The present paper describes a new numerical model called AKTIM (Arc and Kernel Tracking Ignition Model) for describing the flame kernel expansion in SI engines. This model is divided into four submodels. One discribes the secondary side of the electrical inductive system of the spark plug, with the available initial electrical energy, resistance and inductance as input parameters, figure 1(a). The spark is modeled by a set of particles uniformally placed along the spark path, figure 1(b). This spark is elongated in time by the mean flow field. The flame kernel is described by lagrangian marker particles. Each kernel can be seen as the initial flame development of one particular cycle of the engine. It is convected by the mean and turbulent flows and it receives energy from the electrical circuit and loses energy by contact with the spark plug. Finally, a fourth submodel describes the spark plug by a set of discrete particles. This way no remeshing of the chamber mesh is needed. This new ignition model was introduced in KIVA-MB, a multibloc KIVA-II version. It constitutes a major evolution of the actual ignition model used in the Extended Coherent Flame Model (ECFM). The 3D calculations presented in this paper point out the improvements brought by AKTIM and its potential for future developments. First calculations show a good accordance with experimental data and the ability of the model to correctly represent multiple electrical breakdown, convection, and heat loss effects.

View full abstract

Lean-burn gas engines have been operated worldwide in various fields such as the key hardware to the cogeneration system. Lean-burn, however, has crucial problems of unstable ignition and poor flame propagation such as misfire. Although one technical solution to this is the pre-combustion chamber method, even more complex problems could arise depending on the structure of each combustion chamber. For example, excess air ratio in the pre-combustion chamber of lean-burn gas engine is not completely uniform. The ignition method using micro-pilot fuel oil instead of spark plug as an ignition source has been studied. It has been investigated that the effect on the engine performance of above two ignition methods. Micro-pilot ignition method has following characteristics and merits compared with spark ignition method : i. Secure ignition ensured by compression ignition of fuel oil eliminates unstable ignition and misfire in the precombustion chamber. ii. Ignition energy is much larger than that of spark plug, which shortens combustion duration. This characteristic enlarges knock margin and enables higher BMEP output. iii. NOx emission depends on the amount of fuel oil in this micro-pilot ignition. NOx level is low because the amount of the fuel oil is mere about 1% of the total input fuel energy. Therefore, under the same restrictions of NOx concentration, larger pre-combustion chambers are allowed in the engine. It contributes to improve thermal efficiency. Multi-ignition and swirl are key factors to improve thermal efficiency of the engines. Multi-ignition enables to shorten the distance of flame propagation and swirl accelerates the flame propagation in the mixture. Optimizing the direction of a jet hole of pre-combustion chambers against swirl flow should be considered as an important design factor to obtain higher thermal efficiency. A single cylinder engine of a 260mm bore and a 275mm stroke has been used for combustion tests. The engine performance has been investigated at BMEP of 1.47MPa. The effect of turbulent flow to combustion of lean-mixture in the combustion chamber has been also investigated. This paper describes the engineering findings of combustion characteristics of the micro-pilot ignition method and the effect of the direction of a jet hole against the swirl flow on thermal efficiency with low-NOx emission.

View full abstract

A MPI (Multi Point Injection) LPG (Liquefied Petroleum Gas) engine for heavy-duty vehicles such as city buses has been developed, which uses the liquid phase LPG injection (hereafter LPLI) system, since MPI fuel system is able to accomplish the higher power and efficiency, and lower emissions compared to the SPI (Single Point Injection) and mixer type systems. In this fundamental study on the LPLI system, the engine output and combustion performance were investigated with various operating conditions using a single cylinder engine, which has a displacement volume of 1.8L and compression ratio of 10. The experimental results revealed that the volumetric efficiency and engine output increased by approximately 10%, and exhaust gas temperature decreased by 20〜30℃ when the liquid phase LPG is injected. This MPI LPG fuel system also brought an additional advantage of controlling knocking phenomenon via optimizing spark timing and compression ratio. An LPLI engine in this work was also able to be operated either at lean condition of λ=1.5 or with EGR (30%) while achieving stable combustion. The optimized swirl ratio was around 2, and the oil spot method demonstrated that the high swirl ratios above 2.3 did not show merit in breaking swirl flow into small turbulence scale. Among the piston cavities tested in this work, such as bathtub, ellipse, double ellipse and nebula type, the performance of a nebula type showed highest efficiency and engine output under lean mixture conditions. An investigation for various LPG fuel compositions was also carried out, and revealed that the case with 40% propane and 60% butane shows the lowest efficiency at stoichiometry, however, as the mixture became leaner its efficiency increased and became even higher than the case with 100% propane.

View full abstract

Characteristics of combustion and emissions of natural gas direct-injection (CNG DI) were studied by using a rapid compression machine. Results show that CNG DISC has the short overall combustion duration (Fig. 7(d)), high pressure rise due to combustion, and high rate of heat release (Fig. 12), which are considered to come from the charge stratification. The CNG DI can realize extremely lean combustion (Fig. 7(d)). Unburned methane showed almost the same level as that of homogeneous mixture combustion. CO increased steeply with the increase in equivalence ratio φ when φ was greater than 0.8 due to the excessive stratification, and NOx peak value shifted to the region of lower φ. Combustion inefficiency maintains less than 0.08 in the range of φ from 0.1 to 0.9(Fig. 1O(b)) and increases at very low φ due to bulk quenching and at higher φ due to excessive stratification. The combustion efficiency estimated from combustion products shows good agreement with that of heat release analysis(Fig. 12). Thus CNG DI is considered to have much potential to be used in a spark-ignition engine from the point of view of these beneficial characteristics of wide operation range and possibly high energy conversion efficiency.

View full abstract

Cogeneration systems, using natural gas engine, have been increased because of low NOx emission and low particulate emission. But recently electric power company have been reduced the price of electric power, so it is necessary to increase power generation efficiency of gas engine. To improve thermal efficiency of gas engine, we applied the Miller cycle to the lean burn gas engine. And we improved combustion system and turbocharger for Miller cycle. In order to decide compression ratio and expansion ratio, we used a cycle simulation and a single cylinder test engine. Compression ratio was set to 11 and expansion ratio was set to 15. Inlet swirl ratio was reduced to prevent knocking. In case of low swirl, the combustion rate was decreased, so ignition timing could be advanced. In case of Miller cycle, high boost pressure is necessary to keep engine power. So we developed the high pressure ratio and high efficiency turbocharger. And high durability type spark plug and electrical fuel supply system were applied for high reliability. Finally thermal efficiency of Miller cycle gas engine reached to 42%(Power generation efficiency is 40%). We confirmed that 42% of thermal efficiency during 4000Hr. We have already put 280kW cogeneration package in the market.

View full abstract

In order to reduce NOx and smoke simultaneously and also to improve markedly the trade-off between smoke and NOx without deteriorating fuel consumption, natural gas was charged homogeneously into the intake air and was burned igniting by a small amount of gas oil injection in a four cylinder naturally-aspirated DI diesel engine. Combustion tests were carried out by changing the ratio of the amount of natural gas and the amount of gas oil first, secondarily the intake preheating temperature, and thirdly the EGR rate respectively. Effects of the respective parameter on the ignition and the burning rate of natural gas, exhaust emissions and specific fuel consumption were investigated. It is found that significant improvement of Smoke-NOx trade-off can be obtained without deteriorating fuel consumption by the suitable combination between the natural gas charge rate, the EGR rate and the intake preheating temperature for each engine load condition as shown in Fig. 1. Concluding remarks are as follows ; (1) A high burning rate of natural gas results in shortening the combustion duration, then, leads to lower fuel consumption. (2) Too high burning rate of natural gas results in increase of NOx, on the other hand, too low burning rate of natural gas results in increases of fuel consumption and THC. (3) Increases in fuel consumption and THC due to incomplete combustion of natural gas at the low load are improved drastically by raising the intake charge temperature up to 120℃, which increases the burning rate of natural gas. (4) Increase in NOx due to high burning rate of natural gas at the high load is improved by lowering the intake charge temperature below 60℃, which suppresses the burning rate of natural gas. (5) High EGR rate shows suppression effect on the burning rate at the high load, however, the burning rate is hardly affected by EGR at the low load. (6) NOx is reduced effectively by EGR without deteriorating fuel consumption, which is due to the water content brought by EGR as well as the decrease in oxygen concentration of the intake charge.

View full abstract

In a compression ignition engine, using a rich and lean biform mixture composition that avoids near stoichiometric and extremely over-rich regions would be effective to suppress NOx formation without increasing smoke when the overall excess air ratio approaches the stoichiometric ratio. To realize the formation of rich and lean mixtures and the control of ignition timing, a dual-fuel diesel engine with an induced gas with resistance to self-ignition as the main fuel and with a small quantity of diesel fuel for the ignition source has potential merits. However, this method has the problem of knocking and misfiring when the percentage of inducted fuel is increased. In this research smokeless and ultra low NOx combustion without knocking over a wide operating range was established in a dual-fuel diesel engine with induced natural gas as the main fuel. Optimizations of the combustion chamber shape and operating factors, including EGR and intake air throttling, which determine conditions of the in-cylinder gas, were investigated at several IMEP conditions. A lean quasi-homogenous mixture was formed with induced natural gas in the whole combustion chamber while a small quantity of diesel fuel was directly injected for an ignition source, as shown in figure 1. The injection timing of the diesel fuel was set at a relatively early stage in the compression stroke to avoid smoke emissions. A piston cavity divided by a lip in the sidewall, shown in the figure, was suitable to confine diesel fuel into the lower part of the cavity, and this suppressed knocking just after ignition. It was also expected that a slightly richer than stoichiometric mixture is formed in the lower part of the cavity, and a leaner mixture is formed in the upper part of the cavity. This mixture formation process was analyzed with a CFD simulation, which showed the possibility for biform mixtures without near stoichiometric or extremely over-rich regions in the chamber. The results of the experiments showed that a combination of the divided cavity, EGR, and intake air throttling was effective to simultaneously eliminate knocking and reduce THC and NOx over a wide IMEP range. At high IMEP silent and smooth combustion without knocking was achieved even with a large amount of induced natural gas. Moreover, the maximum IMEP increased in comparison with conventional diesel operation with a lower injection pressure system because smoke emission was not a limitation. At medium IMEP it was effective to adopt EGR and intake air throttling for ultra low NOx under relatively lower THC. However, at low IMEP it was difficult to avoid increases in THC and ISEC while realizing ultra low NOx and smokeless operation. At lower overall excess air ratio conditions, NOx reduction was shown with a biform mixture composition without near stoichiometric or extremely rich regions.

View full abstract

In the present study, we proposed a novel fuel design concept in order to achieve low emissions and combustion control in engine systems. The fuel design concept is based on the combustion control that could be realized by using a mixed fuel consisting of a lower boiling point (b.p.) fuel and a higher b.p. fuel as shown in figure A. Mixing of n-pentane (low b.p. fuel), a gasoline component, with n-tridecane (high b.p. fuel), a gas oil component, was the selected fuel design concept of this study. The spray evaporation process of the mixed fuel was analyzed using planar laser-induced fluorescence (PLIF) and Mie scatter imaging. Vaporization of the mixed fuel is promoted through the occurrence of a two phase region. Additionally, combustion experiments were conducted inside a rapid compression and expansion machine (RCEM) in order to confirm combustion and exhaust constituent properties. The ignition delays for several mixture mole fractions obtained from the heat release profiles are presented in figure B. In the case of single component fuels having the same transport properties of mixed fuels, the ignition delay linearly increases with decreasing carbon number. However, in the case of the mixed fuels, it is confirmed that, compared to pure fuels of the same mean carbon number, ignitability is improved, even though the high ignitability fuel (n-tridecane) is diluted by the low ignitability fuel (n-pentane).

View full abstract

The objectives of this study are to clarify the influence of fuel physical properties on injection characteristics by means of numerical simulation of injection system fueled biodiesel fuel, and to clarify the influence of components of the fatty acid methyl ester on the engine performance and characteristics of exhaust emissions ; PM, NOx, etc, including unregulated emissions by experiments. In these experiments, several kinds of biodiesel and methyl oleate and methyl linoleate were investigated, both of which are main chemical components in biodiesel (Table 1). The curves of bulk modulus depending on liquid pressure and temperature are different between gas oil and biodiesel. The computed simulations show that a at lower fuel temperature ; 293 and 313 K, the injection pressure of biodiesel rises earlier than that of gas oil. This is because at lower liquid pressure, the bulk modulus of biodiesel is higher than that of gas oil (Fig. 2), so that the rate of liquid pressurerise goes up and the injection timing is advanced. And peak injection pressure at biodiesel is higher than that at gas oil in the case of lower mean injection pressure. But, it becomes be similar to that at gas oil under higher mean injection pressure. At higher fuel temperatures, there are no differences in injection timing and pressure between biodiesel and gas-oil. From experimental obtained on DI-CI engine, SOF in PM by biodiesel emitted more than that by gas oil on a low engine load (Fig. 3). From the results obtained on spray visualization by the laser-sheet technique, it was found that the spray penetration by biodiesel become shorter than that by gas oil, so that the mixing of air and fuel deteriorated, and a fuel-rich mixture is formed and emitted as SOF. A SOF concentration and ignition delay increase corresponded with a decrease of the fraction of methyl oleate ester in biodiesel fuel. This will be due to low ignitability of methyl linoleate ester.

View full abstract

Considerable amount in 400,000 〜 600,000 tons per year of waste vegetable oil in Japan is still flushed down the drain. Utilization of waste vegetable oil for diesel fuel leads to two advantages for environmental protection, to reduce CO_2 emission from engines and to avoid water pollution of rivers. In this study, combustion characteristics of waste vegetable oil methyl ester (WME) are in detail investigated by not only engine test run but also observation of burning flames in a visual engine. In order to reduce NOx emission of WME, emulsified waste rapeseed oil methyl ester (EME) also is tested. Engine test run shows that smoke emission from WME is lower than gas oil, that thermal efficiency of WME is the same as gas oil, and that NOx from WME is about 13 % higher than gas oil at full load. Though CO and HC emissions from WME are higher than from gas oil, absolute values of them are acceptable. According to the results, it is concluded that WME can be used for diesel engines instead of gas oil. Figure 1 shows the effects of water in EME on the emissions and thermal efficiency at full load. Paying attention to the point of 15 % water content, 18 % of NOx reduction is carried out accompanied by some reduction of smoke emission. Thermal efficiency is also successfully 4.5 % improved with 15 % water content. Figure 2 shows the heat release rate of EME with 15 % water compared with WME and gas oil. As EME contains water, larger quantity of EME must be injected than other two fuels to release the same heat. So, injection duration of EME is longer than the others. And, EME starts to burn later than other two fuels because of longer ignition delay. However, combustion end of EME is not later than the others, which means the burning speed of EME spray is higher. To investigate in detail the improvement of combustion by applying EME, visual study of burning flames is carried out and the photos are introduced in the full paper. According to such investigations, it is considered that the reduction of soot formation and the faster combustion after the end of injection with EME is achieved by improving air entrainment into the spray.

View full abstract

The present study focuses on clarifying the combustion mechanism of the homogeneous charge compression ignition (HCCI) engine in order to control ignition and combustion as well as to reduce HC and CO emissions by calculating the chemical kinetics of elementary reactions. For the calculations and experiments, n-Butane was selected as fuel since it is the fuel with the smallest carbon number in the alkane family that shows two-stage auto-ignition (heat release by low temperature reaction (LTR) and by high temperature reaction (HTR)) similar to higher hydrocarbons such as gasoline. The CHEMKIN code was used for the calculations assuming zero dimensions in the combustion chamber and adiabatic change. Calculations were carried out for various parameters, including : equivalence ratio, initial temperature, initial pressure, compression ratio, engine speed and initial chemical species concentrations. The influences of these parameters on the appearance timing of LTR and HTR, and on the combustion duration of LTR and HTR were clarified. The results of the present study indicate that in order to reduce HC and CO emissions in exhaust gas and/or to maintain high combustion efficiency for the HCCI engine, it is necessary that the maximum temperature of the combustion cycle be over 1500K.

View full abstract

In order to numerically predict the mixture formation and combustion processes in premixed compression ignition engines, submodels for fuel sprays (solid-cone and hollow-cone sprays), ignition and combustion were investigated. As an ignition model, the Livengood-Wu's model and the Schreiber's reduced kinetic model were examined. As a combustion model, the Reitz's model was employed. These submodels were incorporated into the authors' GTT code, and the mixture formation and combustion processes in the two types of premixed compression ignition engines (side-injection and central-injection engines) were numerically analyzed using this code. The validity of the submodels was confirmed by comparing the calculated results with the experimental ones. Figure I shows the mixture formation process in the side-injection engine. It has been found that using the Livengood-Wu's model the curves of heat release rate in the main combustion period in the test engines have been predicted reasonably well with the appropriate model constants. It has been also found that using the Schreiber's reduced kinetic model the curves of heat release rate in both periods of low-temperature oxidation and main combustion for two cases in the side-injection engine have been predicted considerably well with the appropriate parameters for the reaction rate constants. Figure II shows an example of the comparison between the experimental result of the heat release rate in the side-injection engine and the calculated result of the heat release rate using the Schreiber's model and the Reitz's model.

View full abstract

The pre-mixed lean diesel combustion (PREDIC) has the possibility of significant reduction of both NOx and smoke emissions. The experimental results of PREDIC without after treatment have shown a NOx emission less than 20 ppm, a level which the conventional diesel engines have never achieved. PREDIC owes to the adoption of electronically controlled common rail fuel injection equipment, which can inject fuel at a scheduled crank angle and provide a good fuel and air mixture by using a pintle swirl nozzle. However, there is a big problem in PREDIC, that is, its output is only a half load of the naturally aspirated conventional engine. In order to obtain higher load, a supercharged and intercooled system has been adopted in a single cylinder engine for PREDIC. PREDIC's load conditions by high boosting are improved in three times higher than in the naturally aspirated (NA) condition. At the same time, the high boost condition causes the deteriorations of both THC and smoke emissions. The emission deterioration of PREDIC under the high boosting condition has been analyzed by using the high-speed photography of laser shadowgraph at bottom view type single cylinder visualized engine. Both the conventional and PREDIC up to the high boosting condition of 3.4 times of NA engine are studied in order to clarify the difference of boosting effects. The processes of mixture formation in both the conventional combustion and the PREDIC have been investigated by observations of combustion photographs. Under the high boosting condition the PREDIC's spray by pintle swirl nozzle does not spread widely and the mixture formation gets worse due to the high cylinder pressure. The PREDIC's mixture formation is more sensitive than the conventional combustion because PREDIC's injection timings are very early and its cylinder pressures vary fifty times from atmospheric pressure when the injection starts. So a better or more improved injection nozzle is needed than the current pintle swirl nozzle and it expects to make an appropriate mixture in several cylinder pressures, which widely vary fifty times from an atmospheric pressure.

View full abstract

HCCI engines can operate with high thermal efficiency and significant low NOx emission. A new engine concept consists of HCCI combustion in low and middle load and spark ignition combustion in high load was introduced (Fig. 1). Conditions for HCCI was considered to adjust the negative valve overlap and effective compression ratio by intake valve close timing. And also, we investigated the influence of the mixture formation on auto-ignition using a direct injection engine. Fig.2 shows how to control ignition and combustion. As the result, HCCI combustion was achieved with relatively low compression ratio combined with an intake air heat by internal EGR. High thermal efficiency comparable to those of modern diesel engine, almost zero NOx emission and no smoke were realized. The mixture stratification increases the local concentration resulting in higher reactivity. The wide range of combustible A/F ratios explosion enables control of ignition timing of compression ignition. Combustion photographs indicate how the flame fills the entire combustion chamber to produce combustion, and to reduce emissions and fuel consumption.

View full abstract

In diesel engines, most of the fuel is burned in a diffusion combustion phase, which inherently leads to the formation of excessive amount of soot and NOx emissions. In order to decrease soot and NOx emissions simultaneously, the concept of lean premixed compression ignition engine has been proposed. The onset of the combustion of the engine depends on the autoignition of the fuel, so it is quite difficult to control the ignition timing. On the other hand, it has been revealed that Pulsed Flame Jet (PFJ) has a great potential to enhance ignition reliability and burning rate in lean mixtures. In PFJ, the combustion is initiated in the jet issuing from the igniter, that is, the combustion is initiated volumetrically. This volumetric combustion initiation must behave as a trigger for the autoignition of the fuel in the combustion chamber. In this paper, autoignition characteristics of n-butane/air mixtures in a rapid compression machine (RCM) were shown first. The appearance of low temperature flames was observed in autoignition of n-butane in the RCM used here, and it was realized that the final compression conditions in the RCM correspond to the upper end of the low-temperature range of the positive temperature dependence region of ignition delay. Then the combustion tests with PFJ were carried out and it was demonstrated that the onset of combustion can be controlled by PFJ, and it was revealed that PFJ has a potential for the ignition timing control of the compression ignition premixed charge engine.

View full abstract

A homogeneous charge compression ignition (HCCI) engine has been known to have high thermal efficiency and low nitrogen oxide emission. However, the control of ignition and its combustion period over wide range of engine speed and load is one of the barriers of the realization of the engine. In the lean side of equivalence ratio, control of ignition is difficult for its long delay of ignition and there are knock-like problems in the high range of load. Regarding the extension of the range of load, the onset of hot flame and its combustion period were examined for a wide range of fuel input by Chemkin II computation and experiment. In the computation, detailed reaction scheme of dimethyl ether proposed by P. Dagaut et al. in 1998 was used in the study. Also, the computational results were compared to the experimental results from a HCCI engine. The intake temperature and the amount of CO_2 mixing were changed for the control of hot flame onset angle. The onset angles of cool flame from computation corresponded well with experimental results. However, the pressure rise in this stage was well higher than that in experimental results. Therefore, the onset angle of hot flame was more advanced than that was in experiment. From the computational results, a flat region of hot flame onset angle where hot flame onset angle hardly changed with increase in fuel input could be found for a given intake temperature and CO_2 mixing rate in the rich side, which was caused by opposing effects of the raised specific heat of the mixture and advance in ignition delay between cool and hot flame. However, the confirmation of flat region was not realized in the experiment for the progressive advance of cool flame onset angle and knock-like combustion near rich side. In the experiment, the operable range (the possible range of fuel input from just ignitable to knock-occurring fuel input) shifted to rich side with decrease in intake temperature while the range of fuel input was reduced, as the hot flame onset angle advanced more quickly than it did under high intake temperature. And the progressive advance of cool flame due to the increased temperature of residual gas and cylinder wall by raised fuel input is considered to take an important role in the fast advance of hot flame onset angle and generation of knock-like combustion. However, the mixing of CO_2 made the operable range shift to the rich side with keeping up the width of it as shown in Fig. 1, possibly due to the relatively small portion of heat generation in cool flame. The duration of hot flame was examined too, and it was confirmed that the duration was mostly dependent on fuel input and its onset angle, in spite of any variation of intake temperature or rate of CO_2 mixing as shown in Fig. 2.

View full abstract

It has been respected to realize large-scale natural-gas engines. Low flame-propagation speed and poor spatial in-cylinder penetration of gaseous fuels impede the large-scale natural-gas practical. Premixed compression-ignition (homogeneous-charge compression-ignition, HCCI) engines have not such kind of problems, because natural-gas and air are supplied as mixtures, and burn out through speedy knocking-like combustion. However, a very narrow operating range could be established due to a lack of autoignition-timing control procedures. A formaldehyde-assisted premixed-compression-ignition natural-gas engine concept has been proposed previously by the authors to realize large scale natural-gas engines having ignition-timing control procedures. Generation/consumption histories of chemical species during the preflame induction period approaching to the final hot-flame ignition were investigated concerning the formaldehyde-assisted compression-ignition natural-gas engine. Small amount of formaldehyde is supplied as an additive into the premixed intake charge of natural gas and air. The formaldehyde addition has a strong promoting effect for lean mixture ignition of natural gas. When suitable amount of formaldehyde, even a hundreds ppm order of magnitude, is added into the intake fuel/air mixture the piston-compression ignition will occur adequately near the top dead center. This procedure enabled us to obtain ignition or hot-flame explosion even of the fuel/air mixtures nonflammable through a simple piston compression. The formaldehyde acts efficaciously as the ignition-promoting additive for the methane-based gaseous fuels which are weak in cool-flame generation during the preflame induction periods. Natural gas is a typical one of this kind. The in-cylinder gas composition histories were obtained by gas sampling / analyzing processes with a magnetic operating valve and gas chromatographs, concerning mainly to the methane, carbon monoxide, carbon dioxide and formaldehyde. Experiment was carried out using a single cylinder engine and a commercial natural gas 13A. The formaldehyde concentration shows a slight rise followed by prompt decrease at the final stage of the ignition delay period ; so-called blue-flame period, but seemingly stable during almost the whole induction period up to the ignition. A piston compression of a simple charge of natural gas and air with no intake formaldehyde addition showed a gradual formaldehyde generation during the preflame period and sometimes slight consumption at the final stage before the hot-flame occurrence. When the intake air is mixed with formaldehyde only, i. e., no fuel is included, the formaldehyde is consumed briskly near the top dead center, and shows a small but recognizable pressure rise due to a heat release. The effect of formaldehyde added into the fuel/air mixture leading to the ignition would not be an event antecedent to the natural gas preflame reaction but a promoting event of the preflame reaction of the main natural gas fuel. It is demonstrated that during the induction period preflame reaction of main natural gas fuel is under way the effect of formaldehyde addition become extremely efficacious at the final stage just before the heat release of hot flame occurrence.

View full abstract

The advantages of homogeneous combustion in internal combustion (I. C.) engines are well known and many research groups all over the world are working on its practical realization. Recently, the present authors have proposed a new combustion concept that fulfils all requirements to perform homogeneous combustion in I. C. engines using the Porous Medium Combustion Enigine, called "PM-engine". This is an I. C. engine with the following processes realized in a porous medium : internal heat recuperation, fuel injection and vaporization, mixing with air, homogenization, 3D-thermal self-ignition followed by a homogeneous combustion. Figure 1 shows the simplest case of the operation of a PM-engine, where the PM-combustion chamber is mounted in engine head. During the intake stroke it is weak influence of the PM-heat capacitor on the in-cylinder air thermodynamic conditions. Heat exchange process (non-adiabatic compression) increases with continuing compression, and at the TDC the whole combustion air is closed in the PM volume. Near the TDC of compression the fuel is injected in to PM volume and very fast fuel vaporization and mixing with air occur in 3D-structure of PM-engine. The self-ignition process and homogeneous combustion occur in PM volume close to the TDC. The main features of the PM-engine are the following : 1) Very low emissions level due to homogeneous combustion and controlled temperature in the PM-combustion zone (e.g. NOx between 100 and 300 mg/kWh for the (A/F) ratio from 1 to 5 ;, CO can be reduced by several times ; (almost) eliminated soot formation). 2) Theoretically higher cycle efficiency due to similarity to the Carnot cycle. 3) Very low combustion noise due to significantly reduced pressure peaks. 4) Nearly constant and homogeneous combustion temperature field in the PM-volume. 5) Very fast combustion. 6) Multi-fuel system. 7) May operate with homogeneous charge : from stoichiometric to very lean mixture compositions. 8) Weak effect of in-cylinder flow structure, turbulence or spray atomization.

View full abstract

Taking advantage of several of the most modern research and development achievements in combustion technology, material science, control engineering, together with some unconventional thoughts on thermodynamics, a novel concept of a steam engine is presented, which is especially suitable for automotive engine applications. The paper starts by describing the general concept of the so-called Zero Emission Steam Engine, which includes supercritical thermodynamic states and high-pressure steam injection, and further focuses on the porous burner technology as the actual heat source of the process. Following a brief description of the principle of operation, the unique advantages of this novel heat source for the steam engine are outlined. These advantages include a wide, infinite variable power turndown ratio of about 1 : 20, which is roughly a factor of 5 greater than the modulation range of competitive burner technologies. Additionally, the porous burner excels by very low combustion emissions that are independent of the actual thermal load of the burner and are - at any state of operation - clearly below the most stringent Super Ultra Low Emission Vehicle (SULEV) standards. Another advantage of immense importance in mobile applications is the compactness of the porous burner units. Using superior heat transport properties, the porous burner allows power densities of about 3.000 kW/m^2, which is about a factor of 10 greater than the value of other premixed, low emission burner technologies, resulting in very small burner units. Finally, the porous burner technology allows complex combustion chamber geometries to be realized, which may be of special interest in applications where space is limited. Using these advantages, several porous burners for different concepts of ZEE-engines have been developed and are currently undergoing testing from which the results are presented in this paper. Although the presented prototypes are far from being mass-produced, the experimental results obtained from the prototypes clearly indicate that porous medium burners together with a modern steam engine concept can indeed be successfully applied to car engines and offer significant advantages compared to conventional state-of-the-art engines.

View full abstract

Direct injection gasoline engines have been developed for improvement of fuel economy and exhaust emissions. Recently, a new stratified charge combustion concept is proposed. A slit nozzle is adopted to realize the new concept. The nozzle has a rectangular orifice as shown in figure 1 and forms a thin-fan shaped spray. This paper describes the spray characteristics of the slit nozzle. The following results were obtained. (1) The spray tip penetration increases with increasing the slit thickness as shown in figure 2. And the calculated results based on Waguri's spray momentum theory agree with the measured spray tip penetration. (2) The relation between the Sauter mean diameter and the effective factors is described by the following equations. For lower injection pressure ΔP=0.1 MPa to 0.3 MPa : d_<32> ∝ ΔP^<-0.25>・H^<0.3> For higher injection pressure ΔP=6 MPa to 14 MPa : d_<32> ∝ ΔP^<-0.5>・H^<0.1> where d_<32> is Sauter mean diameter, ΔP is injection pressure and H is slit thickness. Next, the spray characteristics of the slit nozzle are compared with that of a swirl nozzle. Both nozzles are commercially available. (3) The Sauter mean diameter of the slit nozzle is slightly smaller than that of the swirl nozzle as shown in figure 3. (4) As shown in figure 4, at 0.1 MPa back pressure, the difference in the spray concentration of both nozzles is small. On the other hand, at 0.5 MPa back pressure, the spray concentration of the slit nozzle is lower than that of the swirl nozzle, especially in the center of the spray. (5) It is clarified from the above mentioned results that the feature of the slit nozzle is high spray penetration, widely diffusion spray and fine atomization.

View full abstract

This paper describes the numerical and experimental approaches that were applied to study swirl injectors that are widely used in direct-injection gasoline engines. As the numerical approach, the fuel and air flow inside an injector was first analyzed by using a two-phase flow analysis method employing a volume of fluid (VOF) model. The calculated results made clear the process from initial spray formation to liquid film formation. Spray droplet formation was then analyzed with a discrete droplet model (DDM). As the experimental approach, particle image velocimetry (PIV) was used to measure the spray velocity distribution. These approaches were applied to test nozzles having a tapered tip geometry at the nozzle exit (Fig. I). The spray shapes that are come out from the nozzles skewed to tapered side (Side A). And the cone angle of skewed side spray does not change not so much even under high ambient pressure conditions (Fig. II). Because of this reason since the tapered nozzle is able to make fuel mixture reach towards the spark plugs on the stratified engine condition, i.e. compression condition, we consider this nozzle as the effective tool. At first, internal flow analysis of the nozzles by using VOF model was done, and the results show that the cone angle at the skewed side is larger than that of non-skewed side (Fig. III). After that, Velocity distribution during injection were measured by using PIV (Fig. IV). And the results show that as it apart from the nozzle exit, skewed spray velocity decline much slower than non-skewed spray velocity does. This phenomena are caused mainly by the reason that the quantity of the fuel of the skewed side (Side A) is larger than that of the iron-skewed side (Side B). Spray calculation was done setting this circumferential fuel distribution as initial conditions, the results show on Fig. II, The calculated results show good agreement with the experimental results. These results above show that the spray shapes are influenced mainly by spray cone angle and circumferential fuel mass distribution at the nozzle exit.

View full abstract

The experiments on the instantaneous fuel flow rate and the spray characteristics, velocity and Sauter mean diameter, are introduced in this paper. Five swirl nozzles, which are used for gasoline direct injection engines and are the same design however differed in the static flow rate as 600, 700, 800, 900 and 1000 cc/min respectively, are utilized. Alternative fuel of normal-heptane instead of gasoline is employed for all experiments. The flow rate measurements have been performed under the injection pressure of 7.0 MPa and injection frequency of 16.7 Hz. The amount of fuel at each injection is set as certain values for all nozzles so that valve-opening duration is changed from 0.7 to 3.51 ms for each nozzle. A laser Doppler anemometer (LDA) is applied to measure the centerline velocity of the test section within a quartz-glass pipe of 3.5 mm inner diameter. The instantaneous fuel flow rate and integrated mass are simulated using measured centerline velocity. For the spray measurement, a phase Doppler anemometer (PDA) is applied to obtain the droplet velocity and diameter distribution pattern. The experiments are made under the fuel injection amount of 24.4mg/cycle for 700 and 1000 cc/min nozzles and 8.6 mg/cycle for all the nozzles. An example of flow rate results is shown in Fig. 1, where the fuel injection amount is set to 8.6 mg/cycle for the nozzles of 700 and 1000 cc/min. In the graph, the abscissa axis is the phase angle. The phase angle of 360 degree is corresponding to one injection period. The result shows that for the nozzle of 700 cc/min, its instantaneous flow rate is larger than that of 1000 cc/min at initial stage (from 53 to 56 degree) but is smaller that that of nozzle of 1000 cc/min as injection processes. After 60 degree, sharp decreasing flow rate is observed for the nozzle of 1000 cc/min. This feature is related with the valve opening duration, valve moving speed and the flow coefficient at the nozzle passage. Negative fuel flow rate during 62 to 67 degree and after 78 degree are observed in the figure. They are caused by the oscillating flow inside the pipe of test section and it is related with the inner diameter and length of the pipe and/or fuel flow line volume. Figure 2 shows the radial distribution of mean velocities for all nozzles. The measuring position is at z = 25 mm and injection amount is 8.6 mg/cycle. It can be seen that the position of the maximum mean velocity is shift to the outer position with increasing the static flow rate of the nozzle except for the 1000 cc/min nozzle. It is considered that the results are related with the spray angle because increasing in the instantaneous fuel flow rate causes the different spray angle and the different maximum velocity. At last, the time dividing analysis for intermittent spray is also presented.

View full abstract

Measurement of the ambient air velocity distributions in and around a D. I. gasoline engine spray was made by combining the laser induced fluorescence technique and the particle image velocimetry (LIF-PIV) technique. The rhodamine and water solution was injected into ambient air by a swirl-type injector to disperse the fine fluorescent liquid particles as tracers for the ambient air motion. A fuel spray was injected into the fluorescent tracer cloud by the D. I. gasoline injector and was illuminated by an Nd : YAG laser light sheet (532nm). The scattered lights from the spray droplets and tracers were cut off by a high-pass filter (>580nm). As the rhodamine absorbs the light of the laser, it fluoresces at a light whose wavelength greater than 600nm. The fluorescence from the tracers, were transmitted through the high-pass filter and the images were captured by using a high-resolution digital CCD camera. The tracer images were analyzed by double frame cross-correlation PIV and the ambient air velocity distribution could be obtained. The ambient air velocity distribution by the rhodamine tracer was compared with the ambient air velocity distribution by the microballoon tracer to validate accuracy of the tests the velocity distribution of the rhodamine tracer was almost the same as that of the microballoon tracer around the spray. Then, this technique was applied to a D. I. gasoline spray. As shown in Fig. A-(a), the ambient air flows up around the spray and enters into the tail of the spray. As shown in Figure A-(b) shows the comparison of the velocities between the spray and ambient air.

View full abstract

Predictions of the mixture formation process inside gasoline DI engines are strongly required to improve both the fuel consumption rate and the exhaust emissions. Swirl-type injectors are commonly used for gasoline DI engines, as its spray characteristics are favorable for gasoline DI engines. The spray formation of a swirl-type injector consists of a liquid sheet and droplets that arise from the breakup of the liquid sheet. Numerical simulations of free sprays formed by a swirl-type injector are carried out on the basis of a method of DDM (Discrete Droplet Model) with employing different models for the breakup of the liquid sheet and that of droplets downstream. The breakup of the liquid sheet is modeled by Reitz's wave breakup model while that of droplets downstream adopts TAB (Taylor Analogy Breakup) model. Schematic diagram is illustrated in Figure 1. In this study, the boundary condition of the ambient pressure is set at two values ; a negative pressure and a high pressure. The droplet deformation calculated by the breakup model is incorporated into the drag force term to take the influence of the drag variations into account. The results of calculated parcels on centered vertical cross-section are shown in Figure 2. Figure 2(a) is the result assuming that the drag force is of a rigid sphere (without droplet deformation), while Figure 2(b) is the result assuming that the drag force is of an ellipsoid (with droplet deformation). As a result, by taking the drag force variation due to deformation of droplets into account, calculations under high ambient pressure can reproduce the change of spray shape. The drag force of droplets and the pressure difference of spray inside and outside cone contribute to the change of the spray shape. This pressure difference is small when the ambient pressure is low.

View full abstract

High pressure fuel injection in a direct injection diesel engine is one of the most effective methods to reduce PM emissions. Pilot injection is also used to reduce NOx emissions and engine noise. Little research has investigated the fuel spray behavior under high pressure injection and pilot injection, and cylinder to cylinder deviations in spray behavior and the correlations with deviations in HC emissions are not clear. This study attempted to determine the relationship between cylinder to cylinder deviations in spray configuration and deviations in THC emissions with high pressure and pilot injections. It was found that an increase in injection pressure causes decreases in hole to hole and nozzle to nozzle spray deviations (Fig. 1). Very high injection pressures cause unstable spray development due to leakage from the magnetic valve. Deviations in THC emissions have a minimum at 40 to 80 MPa injection pressures (Fig. 2). Pilot injection of a very small quantity of fuel causes larger deviations in the spray configuration and mass. The amount of fuel adhering to the walls is less with pilot injection but there is no significant improvement in the THC emission while the HC components in the exhaust gas differ.

View full abstract

Some recent experimental studies have shown that with split injection strategy, the soot and NOx emissions from a diesel engine can be reduced significantly in comparison with the conventional non-split injection strategy. To understand the mechanism of emissions reduction, it is essential to clarify the process of the fuel mixture formation. In order to characterize the droplets and vapor concentration distribution inside the fuel sprays, a dual-wavelength laser absorption-scattering technique was developed by using the second harmonic (532nm) and the fourth harmonic (266nm) of a Nd : YAG laser and using dimethylnaphthalene as the test fuel. It was clarified that dimethylnaphthalene, which has physical properties similar to diesel fuel, is transparent in visible light near 532nm and is a strong absorber of ultraviolet light near 266nm. Based on this result, the vapor concentration in a fuel spray can be determined by the two separate measurements : a transmission measurement at a non-absorbing wavelength to detect the droplets optical thickness and a transmission measurement at an absorbing wavelength to detect the joint vapor and droplets optical thickness. The droplets density can be determined by extinction imaging through the transmission at the non-absorbing wavelength. By applying the ultraviolet-visible laser imaging technique, the distributions of droplets and vapor concentration in the spray, which was injected into a high-temperature and high-pressure nitrogen gas in a constant volume vessel, were measured and quantitatively analyzed. The effect of injection mass ratio of double-pulse injection on distributions of equivalence ratios of vapor and droplets was clarified. In the following figure, the contours of equivalence ratios of vapor and droplets in the diesel sprays with different injection ratios are shown. It has been found that the injection mass ratio has significant effect on the fuel distributions in the diesel spray ; the subsequent injection of a double-pulse injection scheme has a turbulent effect on the fuel-air mixing in the diesel spray injected in the preceding pulse.

View full abstract

The use of in-cylinder computational fluid dynamics (CFD) to model fuel and air interaction is increasingly being used to rapidly design and develop direct injection combustion systems. Use of such CFD techniques has substituted many physical engine testing experiments and hence shortened development times. However, the fundamental propagation of diesel fuel spray is critical in resolving air and fuel mixing characteristics. The lack of realistic measured diesel spray data and inappropriate phenomenological correlation lead the authors to investigate diesel fuel spray at conditions representative of a modern common rail equipped turbocharged and after-cooled HSDI diesel engine. Operating conditions were achieved in an optical rapid compression machine fitted with a common rail fuel injector. The initial stages of these investigations are described within this paper, where both stills and high speed imaging techniques were used. The influences of injector nozzle configuration, injection pressure and air charge conditions on the diesel fuel spray were examined using back-lighting techniques. Qualitative differences in spray structure were observed between tests performed with short and long injection periods. Changes in the flow structure within the nozzle could be the source of this effect. Differences in the fuel spray liquid core were observed between VCO (Valve Covers Orifice) and mini-sac nozzles, with the mini-sac nozzles showing a higher rate of penetration under the same conditions.

View full abstract

Behavior of spray impinged on a recessed-wall was experimentally investigated as a new concept of impingement diesel spray. The spray injected through a diesel nozzle impinged vertically on the center of the circular recess. Spray height as well as spray path penetration was measured at various timings on high speed photographs of the spray, which was injected by a single shot injection system into a high pressure chamber. The spray after impinging on the recessed-wall was reflected from the recess with rolling up motion more than the impingement spray on a flat wall (upper photographs in Fig. 1). However, in a case that the spray width at the wall location was wider than the recess diameter, the spray after impingement expanded along the wall (lower photographs in Fig. 1). Figure 2 shows parameters of spray impinging on a recessed-wall. The spray behavior on the recessed wall was controlled more by the width-to-diameter ratio Ws/Dr than by the recess diameter (Fig. 3). Also, for the purpose of increasing the spray height, it is necessary that the value of Ws/Dr should be around 0.5 (Fig. 4). At this ratio, the rolling-up motion was observed clearly and strongly on the recessed-wall.

View full abstract

Rapid mixing of fuel and air is an essential factor in improving combustion and emissions of diesel engines. Thus extensive investigation has been made to increase turbulence in the diesel combustion process by increasing swirl ratios, injection velocity, and modifying combustion chamber configurations. However, the relationship between the microscopic structure of the heterogeneous distribution of fuel clouds and the local turbulence structure is not well understood. Additionally there is no appropriate index for the analysis of the degree and size scale of heterogeneity. This paper investigates local diffusion phenomena with focusing on scales of fuel cloud and eddies based on a newly developed entropy method. The entropy analysis is based on the concept of statistical entropy, and it identifies the degree of homogeneity in the fuel concentration. The entropy values increase with the progress of uniformity in the diffusion process. Figure A1 is an example of entropy values comparing two different states of heterogeneity. The pictures show tracers mixed in fuel jets and were taken by a laser sheet. The entropy value of the N_2 jet image is higher than the value of the C_2H_4 image, indicating the higher degree of diffusion. By analyzing the speed of change in the entropy values, the diffusion intensity of the fuel cloud can be estimated, and it is also possible to identify size scales of the heterogeneity. Using the entropy analysis, the microscopic structure in turbulent jets and diesel sprays was investigated. Figure A2 is a picture of a turbulent jet simulating a diesel spray. In the experiment, a fluorescent compound is injected in to water. The PIV method was used for the analysis of the velocity distribution, and the diffusion intensity was obtained by the analysis of the local entropy. The results show that the diffusion intensity is the highest in the vicinity of the nozzle exit, and the heterogeneity scale is the smallest here. The heterogeneity scale increases gradually along the spray axis towards the downstream, with smaller size scales in the large clouds. In the downstream region, small-scale structures diffuse and become unclear, while large scale structures clearly remain. The paper details the microscopic structure of the heterogeneity in diesel sprays, and it demonstrates availability of the entropy method.

View full abstract

A spray wall impingement model that conserves mass, tangential momentum, and energy of an impinging parcel is developed for multidimensional spray calculations. This model focuses on spray impact on dry and wet surfaces below the fuel's Leidenfrost temperature, a scenario encountered under typical engine operating conditions [10]. The model is integrated into the KIVA-3V framework. Three splashing parcels and one wallfilm parcel are used to represent the shattering of a splashing droplet upon impact with the surface. It is assumed that the impulsive force on an impinging droplet normal to the surface is dominant allowing one to treat the magnitude of its tangential momentum component constant after impact. The viscous dissipation of an impinging droplet and kinetic energy of the wallfilm are accounted for in the energy conservation equation. The new spray impact model is validated using experimental measurements of a evaporating solid cone spray of n-tridecan at 550K impinging normally onto a flat plate at elevated temperature and pressure [7]. The new model shows improvements over a group of previous impingement models in predicting the spread rate of the liquid and vapor phases the spray. An energy balance of a splashing impinging parcel illustrates that more than half of its incoming energy is lost to viscous dissipation at high Weber Numbers (We). Most of the remaining energy is consumed by the kinetic energy of the smaller secondary parcels. At lower We, viscous dissipation decreases in magnitude, while the energy used to form a wallfilm on the surface becomes predominant. Finally, the effects of injection velocity and wall temperature on the radial penetrations of the liquid and vapor are investigated.

View full abstract

The deformation, breakup, and evaporation processes of liquid droplets after impinging on hot and cold solid walls are analyzed with the cyto-fluid dynamic theory on liquid droplet deformations (Naitoh, K. and Takagi, Y, SAE Paper 962017,1996, K. Naitoh, Oil&Gas Science and Technology, Vol. 54, No. 2, 1999), the partial differential equation describing heat transfer process inside the droplet, and experimental visualizations. It should be emphasized that the cyto-fluid dynamic theory can predict the number of the child droplets caused by wall-impingiment and the mass of liquid remaining on the cold wall, while varying parent droplet sizes and parent droplet speeds. Then, the breakup processes at cold walls are analyzed with the experimental photographs. Figure 1 shows the experimental photograph of the child water droplets after the impingement on cold wet steel wall. The steel wall shows us a beautiful flower shape of child droplets. Next, the Leidenfrost phenomenon, is analyzed, that evaporation time of liquid droplet after hot-wall-impingiment is drastically changed according to increasing wall temperature. Figure 2 shows the calculated evaporation time plotted against wall temperature, when a droplet of n-heptane is employed. We can observe that the evaporation time drastically increases around the Leidenfrost temperature. The calculated results shown in Fig. 2 agree well with the experimental data (Chandra, S. and Avedisian, C. T., Proc., R. Soc., London A, 1991). The present analysis shows that, when the wall temperature is around the Leidenfrost point, the speed of expansion flow of vapor will become comparable to the deformation speed of droplet, that is, the speed of gravity center of the droplet on wall.

View full abstract

To clarify the correlation between the motion of large-scale vortices and the uneven distribution of fuel in diesel sprays, a transient non-evaporating spray was analyzed by Large Eddy Simulation (LES). Because LES can describe large-scale vortices formed in turbulent shear flow of high Reynolds number, the numerical method was used for the simulation of transient spray phenomena. The calculated distributions of fuel mass concentration and vorticity of gas phase showed that high concentration region is distributed in periphery of large-scale vortices. This result indicates that the branch-like distribution that has been observed in an actual diesel spray is induced by the large-scale vortices formed in the spray. The predicted distribution of local Sauter mean diameter of fuel droplets showed that the larger fuel droplets are gathered in spray periphery. This trend in the distribution of mean size of droplets corresponds with experimental data obtained by Yeh et al. [1]. The comparison among the distribution of fuel mass concentration in different ranges of Stokes number (St) showed that fuel droplets in a range of St < 1 are distributed inside the vortices, while fuel droplets in a range of St > 1 tends to gather in the periphery of vortices. The correlation between the motion of large-scale vortices and fuel droplets distribution are weaker in the case of St > 100 than that of St < 100. This is due to that the fuel droplets of larger Stokes numbers are transported outside by centrifugal force of large-scale vortices. These results indicates that the distribution of fuel droplets is determined by the effect of the distribution of droplets size on the momentum interaction between gas and droplets.

View full abstract

Laser spectroscopy is a powerful tool for performing engine diagnostics with high temporal and spatial resolution. The conventional technique for generating cycle-resolved measurements of species concentration is to collect data from independent engine cycles and to successively delay the laser shot from measurement to measurement to match the desired range of crank angles. However, due to cycle-to-cycle variations in combustion engines this approach will only produce either data sets of physically uncorrelated events or averaged results depending on data acquisition strategy. Ideally therefore measurement data should be recorded within a single cycle. Until now this has not been possible for species selective visualisation, owing to a lack of high power, high-speed tuneable laser sources and detectors. This paper presents true single cycle resolved measurements of fuel distribution, OH-distribution and chemiluminescence in a laboratory spark ignition (SI)-engine. A unique laser and detection system for high-speed imaging was used for this purpose. The laser source consists of four individual Nd : YAG lasers, whose beams are combined to provide a single output beam. Each laser can be operated in a double pulse mode, allowing for a maximum number of eight pulses to be produced per burst sequence. The time delay between the pulses can be arbitrarily varied from a few nanoseconds to 100 milliseconds. A framing camera, capable of recording eight images with maximum repetition rates of 100 MHz, was used as detector. Planar Laser Induced Fluorescence (PLIF) was used both for the fuel visualisation and for the OH measurements. 3-pentanone was used as tracer species for the fuel. The fundamental YAG-wavelength was quadrupled for excitation of 3-pentanone at 266 nm. Since it is rapidly pyrolised, 3-pentanone acts as a marker for unburned regions after onset of combustion. To investigate the changes of the reaction zone during a single cycle, OH fluorescence was used. For excitation at 283 nm a dye-laser, pumped by the Nd : YAG-cluster, was employed. The framing camera recorded the corresponding fluorescence at 309 nm. These images visualise the flame front as well as hot post flame regions. Integral aspects of flame propagation could be observed on line-of-sight chemiluminescence images, also recorded by the fast framing camera.

View full abstract

In this work we study mixture preparation inside the cylinder of an optically accessible DI gasoline-engine. Crank angle resolved investigation of the atomization process of different model-fuels is carried out by means of planar Mie-scattering. The vaporization of the liquid fuel and the transport processes inside the cylinder are studied by Planar Laser-Induced-Exciplex-Fluorescence (PLIEF) using two different exciplex-systems and two different model-fuels. Besides the standard mono-component model fuel isooctane a newly developed multi component fuel (MCF) was examined. It represents the boiling behaviour of standard gasoline quite well and shows almost the same spectroscopic characteristics as isooctane. The simultaneously recorded crank angle resolved distributions of liquid and vaporized fuel for a stratified and homogeneous operation are visualized for both model fuels and both exciplex-tracers. The corrected intensity values of the fluorescence emissions of the two exciplex-systems TMPD/naphthalene and TEA/benzene were compared under same operation conditions. Furthermore cyclic variations of vaporized fuel distributions near the spark plug are investigated. Simultaneous detection of the in-cylinder pressure allowed the correlation between the fuel vapor distribution near the spark plug and the combustion quality of air-fuel-mixture.

View full abstract

A new correction method for laser absorption in LIF measurements inside engine cylinders is proposed. Frequently we have a problem of absorption of excitation laser by high-density molecules in LIF measurements at high concentration media such as in-cylinder gas. We developed a method to correct the attenuation effect by use of two counter propagating laser beams. Corrected information on molecule distribution is obtained by taking square root of data after multiplying two data which were from measurements by each excitation beam. A couple of laser light pulse A and B propagating in counter directions is considered. Pulse A is propagating in ^+x direction and pulse B is going in -x direction. Pulse A induces a fluorescence light [numerical formula] (1) where we considered laser light absorption using Lambert-Beer's law. Here, φ_A, n, η, B, I_A and α(x,y) are detection efficiency, density of the molecule, quantum yield which includes ro-vibronic state population, Einstein's B coefficient, excitation laser light intensity, and absorption coefficient (=σ×n(x, y), σ is absorption cross section). Similarly, fluorescence signal induced by pulse B is described as [numerical formula] (2) Multiplying and taking square root of these two equations, we obtain [numerical formula] (3) This Eq. (3) does not include α. This means that effect from laser absorption was canceled out. It is emphasized that α does not have to be constant, but can be a function of space. That is, this method is applicable to any fields including heterogeneous fields in principle. Eq. (3) means that we need intensity profiles of I_A and I_B at x=0. The origin for x-axis can be determined freely, but I_A and I_B must be known at a same position. Namely, if I_A was measured at the location before incidence to the measurement field (as described in Fig. 1), I_B has to be measured after it propagated through the measurement field ( after the attenuation). In order to conduct absolute measurements, one will need η(x, y) which includes quantum yield and statistical ro-vibronic population determined by temperature. We should note that re-absorption of fluorescence light is not counted into this method, and generally we should be careful on this issue. However, in case of LIF of ketone, which we will demonstrate later in this paper, the fluorescence spectrum is shifted to region from 350 to 450 nm because of vibrational relaxation inside S_1 state and it does not overlap with absorption spectrum which located from 250 to 310 nm. This means that re-absorption of fluorescence light does not occur in case of ketone LIF. Another method using two counter propagating laser beams had been presented by Versluis et al [1], in 1997. Our method is totally new one and is physically different from the method proposed by Versluis et al.. In this paper, differences between Versluis' method and our method are discussed, and it is clarified that our method has a considerable advantage in case of noisy measurement fields. We also demonstrate a fuel distribution measurement by our new method.

View full abstract

Beginning in 2004, all passenger cars and light-duty trucks sold in the U. S. will be required to meet the same emission regulations regardless of the fuel type (EPA Tier 2). Thus, for the first time gasoline-fueled vehicles will have to meet particulate matter emission standards that in the past have only been required of diesel-fueled vehicles. It is expected that contemporary port fuel-injected (PFI) engines with catalysts will be able to meet the requirements, which are currently based on particulate mass. In contrast, it is uncertain at this time whether the new generation of direct-injection gasoline engines can achieve certification. Furthermore, health risks are more strongly correlated with particle size and number density than with mass concentration. If new regulations are based on these parameters, gasoline engines will come under closer scrutiny. To help meet these new and anticipated regulations, improved particulate measurement techniques are needed. The two capabilities most urgently required are fast response to follow transients, and high-sensitivity to measure small amounts in short times. Laser-induced incandescence (LIT) offers these capabilities. The purpose of this paper is to demonstrate the use of LII for making time-resolved measurements of soot volume fraction during engine transients. The measurements presented were made at the exhaust-port exit of a production, four-valve, PFI spark-ignition engine with only one active cylinder. Optical access to the exhaust flow between the head and exhaust manifold was achieved by an optical flow channel with three windows, which permitted a laser beam to cross the exhaust steam while observed from above. Planar LII imaging was used to investigate the spatial distribution of soot across the exhaust port ; analog LII measurements were made to measure the soot volume fraction during transient engine operation. The planar LII images obtained in the exhaust port show that soot concentration varies widely both spatially and temporally, from cycle-to-cycle. For some cycles the soot distribution was relatively uniform, whereas for others it varied greatly between the two valves. On average, the amount of soot emitted from each valve was different. For open-valve injection, the soot volume fraction increased rapidly over the first 20 cycles of a simulated cold start, and then continued to increase at a slower rate for the duration of the 300 cycle test. In contrast, for closed-valve injection the soot volume fraction rose very quickly to a maximum in less than 10 cycles, and then fell rapidly to a low concentration by about 50 cycles, after which it continued to fall gradually for the remainder of the test. We believe the large differences in the soot behavior for these two cases is related to the dynamics of the in-cylinder wall films, and in particular, to the changing composition of the films as wall temperatures increase. For a snap-throttle transient, where the manifold pressure was suddenly increased from 40 to 98 kPa, the soot volume fraction increased steadily during the throttled period, fell to near zero at the time of the snap-throttle, and then increased steadily to a value approximately ten times greater than for the throttled period. Severe window fouling occurred during the wide-open-throttle portion of this test, resulting in a 30% loss in laser beam transmission after engine shutdown. This fouling problem limited this study to qualitative observations, and represents the main disadvantage of optical techniques for engine exhaust studies. Fortunately, it is a design problem that can be corrected, although at the expense of added complexity.

View full abstract

Diesel engine emissions have gained increased attention during the last two decades. Especially participate matter and volatile components absorbed on the particles surface, like polycyclic aromatic hydrocarbons (PAH) and sulfate components are suspected to cause serious health hazards. Therefore, future emission regulations are getting stricter and force the engine manufacturers to a drastic reduction of particle emission levels. Besides the limitation of the total mass emission different particle size classes may be differently weighted. More realistic test conditions, like transient test cycles have been demanded and exhibit severe deficiencies of current particle measurement techniques. Conventional measurement systems are particularly inapplicable for ultra low concentrations or the investigation of transient behavior. In this study, the performance of time-resolved laser-induced incandescence (TIRE-LII) as a favorable on-line measurement tool for soot investigation within the exhaust pipe of low emission engines is described in the form of a prototype soot sensor. Based on a theoretical model, a favorable laser-optical method to measure both the soot mass concentration and the primary particle size utilizing the thermal radiation of laser heated soot is demonstrated. Particularly, the potential of LII is demonstrated with the help of three medium duty diesel engines. Several operation points have been investigated to study the effect on the soot mass concentration and primary particle size by varying engine parameters like rail and boost pressure and injection timing. The measured mass concentration has partly been compared with the results of conventional measurement systems. Finally, some ESC-tests (European Stationary Cycle) and, furthermore, first transient tests have been performed.

View full abstract

Participate matter (PM) emissions have been simultaneously measured by laser-induced incandescence (LII) and the standard gravimetric procedure in a mini dilution tunnel connected to the exhaust of a single-cylinder DI research diesel engine. The engine used in this study incorporates features of contemporary medium- to heavy-duty diesel engines and is tuned to meet the U. S. EPA 1994 emission standards. The engine experiments have been run using the AVL 8-mode steady-state simulation of the U. S. EPA heavy-duty transient test procedure. Results of the PM concentrations measured using the two methods are compared, the primary particle sizes are determined on a mode-by-mode basis, and the use of LII for comparing the PM emissions from four different fuels is demonstrated. Results have shown that : 1. The use of three wavelength detection to determine particle surface temperature, combined with absolute sensitivity calibration, provides a sensitive, precise, and repeatable measure of the particulate concentration over a wide dynamic range. 2. The LII technique produces good correlation with the gravimetric filter method measurements on a mode-by-mode basis over a wide range of operating conditions. 3. The primary particle size can be determined from the LII signals, and that this method is precise enough to distinguish particle sizes for different operating conditions. 4. Once the particulate concentration and primary particle size are known, it is possible to determine the number density of primary particles. 5. LII has also been shown to be sensitive in differentiating the PM performance between four different fuels, predicting the same trends in brake specific PM emissions as the gravimetric filter method. The LII technique is capable of real-time particulate matter measurements over any engine transient operation, making it a valuable tool in tuning diesel engine PM emissions performance. The wide dynamic range and lower detection limit of LII make it a potentially preferred standard instrument for PM measurements. Further development of the LII technique has the potential to give information about extensive aspects of the morphology of the particulate matter. Use of LII also provides a significant time advantage over the gravimetric procedure, both in the collection and processing of data.

View full abstract

Soot cloud structure is studied using Laser Induced Incandescence. A single hole Common Rail Diesel injector is used to enable high injection pressures up to 150MPa. The spray is observed in a high pressure, high temperature cell that reproduces the conditions existing in a combustion chamber of a Diesel engine during injection. LII is used to visualize soot volume fraction and is combined with a multi-layered technique in order to quantify LII signal attenuation by soot lying between the incandescent layer and the camera. Several laser sheet and camera positions are used, including vertical and horizontal sheets with respect to the vertical axis of the spray. Results analysis shows that strong LII signal attenuation occurs in the spray because of very high soot density, preventing any quantitative analysis of soot volume fraction. However, the multi-layered procedure enables a detailed estimation of the attenuation that leads to a qualitative characterization of the soot cloud structure. A database of soot visualizations is built with varying injection parameters. It is found that the first soot particles appear at the periphery of the vapor cloud. Later during injection, a high soot density area is formed at the tip of the spray, the maximum concentration being in the center of the cloud. The very tip of the soot cloud exhibits a very abrupt decay of soot volume fraction while the lateral sides show more progressive gradients. Finally, the effect of injection pressure, ambient temperature and dilution are discussed, and their effect on global soot density during injection is compared. Pressure, temperature and dilution effects do not change the global structure of the soot cloud, but they have an influence on the timing of appearance and on the global location of the soot cloud.

View full abstract

To investigate the structure of the reaction zone and the early process of young soot formation in a diesel spray flame, CH radical and polycyclic aromatic hydrocarbons (PAH) formed in the diffusion burner flames and the diesel spray flames were visualized two-dimensionally. CH radical was used as the indicator of the reaction zone in the flame because CH is produced and destroyed in the fuel decomposition zone and exists in the thin layer near the stoichiometric contour in the flame. For the visualization of early soot formation zone in the flame, PAH were used as tracers in LIF technique. PAH are well known as the precursors of young soot. An Nd : YAG pumped dye laser was used as a light source. The wavelength of the incident light was tuned to 390.58 nm to excite CH with the Q1(8.5) transition of B^2Σ-Χ^2 Π (v'=0, v''=0) band. The sheet of light passed through the mid-plane of the flame and the LIF from the species in the flame were imaged by an ICCD camera. A narrow band optical filter with a central wavelength of 430 nm and a FWHM of 10 nm was used to detect the fluorescence from CH with the transition of A^2Σ-Χ^2 Π (v'=0, v''=0) band. PAH formed in the flame were excited by the incident lights with the wavelengths of 266, 355, and 391 nm and the LIF spectra from PAH were detected by a spectrometer. For the experiments with diffusion burner flames, methane and propane were used as the fuel. The LIF images excited at the excitation wavelength of 390.58 nm showed that very thin layers of LIF from CH with the width of less than 1 mm is located in the jet periphery. In the central region surrounded by the CH thin layer, strong LIF from PAH were observed. However, the LIF signals of both species did not coexist with each other. The spectra of LIF observed in the central region of the jet had broad band structure which had a peak around 440 nm and spread over 600 nm. This spectra structure is similar to the that of some 3-6 rings PAH such as anthracene, fluoranthene, benzo(ghi)fluoranthene, anthanthrene, and benzoperylene, which were reported in the past studies. The LIF images of free diesel spray flames in a rapid compression machine showed that the fluorescence appears in the whole cross section of leading portion of spray immediately after the ignition. The thin layer of LIF from CH was not observed. The structure of the LIF observed in diesel spray flame was broader than that of diffusion burner flames of methane or propane. This LIF intensity increased monotonously with the increase in wavelength in a range between 350 nm and 650 nm. The shape of LIF spectra from diesel flame was relatively insensitive to the excitation wavelength.

View full abstract

Local chemiluminescence emission intensities of OH^*, CH^*, and C_2^* were measured at points within an 125-cc two-stroke optical engine, cassegrain optics device was used to detect the local chemiluminescence intensity at the flame front. Figure 1 shows an example of a instantaneous flame image. Time-resolved optical multi-channel analyzer (OMA) measurements at the flame front, as shown in Fig. 2, indicate that the spectrum intensity profile of the radicals persists when a flame front is passing the measurement volume determined by cassegrain optics. The tendency was similar at different engine speeds, slowing a constant peak radical intensity ratio at the preset engine condition with the throttle 20% open. In addition, a monochrometer was used to investigate the characteristics of flame emission. Spectrum-resolved analyze was applied to the detailed spectra of each chemiluminescent radical. As indicated in Fig. 3, the CH^* spectra at three different engine speeds are the same as that of a CH_4 laminar flame, this information can be used to develop a better chemical kinetics model for SI engine combustion.

View full abstract

In order to estimate the spray characteristics using a swirl-type injector, the spatial distribution of droplet size is one of the most important factors. In this study, two-dimensional measurements of spray droplet size distribution caused by a swirl-type injector are performed using an improved ILIDS (Interferometric Laser Imaging Droplet Sizing) method, which was developed by the authors. Also, measurements of the velocity of droplets using a PIV (Particle Image Velocimetry) method are carried out. The defect of the ILIDS method is the difficulty to distinguish overlapped interference images by droplets. In order to overcome this issue, lenses compressing images are set between a collecting lens and an image plane (CCD sensor). And the circular images are concentrated together with fringes retaining the information for circle diameter and involved fringe number. The experimental setup is illustrated schematically in Fig. 1. The fuel injection system consists of a pressure pump, a feeding tank, a swirl-type injector and a driver. As a light source, a double-pulse YAG laser is employed. A laser sheet with a thickness of 0.5mm is formed at measuring field by transmitting lenses. The scattered lights are collected by collecting lenses. The instantaneous images are captured by a mega-pixel CCD camera synchronizing the light source emissions. As a result, the spatial distributions of droplet-diameter were widely measured at several regions on a vertical cross section including the center-axis. By increasing the injection pressure, the mode diameter shifted to a small value. The distribution was curve-fitted to the Nukiyama-Tanasawa's distribution function and Rosin-Rammler's distribution function by changing the parameters. Estimation of the parameters was experimentally performed.

View full abstract

Measurements of the integral length scale in a motored gasoline engine using different techniques and methods are described. An attempt is made to compare the length scales with and without contributions from cycle to cycle variations of the in-cylinder flow. A special evaluation technique developed for this purpose was applied using two-point LDV, Laser Doppler Velocimetry, and PIV, Particle Imaging Velocimetry. Length scales were obtained using both longitudinal and transversal correlation coefficients. Of particular interest was to investigate whether PIV could be used to generate similar results as the more established but rather elaborate two-point LDV technique. Also turbulence intensities obtained from the two techniques were compared. The LDV and PIV measurements were performed in the top of a motored 4-valve optical gasoline engine with pentroof design. It was shown that velocity fluctuations are substantially lower if cyclic variations of the cylinder flow are excluded and that results obtained for the turbulence intensity are similar using PIV and LDV. Using the special evaluation technique also measured length scales are very similar using PTV and LDV indicating that the less elaborate PFV technique is a useful tool for length scale measurements.

View full abstract