Bio-diesel effects on combustion processes in an HSDI diesel engine using advanced injection strategies

Bio-diesel effects on combustion processes in an HSDI diesel engine using advanced injection strategies

Available online at Proceedings of the Proceedings of the Combustion Institute 32 (2009) 2785–2792 Combustion Institute www.e...

897KB Sizes 6 Downloads 57 Views

Available online at

Proceedings of the

Proceedings of the Combustion Institute 32 (2009) 2785–2792

Combustion Institute

Bio-diesel effects on combustion processes in an HSDI diesel engine using advanced injection strategies Tiegang Fang a,*, Chia-fon F. Lee b a

Department of Mechanical and Aerospace Engineering, North Carolina State University, 3182 Broughton Hall, 2601 Stinson Drive-Campus Box 7910, Raleigh, NC 27695, USA b 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 was used to investigate the combustion process using different fuels including European low sulfur diesel and bio-diesel fuels with advanced multiple injection strategies. Influences of injection timings and fuel types on combustion characteristics and emissions were studied under similar loads. In-cylinder pressure was measured and used for heat release analysis. High-speed combustion videos were captured for all the studied cases using the same frame rate. NOx emissions were measured in the exhaust pipe. Different combustion modes including conventional diesel combustion and low-temperature combustion were observed and confirmed from the heat release rates and the combustion images. Natural luminosity was found consistently lower for bio-diesel than the European low sulfur diesel fuel for all the cases. However, for NOx emissions, under conventional combustion cases such as cases 2 and 3, it was found that bio-diesel leads to increased NOx emissions. Under a certain injection strategy with retarded main injections like case 4 and 5, it is possible to have up to 34% lower NOx emissions for B100 than B0 for case 4 with low-temperature combustion mode. Simultaneous reduction of NOx and natural luminosity was achieved for advanced low-temperature combustion mode. It is hypothesized based on the results that the lower soot generation for bio-diesel fuel is believed due to a lower soot formation rate and a higher soot oxidation rate. The NOx increase problem for bio-diesel fuel can be amended by employing advanced injection strategies with low-temperature combustion modes. Ó 2009 The Combustion Institute. Published by Elsevier Inc. All rights reserved. Keywords: Diesel combustion; Advanced injection strategy; Natural luminosity; Bio-diesel; Low-temperature combustion

1. Introduction Due to limited fossil fuels in the world, finding alternative fuels and reducing the fuel consumption of current internal combustion engines become more and more important. Meanwhile, *

Corresponding author. Fax: +1 919 515 7968. E-mail address: [email protected] (T. Fang).

worldwide environmental concerns make the emission regulations more stringent. Exhaust emissions like oxides of nitrogen (NOx) and particulate matter (PM) must be reduced for diesel engines to meet future emission standards. Multiple-injection strategies have been reported for simultaneous reduction of NOx and PM in large-bore DI diesel engines [1–3] and small-bore HSDI diesel engines [4–6]. Nehmer and Reitz [1] showed that pulsed injections

1540-7489/$ - see front matter Ó 2009 The Combustion Institute. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.proci.2008.07.031


T. Fang, C.F. Lee / Proceedings of the Combustion Institute 32 (2009) 2785–2792

might provide a method to reduce both PM and NOx. The effectiveness of double, triple, and rate shaped injection strategies to simultaneously reduce NOx and PM was evaluated [2]. Numerical simulations were conducted to explore the mechanism of soot and NOx reduction for multiple injection strategies [3]. Reduced soot emissions are due to the fact that the soot producing rich region is not replenished when the injection pulse is terminated and restarted. Zhang [4] investigated the effect of a pilot injection on NOx, soot emissions, and combustion noise in a small diesel engine. By optimizing the EGR rate, pilot timing and quantity, main timing, and dwell between the main and pilot injections, simultaneous reduction of NOx and PM was obtained in an HSDI diesel engine [5]. Another study on pilot injection was done by Tanaka et al. [6]. It was shown that simultaneous reduction of combustion noise and emissions was possible by minimizing the pilot fuel quantity and advancing the pilot injection timing. Homogeneous charge compression-ignition (HCCI) combustion has been shown to be effective for NOx and PM reduction. Early investigations on this combustion mode [7,8] showed that it was a new combustion mode different from traditional spark-ignition (SI) and compressionignition (CI) combustion. The studies in fourstroke SI-engines were done by Najt and Foster [9] and Thring [10]. HCCI combustion in diesel engines was reported much later than SIengines. Because of the flexibility of the multiple injection strategies in controlling the mixing and combustion processes, they were also employed in DI diesel engines for HCCI combustion. Hashizume et al. [11] proposed MULtiple stage DIesel Combustion (MULDIC) for higher load operating conditions with low soot emissions. A multi-pulse HCCI combustion study was done by Su et al. [12]. For very early injection, hydrocarbon (HC) increased. Hasegawa and Yanagihara [13] employed two injections in their UNIform BUlky combustion System (UNIBUS). The ignition of the premixed gas could be controlled by the second injection when the early injection maintained a low-temperature reaction. A dual mode operation was used in a narrow-angle direct-injection (NADITM) concept [14]. The engine was operated in HCCI combustion under partial loads and in conventional diesel combustion at full-load conditions. HCCI combustion in a small-bore HSDI diesel engine was investigated by using early multiple short injection pulses during the compression stroke [15]. Results showed dramatic NOx and smoke reductions, while HC and CO substantially increased. A two-stage premixed charge compression ignition (PCCI) combustion was studied by Kook and Bae [16]. A large fuel frac-

tion was injected very early before TDC. A second injection with a small amount of fuel was injected near the compression TDC to ignite all the air–fuel mixtures. Agricultural fat and oils, in raw or chemically modified forms, have the potential to supplant a significant proportion of petroleum-based fuels. Bio-diesel is of particular interest because it significantly reduces PM, HC and 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 [17]. In addition to its reduced emissions of PM, HC and CO, bio-diesel contributes less to global warming than fossil fuels due to its closed carbon cycle. Bio-diesel is also the only alternative fuel that has passed the EPA-required Tier I and Tier II Health Effects testing requirements 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. Bio-diesel has been criticized for its higher brake specific NOx emissions comparing to diesel fuels [18]. Many studies and measurements of NOx emissions from diesel engines fueled with bio-diesel have been published [19–26] and most of them were focused on conventional diesel combustion. Low-temperature combustion has its unique advantage to reduce NOx [27]. The lowtemperature combustion mode of bio-diesels has also been reported recently [28–35]. Advanced injection strategy was used to implement better emissions of bio-diesel for HSDI diesel engines [28,34]. Simultaneous reduction of NOx and soot emission was achieved for bio-diesel–diesel blend with high-efficiency clean combustion (HECC) in a diesel engine [29]. Zheng et al. [30] investigated low-temperature combustion mode for bio-diesel with high EGR rate. The results showed that it was possible to achieve NOx and soot reductions for low-temperature combustion modes with biodiesel. Bio-diesel HCCI combustion was investigated in a diesel engine by fueling in the intake port [31] and numerical simulation was used to study the detailed reaction in the engine cylinder. Both conventional and PCCI combustion modes were numerically investigated for bio-diesel to study the effects of fuel properties [32]. HCCI combustion was experimentally studied in a small HSDI diesel engine using early and late injection strategies [33,35]. Low-temperature combustion modes were implemented by employing high EGR rate and/or later injection timing. However, most of the above mentioned studies were in metal engines with no detailed information about the combustion flame inside the engine cylinder. Incylinder visualization of the combustion process is important in understanding the mixing, combustion, and the soot generation processes. There-

T. Fang, C.F. Lee / Proceedings of the Combustion Institute 32 (2009) 2785–2792

fore, the objective of this work is to investigate the effects of different fuels and injection strategies on combustion processes in an optical engine with realistic piston geometry using advanced injection strategy. The mechanism controlling soot and NOx emissions will be analyzed and discussed.

Side Side Window


Quartz Piston

2. Optical engine facility and visualization technique A complete description of the optical engine design can be found in a previous publication [36], so only a brief introduction of the setup will be provided here. Ford Motor Company provided a single-cylinder equivalent of the 1.2-l, 4-cylinder Ford Direct-Injection Aluminum Through-bolt Assembly (DIATA) engine for conversion into an optical engine. Typical engine specifications are listed in Table 1. Optical access to the combustion chamber is attained from the side window or from the fused silica piston. The fused silica piston mimics all the features of the metal piston. A schematic of the design for optical access is shown in Fig. 1 and the pictures of the optical piston and metal piston are illustrated in Fig. 2. The engine is equipped with a Bosch common-rail electronic injection system, capable of injection pressures up to 135 MPa. A Valve-Covered-Orifice (VCO) injector with six 0.124 mm holes was used. The injector body was fitted with a needle lift sensor monitoring the needle motion throughout injection. A Phantom V7.1 high-speed camera was used to capture the natural luminosity. An MEXA-720 NOx analyzer from Horiba was used to measure the diluted NOx concentration from the exhaust pipe. This non-sampling NOx analyzer provided a faster response with a NOx sensor. The sensor had a response time of about 0.7 s. The optical engine was operated in an injection pattern of 3 injection cycles followed by 10

Fig. 1. Assembly cross-section of optical engine design with drop-liner raised.

Fig. 2. Optical piston top (left) and metal piston (right).

motoring cycles for a certain time. The analyzer was calibrated according to the operation manual before emission measurement. The final NOx emission values were corrected based on the duty-cycle of the operation pattern. Natural luminosity was employed to visualize combustion processes. The natural luminosity comes from chemiluminescence and soot luminosity. In hydrocarbon flames, chemiluminescence is typically seen in the visible and near ultraviolet from the OH, CH, and C2 radicals [37]. Soot luminosity is thermal radiation from soot particles within the combustion flame with a broadband emission spectrum. The perceived intensity depends on the soot concentration and temperature [38–40].

Table 1 Specifications for the optical engine Bore Stroke Displacement/cylinder Compression ratio Swirl ratio Valves/cylinder Intake valve diameter Exhaust valve diameter Maximum valve lift Intake valve opening Intake valve closing Exhaust valve opening Exhaust valve closing

70 mm 78 mm 300 cc 19.5:1 2.5 4 24 mm 21 mm 7.30/7.67 mm (intake/exhaust) 13 CAD ATDC (at 1 mm valve lift) 20 CAD ABDC (at 1 mm valve lift) 33 CAD BBDC (at 1 mm valve lift) 18 CAD BTDC (at 1 mm valve lift)

3. Engine operating conditions Different combustion modes with multipleinjection strategies were studied for pure diesel and soybean bio-diesel fuels. The injection strategies included a small first injection with an early pre-TDC timing and a main injection at or after TDC. The first injection fuel quantity was fixed at 1.5 mm3 for both fuels. The injection timing was changed to achieve different combustion modes. The first injection timing changed from 40 CAD to 20 CAD ATDC by a step of 10 CAD. The main injection timings were chosen at TDC and 10 CAD ATDC. There are five conditions for each fuel. The operating conditions are tabulated in Table 2. The combustion videos were


T. Fang, C.F. Lee / Proceedings of the Combustion Institute 32 (2009) 2785–2792

Table 2 Engine operating conditions Fuel type

Case number

B0 B0 B0 B0 B0

1 2 3 4 5

B100 B100 B100 B100 B100

1 2 3 4 5

First injection fuel quantity (mm3)


Main injection fuel quantity (mm3)

IMEP (bar)

40 30 20 30 20

1.5 1.5 1.5 1.5 1.5

0 0 0 10 10

6.8 6.3 6.6 7.3 7.0

3.98 3.98 3.99 3.98 4.00

40 30 20 30 20

1.5 1.5 1.5 1.5 1.5

0 0 0 10 10

8.8 8.5 8.5 8.9 8.7

4.00 4.08 4.07 3.93 4.02

First injection SOI (CAD ATDC)

The pressure traces are shown in Fig. 3 grouped by the two fuels. For comparison, the motoring pressure is also plotted. The in-cylinder pressure shows quite different features for different injection timings. However, the fuel effects on the in-cylinder pressure are not so obvious. For different injection strategies, the in-cylinder pressures are quite similar and closely follow the motoring pressure before 20 CAD ATDC. Then with heat release from the first injection, the pressure starts to deviate from the motoring curves. Some pressure drop due to fuel evaporation can be seen for the cases with first injection timing at 20 CAD ATDC. The heat release rate from the first injection at 40 CAD ATDC is small leading to a small increase in the TDC pressure compared with the motoring pressure. The cases with an injection timing at 30 CAD ATDC have higher TDC pressure than those with first injec-

tion timing of 40 CAD ATDC due to more heat release from the first injection. This is due to the fact that the air–fuel mixture is too lean to combust completely for the first injection of case 1. But for cases 3 and 5 with a first injection timing at 20 CAD ATDC, a significantly higher heat release is seen than cases 1, 2, and 4. This results in higher ambient temperature and pressure near TDC for cases 3 and 5. Due to the difference in the heat release rate of the first injection, the ignition delay of the main injection is greatly different. At the same main injection timing, an early first injection leads to longer ignition delay and more rapid pressure rise. By delaying the main injection timing, the ignition delay is elongated for cases 4 and 5 compared with cases 2 and 3. A lower pressure rise rate is seen for the retarded main injection cases. No obvious difference is found for the two fuels. The ignition delay and peak pressures are quite similar for the same injection strategy. The computed heat release rate curves are shown in Fig. 4. The injection strategy greatly influences the heat release pattern and the combustion mode. The ambient temperature at the main injection timing is lower for case 1, which leads to a longer ignition delay for the main injection. As a consequence, a premixed combustion mode is observed. By retarding the first injection

Fig. 3. In-cylinder pressure.

Fig. 4. Heat release rate.

taken with the high-speed digital camera. For each case, five 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 when taking in-cylinder pressure and combustion videos. 4. In-cylinder pressure and heat release results

T. Fang, C.F. Lee / Proceedings of the Combustion Institute 32 (2009) 2785–2792

timing, more heat is released from the first injection and the ambient temperature becomes higher. The ignition delay becomes shorter for later first injection. Relatively high heat release rate peaks are seen for the first injections of cases 3 and 5. When the ignition delay becomes shorter, the combustion mode becomes more diffusion combustion dominated. This diffusion combustion mode is apparently observed for case 3 with a small portion of premixed combustion and a large portion of mixing controlled combustion. The heat release rate peak is much smaller than that of case 1. The combustion mode of case 2 is between case 1 and case 3. For case 5, there is more premixed combustion compared with case 3, but the heat release rate has a long tail indicating slow diffusion combustion after the premixed combustion phase. With a retarded main injection for case 4, the ignition delay is greatly elongated. The heat release rate shows a single peak premixed combustion mode. Again, the fuel effects on heat release rate curves are hardly seen. 5. Natural luminosity and NOx emissions The combustion process was visualized by using the high-speed camera with a three-dimensional imaging setup [40–42]. The frame rate is 12,000 frames/s with an exposure time of 2 ls at a resolution of 512  256. The Spatially Integrated Natural Luminosity (SINL) was obtained by summing up the pixel values of the bottomview combustion images. The SINL is shown in Fig. 5. Each trace is an average of five sets of data. The data are grouped by fuel type in two subplots. For a certain fuel, a large difference is observed for differing injection strategies. For B0, the highest natural luminosity is seen for case 3. Significantly lower natural luminosity is seen for case 4 with moderately early first injection and retarded main injection timings. But for case 5, although it has a retarded main injection, the natural luminosity is comparable to that of case 3. SINL for cases 1 and 2 are between

Fig. 5. Spatially integrated natural luminosity (SINL).


those of case 4 and case 5. High natural luminosity would imply high temperature and/or high soot concentration in the combustion flame. For B100, the natural luminosity is lower than B0. The influence of injection strategies on the natural luminosity is similar to B0. The main difference between B100 and B0 is the flame existence time duration. A shorter time is seen for B100 than B0 indicating a faster combustion process. The natural luminosity variation rate was calculated to find the natural luminosity increase rate or the decrease rate. The variation rate is obtained by calculating the time derivative of the natural luminosity with time. The decrease rate is the absolute values of the negative natural luminosity variation rate. A larger increase rate indicates faster early combustion. A larger decrease rate implies faster soot oxidation and/or combustion gas temperature decreasing process. The peak values of the natural luminosity (NL), natural luminosity increasing rate (NLIR), and natural luminosity decreasing rate (NLDR) are plotted in Fig. 6 to illustrate the influence of injection strategies and fuels. The peak NL increases with retarding the first injection timing and decreases with retarding the main injection timing. But the peaks are comparable for case 3 and case 5 and it is believed due to strong soot formation in both cases. For B100, the peak NL values are less than B0. The lower peak NLIR for B100 might be attributed to a lower soot formation rate for bio-diesel. The peak NLDR is higher for B100 for cases 1–3, which might imply a faster soot oxidation process for bio-diesel in the late combustion cycle if the in-cylinder gas temperature is close for B0 and B100. A close or slightly lower peak NLDR is seen for B100 than B0 for cases 4 and 5. This lower NLDR might be attributed to the lower soot concentration in B100 than B0 leading to a lower soot oxidation rate [43]. Based on the observations, it is hypothesized that bio-

Fig. 6. Peak values for natural luminosity (NL), NL increase rate (NLIR), and NL decrease rate (NLDR).


T. Fang, C.F. Lee / Proceedings of the Combustion Institute 32 (2009) 2785–2792

diesel might have a lower soot formation rate during the early combustion stage and a higher soot oxidation rate during the late cycle crank angles. These two factors could determine the lower soot emissions for bio-diesel in the exhaust compared to petro-diesel. Retarding the main injection timing could reduce soot formation. But it could also reduce soot oxidation rate due to a lower combustion temperature. The NOx emissions are shown in Fig. 7. Case 1 has similar NOx emissions for B0 and B100 with B100 slightly smaller than B0, but the difference is within the error bar. But for cases 2 and 3, NOx emissions are higher for B100 than B0, which is consistent with the previous results with conventional diesel combustion [17,19–26]. For cases 4 and 5, B100 has less NOx emissions than B0 for injection strategies with retarded main injections. In order to compare the natural luminosity of different injection strategies and fuels, a number, called natural luminosity parameter (NLP) thereafter, was defined as the ratio of mean natural luminosity to the total heat released. A small value showed lower temperature and/or lower sooting combustion. In addition, to compare the NOx emissions, a NOx parameter (NP) was defined as the NOx emissions divided by the released heat for the combustion process. The unit for the NOx parameter is ppm/J. A smaller NP value indicates lower NOx emissions. The NLP and NP data are plotted in the NP–NLP plane in Fig. 8. Case 4 offers the best performance with the data points closest to the origin. The performance of case 3 is most non-preferable with high NOx and natural luminosity. For cases 1, 2, and 5, there are some tradeoff between the natural luminosity and NOx emissions. It is not easy to achieve simultaneous reduction of natural luminosity and NOx emission for these conditions. But compared with case 3, the natural luminosity and NOx emissions are reduced simultaneously for cases 1, 2, and 5. The fuel effects are not as important as injection strategies. The fuel blends can be used to fine-tune the emission data points in the NP–

Fig. 8. Natural luminosity parameter (NLP) and NOx. Parameter (NP) in NP–NLP plane.

NLP plane to obtain better combustion performance. Based on the above discussion, it is expected that a multiple-injection strategy can greatly change the emission behavior for different fuels under similar load conditions. There is not a certain trend for the bio-diesel effects on the NOx emissions compared with petro-diesel. This is quite different from the single-injection strategy cases with conventional and late injection timings [17,19–26], where NOx emissions increase with increasing bio-diesel content. A multiple-injection strategy could change the NOx emission behavior. Under a certain injection strategy with retarded main injections, it is possible to have lower NOx emissions for bio-diesel. Bio-diesel generally reduces soot generation compared with petro-diesel. Simultaneous reduction of soot and NOx emissions is possible by using a multiple-injection strategy with a longer dwell time between the two injections. However, due to the oxygen content in B100, its energy density is lower than B0 and the fuel consumption increases when using bio-diesel as shown by the injected fuel quantities in Table 2. 6. Combustion flame images

Fig. 7. NOx emissions in ppm.

Combustion images of both fuels for case 3 are shown in Fig. 9 to illustrate the effect of bio-diesel on a typical diffusion flame combustion mode. The combustion is poor in terms of soot emissions with a significantly high soot concentration formed in the flame. Ignition delay is very short for the main injection of case 3, namely about 1.5 CAD after the liquid fuel comes out of the nozzle. Ignition occurs at 2.75 CA ATDC for both fuels. At the end of injection, namely at about 8.00 CAD ATDC, strong luminous flames fill in the squish region and the near wall region in the bowl. The jet structure corresponding to the six spray jets is obviously observed. The longer spray-flame

T. Fang, C.F. Lee / Proceedings of the Combustion Institute 32 (2009) 2785–2792



B0: -20, 0


B100: -20, 0

Case 1, -40, 0 20.00 ATDC





B0: -20, 0

B100: -20, 0








B100: - 30, 10


Fig. 9. Combustion images for case 3 of B0 and B100, and case 4 of B100. All times shown are CAD ATDC.

overlap greatly inhibits the air–fuel mixing and could increase the soot formation during the early combustion stage. The liquid penetration is quite short for both fuels. Comparing the combustion images at 14, 20, 26 and 38 CAD ATDC, it is apparent that B100 has a higher flame burnt-out rate than B0. Due to the high soot concentration in the conventional combustion flame, it takes longer to burn out during late cycle crank angles. Based on these results, the combustion process for the bio-diesel are not only determined by the mixing rate but also influenced by the oxygen availability in the combustion flame. When the mixing rate is limited, the role of the oxygen content becomes more evident. With retarded main injection for case 4, the combustion flame is different from case 3 as seen also in Fig. 9. Due to a lower ambient temperature, ignition occurs near the end of injection between 17.00 and 17.75 CAD ATDC. The difference of the early flame in case 4 from case 3 is that the early flame develops more slowly with a weak natural luminosity. This lower temperature flame elongates the air–fuel mixing time without soot formation. A weak and more uniform combustion flame is observed in case 4. The combustion is similar to HCCI combustion mode. The natural luminosity is significantly lower than case 3.

Case 2, -30, 0 23.00 ATDC

Case 3, -20, 0 29.00 ATDC

Case 4, -30, 10 32.00 ATDC

Case 5, -20, 10 38.00 ATDC

Fig. 10. Typical combustion images for cases 1–5 (from left to right) for both fuels. All times shown are CAD ATDC.

From the combustion images, it is found that besides fuel effects injection strategies greatly influence the combustion mode. Some typical flame images with high luminosity are shown in Fig. 10 for the ten cases. The images further confirm the observations in the SINL results. Lower natural luminosity is seen for bio-diesel than B0. Increasing the gap between the two injections reduces natural luminosity. Conventional combustion modes, namely cases 2 and 3, have higher natural luminosity than the lowtemperature combustion mode in cases 1, 4, and 5. There are two important factors controlling the combustion process. One is the mixing process, which dominantly determines the combustion mode transition from premixed combustion to diffusion flame combustion. The mixing process is greatly influenced by injection parameters. The other is the oxygen content in the fuel. Higher oxygen content could help reduce soot formation and increase soot oxidation. Therefore, soot emissions are expected to be lower for bio-diesel. 7. Conclusions In this paper, the effects of European low sulfur diesel and bio-diesel on the combustion process were experimentally investigated in a smallbore HSDI diesel engine using advanced multiple injection strategies. Five injection strategies were studied showing the injection timing influences on the combustion modes. Less luminous flame was observed for bio-diesel than the diesel fuel. Compared with conventional combustion mode, low-temperature premixed combustion modes resulted in lower natural luminosity. For conventional-like mode, namely cases 2 and 3, higher NOx emission was seen for B100 than B0. However, for cases 4 and 5 with retarded main injection, B100 led to lower NOx emissions. For a certain type of fuel, retarding injection timing resulted in significant reduction in NOx emissions. A multiple-injection strategy can change the emission behavior for different fuels under similar load


T. Fang, C.F. Lee / Proceedings of the Combustion Institute 32 (2009) 2785–2792

conditions. Fuel effects can be used to fine-tune the combustion performance. There is no certain trend obtained for the effects of bio-diesel on the NOx emissions. This is quite different from the single-injection strategy cases with conventional combustion, where NOx emissions increase for bio-diesel. Under a certain injection strategy with retarded main injections, it is possible to have up to 34% lower NOx emissions for B100 than B0 for case 4 with low-temperature combustion mode. The natural luminosity is always lower for B100. Simultaneous reduction of soot and NOx emissions is possible by using a multiple-injection strategy with a longer dwell time between the two injections. The lower soot generation for bio-diesel is hypothesized due to a lower soot formation rate and a higher soot oxidation rate. The NOx increase problem for bio-diesel can be amended by employing advanced injection strategies with low-temperature combustion modes.

Acknowledgements This work was supported in part by the Department of Energy Grant No. DE-FC2605NT42634, by Department of Energy GATE Centers of Excellence Grant No. DE-FG2605NT42622, 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] D.A. Nehmer, R.D. Reitz, SAE Paper 940668, 1994. [2] T.C. Tow, D.A. Pierpont, R.D. Reitz, SAE Paper 940897, 1994. [3] Z. Han, A. Uludogan, G.J. Hampson, R.D. Reitz, SAE Paper 960633, 1996. [4] L. Zhang, SAE Paper 1999-01-3493, 1999. [5] S.K. Chen, SAE Paper 2000-01-3084, 2000. [6] T. Tanaka, A. Ando, K. Ishizaka, JSAE Rev. 23 (2002) 297–302. [7] S. Onishi, S. Hong Jo, K. Shoda, P. Do Jo, S. Kato, SAE Paper 790501, 1979. [8] M. Noguchi, Y. Tanaka, T. Tanaka, Y. Takeuchi, SAE Paper 790840, 1979. [9] P.M. Najt, D.E. Foster, SAE Paper 830264, 1983. [10] R.H. Thring, SAE Paper 892068, 1989. [11] T. Hashizume, T. Miyamoto, H. Akagawa, K. Tsujimura, SAE Paper 980505, 1998.

[12] W. Su, T. Lin, Y. Pei, SAE Paper 2003-01-0741, 2003. [13] R. Hasegawa, H. Yanagihara, SAE Paper 2003-010745. [14] B. Walter, B. Gatellier, SAE Paper 2002-01-1744, 2002. [15] A. Helmantel, I. Denbratt, SAE Paper 2004-010935, 2004. [16] S. Kook, C. Bae, SAE Paper 2004-01-0938, 2004. [17] C.A. Sharp, S.A. Howell, J. Jobe, SAE Paper 200001-1967, 2000. [18] M. Lapuerta, O. Armas, J. Rodriguez-Fernandez, Prog. Energy Combust. Sci. 34 (2008) 198–223. [19] K.W. Scholl, S.C. Sorenson, SAE Paper 930934, 1993. [20] M.S. Graboski, J.D. Ross, R.L. McCormick, SAE Paper 961166, 1996. [21] C.Y. Choi, G.R. Bower, R.D. Reitz, Final Report for the National Biodiesel Board, 1997. [22] Y. Yoshimoto, M. Onodera, H. Tamaki, SAE Paper No. 1999-01-3598, 1999. [23] R.L. McCormick, M.S. Graboski, T.L. Alleman, A.M. Herring, K.L. Tyson, Environ. Sci. Technol. 35 (9) (2001) 1742–1747. [24] C.N. Grimaldi, L. Postrioti, M. Battistoni, F. Millo, SAE Paper 2002-01-6085, 2002. [25] M.E. Tat, J.H. Van Gerpen, ASAE Paper 036034, St. Joseph, MI, ASAE, 2003. [26] W. Yuan, A.C. Hansen, Q. Zhang, Proceedings of the ASAE 2005 Annual International Meeting, Tampa, Florida, July 17–20, 2005. [27] K.Akihama, Y. Takatori, K. Inagaki, S. Sasaki, A. Dean, SAE Paper No. 2001-01-0655, 2001. [28] L. Postrioti, M. Battistoni, C.N. Grimaldi, F. Millo, SAE Paper 2003-01-0768, 2003. [29] C.S. Sluder, R.M. Wagner, S.A. Lewis, J.M.E. Storey, SAE Paper No. 2006-01-0080, 2006. [30] M. Zheng, M.C. Mulenga, G.T. Reader, M. Wang, D. S-K. Ting, SAE 2006-01-3281, 2006. [31] J.P. Szybist, J. McFarlane, R.G. Bunting, SAE 2007-01-4010, 2007. [32] K. Chakravarthy, J. McFarlane, S. Daw, Y. Ra, R.D. Reitz, J. Griffin, SAE 2007-01-4030, 2007. [33] A. Weall, N. Collings, SAE 2007-01-2020, 2007. [34] Y. Zhang, A.L. Boehman, Energy Fuels 21 (2007) 2003–2012. [35] T. Fang, Y.C. Lin, T.M. Foong, C.F. Lee, SAE Paper 2008-01-1390, 2008. [36] W.S. Mathews, R.E. Coverdill, C.F. Lee, R.A. White, SAE Paper 2002-01-2666, 2002. [37] J.E. Dec, C. Espey, SAE Paper 982685, 1998. [38] P.C. Miles, SAE Paper 2000-01-1829, 2000. [39] H. Zhao, N. Ladommatos, Prog. Energy Combust. Sci. 24 (1998) 221–255. [40] T. Fang, R.E. Coverdill, C.F. Lee, R.A. White, SAE Paper 2005-01-0919, 2005. [41] T. Fang, R.E. Coverdill, C.F. Lee, R.A. White, SAE Paper 2006-01-0078, 2006. [42] T. Fang, R.E. Coverdill, C.F. Lee, R.A. White, SAE Paper 2007-01-0203, 2007. [43] J.B. Heywood, Internal Combustion Engine Fundamentals, McGraw-Hill, Inc., New York, 1988.