22

NOx and PM emissions reduction on an automotive HSDI Diesel enginewith water-in-diesel emulsion and EGR: An experimental study Alain Maiboom ⇑ , Xavier Tauzia Laboratoire de Mécanique des Fluides, UMR CNRS 6598, Internal Combustion Engine Team, Ecole Centrale de Nantes, BP 92101, 44321 Nantes Cedex 3, France a r t i c l e i n f o Article history: Received 7 June 2010 Received in revised form 9 June 2011 Accepted 10 June 2011 Available online 29 June 2011 Keywords: Automotive Diesel engine Water-in-diesel emulsion Exhaust gas recirculation Combustion Heat release a b s t r a c t Automotive Diesel engines exhaust emissions must constantly be reduced to comply with more and more stringent regulations, all over the world. The introduction of water in the combustion chamber is already used on some large marine diesel engines to cut down NOx emission. In this paper an experimental study is conducted on a modern automotive 1.5 l HSDI Diesel engine while injecting a water-in-diesel emulsion (WDE) with a volumetric water-to-fuel ratio of 25.6%. Four injection strategies are considered with and without pilot injection, with two levels of injection pressure. First, the injection of WDE is compared to diesel-fuel in terms of combustion and NOx and PM emissions without using exhaust gas recirculation (EGR). Depending on the WDE fuelling rate and injection strategy (with or without a pilot injection before main injection), NOx emissions are most often reduced (of up to 50%), and PM emission are most often decreased as well (the maximum relative reduction being 94%). The combustion is largely affected by the injection of WDE as compared with pure diesel-fuel, the main observations being an increased of the ignition delay and an improved mixing-process between the fuel and the surrounding gases. After that, the use of WDE in parallel with EGR (with various EGR rates) is tested with the aim at improving the NOx–PM trade-off (reduction of NOx emission at a given PM emission level or reduction of PM emission at a given NOx emission level). The results show that this method is an effective way for NOx and PM emission reduction on an automotive Diesel engine. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction In light of the current requirements as regards the reduction of pollutant emissions of automotive Diesel engines such like EURO 6 in Europe, manufacturers have to develop new in-cylinder strate- gies and/or aftertreatment devices [1]. With the upcoming pollu- tant regulations, NOx emission will become particularly critical on automotive Diesel engines. As regards the in-cylinder strategies aiming at reducing NOx emission, exhaust gas recirculation (EGR) into the engine intake is the most used and studied technology. The decrease of NOx emission with EGR is the result of complex and sometimes oppo- site phenomena occurring during combustion [2–15]. The main effect is the decrease of local temperatures in the combustion chamber, in particular those corresponding to zones where NO is produced (on the lean side of the diffusion flame during fuel injec- tion [16] and in the combustion products after the end of injec- tion). The main drawback of EGR is the increase of PM emission in the classical high temperature diesel combustion (HTC) and the need to increase boost pressure at middle and high loads when using EGR to maintain the air–fuel ratio (AFR) at a suitable level [5,6]. Another in-cylinder strategy to reduce local temperatures and consequently the NO production rate is the injection of water (WI), either into the engine inlet [17–25], directly in the combus- tion chamber [26–33], or in emulsion with the fuel [18,19,21–23, 34–56]. One advantage of WI as compared with EGR is the possible reduction of NOx emission either at low loads and high loads with- out a substantial increase in PM emission. The probably easiest way to inject water in the engine is inlet WI [17–25]. This technique has been used on some large marine Diesel engines [18] and various strategies to inject water in the inlet air are presented in the literature. The main drawback of inlet WI is that a water mass of about 60–65% of the fuel is needed to achieve a 50% NOx reduction, and is very high (up to four times the amount of fuel mass) if trying to drastically reduce NOx emission [25]. Different strategies have been also proposed to inject water directly into the combustion chamber, with the aim at reducing NOx emission while limiting the water quantity as compared with inlet WI. One advantage of direct WI as compared with water-in- diesel emulsion is the possibility to change the water-to-fuel ratio, while varying engine parameters (speed and load) or during engine warm-up (cold start) [28–30]. 0016-2361/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.fuel.2011.06.014 ⇑ Corresponding author. Tel.: +33 2 40 37 68 80; fax: +33 2 40 37 25 56. E-mail address: [email protected] (A. Maiboom). Fuel 90 (2011) 3179–3192 Contents lists available at ScienceDirect Fuel j our nal homepage: www. el sevi er . com/ l ocat e/ f uel The last method to inject water is the use of water-in-diesel emulsion or microemulsion [18,19,21–23,34–57]. Most engine experiments and numerical studies using WDE technique showthat the NOx reduction is accompanied with a large reduction of PM and soot emissions. Main effects of WDE on PMemission are as follows: At a given fuel injection rate, the use of WDE leads to an increase of the total injected mass, of which a consequence is an increase of the mixing rate between fuel and air, thus reduc- ing local fuel–air ratios and consequently PM production [51]. The vaporisation of water and the dilution effect of water lead to a decrease of the temperatures within the core spray where soot is produced, thus reducing the soot production rate [35,51]. The increased presence of water within the combustion jet may affect the chemical kinetic mechanisms of soot formation (in the core spray) and soot oxidation (at the jet periphery) [35,51,56]. The presence of water in the emulsion has a tendency to increase the ignition delay. Thus, the mass fraction of fuel that burns under a premixed combustion is increased, which a direct consequence is the decrease of soot production rate [36,38,39,52,53]. It has been shown that the location of flame lift-off on diesel fuel jet plays an important role in the soot formation process, by allowing fuel and air to mix upstream of the lift-off length (i.e. prior any combustion) [58–61]. Just downstream of the lift-off length, the partially premixed air–fuel mixture under- goes a premixed combustion that generates a significant local heat release and fuel-rich product gas that becomes the ‘fuel’ for the diffusion flame at the jet periphery. The soot formation was shown to be directly dependant of the equivalence fuel– air ratio at the lift-off length [58–61]. Local studies in an optically-accessible engine show that that flame lift-off is sig- nificantly increased with WDE, leading to leaner mixtures within the jet during the diffusion stage of combustion [39]. The water droplets contained in the emulsion have a lower boil- ing point than diesel-fuel. According to some researchers, the sudden and dramatic expansion of vaporising water (called micro-explosion) would enhance the mixing process between air and fuel [21,43–45,51,53] and lead to the combustion of smaller diesel droplets. Polycyclic aromatic hydrocarbons (PAH) have been shown to play an essential role as precursors for soot particles during their formation and growth [51,61]. The combustion of WDE leads to lower PAH concentration within the core spray, thus reducing the soot production rate [51]. Finally, the vaporisation of water and the dilution effect of water lead to a decrease of the temperatures at the jet periphery where soot is partially oxidised [61,67,68] that may reduce the soot oxidation rate. As regards unburned hydrocarbons (UHC) and carbon monoxide (CO) emissions, results are contradictory with some tests showing reduction [21,47] and other tests showing increase [22,34,35]. WDE is used on some large marine Diesel engines to cut down NOx emission [62–64] and some on-road tests have been carried out on captive fleets of trucks and buses [65]. As mentioned above, the most efficient WI technologies to re- duce NOx emission are WDE or direct WI, because the water is in- jected directly into the combustion zone, allowing a large decrease of combustion temperatures [18,23,51]. As a consequence, for a gi- ven quantity of injected water, the NOx reduction with direct water injection or WDE is around twice as high as with inlet water Nomenclature AFR air–fuel ratio (–) AFR st stoichiometric air–fuel ratio (–) AMF air mass flow (kg/h) FMF fuel mass flow (kg/h) M NO 2 molar mass of NO 2 (g/mol) NOx (ppm) nitrogen oxides concentration (ppm) NOx (g/h) nitrogen oxides emission (g/h) P pressure (bar) PM particulate matter (g/h) Q exhaust exhaust gas flow (non-condensed) (m 3 /h) Q exhaust_dry exhaust gas flow (dry) (m 3 /h) Smoke (mg/m 3 ) smoke emissions in mg/m 3 Smoke (g/h) smoke emissions in g/h T temperature (°C) V m molar volume (l/mol) X EGR EGR ratio (%) X CO2 CO 2 concentration (percentage in volume, dry) (%) k excess air/fuel ratio q air air density (kg/m 3 ) q burned-gas burned gas density (kg/m 3 ) Subscripts 1 ambient conditions 2 post-compressor 2 0 post-intercooler 2 00 post-EGR air air m relative to main injection p relative to pilot injection rail fuel rail Abbreviations ATDC after top dead centre BMEP brake mean effective pressure BSFC brake specific fuel consumption CA crank angle (deg) DOC Diesel Oxidising Catalysts EGR exhaust gas recirculation FSN filter smoke number HP high pressure HSDI high speed direct injection HTC high temperature combustion ID ignition delay IGR internal gas residual IVC intake valve closing NEDC new European driving cycle PM particulate matter ROHR rate of heat release (W) SOC start of combustion (CA deg) SOI start of injection (CA deg) TDC top dead centre UHC unburned hydrocarbons VGT Variable Geometry Turbine WDE water-in-diesel emulsion WI water injection 3180 A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 injection [18,23]. As regards PM emission, the use of WDE seems to show better results than other WI techniques. Thus, the aim of this paper is to experimentally study the poten- tial of WDE in parallel with EGR as an in-cylinder strategy for both NOx and PM emissions reduction on a modern automotive com- mon-rail DI Diesel engine for future emissions standards. 2. Material and methods 2.1. Description of the engine The engine used for the experiment is a 1.5 l water-cooled HSDI 4-cylinders diesel engine, with two valves per cylinder and which conforms to Euro III standards. It is equipped with a common-rail solenoid injection system with a maximum injection pressure of 1600 bar, re-entrant bowl-in-piston combustion chambers, a Vari- able Geometry Turbine (VGT) turbocharger, an intercooler, and a Diesel Oxidising Catalyst (DOC) (see Fig. 1). Engine specifications are given in Table 1. The engine is originally equipped with a high pressure (HP) un- cooled EGR loop. An EGR cooler has been added and an indepen- dent water circuit on the EGR-cooler was used in order to control the temperature of the recirculated gases, and thus the tempera- ture T 2 00 of inlet gases after mixing with EGR. The mean EGR rate is defined as follows: X EGR ð%Þ ¼ 100 Á X CO2 inlet X CO2 exhaust ð1Þ where X CO2_inlet and X CO2_exhaust are measured CO 2 concentrations in inlet and exhaust manifolds respectively. The original air/air intercooler was turned into a water/air inter- cooler allowing the air temperature T 2 0 and/or T 2 00 to be controlled separately. 2.2. Water-in-diesel emulsion Many experiments have been done to find a stable WDE (over some days or weeks). Various concentrations of different types of emulsifiers were tested in the WDE. The Hydrophilic Lipophilic Balance (HLB) methodology was used to help finding a stable WDE. During the tests, the more stable emulsion was obtained by using two emulsifiers, Span 80 (sorbitan monooleate) and Tween 85 (polyoxyethylenesorbitan trioleate), the volumetric ra- tios being respectively 1.3% and 0.7% and the HLB number being equal to 6.5. Also, various water quantities in the WDE have been tested on the test bench (with a volumetric water-to-fuel ratio of up to 43%), in order to reduce NOx and PM emissions as much as possible while limiting the negative impacts of water (high cycle-to-cycle dispersions and misfired cycles at very low load conditions, large increase of CO and UHC emissions at low load conditions or during engine warm-up, misfire problems at engine start-up in particular in cold conditions). A volumetric water- to-fuel ratio of 25.6% was found to be a good compromise (the volumetric fraction of water in the total WDE being thus equal to 20%). The volumetric composition of the WDE tested in this study is as follows: Diesel-fuel: 78%. Water: 20%. Span 80: 1.3%. Tween 85: 0.7%. A photography of the WDE has been obtained with a micro- scope (Fig. 2), showing that most water bubbles have a diameter lower than 40–50 lm, but with some slightly larger water bubbles. It must be underlined that the emulsion was not optimised in terms of stability or bubbles diameter dispersions (such like WDE commercialized by some manufacturers), but was found to be good enough to undertake experimental studies on the test bench. 2.3. Evaluation of the mean gross rate of heat release (ROHR) The gross ROHR is obtained for each operating condition thanks to a calculation procedure developed at the laboratory. The calcu- lation is classically based on the in-cylinder pressure. It is mea- sured with a Kistler 6055BB piezo electric pressure transducer and an encoder with a resolution of 0.36°CA. The cylinder pressure used is the mean value over 100 consecutive cycles, which was found to be enough for the mean value to be reliable. It must be underlined that the mean ROHR gives no indication on eventual cy- cle to cycle dispersions. The gross ROHR was extracted fromthe net ROHR by calculating the heat transfer to the combustion chamber walls with Hohenberg’s model [66]. Since valve overlap is negligible on this engine, trapped mass of fresh air can be estimated directly from air mass flow measured at Fig. 1. Engine configuration. Table 1 Engine specifications. Compression ratio 17:1 Number of cylinders 4 Number of valves per cylinder 2 Combustion chamber Re-entrant bowl-in-piston Injection system Common-rail solenoid Maximum injection pressure 1600 bar Number of injection holes 6 Fig. 2. Photography of the emulsion obtained with a microscope. A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 3181 the engine inlet. The mass of residual gas is calculated with perfect gas law applied at EVC, from volume, measured in-cylinder pres- sure and in-cylinder temperature estimated from exhaust gas tem- perature. The residual gas composition is deduced from exhaust gas analysis. When EGR is used, EGR mass flow is calculated from CO 2 concentrations measurements at the engine inlet and engine exhaust (that gives the ratio between EGR mass flow and total air + EGR mass flow, the air mass flow being measured at the en- gine inlet). EGR composition is deduced from exhaust gas analysis. 2.4. Emissions measurement NOx emission is measured with an ECO PHYSICS CLD 700EL gas analyser, which uses the chemical luminescence detector (CLD) method. NOx emissions are converted from ppm to g/h: NOx ðg=hÞ ¼ M NO2 Á NOx ðppmÞ Á Q exhaust dry 10 3 Á V m ð2Þ where M NO2 = 46.005 g/mol and V m = 22.41 l/mol at standard tem- perature and pressure. PM emission at the exhaust is measured with an AVL 415S smoke-meter. The relation between filter smoke number (FSN) and smoke emissions in mg/m 3 is given in the AVL 415S operating manual and is as follows: Smoke ðmg=m 3 Þ ¼ 1 0:405 Á 5:32 Á FSNÁ expð0:3062 Á FSNÞ ð3Þ Smoke emissions in g/h are thus given by: Smoke ðg=hÞ ¼ 10 À3 Á Smoke ðmg=m 3 Þ Á Q exhaust ð4Þ As the value of air excess is more than 1 for all tested operating conditions, exhaust gases are composed of stoichiometric burned gases and non-consumed air. Thus, the non-condensed exhaust gas flow Q exhaust is given by: Q exhaust ¼ FMF Á 1 þAFR st q burned gas þ ðk À1Þ Á AFR st q air " # ð5Þ where AFR st = 14.4 is the stoichiometric air–fuel ratio, q burned_gas = 1.33 kg/m 3 is the burned gas density of a stoichiometric mixture, q air = 1.293 kg/m 3 is the air density, and k is the air excess. Similarly, the dry exhaust gas flow Q exhaust_dry is given by: Q exhaust dry ¼ FMF Á ð1 þAFR st Þ Á a q burned gas þ ðk À1Þ Á AFR st q air " # ð6Þ where a is the mass of dry exhaust gases in 1 kg non-condensed ex- haust gas (equal to 0.924 kg for diesel fuel). The air excess is calculated with measured air mass flow AMF and fuel mass flow FMF: k ¼ AMF AFR st Á FMF ð7Þ Inlet and exhaust CO 2 concentration are measured with a CAP- ELEC CAP3200 gas analyser and a SIEMENS ULTRAMAT 23 gas ana- lyser respectively, which use the non-dispersive infrared (NDIR) measurement technique. Each gas analyser is calibrated every 4 h of experiments with specific gas standards. If the necessary shift is under 0.3%, then the experiments done since the previous cali- bration are validated. 2.5. Evaluation of fuel proportion injected during ignition delay In order to interpret NOx and soot emissions it can be helpful to know the proportion of fuel that burns during the premixed phase and that which burns during the diffusion phase. In fact, it is rather impossible to access directly this proportion. Thus, the proportion r of fuel injected during ignition delay (ID) was calculated for each operating condition tested. The injection rates were obtained thanks to a simplified model of the injector (calibrated with mea- surements of instantaneous injection rates with diesel fuel made on a standard injection test bench [69]) for each operating point, thus providing the instantaneous proportion of injected fuel at each time step. The ignition delay is supposed to end when com- bustion acceleration (differential of ROHR, in W/s) reaches a critical value, which was arbitrarily fixed at 2 Â 10 8 W/s. It should be no- ticed that the air–fuel mixture formed during ID is generally rich [68], such that all the diesel-fuel injected during ID will not burn under a premixed combustion mode. The ratio r can be thus con- sidered as an upper bound of the proportion of fuel that burns un- der a premixed combustion mode. 2.6. Error analysis Table 2 sums up the measurement technique, calibrated range, accuracy and relative error of various instruments involved in the experiment for various parameters. Errors in experiments can arise from instrument conditions, calibration, environment, observation, reading and test planning. The accuracy of the experiments has to be validated with an error analysis. That was performed here using the differential method of propagating errors based on Taylor’ the- orem. It gives the maximum error u of a function f(x 1 , x 2 , . . . , x n ) as follows: uðf ðx 1 ; x 2 ; . . . ; x n ÞÞ ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi X ðc i Á uðx i ÞÞ 2 q ð8Þ As a result, the maximum relative errors for X EGR , NOx (g/h), PM (g/h) are 1.4%, 1.5%, and 2.3% respectively. 2.7. Operating points The study is conducted for five various operating conditions (load and speed). The engine speed, torque, pilot and main die- sel-fuel injections quantities, pilot and main start of injections (SOI) as well as approximate brake mean effective pressure (BMEP) are given in Table 3. Operating points 1–3 are operating points such as those encountered in the European emissions test cycle (NEDC) for a classical vehicle (equipped with a moderated down- sized Diesel engine). The whole cycle is composed of four urban driving cycles and one extra urban driving cycle. Operating points 4 and 5 are operating points with higher loads, thus representative of operating points that could be encountered in the extra urban driving cycle with a much downsized Diesel en- gine. Such operating points are particularly critical in terms of NOx and PM emissions in this test cycle. Actually, on the one hand, the use of EGR is limited to a low level with the HP EGR loop at these loads because an increase of EGR results in both a decrease of O 2 concentration in the inlet and a decrease of boost pressure, thus reducing the in-cylinder O 2 content. At part load operation, this would irremediably lead to a low value of the air excess, and thus to large amounts of PM, CO, and UHC emissions, as well as a large increase of brake specific fuel consummation (BSFC). On the other hand, there is a clear tendency to further downsize automotive Diesel engines, and thus there is a need to reduce pollutant emis- sions at higher loads. It is why operating points 4 and 5 have been studied in this paper. Quantities of injected diesel fuel are held constant for each operating point. Thus, BMEP is little modified with various tested modifications (use of WDE, EGR rate, inlet temperature). Initial val- ues of boost pressure for each operating point were held constant in all tests presented hereafter; when using HP EGR, the decrease of boost pressure while opening the EGR valve is compensated 3182 A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 by closing VGT vanes. Pilot and main SOI are maintained constant as well. 3. Influence of water-in-diesel emulsion without EGR A first study has been carried out to investigate some of the ef- fects of WDE on combustion, NOx and PM emissions (without EGR). Actually, as described in the introduction, most of experi- mental studies dealing with the use of WDE presented in the liter- ature have been done on large Diesel engines. The aim of this first section is thus to present the effects of WDE on a recent automo- tive HSDI Diesel engine, equipped with a common-rail high pres- sure injection system. Two injection strategies have been tested: The first strategy consists in injecting WDE with the same max- imum instantaneous mass injection rate through the injector as with diesel fuel. This is achieved with a slight reduction of rail pressure (because of the higher density of WDE as compared with diesel fuel, which a consequence is an increase of mass flow rate through the injector if maintaining the same rail pres- sure). With this first strategy (‘‘WDE – Prail 1’’ on the following figures), to ensure that same fuel mass is injected per cylinder and per cycle, the injection duration has to be increased. The instantaneous injection rate of the fuel part in the WDE is thus lower than with diesel fuel being injected alone. Second strategy consists in maintaining approximately the same diesel fuel injection rate with WDE as compared with die- sel fuel alone. The WDE must thus be injected faster into the combustion chamber to maintain the same instantaneous fuel introduction. This is achieved by increasing rail pressure (‘‘WDE – Prail 2’’ on the following figures). The injection dura- tion is kept approximately constant as compared with diesel fuel alone. The calculation of the injection pressures Prail1 and Prail2 was obtained thanks to the simplified model of the injector. Rail pres- sures Prail1 and Prail2 for each operating point are given in Table 4. It can be noticed that the injection pressure Prail2 is much high- er than Prail1, especially for operating points 2–4. Moreover, the influence of WDE on combustion and pollutant emissions has been studied with both a single fuel injection, or with a double-shot injection (one pilot injection before main injec- tion). One or two pilot injections before main injection are classi- cally used to decrease the ID of main injection such that the premixed part of main combustion is lower, thus limiting the high in-cylinder pressure derivative and the corresponding combustion noise. When using only one injection (without pilot injection), the fuel mass of pilot injection is added to the fuel mass injected dur- ing main injection, such that the total mass of fuel injected per cy- cle is constant for each injection strategy (with or without pilot injection, diesel fuel alone, WDE with injection pressure Prail1, WDE with injection pressure Prail2). Main results in terms of combustion and NOx–PM emissions with these various injection strategies are presented hereafter. 3.1. Influence on combustion 3.1.1. Without pilot injection Before studying a classical pilot + main injection, it was found interesting to study the influence of WDE on a single injection. Mean gross ROHR for operating points 4 and 5 are presented in Fig. 3. The standard analysis of the rate of heat release [2,68] shows several effects. Without pilot injection, the ROHR diagram is com- posed of two phases: a ‘‘premixed peak’’ due to the fast combus- tion of some of the fuel injected during ID, followed by a diffusion part during which combustion is controlled by the mix- ing-process between diesel-fuel and the surrounding gases (see Refs. [2,68] for more details about the ROHR analysis). Table 5 gives the ID and the proportion of diesel-fuel injected during ID. The very first effect of WDE is the increase of ID (Table 5 and Fig. 3). This is in agreement with previous researches [21,35, 36,52]. The ID increase is probably due to the evaporation of water contained in the WDE that causes local temperatures in the fuel jet to decrease. The ID increase is observed for both injection strategies, but the second strategy (that maintains the diesel-fuel introduction rate into the combustion chamber owing to a higher injection pressure) results in a lower ID increase. For instance, for operating point 4, the ID increase with first WDE injection strategy is about 27% whereas it is about 4% for the second WDE injection strategy. Same tendencies are obtained for the others operating points. The lower ID increase with the second WDE injection strategy may be due to a higher WDE injection rate and better atomisation as compared with the first strategy, such that some fuel is mixed with air faster and can ignite earlier. Given experimental results, it seems that the first effect is pre- dominant, such that the second WDE injection strategy results in a Table 2 Relative measurement error. Instrument Calibrated range Accuracy Relative error (%) Inlet gas temperature (k-type thermocouple) 0–1000 °C ±1 °C ±0.75 Inlet gas pressure (2 bar piezoresistive relative pressure sensor HCS Sensor Technics) 0–2 bar ±5 mbar ±0.25 Air mass flow (hot wire air flow meter) 0–800 mg/str ±4 mg/str 1 Fuel consumption (ROTRONICS DMC202) 0–40 kg/h ±40 g/h ±0.1 NOx (ECO PHYSICS CLD 700EL) 0–1000 ppm ±5 ppm 1 Smoke (AVL 415S) 0–10 FSN ±0.1 FSN 2 Inlet CO 2 (CAPELEC CAP 3200) 0–20% ±0.15% 1.5 Exhaust CO 2 (SIEMENS ULTRAMAT 23) 0–20% ±0.1% 1 In-cylinder pressure (Kistler 6055BB) 0–200 bar ±0.5 bar 1 Table 3 Operating points. Point 1 2 3 4 5 Engine speed (rpm) 1480 2035 1480 2065 1460 Torque (Nm) 21.0 65.2 78.7 116.5 163.9 BMEP (bar) 1.32 4.11 4.95 7.33 10.31 Pilot quantity (mg diesel/stroke) 1.3 1.7 1.6 1.9 2.0 Principal quantity (mg diesel/ stroke) 5.3 12.1 14.8 21.1 28.8 Boost pressure P 2 00 (mbar) 1040 1250 1230 1690 1630 Pilot SOI (°CA ATDC) À27.0 À31.5 À32.6 À36.5 À40.0 Main SOI (°CA ATDC) 3.2 3.1 3.5 1.7 À0.3 Table 4 Rail pressures for various injection strategies. Point 1 2 3 4 5 Diesel 405 655 520 880 900 WDE – Prail 1 392 633 502 850 869 WDE – Prail 2 490 1032 813 1400 1430 A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 3183 large decrease of ID as compared with the first WDE injection strategy. Another effect of WDE on combustion is the modification of the premixed part of combustion. For operating points 4 and 5, the pre- mixed part of combustion (that results from the introduction of diesel-fuel during ID) is largely lower than the diffusion part of combustion. Actually, the proportion r of diesel-fuel injected dur- ing ID does not exceed 20% (see Table 5). For these two operating points, the premixed part of combus- tion results in a ‘‘traditional’’ premixed peak on the ROHR diagram. Main factors that influence the magnitude of the premixed peak are as follows: The ID directly influences the quantity of fuel that is injected during ID, and thus that can mix with the surrounding air to burn in a premixed combustion mode [2,68]. In case of dilution of the diesel-fuel (by water in this study), for a given quantity of injected mass during ID, the dilution of the fuel decreases the quantity of diesel-fuel injected during ID. The intensity of the mixing rate between the injected fuel and the surrounding air during ID in the combustion chamber also influences the premixed peak, a higher mixing rate resulting in a higher quantity of injected fuel that is mixed with air and thus that burns in a premixed combustion mode [2,68]. With the second WDE injection strategy, the increased injection pres- sure increases the mixing rate as well. Moreover, a premixed combustion is kinetically-controlled, such that higher temperatures in the air–fuel mixture formed during ID will result in a higher premixed peak [2]. Finally, the gas composition (in case of EGR and/or IGR) clearly influences the premixed peak as well. Actually, for a given mix- ing rate between injected fuel and surrounding gases during ID, the dilution of fresh air by burned gases – owing to the external EGR loop or because of residual gases at intake valve closing (IVC) – decreases the mixing rate between fuel and fresh air. In our case, for both WDE injection strategies, the first four items are affected. Only the dilution of the fresh air is unchanged since EGR valve is closed and there is no reason for IGR to change. With the first WDE injection strategy, the premixed peak on ROHR diagram is increased as compared with diesel-fuel. The in- creased ID results in a slight increase of the ratio r, from 15% to 19% for operating point 4 and from 14% to 19% for operating point 5. Although the diesel-fuel introduction rate is reduced with first WDE injection strategy as compared with diesel-fuel, there is slightly higher diesel-fuel injected during ID because of a longer ID. The local temperatures in the premixed zone are certainly re- duced because of the water evaporation process, but this thermal effect seems to have less influence than the increase of the propor- tion r such that the premixed peak on the ROHR is increased. With the second WDE injection strategy, the ID increase is low- er as compared with the first WDE strategy, but the proportion r of diesel-fuel injected during ID is approximately constant as com- pared with first WDE injection strategy (r is equal to 20% against 19% for both operating points 4 and 5). For these operating points, the ID decrease with Prail2 as compared with Prail1 compensates for the higher diesel-fuel injection rate with Prail2. Although the fuel proportion r is approximately the same with both WDE injec- tion strategies, the second strategy results in a largely higher pre- mixed peak on the ROHR diagram. This can be explained by a higher mixing process between the WDE jet and the surrounding gases. Actually, since the mass introduction rate of water + die- sel-fuel is increased as compared with diesel-fuel alone, the air entrainment by the jet is increased, resulting in an increased air entrainment per unit of diesel-fuel in the WDE jet, in particular during ID, resulting in a less rich air–fuel mixture formed during ID. Again, as mentioned previously, the improved mixing-process between surrounding gases and diesel-fuel is also a consequence of the increased injection pressure with second WDE injection strategy; this effect would also occur if using only diesel-fuel with increased injection pressure. As regards the diffusive part of combustion, the use of WDE also produces noticeable effects. For first WDE injection strategy, the ROHR diagram shows a slower diffusive combustion speed as com- pared with diesel-fuel. This is due to a lower introduction rate of diesel-fuel with first WDE injection strategy. For the second WDE injection strategy, while the diesel-fuel introduction rate is approx- imately identical as compared with diesel-fuel being injected alone, the combustion speed during diffusion combustion is largely in- creased, showing again that there is an enhanced mixing of sur- rounding gases per unit of injected diesel-fuel. One consequence is a positive impact on PM emission as will be shown later. 3.1.2. With pilot injection The influence of WDE on the combustion with a ‘‘classical’’ dou- ble-shot pilot + main injection is given in Fig. 4 for operating points 2–5. Also given in Fig. 4 are the SOI for pilot and main injections. Fig. 3. ROHR comparison of diesel-fuel and WDE, without pilot injection. Table 5 Ignition delay and proportion of fuel injected during ignition delay (r). Operating condition 4 4 4 5 5 5 Diesel WDE WDE Diesel WDE WDE Prail1 Prail2 Prail1 Prail2 ID (ms) 0.29 0.37 0.30 0.33 0.42 0.34 ID increase – +27% +4% – +29% +5% r (%) 15 19 20 14 19 20 3184 A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 It can be noticed that the fuel injected during pilot injection has a two-stage autoignition, with a ‘‘cool-flame’’ low-temperature heat-release (LTHR) phase followed by a main high-temperature heat-release (HTHR) phase (see annotations in Fig. 4a). Such a com- bustion mode is classically observed in compression ignition en- gines with fuel being injected sufficiently early in the cycle when the in-cylinder air temperature is too low for the fuel to ignite di- rectly, thus letting fuel enough time to mix with air. The air–fuel mixture ignites later in the compression stroke (ignition by com- pression) when in-cylinder temperature is higher and the corre- sponding combustion is kinetically controlled. Such combustion regimes are also involved in homogeneous charge compression ignition (HCCI) engines where all the fuel is injected very early in the cycle (at the beginning of the compression stroke or in the in- take manifold). For operating points 1–3, the use of WDE has a tendency to in- crease the ID of pilot combustion, for both WDE injection strate- gies, but with no remarkable difference on the ROHR diagram between the two WDE injection strategies. At higher loads (operat- ing points 4 and 5), the second WDE injection strategy leads to a very low ROHR for pilot combustion, showing that a large amount of fuel injected during pilot injection is not burning before main injection. This is probably due to the fact that with the second injection strategy the increased injection pressure leads to a faster fuel jet penetration in the combustion chamber such that the fuel can reach the combustion chamber walls before ignition. In that case, the air–fuel mixture is cooled by the cold walls and will not burn before main injection. Another explanation for the lack of combustion for pilot injection with second WDE injection strategy could be the formation of mixtures that are too fuel-lean to ignite because of the fast mixing at high injection pressure. As regards main combustion, when comparing Figs. 3 and 4, there is a little decrease of ID for main injection when using a pilot injection, as traditionally observed. One consequence is a decrease of the premixed peak on the ROHR diagram. The use of WDE has same consequences on main combustion than in the case of a sin- gle injection (without pilot injection), described in details in previ- ous paragraph. Only for operating point 4, the fact that the fuel injected during pilot injection does not burn in the second WDE injection strategy leads to a higher ID of main combustion of sec- ond WDE injection strategy as compared with first WDE injection strategy, which is opposite to the tendencies observed with a sin- gle injection. 3.2. Influence on NOx and PM emissions 3.2.1. Without pilot injection The influence of WDE on NOx and PM emissions with a single injection is given in Fig. 5 for operating points 4 and 5. For operating points 4 and 5, there is a noticeable decrease of NOx emission with first WDE injection strategy as compared with diesel-fuel (32% and 30% for operating points 4 and 5 respectively). This can be explained as follows: The evaporation of water causes local temperatures in the spray to decrease, including those of zones where NO is produced (on the lean side of the diffusion flame during injection and in the combustion products after the end of injection), resulting in a decrease of NO production rate. Moreover, for a given mass of burned fuel, the heat is released in a higher mass of gas (because of the dilution of the fuel by water). The dilution by water has thus a local thermal effect. Fig. 4. ROHR comparison of diesel-fuel and WDE, with pilot injection. A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 3185 Another effect is the combustion off-phasing by WDE (described in Section 3.1.1.) that shifts combustion further into the expansion stroke and thus results in a global decrease of in- cylinder temperatures including local temperatures in the fuel jet. Finally, as described by some researchers, the presence of water can form OH radicals which are involved in the formation of thermal NO [2] and could thus have a chemical effect on NO formation. With second WDE injection strategy, there is a large increase of NOx emission as compared with first WDE injection strategy. This should be mainly due to an increase of instantaneous ROHR that may result in higher local temperatures. The final result of second WDE injection strategy as compared with diesel-fuel on NOx emis- sion depends on operating point: NOx emission increases for oper- ating point 4 and decreases for operating point 5 (Fig. 5a). As regards PM emission, different tendencies are also observed between the two WDE injection strategies. For first WDE injection strategy, there is a very slight decrease of PM emission for oper- ating point 4 and a large relative increase for operating point 5 (Fig. 5b). An increase of PM emission is contradictory with most experimental studies described in the literature. It shows that there are opposite phenomena on PM formation and oxidation processes occurring in the combustion chamber when injecting WDE. For operating point 5, the decrease of local temperature at the jet periphery that decreases the soot oxidation rate seems to have more impact on PM emission than all others effects that decrease PM emission (given in Section 1). With second WDE injection strategy, there is a large decrease of PM emission as compared with diesel-fuel (À90% and À28% for operating points 4 and 5 respectively). This is most likely due to the increased injection pressure that promotes the atomization of the fuel and the mixing with surrounding air, and to the increase of the pre- mixed part of combustion. It must be noticed that the increased injection pressure would almost certainly reduce PM emissions with diesel-fuel. It must be underlined that PM emission is very low in each case, because of the high air–fuel ratio since no EGR is used. It will be seen in Section 4.2 that when EGR is used, a decrease of PM emis- sion with WDE was always observed. 3.2.2. With pilot injection The influence of WDE on NOx and PM emissions with a ‘‘classi- cal’’ double-shot pilot + main injection is given in Fig. 6 for operat- ing points 1–5. Fig. 5. Comparison of WDE and diesel-fuel on NOx and PM emissions, without pilot injection. Fig. 6. Comparison of WDE and diesel-fuel on NOx and PM emissions, with pilot injection. Fig. 7. Influence of EGR on ROHR with WDE. 3186 A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 For operating points 1–4, same tendencies are observed than without pilot injection, from a qualitative point of view: The first WDE injection strategy results in a large decrease of NOx emission, whereas second WDE injection strategy results in either a little decrease or increase of NOx emission as com- pared with diesel-fuel. With first WDE injection strategy there is a decrease of PM emission (that is particularly drastic for operating points 1–3) except for operating point 5, as it is the case without pilot injec- tion. The second strategy results in a PM emission reduction for all operating points, which is drastic for operating points 1–4 (of up to 94%). Again, there are opposite phenomena occurring in the combustion chamber as regards PM formation and oxida- tion processes that can result in opposite tendencies on the final PM level at the exhaust; in most cases there is a decrease of PM emission. The combustion analyse has shown than with both WDE injection strategies, there is an increase of the air entrain- ment per unit of diesel-fuel mass; this effect results in a decrease of PM emission. Fig. 8. Influence of EGR on in-cylinder pressure with WDE. Fig. 9. Comparison of WDE and diesel-fuel on NOx emission for various EGR rates. A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 3187 4. Influence of water-in-diesel emulsion with EGR As described in previous section, the use of WDE results in most cases in a large decrease of PM emission and a noticeable decrease of NOx emission. The aim of this section is to study the use of ex- haust gas recirculation in parallel with WDE to further decrease NOx emission while maintaining PM emission at a low level. It has been shown than second WDE injection strategy results in a large increase of instantaneous ROHR for main combustion and can have negative impact on pilot combustion, as well as on NOx emission. As a matter of fact, the first WDE injection strategy has been used here, with a double injection pilot + main. 4.1. Influence on combustion The influence of EGR in parallel with WDE on combustion and in-cylinder pressure is given in Figs. 7 and 8 respectively, for oper- ating point 2. Same tendencies have been observed for other oper- ating points. It is shown that the use of EGR has a slight impact on ROHR for both pilot and main combustion. First, for pilot combustion, there is a decrease of ID while increasing EGR rate (Fig. 7a). This is con- tradictory with most experimental studies on the influence of EGR on combustion. In fact, the increase of EGR rate has two opposite consequences on ID: The dilution by EGR results in an increase of ID [4–6,8]. The increase of EGR rate results in an increased inlet tempera- ture T 2 00 , which a consequence is an increase of the temperature at pilot SOI. Since pilot combustion ID is mainly kinetically con- trolled, this tends to decrease ID. The EGR cooler used here is not effective enough to maintain inlet temperature T 2 00 at a low level while increasing EGR rate, such that the increase of inlet temperature has a predominant effect on ID in theses tests. Fig. 10. Comparison of WDE and diesel-fuel on PM emission for various EGR rates. 3188 A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 As regards main combustion, there is a slight increase of ID that results in a slight increase of the premixed part of combustion. This may be accompanied with an increase of ROHR peak (as for in- stance for operating point 2 while increasing EGR rate of up to 8.2%: see Fig. 7a). For main combustion, the dilution effect of EGR has a predominant influence on ID over the temperature in- crease at main SOI. The slight off-phasing effect of EGR on main combustion results in a decrease of in-cylinder pressure during main combustion. It has thus a negative impact on cycle efficiency and on specific diesel-fuel consumption. A readjustment of main injection by advancing main SOI is thus needed if trying to main- tain approximately the same in-cylinder pressure evolution. From a qualitative point of view, main tendencies observed while vary- ing EGR rate are not impacted by the use of WDE instead of die- sel-fuel alone. 4.2. Influence on NOx and PM emissions 4.2.1. NOx emission NOx emission while increasing EGR rate with diesel-fuel or WDE is given in Fig. 9. Whether with diesel-fuel or WDE, there is a decrease of NOx emission while increasing EGR rate. The decrease of NOx emission with EGR has been largely described and documented in the liter- ature. The main effect of EGR is the decrease of combustion temperatures, in particular those of zones where NO is produced [4–13]. For each operating point, there is a decrease of NOx emission with WDE for a given EGR rate, from 14% of up to 75%. At low load conditions (operating points 1 and 2), very low NOx emission can be achieved with EGR and WDE. The decrease of NOx emission at Fig. 11. NOx–PM trade-offs comparison with diesel-fuel and WDE. A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 3189 higher load conditions is less interesting, in particular for operating point 5 (Fig. 9e). These tests conducted at a given water-to-fuel ra- tio show that the variation in NOx emission is not only a function of the water content in the WDE, as depicted in some studies on large Diesel engines, but strongly depends on the load, and proba- bly on some other engine parameters. 4.2.2. PM emission PM emission while increasing EGR rate with diesel-fuel or WDE is given in Fig. 10. With diesel-fuel, while increasing EGR rate, there is a large in- crease of PM emission, as traditionally observed in the classical HTC mode [4–6,11–14]. Main explanations for the increase of PM emission with EGR are the decrease of local temperature in zones where PM is partially oxidised (at the jet periphery) and the reduc- tion of in-cylinder oxygen quantity that decrease PM oxidation rate as well [61,67,68]. The higher the load, the higher is PM emission level for a given EGR rate, because of a reduced oxygen quantity in the combustion chamber. When using WDE, there is a remarkable decrease of PM emis- sion (of up to 90% in these tests) as compared with diesel-fuel, in particular at high EGR rates. Only for operating point 5 the de- crease of PM emission with WDE is moderate (around 30–45%). Main explanations for the decrease of PM emission with WDE have been listed in Section 1. The separation of these effects would be very interesting but is very hard to perform, in particular on a non-optically accessible engine. As described earlier, these tests show an evident increase of air entrainment per unit of mass of diesel-fuel, of which a consequence is a decrease of PM emission. 4.2.3. NOx–PM trade-off The influence of WDE on the NOx–PM trade-off while varying EGR rate is finally depicted in Fig. 11. Since there is a decrease of both NOx and PM emissions with WDE, the NOx–PM trade-off is largely improved with WDE as com- pared with diesel-fuel (decrease of NOx emission level for a given PM emission level, or decrease of PM emission level for a given NOx emission level). The use of WDE in parallel with EGR is thus an interesting in-cylinder method for both NOx and PM emissions. The NOx emission decrease is mainly due to EGR, while the use of WDE allows PM emission to be maintained at a low level while increasing EGR rate and thus compensates for the negative effect of EGR on PM emission observed with pure diesel-fuel. The improve- ment of NOx–PM trade-off is less interesting at higher load condi- tions (operating point 5). 4.3. Influence on brake specific diesel-fuel consumption (BSFC) Brake specific diesel-fuel consumptions with pure diesel-fuel and WDE while varying EGR rate are given in Fig. 11. In all cases, there is an increase of BSFC while increasing EGR rate, whether with diesel-fuel or WDE. The increase of BSFC with increased EGR rate while maintaining injection parameters (pilot and main SOI, pilot and main quantities and injection pressure) is mainly due to the off-phasing effect of EGR on combustion that results in lower in-cylinder pressures during the compression stroke. Some tests (not presented here) have been done to try to main- tain the combustion phasing while increasing EGR rate by advancing the injections events and have shown that BSFC is kept at the same level than without EGR. Only for high EGR rates there is an increased BSFC that cannot be compensated by advancing the injections, because of a reduction of the combustion efficiency. As regards the difference between diesel-fuel and WDE, oppo- site tendencies are observed between the different operating points tested here. There is a slight decrease of BSFC with WDE for operating points 3 and 5, a slight increase for operating points 2 and 4, and a large increase for operating point 1, at a given EGR rate. Opposite phenomena can explain this variation between the different operating points and have been also observed in some previous studies [22,57]: Since injection parameters are kept constant, in particular the SOI and the injection pressure, WDE has a tendency to delay the combustion process, thus lowering cylinder pressure (described in Section 4.1). This off-phasing effect of WDE has thus a negative effect on BSFC. On the other hand, the cooling of in-cylinder content due to evaporation tends to diminish the temperature difference between the gases and cylinder wall, thus reducing the heat transfer. This, in turn, may lead to an increase of thermal effi- ciency and a slight reduction in BSFC. An increase of thermal efficiency has been observed for instance by Ghojel et al. [47]. Since the off-phasing effect of WDE on combustion is very marked for operating point 1 (low load condition), this effect is predominant and results in a large increase of BSFC for operating point 1. For the other operating points, the variation of BSFC is low (around 2–3%). 5. Conclusion An experimental study on the use of water-in-diesel emulsion has been conducted on a common-rail high injection pressure automotive Diesel engine, with a volumetric water-to-fuel ratio of 25.6%. Main conclusions are as follows: The effect of WDE on combustion and emissions depends on injection strategy and operating conditions (engine speed and load). Most often, WDE increases ID (of up to 29%), since water evap- oration lowers in-cylinder temperature. Consequently, pre- mixed part of combustion increases whereas diffusion combustion rate can be slightly decreased when injection total mass flow rate is kept constant (in that case, fuel mass flow rate decreases when WDE is used and injection duration increases). When injection pressure is increased with WDE, so that fuel mass flow rate is approximately the same with WDE as with diesel fuel, diffusion combustion rate increases due to air entrainment enhancement (water addition increasing spray momentum). With the lower injection pressure NOx emission is always reduced when using WDE, compared to pure diesel fuel (the NOx emission relative reduction can vary from 30% to 50%). This can be explained by flame temperature reduction caused by water evaporation, thermal dilution with water and combustion off-phasing. When injection pressure is increased, NOx emission can even increase (of up to 24%) with WDE compared to pure diesel, due to faster combustion. The use of WDE usually reduces PM emission compared to pure diesel fuel. This is especially true at low load conditions (the maximum observed PM relative reduction being 94%) and when injection pressure is increased (this result is also classical with pure diesel fuel). However, PM emission sometimes increases when WDE is used at higher loads. Indeed WDE has several con- trary effects on both PM production and oxidation. Finally, when used in combination with EGR, WDE allows reducing both NOx and PM emissions, the relative reduction being approximately the same whatever the EGR rate. Thus, the traditional NOx-PM trade-off is largely improved and very low emission level can be achieved. 3190 A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 By the way, even if the concept seems promising, several issues should be addressed before any industrial use: Cold-start may be impossible with WDE, so that injection sys- tem should be purged before engine stop in order to use pure diesel fuel for engine start. There might some reliability trouble with some injection com- ponents, in particular high pressure pumps that might need some specific modifications. Long term emulsion stability could be problematic if a WDE dis- tribution network would be developed. A possible alternative is an on-board emulsion fabrication, which could allow water/fuel ratio to vary. Acknowledgments The authors would like to thank Mr. Sébastien Trébuchère, Mr. Anthony Pelletier, Ms. Carole Querel, and Ms. Sheddia Didorally, students at the Ecole Centrale de Nantes, for their valuable partic- ipation to this project, and Dominique Tarlet and his colleagues for introducing us to the fabrication of WDE. References [1] Knetch W. Diesel engine development in view of reduced emissions standards. Energy 2008;33:264–71. [2] Heywood JB. Internal combustion engine fundamentals. New York: McGraw- Hill; 1988. [3] Dickey DW, Ryan TW, Matheaus AC. NOx Control in heavy-duty diesel engines – what is the limit? SAE paper no 980174; 1998. [4] Zheng M, Reader GT, Hawley JG. Diesel engine exhaust gas recirculation – a review on advanced and novel concepts. Energy Convers Manage 2004;45:883–900. [5] Maiboom A, Tauzia X, Hétet JF. Experimental study of various effects of exhaust gas recirculation (EGR) on combustion and emissions of an automotive direct injection diesel engine. Energy 2008;33:22–34. [6] Maiboom A, Tauzia X, Hétet JF. Influence of high rates of supplemental cooled EGR on NOx and PMemissions of an automotive HSDI diesel engine using an LP EGR loop. Int J Energy Res 2008;32:1383–98. [7] Ladommatos N, Abdelhalim SM, Zhao H, Hu Z. The dilution, chemical, and thermal effects of exhaust gas recirculation on diesel engine emissions – Part 4: Effects of carbon dioxide and water vapor. SAE paper no 971660; 1997. [8] Ladommatos N, Abdelhalim SM, Zhao H, Hu Z. Effects of EGR on heat release in diesel combustion. SAE paper no 980184; 1998. [9] Musculus MPB. On the correlation between NOx emissions and the diesel premixed burn. SAE paper no 2004-01-1401; 2004. [10] Nitu B, Singh I, Zhong L, Badreshany K, Henein NA, Bryzik W. Effect of EGR on auto ignition, combustion, regulated emissions, and aldehydes in DI diesel engines. SAE paper no 2002-01-1153; 2002. [11] Ladommatos N, Abdelhalim S, Zhao H. Control of oxides of nitrogen from diesel engines using diluents while minimizing the impact on particulate pollutants. Appl Therm Eng 1998;18:963–80. [12] Kouremenos DA, Hountalas DT, Binder KB. The effect of EGR on the performance and pollutant emissions of heavy-duty diesel engines using constant and variable AFR. SAE paper no 2001-01-0198; 2001. [13] Hountalas DT. Controlling nitric oxide and soot in heavy duty diesel engines using internal measures. In: FISITA 2004 world automotive congress, 2004. [14] Hountalas DT, Benajes J, Pariotis EG, Gonzalez CA. Combination of high injection pressure and EGR to control nitric oxide and soot in DI diesel engines. In: THIESEL 2004 conference on thermo- and fluid dynamic processes in diesel engines; 2004. [15] Hountalas DT, Mavropoulos GC, Binder KB. Effect of exhaust gas recirculation (EGR) temperature for various EGR rates on heavy duty DI diesel engine performance and emissions. Energy 2008;33:272–83. [16] Dec JE, Canaan RE. Plif imaging of NO formation in a di diesel engine. SAE paper no 980147; 1998. [17] Odaka M, Koike N, Tsokamoto Y, Narusawa K, Yoshida K. Effects of EGR with a supplemental manifold water injection to control exhaust emissions from heavy-duty diesel powered vehicles. SAE paper no 910739; 1991. [18] Imahashi T, Hashimoto K, Hayashi JI, Yamada T. Research on NOx reduction for large marine diesel engines. ISME Yokohama; 1995. [19] Samec N, Dibble RW, Chen JH, Pagon A. Reduction of NOx and soot emission by water injection during combustion in a diesel engine. In: FISITA 2000 world automotive congress; 2000. [20] Brusca S, Lanzafame R. Evaluation of the effects of water injection in a single cylinder CFR cetane engine. SAE paper no 2001-01-2012; 2001. [21] Kegl B, Pehan S. Reduction of diesel engine emissions by water injection. SAE paper no 2001-01-3259; 2001. [22] Nazha MAA, Rajukaruna H, Wagstaff SA. The use of emulsion, water induction and EGR for controlling diesel engine emissions. SAE paper no 2001-01-1941; 2001. [23] Hountalas DT, Mavropoulos GC, Zannis TC, Mamalis SD. Use of water emulsion and intake water injection as NOx reduction techniques for heavy-duty diesel engines. SAE paper no 2006-01-1414; 2006. [24] Hountalas DT, Mavropoulos GC, Zannis TC. Comparative evaluation of EGR, intake water and fuel/water emulsion as NOx reduction techniques for heavy- duty diesel engines. SAE paper no 2007-01-0120; 2007. [25] Tauzia X, Maiboom A, Shah SR. Experimental study of inlet manifold water injection on combustion and emissions of an automotive direct injection diesel engine. Energy 2010;35:3628–39. [26] Miyano H, Yoshida N, Nakai T, Nagae Y, Yasueda S. Stratified fuel-water injection system for NOx reduction of diesel engine. ISME Yokohama; 1995. [27] Psota MA, Easley WL, Fort TH, Mellor AM. Water injection effects on NOx emissions for engines utilizing diffusion flame combustion. SAE paper no 971657; 1997. [28] Bedford F, Rutland C, Dittrich P, Raab A, Wirbeleit F. Effects of direct water injection on DI diesel engine combustion. SAE paper no 2000-01-2938; 2000. [29] Stanglmaier RH, Dingle PJ, Stewart DW. Cycle-controlled water injection for steady-state and transient emissions reduction from a heavy-duty diesel engine. J Eng Gas Turb Power 2008:130. [30] Chadwell CJ. Effect of diesel and water co-injection with real-time control on diesel engine performance and emissions. SAE paper no 2008-01-1190; 2008. [31] Tanner FX, Brunner M, Weisser G. A computational investigation of water injection strategies for nitric oxide reduction in large-bore diesel engines. SAE paper no 2001-01-1069; 2001. [32] Nishijima Y, Asaumi Y, Aoyagi Y. Impingement spray system with direct water injection for premixed lean diesel combustión control. SAE paper no 2002-01- 0109; 2002. [33] Ahern B, Djutrisno I, Donahue K, Haldeman C, Hynek S, Johnson K, et al. Dramatic emissions reductions with a direct injection diesel engine burning supercritical fuel/water mixtures. SAE paper no 2001-01-3526; 2001. [34] Barnaud F, Schmelzle P, Schulz P. AQUAZOLE™: an original emulsified water– diesel fuel for heavy-duty applications. SAE paper no 2000-01-1861; 2000. [35] Matheaus AC, Ryan TW, Daly D, Langer DA, Musculus MPB. Effects of PuriNOx™ water–diesel fuel emulsions on emissions and fuel economy in a heavy-duty diesel engine. SAE paper no 2002-01-2891; 2002. [36] Song KH, Lee YJ, Litzinger TA. Effects of emulsified fuels on soot evolution in an optically-accessible DI diesel engine. SAE paper no 2000-01-2794; 2000. [37] Tajima H, Takasaki K, Nakashima M, Kawano K, Ohishi M, Yanagi J, et al. Visual study on combustion of low-grade fuel water emulsion. In: COMODIA 2001 5th international symposium on diagnostics and modeling of combustion in internal combustion engines; 2001. [38] Kee SS, Mohammadi A, Hirano H, Kidoguchi Y, Miwa K. Experimental study on combustion characteristics and emissions reduction of emulsified fuels in diesel combustion using a rapid compression machine. In: JSAE/SAE 2003 international spring fuels & lubricants meeting. SAE paper 2003-01-1792; 2003. [39] Musculus MPB, Dec JE, Tree DR, Daly D, Langer D, Ryan TW, et al. Effects of water–fuel emulsions on spray and combustion processes in a heavy-duty DI diesel engine. SAE paper no 2002-01-2892; 2002. [40] Hall D, Thorne C, Goodier S. An investigation into the effect of a diesel/water emulsion on the size and number distribution of the particulate emissions from a heavy-duty diesel engine. SAE paper no 2003-01-3168; 2003. [41] Gonzalez MA, Rivas H, Gutierrez X, Leon A. Performance and emissions using water in diesel fuel microemulsion. SAE paper no 2001-01-3525; 2001. [42] Selim MYE, Ghannam MT. Performance and engine roughness of a diesel engine running on stabilised water diesel emulsion. SAE paper no 2007-24- 0132; 2007. [43] Lee CFF, Wang KT, Cheng WL. Atomization characteristics of multi-component bio-fuel systems under micro-explosion conditions. SAE paper no 2008-01- 0937; 2008. [44] Jeong IC, Lee KH. Auto-ignition and micro-explosion behaviours of droplet arrays of water-in-fuel emulsion. Int J Automot Technol 2008;9:735–40. [45] Tarlet D, Belletre J, Tazerout M, Rahmouni C. Prediction of micro-explosion delay of emulsified fuel droplets. Int J Therm Sci 2009;48:449–60. [46] Selim MYE, Elfeky SMS. Effects of diesel/water emulsion on heat flow and thermal loading in a precombustion chamber diesel engine. Appl Therm Eng 2001;21:1565–82. [47] Ghojel J, Honnery D, Al-Khaleefi K. Performance, emissions and heat release characteristics of direct injection diesel engine operating on diesel oil emulsion. Appl Therm Eng 2006;26:2132–41. [48] Lin CY, Chen LW. Engine performance and emissions characteristics of three- phase diesel emulsions prepared by an ultrasonic emulsification method. Fuel 2006;85:593–600. [49] Alahmer A, Yamin J, Sakhrieh A, Hamdan MA. Engine performance using emulsified diesel fuel. Energy Convers Manage 2010. doi:10.1016/ j.enconman.2009.11.04. [50] Nadeem M, Rangkuti C, Anuar K, Haq MRU, Tan IB, Shah SS. Diesel engine performance and emission evaluation using emulsified fuels stabilized by conventional and Gemini surfactants. Fuel 2006;85:2111–9. [51] Kadota T, Yamasaki H. Recent advances in the combustion of water fuel emulsion. Prog Energy Combust Sci 2002;28:385–404. A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192 3191 [52] Lif A, Skoglundh M, Gjirja S, Denbratt I. Reduction of soot emissions from a direct injection diesel engine using water-in-diesel emulsion and microemulsion fuels. SAE paper no 2007-01-1076; 2007. [53] Ochoterena R, Lif A, Nyden M, Andersson S, Denbratt I. Optical studies of spray development and combustion of water-in-diesel emulsion and microemulsion fuels. Fuel 2010;89:122–32. [54] Lif A, Stark M, Nyden M, Holmberg K. Fuel emulsions and microemulsions based ion Fischer–Tropsch diesel. Colloids Surf A: Physicochem Eng Aspects 2010;354:91–8. [55] Qi DH, Chen H, Matthews RD, Bian YZH. Combustion and emissions characteristics of ethanol–biodiesel–water micro-emulsions used in a direct injection compression ignition engine. Fuel 2010;89:958–64. [56] Roberts CE, Naegeli D, Chadwell C. The effect of water on soot formation chemistry. SAE paper no 2005-01-3850; 2005. [57] Bertola A, Boulouchos K, Li R. Influence of water–diesel fuel emulsions and Egr on combustion and exhaust emissions of heavy-duty, di-diesel engines equipped with common-rail injection system. SAE paper no 2003-01-3146; 2003. [58] Siebers DL, Higgins B, Pickett LM. Flame lift-off on direct-injection diesel fuel jets: oxygen concentration effects. SAE paper no 2002-01-0890; 2002. [59] Pickett LM, Siebers DL. Non-sooting, low-flame temperature mixing-controlled DI diesel combustion. SAE paper no 2004-01-1399; 2004. [60] Pickett LM, Siebers DL. Soot in diesel fuel jets: effects of ambient temperature, ambient density, and injection pressure. Combust Flame 2004;138: 114–35. [61] Tree DR, Svensson KI. Soot processes in compression ignition engines. Prog Energy Combust Sci 2006;33:272–309. [62] Woodyard D. Pounder’s marine diesel engines and gas turbines. 9th ed. Elsevier; 2009. [63] Fuel water emulsion. <www.mandieselturbo-greentechnology.com> [64] Wärtsilä fuel flexibility. <www.wartsila.com> [65] National bus experimentations. <www.starbus-project.eu> [66] Hohenberg JF. Advanced approaches for heat transfer calculations. SAE paper no 790825; 1979. [67] Maiboom A, Tauzia X, Shah SR, Hétet JF. New phenomenological six-zone combustion model for direct-injection diesel engines. Energy Fuels 2009;23:690–703. [68] Dec JE. A Conceptual model of di diesel combustion based on laser-sheet imaging. SAE paper no 970873; 1997. [69] Mollenhauer K, Tschöke H. BOSCH handbook of diesel engines. Springer; 2010. 3192 A. Maiboom, X. Tauzia / Fuel 90 (2011) 3179–3192