normalized number express the variation in the burning velocity by the flame stretch. In this study, the burning velocity was obtained not from the flame radius of the photographic image in the literature, but from the pressure history. It is related to the rate of mass burning rather than the rate of propagation of the flame. The unstretched laminar burning velocity decreased as the initial pressure increased. The Markstein number decreased with increasing the initial pressure at all the equivalence ratios irrespective of fuel. They were negative for the lean methane and the rich propane mixtures at elevated pressures. The burning velocity of the stretched laminar flame was affected by the pressure in the two ways, the variation in the unstretched laminar burning velocity and the variation in the sensitivity of the burning velocity to the flame stretch. The properties of the flame instability were investigated next. The flame became unstable with the increase in the mixture pressure. Flame was unstable under the condition of low Markstein number. Cellular flame structure due to the flame instability developed earlier in such cases. The cells became small as the Markstein number decreased. The flame instability might affect the burning velocity.

View full abstract

A recent modification of the so-called Flame Speed Closure (FSC) model of turbulent combustion is tested against recently published experimental data on the speed and structure of statistically spherical flames that expand in a turbulent premixed gas after spark ignition. The measurements are simulated by solving the mass conservation equation, the κ - ε equations, the energy equation, and the balance equations for the Favre-averaged mass fractions of fuel and oxidant, which are closed by the FSC model. Results show the ability of the model (1) to predict the mean structure of developing turbulent flames very well and (2) to yield good approximations of turbulent flame speed for different mixtures at different pressures and rms turbulent velocities.

View full abstract

Combustion properties of dimethyl ether (DME) and air mixtures have been elucidated in a spherical bomb by using schlieren images with high-speed camera, for equivalence ratios ranging from 0.8 to 2.0 at an initial pressure of 1 atm and a temperature of 298 K. The unstretched laminar burning velocities were deduced and compared with other study and calculation. Good and partial agreements were obtained with the other study and calculation, respectively. Flame response to stretch (represented by the Markstein number) were also considered both experimentally and computationally and were compared with methane, ethane, propane-air mixture. For our experimental conditions, stretched laminar burning velocities were found to vary linearly with flame stretch, yielding Markstein numbers for particular reactant conditions. In addition, turbulent burning velocities were measured for the above mixtures. In a given explosion, the burning velocity increased with time and radius, as a consequence of the continual broadening of the effective spectrum of turbulence and decreasing of the flame stretch to which the flame was subjected. A decrease in the Markstein number of the mixture increased the turbulent burning velocity.

View full abstract

A premixed turbulent flame stabilized behind a bluff body and oscillating due to oncoming flow pertur- bations is studied. A simple kinematic model of such a flame was developed by Bowling (J. Fluid Mech. 394:51, 1999) who reduced the combustion process to the propagation of an infinitely thin flame at a constant speed. The goal of this work is to extend the model by taking into account the real structure of premixed turbulent flames and the development of turbulent flame speed and thickness. For these purposes, the so-called Flame Speed Closure model for multi-dimensional simulations of premixed turbulent combustion is adapted and combined with the aforementioned Bowling model. Simulations of the heat release rate dynamics have been performed. Typical results show that the oscillations in the integrated heat release rate follow the velocity oscillations with certain time delay. The delays computed using the Bowling and the above approaches are different, thus indicating the importance of resolving flame structure when modeling ducted flame oscillations.

View full abstract