ILASS Americas, 20th Annual Conference on Liquid Atomization and Spray Systems, Chicago, IL, May 2007

Fuel Effects on the Spray and Combustion Processes Within an Optical HSDI Diesel Engine

Tiegang Fang, Tien Mun Foong, Yuan-chung Lin, and Chia-fon F. Lee*

Department of Mechanical Science and Engineering University of Illinois at Urbana-Champaign

Urbana, IL 61801 USA

Abstract An optically accessible single-cylinder high speed direct-injection (HSDI) Diesel engine equipped with a Bosch common rail injection system was used to study the spray and combustion evolution using different fuels including European low sulfur diesel and biodiesel fuels. Influences of injection timing and fuel type on liquid fuel evolution and combustion characteristics were studied under similar loads. High-speed Mie-scattering was employed to inves-tigate the liquid distribution and evolution. High-speed combustion video was also captured for all the studied cases using the same frame rate. NOx emissions were measured in the exhaust pipe. It is found that biodiesel fuel leads to longer liquid penetration and low soot formation with increased NOx emissions, and injection timings play impor-tant roles in soot and NOx emissions.

* Corresponding author: cflee@uiuc.edu

Introduction Direct Injection (DI) diesel engines offer higher

thermal efficiency and better reliability compared with gasoline engines. Diesel engines are considered attrac-tive power stations not only for heavy-duty vehicles but also for light-duty ones. Because of the increasing threat on limited fossil fuel resources and worldwide concern of environmental issues, the regulation of emissions on current engines are becoming more and more stringent. Exhaust emissions, for example oxides of nitrogen (NOx) and Particulate Matter (PM), must be reduced for diesel engines to meet the stricter emission standards. However, due to the well-known trade-off law of NOx and PM emissions for diesel engines, it is difficult to reduce these two emissions simultaneously. New techniques or new combustion concepts must be developed to solve the problems. Homogeneous Charge Compression Ignition (HCCI) combustion concept is a promising technique to meet the requirements.

The HCCI concept was proposed by some re-searchers in the late 70’s of the last century [1-2] for two-stroke engines. The basic idea is to form a more uniform, ideally homogeneous, air-fuel mixture before auto-ignition. Thus the conventional diffusion flame will be eliminated from the chemical reaction and the fuel is mostly burnt in a premixed combustion mode. The HCCI study in four stroke engines was first re-ported by Najt and Foster [3] in 1983. The name of HCCI combustion is attributed to Thring [4]. Although there have been many investigations [5-10] about HCCI combustion mode in diesel engines in the past decade, only a few practical techniques were implemented in a real engine. Modulated Kinetics (MK) combustion [8-10] is one of the HCCI combustion modes used in commercial diesel engines. MK combustion was ac-complished by using heavy EGR rate, retarding injec-tion timing, enhancing air motion, and increasing injec-tion pressure. MK combustion provides a practical ap-proach to implement HCCI combustion in diesel en-gines without massive modifications in combustion chamber and fuel injection systems. This new combus-tion concept results in reduction of NOx and smoke simultaneously due to low-temperature and premixed combustion without increasing fuel consumption.

Agricultural fat and oils, in raw or chemically mod-ified forms, have the potential to supplant a significant proportion of petroleum-based fuels. Bio-diesel is of particular interest to the automobile industry and other areas in energy and environment because it signifi-cantly reduces particulate matter (PM), hydrocarbon (HC) and carbon monoxide (CO) emissions. The engine testing from three different engines, a Cummins N-14 engine, a Cummins B5.9 engine, and a DDC Series 50 engine showed average reductions of 84.4% in HC, 40.5% in CO, and 38.0% in PM emissions [11].

In addition to its benefits to Environmental Protec-tion Agency (EPA) regulated exhaust emissions of PM, HC and CO, bio-diesel contributes less to global warm-ing than fossil fuels due to its closed carbon cycle. There is almost no net increase of carbon dioxide (CO2) emission from bio-diesel combustion. Bio-diesel is also the only alternative fuel that has passed the EPA-required Tier I and Tier II Health Effects testing re-quirements of the Clean Air Act Amendments of 1990. Moreover, bio-diesel is particularly attractive because it is a renewable fuel that can be replenished through the growth of plants or production of livestock, and it has the potential to supplant a fraction of petroleum-based fuels.

One of the main factors impacting the use of bio-diesel is its NOx emission. Bio-diesel has been criti-cized for its up to 15% higher brake specific NOx emis-sions comparing to diesel fuels. In the last ten years, numerous studies and measurements of NOx emissions from diesel engines fueled with bio-diesel have been published [11-19]. However, most of the reports were focused on conventional diffusion combustion of bio-diesel. Low temperature combustion has its unique ad-vantage to inhibit the formation of NOx during the combustion process. In addition, compared to petro-leum diesel fuel, the cetane number of bio-diesel is generally higher, which results in easier auto-ignition and is beneficial to the low temperature combustion process. The low temperature combustion of bio-diesels has not been reported in the literature, it is of practical interests to investigate the combustion characteristics of bio-diesels in a small bore HSDI diesel engine and compare with the low sulfur European diesel.

Air-fuel mixing process is crucial for combustion in diesel engines. Liquid fuel distribution visualization in engine cylinder provides useful information on the evolution of diesel spray evaporation and fuel droplet dispersion. Mie-scattering technique is a commonly used method for liquid phase visualization [20-23]. The scattered signal is seen proportional to the second pow-er of fuel droplet diameter [24], which can be used qua-litatively for liquid fuel distribution without calibration. Using a high-speed video camera synchronized with a high-speed laser light source, the transient liquid fuel evolution can be obtained for a whole injection event.

Combustion visualization gives a qualitative feel for the effects of differing injection strategies. Imaging of the natural flame luminosity from the combustion event through the use of an optical engine has been a technique that has garnered widespread use [25-31]. These works identified ignition locations, flame tem-peratures, evidence of flame wall interaction and late cycle events such as soot oxidation.

In the current work, the effects of different fuels, namely European low sulfur diesel fuel and bio-diesel fuel, on combustion processes in an optical engine with

realistic piston geometry using different injection tim-ings will be presented.

Optical Engine and Facility The optical engine was built using a single cylinder

DIATA research engine supplied by Ford Motor Com-pany. Key aspects of the DIATA engine are listed in Table 1. The design is based on the drop-liner design employed at Sandia National Labs in Livermore, CA. Optical access to the combustion chamber is attained from the side through a window just below the head, or from below through the fused silica piston top, which is attached to a Bowditch-type piston extension as shown in Fig. 1. The optical engine design maintains the ge-ometry of the ports and combustion chamber of the original engine. A complete description of the optical engine used in this study can be found in a previous publication [32].

A Bosch common-rail electronic injection system is used on the research engine, capable of injection pressures up to 1350 bar. A Valve-Covered-Orifice (VCO) injector with six 0.124 mm holes placed sym-metrically in the nozzle tip was used. The injector body is fitted with a needle lift sensor monitoring the needle motion throughout injection.

Bore 70 mm Stroke 78mm Displacement/Cylinder 300 cc Compression Ratio 19.5:1 Swirl Ratio 2.5 Valves/Cylinder 4 Intake Valve Diameter 24 mm Exhaust Valve Diameter 21 mm Maximum Valve Lift 7.30/ 7.67 mm (Intake/ Exhaust) Intake Valve Opening 13 CAD ATDC (at 1 mm valve lift) Intake Valve Closing 20 CAD ABDC (at 1 mm valve lift) Exhaust Valve Opening 33 CAD BBDC (at 1 mm valve lift) Exhaust Valve Closing 18 CAD BTDC (at 1 mm valve lift)

Table 1. Engine specifications of the single cylin-der DIATA research engine.

A 12-bit un-intensified high-speed video camera, Phanton v7.0, built by Vision Research, Inc. is used to obtain the injection and combustion videos. A Copper-Vapor laser synched with the camera is used to illumi-nate the liquid fuel during the injection videos. The camera was operated at 12,000 fps for both injection and combustion studies. During the combustion videos, the same lens f-stop was used for the four cases; there-fore perceived intensities are directly comparable.

National Instruments LabView version 6.0 is used as the data acquisition and timing software for the en-gine. An optical shaft encoder with 0.25 crank angle resolution is used to provide the time basis on which all data acquisition timing systems are operated.

Experimental Operating Conditions An operating speed of 1500 RPM and a nominal 2.0

bar gross Indicated Mean Effective Pressure (IMEP)

load were chosen for the current study. Intake tempera-ture and pressure were increased for the optical engine to match the metal engine TDC conditions. Single in-jection strategies with different injection timings were considered. The injection angle is 150 degree for all conditions. The tested rail pressure was maintained at 600 bar. The injection duration was adjusted to match 2.0 bar IMEP. Three injection timings were chosen including -25 CAD ATDC, -10 CAD ATDC, and 3 CAD ATDC for both fuels. These injection timings represent different combustion mode with an early in-jection strategy, a conventional diesel combustion strat-egy, and a late low temperature combustion strategy. A summary of operating conditions is tabulated in Table 2. The fuels used include a low-sulfur European Diesel fuel and a soybean bio-diesel fuel. Selected properties of the fuels are listed in Table 3. Pictures of the two fuels are shown in Fig. 2. It is obvious that bio-diesel fuel has a dark color than the European low sulfur die-sel fuel.

Hydraulic Assembly

Drop Liner Assembly

InvarPiston Sleeve

Side Window

Quartz Piston

Piston Extension

Hydraulic Assembly

Drop Liner Assembly

InvarPiston Sleeve

Side Window

Quartz Piston

Piston Extension

Figure 1. Assembly cross-section of optical

engine design with drop-liner raised.

Fuel Type

Case Number

Main SOI [CAD

ATDC]

Main Duration

[µs]

Fuel Quan-tity

[mm3] IMEP[bar]

B0 1 -25 165 8.4 2.00 B0 2 -10 125 4.9 2.06 B0 3 3 120 4.4 2.03

B100 1 -25 175 9.3 2.02 B100 2 -10 135 5.6 2.00 B100 3 3 140 6.0 2.04

Table 2. Summary of the selected operating condi-tions.

Both Mie-scattering and combustion videos were taken with the high-speed digital camera operating at a framing rate of 12,000 fps. This frame rate corresponds to 0.75 CAD intervals for two consecutive images. For each case, 5 sets of movies were taken and a typical set of images will be presented. The engine was operated in a skip-fire mode with one injection cycle followed by

12 flushing cycles. Three-dimensional-like videos were obtained by imaging through the piston bottom and the side window [33]. The image resolution is 512x256 pixels.

Fuel Property

European Low Sulfur Diesel Fuel

Soybean Bio-diesel Fuel

Specific Gravity 0.837 0.877

Sulfur (ppm) 196 ~0 Flash Point

(°F) 130.4 >201

Boiling Point (°F)

396.3 (IBP) 518.0(50%) 671.9 (EP) >600

Viscosity (cps) 3.2 (@40°C) 7(@25°C)

Cetane Number 54.0 50.9[ref. 100]

Table 3. Summary of the selected properties of the

two fuels.

Figure 2. Pictures of different fuels: European low sulfur diesel (left) and soybean bio-diesel fuel (right).

Results and Discussion The in-cylinder pressure traces are shown in Fig.

3(a) with the heat release rate data illustrated in Fig. 3(b). The influence of injection timings on the in-cylinder pressure is quite similar for different fuels. For an early pre-TDC injection timing, the ignition takes place quite early before TDC. The in-cylinder pressure is relatively higher than other injection timings. At the same time, because of the early combustion process, the high temperature combusted gas mixture is further compressed by the piston resulting in even higher in-cylinder temperature. This high in-cylinder temperature will possibly increase the NOx emissions. For an injec-tion timing at –10 CAD ATDC, ignition occurs around TDC. This ignition timing is helpful in obtaining higher thermal cycle efficiency by making full use of the geo-metric compression ratio of the engine. The in-cylinder pressure peak is still quite high, but it is lower than the early injection timing case. With a post-TDC injection timing at 3 CAD ATDC, the late injection strategy case has a much longer ignition delay than the case with –10 CAD ATDC injection timing. The in-cylinder pressure is much lower than the other two injection timings. For

some cases, the combustion pressure peak is lower than the motoring pressure peak.

3(a)

3(b)

Figure 3. In-cylinder pressure and heat release rate for the chosen operating conditions.

The effects of different fuels on the in-cylinder pressure are quite apparent. It is observed that the igni-tion delay is longer bio-diesel fuel. This can be ex-plained from the properties of different fuels. From Table 3, the bio-diesel fuel has a relatively higher boil-ing point. This high boiling point leads to a slower droplet evaporation rate and as a consequence, the preparation of the ignitable air-fuel mixture is delayed by a certain time. In addition, the cetane number for the bio-diesel fuel is expected to be similar or less than the European low sulfur diesel fuel. This may have similar effects on the auto-ignition processes. By combining these two factors, it is reasonable that higher bio-diesel content results in a longer ignition delay.

The heat release rates for these conditions are illustrated in Fig. 3(b). The heat release rate information further confirms the observation in the in-cylinder pres-sure. It is worth mentioning that most of the conditions have premixed dominated combustion heat release rate patterns. For a certain type of fuel with different injec-tion timings, an early pre-TDC injection timing, namely at –25 CAD ATDC, results in a higher heat release rate peak followed by the conventional injection timing and then by the retarded post-TDC injection timing case. A higher heat release rate peak value leads to noisy com-bustion. The heat release duration is shorter for the ear-ly injection timing case than the post-TDC late injection timing case. The heat release duration for a late post-TDC injection timing is also longer and the rate is low-er than the early pre-TDC injection case. This is a pre-ferred heat release rate pattern in terms of noise reduc-tion. For different fuels, the ignition delay is longer for the bio-diesel fuel as explained in the previous para-graph. For Case 1 of both of the fuels, the heat release rate curves are quite similar with approximately the same duration. But there is a slight difference in the

heat release rate peak values. A lower value is observed for bio-diesel fuel, which is attributed to the higher boiling point of the bio-diesel fuel. For the conventional injection timing case, namely Case 2, B0 has a lower heat release rate peak than bio-diesel with some evi-dence of diffusion combustion after the dominant pre-mixed heat release event. But for bio-diesel fuel, little diffusion combustion can be found. This is believed to be a result of the longer ignition delay for the bio-diesel blends than the pure low sulfur diesel fuel. For retarded post-TDC injection timings, namely Case 3, the heat release rate curves become lower and broader for the bio-diesel fuel.

European low sulfur diesel fuel (Case 1)

Soybean bio-diesel fuel (Case 1)

Figure 4. Mie-scattering images of the two fuels for early injection strategy at different crank angles.

The liquid spray images are shown in Figs.4-6 ac-cording to the injection timing respectively. The first two images are the very first two images after the start of injection; the third shows the images with maximum liquid penetration; and the other 3 images depict the spray shutting off process.

With an injection timing at –25 CAD ATDC, fuel comes out of the injector nozzle at about 1.5 CAD ASOI with very little difference among the different fuel blends. The spray development process is also sim-ilar for different fuels. Due to the fuel quantity differ-ence as seen in Table 3, a slightly longer injection dura-tion is seen for fuels with more bio-diesel content. But the difference is less than 0.75 CAD based on the spray images. For both fuels, the injection is done well before the ignition timing indicating a premixed combustion mode for these conditions. The spray piston wall im-pingement is observed for both fuels. If careful atten-tion is paid to the side-window Mie-scattering signal from the fuel droplet or film near the bowl wall, a slightly stronger signal is found for B100 than B0, which shows more fuel impingement for B100. The stronger fuel impingement for B100 than B0 is attrib-uted to a few factors. First of all, bio-diesel fuel has a higher boiling point leading to a slower fuel evapora-tion rate causing longer penetration and more fuel im-pingement. Secondly, the density of bio-diesel is slightly higher than that of the diesel fuel, which causes longer penetration. Thirdly, the fuel quantity for the bio-diesel blends is slightly higher than the diesel fuel for the load conditions with the same IMEP due to a lower heat content of the bio-diesel fuel. The combined effects of these factors make the fuel impingement of bio-diesel stronger than B0.

The spray images for Case 2 with different fuels are shown in Fig. 5. The overall process is quite similar to that of Case 1. Fuel comes out of the nozzle at about 1.5 CAD ASOI. Not much difference is seen for differ-ent fuels. B100 has more fuel impingement than B0. For B0, an early flame occurs near the end of injection but before the end of injection, which implies that there is some liquid fuel injected into this flame. This proves the existence of some weak diffusion flame combustion observed from the heat release rate curves. For the bio-diesel fuel, there is no flame-spray overlap indicating a premixed combustion mode. Compared with the spray images of Case 1 with an early pre-TDC injection tim-ing, although the spray develops in a similar manner, fuel impingement is seen to be stronger for Case 1 than Case 2 for both fuels. This difference is attributed to the difference in ambient conditions. For a later injection timing, the ambient gas density is higher and the tem-perature is also higher. A high ambient air density re-duces the liquid spray penetration while a high air tem-perature increases the fuel evaporation rate with less

liquid penetration. Both factors make the fuel impinge-ment stronger for Case 1 than Case 2.

European low sulfur diesel fuel (Case 2)

Soybean bio-diesel fuel (Case 2)

Figure 5. Mie-scattering images of the two fuels for conventional diesel combustion injection strategy at different crank angles.

With retarded injection timings, the spray images shown in Fig. 6 do not show very different spray evolu-tion. Fuel appears out of the nozzle at about 1.25 CAD ATDC. The injection duration is about 0.75 CAD shorter than that of Case 2 for B0. This is consistent with the fact that less fuel is injected for Case 3 than Case 2 as listed in Table 2 for B0. The less fuel quantity may be explained by the reduced heat loss and more premixed combustion. As seen in the combustion im-

ages for Case 2, combustion flame contacts with the piston wall in a relatively large area. But for Case 3, the flame has much less contact with the piston wall. This reduced flame-wall contact greatly decreases the heat loss to the wall. The other factor is due to the radiation heat loss to the wall. As to be seen in the combustion images later, a significantly higher soot concentration is observed for Case 2 than Case 3. Soot particles are good heat radiators. Lower soot concentration in Case 3 greatly reduces the heat loss. The fuel impingement for Case 3 is seen to be stronger than that of Case 2. Stronger fuel impingement is seen for B100.

European low sulfur diesel fuel (Case 3)

Soybean bio-diesel fuel (Case 3)

Figure 6. Mie-scattering images of the two fuels for low temperature combustion late injection strategy at different crank angles.

B0: Case 1 Case 2 Case 3

B100: Case 1 Case 2 Case 3

Figure 7. Comparison of the fuel films on the pis-ton bowl for B0 and B100 right after the end of injec-tion under different injection timings

In order to further elaborate the fuel impingement and fuel film deposition on the wall, amplified bottom-viewed spray images for B0 and B100 at different in-jection timings are illustrated in Fig. 7. For B100, there are obvious fuel film ripples on the bowl wall as indi-cated by the ovals in the figure. For B0, no obvious fuel film ripples can be found in the images. The fuel film is a liquid sheet that does not scatter as much light as the droplet clouds. Only some signal from the film ripples can be observed. However, this observation is sufficient to confirm that more fuel film deposition on the wall occurs for B100 than B0. The fuel film of Case 1 for B100 is the most obvious followed by Case 3, then Case 2. This fuel film further confirms the strength of fuel impingement for different injection timings.

The combustion images for an early pre-TDC in-jection timing at –25 CAD ATDC are shown in Fig. 8 for both fuels. Ignition for B0 occurs at about –13.75 CAD ATDC, about 2 CAD later than the end of injec-tion. The combustion mode is dominated by premixed combustion. Combustion flame is mostly confined in the bowl with little in the squish region. As a matter of fact, the flame is mostly near the bottom of the bowl. Luminous flame is observed for some crank angles in-dicating soot formation in the combustion flame. How-ever, this highly luminous flame is oxidized after a few crank angle degrees. Late cycle flame shows a local structure on the piston wall due to the fuel impingement during the fuel injection process. Up to 14 CAD ATDC, the flame is almost invisible in the engine cylinder with very low flame luminosity. The differences between B100 and B0 are ignition timing and flame intensity. The early flame occurs later for B100. The flame inten-sity B100 is much lower than that of B0. Besides the low flame intensity for bio-diesel, the flame has some local structure. The flame is not as distributed as that of B0. The locally strong flame intensity in the bio-diesel fuel is mainly due to the slow evaporation rate of the bio-diesel fuel. As discussed in the previous section, a later ignition timing for B100 is possibly attributed to

the lower volatility of bio-diesel fuel. The cetane num-ber for bio-diesel is close or less than the European low sulfur diesel, which may partly contribute to the igni-tion timing too. The fact that B0 has the most soot for-mation in the combustion flame is due to that pure die-sel fuel has no oxygen in the fuel compared with bio-diesel fuel and the combustion occurs in a relatively richer environment than bio-diesel fuel. From late cycle flames, there are some local flames in the bowl wall for B100 fuel. This observation confirms that the fuel im-pingement for B100 is stronger than B0 as seen in the spray images.

European low sulfur diesel fuel (Case 1)

Soybean bio-diesel fuel (Case 1)

Figure 8. Combustion images of the two fuels for early injection strategy at different crank angles.

As the injection timing is retarded, the combus-tion mode transits to a more conventional diffusion combustion mode for B0 as seen in Fig. 9 with an injec-tion timing at –10 CAD ATDC. Ignition occurs early

during the fuel injection process at about –2.5 CAD ATDC. The fuel injection ends at about –1.0 CAD ATDC. There is obvious flame spray overlap for B0. Some evidence of diffusion combustion can be seen in the combustion images with apparent liquid jets in early flames. Some flame is observed in the squish region and most of the flame is within the bowl region. This flame structure is due to the spray impingement on the bowl lip with fuel split into the squish region and the piston bowl. The flame is highly luminous showing a large amount of soot generation. There is obvious flame contact with the bowl wall for B0, which leads to the heat loss to the wall. The late cycle flame becomes weak due to soot oxidation. At 32 CAD ATDC, there is still some flame observed in the bowl.

European low sulfur diesel fuel (Case 2)

Soybean bio-diesel fuel (Case 2)

Figure 9. Combustion images of the two fuels for conventional combustion injection strategy at different crank angles.

European low sulfur diesel fuel (Case 3)

Soybean bio-diesel fuel (Case 3)

Figure 10. Combustion images of the two fuels for low temperature combustion late injection strategy at different crank angles.

Compared with B0, bio-diesel fuel has a longer ig-nition delay. This longer ignition delay leads to a pre-mixed combustion mode with –10 CAD ATDC injec-tion timing. Even though there are some local early flames near the spray tip location in the bowl, no appar-ent jet flames are seen. Flame intensity is weaker than B0 indicating less soot formation in the combustion flame. Another difference is that the late cycle flame is oxidized much faster for B100 than B0. This provides a proof that the bio-diesel fuel has a higher late cycle soot oxidation rate than pure diesel fuel. The reason is the oxygen content in the fuel. Based on the current results, it is concluded that bio-diesel fuel greatly helps reduce the soot emissions due to the fact that soot formation is less and soot oxidation is faster compared with pure

diesel fuel. Both factors lead to the lower soot emis-sions in the exhaust for the bio-diesel fuel.

With retarded post-TDC injection timings, the combustion mode transits to premixed combustion for all the four fuels as seen in Fig. 9. For B0, the fuel in-jection is completed at about 10.25~11.00 CAD ATDC. Early flame occurs at 11.75 CAD ATDC. There is no flame-spray overlap indicating a premixed combustion mode. However, due to the early ignition timing, the air-fuel mixing process is not sufficient to have an HCCI-like combustion mode. The flame luminosity is still quite high, which is comparable to that of Case 1. Luminous flame lasts a long time. At 38 CAD ATDC, the flame is still observable. But for bio-diesel fuel, the combustion process is significantly different. Due to much later appearance timing of the early flame, name-ly at about 18.5 CAD ATDC for B100, the air-fuel mix-ing process is much longer for these fuels than B0. Consequently, there is sufficient time to obtain a more distributed air-fuel mixture with an HCCI or PCCI-like combustion mode. The flame intensity is much weaker compared with that of B0. The weak flame has little contact with the bowl wall with less heat loss from the wall. Compared with Case 2, the flame contact on the wall is a lot less for Case 3. This, to some extent, ex-plains the lower fuel quantity for Case 3 than Case 2 for B0.

Figure 11. The time integrated SIFL and NOx

emissions for the six conditions.

In order to evaluate the emission performance of different operation conditions, the NOx emissions and the time integrated Spatially Integrated Flame Luminos-ity (SIFL) data are plotted in Figs. 10. The time integra-tion of SIFL is a parameter that can show the soot emis-sion characteristics and the NOx emissions directly rep-resent the NOx concentration in the exhaust pipe. The time integration of SIFL is consistent with the flame luminosity observation. For Case 1, B0 has a larger value that B100. For Case 2 with an injection timing causing a conventional diffusion combustion mode, B0 has a much higher value of the flame luminosity than

B100. Case 2 has a higher value of the time integration of SIFL than Cases 1 and 3 for both fuel types. For Case 3 with retarded injection timing, significantly lower values are observed for B100 than B0. Based on these results, it is concluded that bio-diesel fuel greatly helps reduce the soot emissions due to higher oxygen content in the fuel with less soot formation and high soot oxidation during the combustion process. Early injection and retarded injection are helpful in reducing soot emissions compared with the conventional injec-tion timing for both fuels. NOx emissions show interest-ing variation trends. For Cases 2 and 3, NOx emissions are higher for B100 than B0, which is consistent with the previously published results. However, for early pre-TDC injection timings, NOx emissions are lower for B100 than B0. For this case, there is a trade-off be-tween the ignition delay and the oxygen content. For ignition occurring very early before TDC, early ignition timing leads to a high in-cylinder global temperature. This in-cylinder temperature becomes even higher than the other two injection timings due to more piston com-pression work. Consequently, there is a balanced effect of the high in-cylinder temperature resulting from early ignition and the high oxygen content in the bio-diesel fuel. The tradeoff of ignition delay and bio-diesel con-tent for different fuel blends leads to the observed NOx emissions for Case 1. For a certain type of fuel, retard-ing injection timing significantly reduces the NOx emis-sions. Especially for a retarded post-TDC injection strategy, low NOx and soot emissions are observed due to PCCI or HCCI-like low temperature combustion. Simultaneous reduction of NOx and soot is feasible for the low temperature HCCI-like combustion mode. Without other measures, for instance EGR, an early pre-TDC injection strategy is not preferable to be used in a practical engine. It is concluded that a retarded post-TDC injection strategy leads to the best perform-ance in terms of soot and NOx emissions. Meanwhile, the high NOx emission problems of bio-diesel fuel can be solved by applying this injection strategy with a si-multaneous reduction in soot and NOx emissions com-pared with the conventional combustion with pure European low sulfur diesel fuel. Low temperature HCCI or PCCI-like combustion is a promising solution for low emission bio-diesel engines.

Conclusions In this paper, the effects of European low sulfur di-

esel fuel and bio-diesel fuel on the combustion process were experimentally investigated in a small-bore HSDI diesel engine using single injection strategies. Three injection strategies were studied showing the injection timing influences on the combustion modes. Results show that bio-diesel fuel leads to stronger fuel im-pingement on the wall and longer ignition delay than the low sulfur diesel fuel for all three injection strate-

gies. Less luminous flame was observed for biodiesel fuel than the European low sulfur diesel fuel indicating lower soot concentration in the combustion flame. Compared with conventional injection timing, early injection timing and late injection timing resulted in lower soot formation. For conventional injection timing and late injection timing, higher NOx emission was seen for bio-diesel fuel than the European low sulfur diesel fuel. However, for early injection strategy, bio-diesel fuel led to lower NOx emissions. For a certain type of fuel, retarding injection timing resulted in sig-nificant reduction in NOx emissions. The retarded post-TDC injection timing results in a further reduction in soot formation on the basis of bio-diesel fuel. By com-bining the low sooting characteristics of bio-diesel with the low temperature combustion feature of the retarded post-TDC injection strategy, ultra-low soot and NOx emissions was achieved for bio-diesel engines. It is concluded that without other measures such as EGR, a retarded injection strategy offers the best performance in terms of emissions and fuel consumption among the investigated three injection strategies. This is a promis-ing combustion mode for practical applications in the low emission diesel engines operated with both petro-diesel or bio-diesel fuels in the near future.

Acknowledgements

This work was supported in part by the Department of Energy Grant No. DE-FC26-05NT42634, by De-partment of Energy GATE Centers of Excellence Grant No. DE-FG26-05NT42622, and by the Ford Motor Company under University Research Program. We also thank Paul Miles of Sandia National Laboratories, Evangelos Karvounis and Werner Willems of Ford for their assistance on the design of the optical engine and on the setup of the experiments.

References 1. Onishi, S., Hong Jo, S., Shoda, K., Do Jo, P., Kato,

S., SAE Paper 790501, (1979) 2. Noguchi, M., Tanaka, Y., Tanaka, T., Takeuchi, Y.,

SAE Paper 790840, (1979) 3. Najt, P.M., Foster, D.E., SAE Paper 830264, (1983) 4. Thring, R.H., SAE Paper 892068, (1989) 5. Takeda, Y., Keiichi, Na., Keiichi, Ni., SAE Paper

961163, (1996) 6. Gray III, A.W., Ryan III, T.W., SAE Paper 971676,

(1997) 7. Akagawa, H., et al., SAE Paper 1999-01-0183,

(1999) 8. Kimura, S., et al., SAE paper 1999-01-3681, (1999) 9. Kimura, S., et al., SAE paper 2001-01-0200, (2001) 10. Lee, T., Reitz, R.D., JSME International Journal, 46

(3), 451-459, (2003) 11. Sharp, C.A., S.A. Howell, and J. Jobe. SAE Paper

No. 2000-01-1967. Warrendale, PA: SAE, 2000

12. Scholl, K.W. and S.C. Sorenson. SAE Paper No. 930934. Warrendale, PA: SAE, 1993.

13. Graboski, M.S., J.D. Ross, and R.L. McCormick. SAE Paper No. 961166. Warrendale, PA: SAE, 1996.

14. Choi, C.Y., G.R. Bower and R.D. Reitz. Final Re-port for the National Biodiesel Board, 1997.

15. Yoshimoto, Y., M. Onodera and H. Tamaki. SAE Paper No. 1999-01-3598. Warrendale, PA: SAE, 1999.

16. McCormick, R.L., M.S. Graboski, T.L. Alleman, A.M. Herring and K.L. Tyson. Environmental Sci-ence & Technology 35(9): 1742-1747, 2001.

17. Grimaldi, C.N., L. Postrioti, M. Battistoni, and F. Millo. SAE Paper No. 2002-01-6085. Warrendale, PA: SAE, 2002.

18. Tat, M.E. and J.H. Van Gerpen. ASAE paper No. 036034, St. Joseph, MI: ASAE, 2003.

19. Yuan, W., A.C. Hansen, and Q. Zhang. 2005. Paper to be presented at ASAE 2005 Annual International Meeting, July 17-20, 2005, Tampa, Florida.

20. Dec, J.E., Espey, C. SAE paper 922307, (1992) 21. Honig, R., Kappler, G., Andresen, P., Brehm, N.,

Combustion Science and Technology, 102(1-6), 255-272, (1994)

22. Zhao, F.Q., Lai, M.C., Amer, A.A., Dressler, J.L., Atomization and Sprays, 6(4), 461-483, (1996)

23. Becker, J., Hassa, C., Atomization and Sprays, 12(1-3), 49-67, (2002)

24. Siegel, R., Howell, J.R., Thermal radiation heat transfer, 3rd edition, p. 581, Taylor & Francis Inc., Washington, DC, (1992)

25. Miles, P. C., SAE Paper 2000-01-1829, (2000) 26. Arcoumanis, C., Cho, S. T., Gavaises, M. and Yi, H.

S., SAE Paper 2000-01-1183, (2000) 27. Winterbourne, D. E., Yates, D. A., Clough, E., Rao,

K. K., Gomes, P. and Sun, J-H., Proc. Instn. Mech. Engrs., Pt. C, Vol. 208, pp. 223-240, (1994)

28. Shiozaki, T., Nakajima, H., Yokota, H. and Miya-shita, A., SAE Paper 980141, (1998)

29. Wang, T-C., Han, J-S., Xie, X., Lai, M-C. and Hei-nen, N. A., SAE Paper 1999-01-3496, (1999)

30. Bakenhus, M. and Reitz, R. D., SAE Paper 1999-01-1112, (1999)

31. Zambare, V. V. and Winterbourne, D. E., SAE Pa-per 1999-01-1501, (1999)

32. Mathews, W. S., Coverdill, R. E., Lee, C. F., and White, R. A., SAE Paper 2002-01-0266, (2002)

33. Fang, T., Coverdill, R. E., Lee, C. F., White, R. A., SAE Paper 2005-01-0919, (2005)