Effect of Ambient Temperature and Density on Diesel-Spray-Generated Shock Waves

Transcription

1 ILASS Americas, 21 st Annual Conference on Liquid Atomization and Spray Systems, Orlando, Florida, May Effect of Ambient Temperature and Density on Diesel-Spray-Generated Shock Waves Sanghoon Kook * and Lyle M. Pickett Sandia National Laboratories 7011 East Ave., PO Box 969, MS9053, Livermore CA94551, USA Abstract Shock wave generation by high-pressure diesel sprays has been investigated over a range of ambient temperature and density in a high-temperature, high-pressure vessel. In the past, shock wave generation was considered unlikely for diesel sprays because of the high-temperature environment and low fuel injection pressure. However, recent trends towards very high injection pressures, and earlier (lower temperature) injection in modern diesel engines, can potentially produce shock waves. Through high-speed schlieren imaging, injection-resolved shock waves have been captured at injection pressures of 150 MPa. Marked changes in shock wave generation are shown to occur with variation in ambient conditions that are experienced in an engine, including: (i) Shock waves diminish in strength with increasing temperature and eventually disappear due to the increasing speed of sound. (ii) At low ambient temperature, high-density ambient gas quickly decelerates the tip of the spray, causing shock waves to detach from the spray. (iii) At high ambient temperature, the high density decelerates the spray tip during injection-rate ramp-up before the spray reaches the speed-of-sound, such that no shock waves are produced. * Corresponding author

2 Introduction Modern diesel engines feature fuel injectors that operate with fuel pressures near 200 MPa or higher during engine operation. Due to the high injection pressure, the ideal maximum velocity of a diesel spray can be supersonic at normal operating conditions. A supersonic velocity has the potential to generate a shock wave, a type of propagating disturbance that travels through a medium and carries energy [ 1]. If generated, shock waves may affect the penetration and break up of diesel sprays. Shock waves have been observed in previous studies using similar fuel injection pressure conditions as a diesel engine, but in chambers that were at room temperature [ 2-5]. Different types of shock waves were observed depending on the diesel spray Mach number (M). Using an injection pressure of 150 MPa and a vessel filled with nitrogen gas (330 m/s speed of sound at room temperature) to a pressure of 2 MPa, Nakahira et al. [ 2] observe bow shock waves detached from the penetrating diesel spray. By contrast, attached oblique shock waves are produced when using an injection pressure of 135 MPa and an ambient gas of sulfur hexafluoride at room temperature and pressure (136 m/s speed of sound, M = 2.5) [ 3-4]. Much faster diesel sprays (2000 m/s, M > 6) also generate oblique shock waves with atmospheric back pressure [ 5]. However, none of these studies address shock wave formation at the high ambient temperature and density conditions which are more relevant to an operating engine. By the ideal gas assumption, the speed of sound correlates with the ambient temperature such as: c i = p γ R T γ = (1) ρ M where c is the speed of sound in an ideal gas, i γ is the ratio of specific heats, p is the ambient pressure, R is the universal gas constant, T is the ambient temperature, and M is the molecular weight of the ambient gas. a Figure 1 shows the computed speed of sound during the compression stroke of an engine. Using Eq. 1, the charge gas speed of sound is calculated at a given crank angle before top-dead center (TDC). The temperature and density corresponding to the crank angle are given at the right and top of the figure. Assuming the injected diesel spray penetrates at the Bernoulli velocity, typical diesel injection pressures would be high enough to generate supersonic sprays. Note that at an injection pressure of 150 MPa the spray velocity exceeds the speed of sound even at TDC (0 crank angle). In addition, modern diesel engines have a a Figure 1. Computed speed of sound during the compression stroke of the engine. The temperature and density corresponding to the crank angle are calculated at bottom-dead-center temperature of 348 K and compression ratio of 16. trend toward earlier injection timing, at lower ambient temperature and density, to enhance fuel-air mixing. Based on the assumption that sprays would penetrate at the Bernoulli velocity, shock waves generation would therefore be likely in a diesel engine. Figure 1 shows that ambient density and temperature increase during the piston compression. As Eq. 1 shows, however, ambient density has no effect on the speed of sound at fixed ambient temperature. Accordingly, one might think that shock wave generation would not be affected by the ambient density. Although the speed of sound is not changed, the spray velocity can be affected by increasing ambient density, if the density causes deceleration of the spray. Therefore, it is important to consider how ambient density effects diesel spray penetration. Our objective is to determine how shock wave generation from diesel sprays is affected by changes in ambient temperature and density that are typically experienced in practical engines. To this end, ambient temperature and density were varied in an optically accessible spray vessel. Experiments Figure 2 shows the high-pressure, high-temperature vessel and the optical setup used in this study. The vessel has a cubical-shaped combustion chamber of 108 mm on each side. Each side of the combustion chamber has a round optical or metal port. More details about the vessel geometry may be found in [ 6]. High-temperature, high-pressure conditions typical of a diesel engine are simulated in the vessel by spark ignition and combustion of a premixed combustible gas mixture. Once the products of this combustion event

3 Parabolic Mirror 115 mm diameter, f/8 1 mm Aperture f/1.2 Nikkor Condensing Lens Xe Arc Lamp Combustion Vessel Mixing fan High-Speed CMOS Camera Nikkor Lens f=50 mm Knife Edge Parabolic Mirror 115 mm diameter, f/8 Diesel Spray (front view) Figure 2. Constant-volume combustion vessel and high-speed schlieren imaging setup. cool to a pre-selected pressure and temperature state, the diesel fuel is injected [ 6]. A mixing fan, shown in Fig. 2, is used to produce a core region within the vessel of near-uniform temperature during the cool down. A second-generation Bosch common-rail fuel injector with an axial-hole nozzle is installed on a sideport of the vessel such that a single diesel spray is directed into the center of the chamber. The injection pressure is set to 150 MPa, which Fig. 1 shows can potentially generate shock waves even at high ambient temperature and density conditions. Ambient and injection conditions are listed in Table 1. High-speed schlieren imaging was performed to image shock waves formed during a single injection as shown in Fig. 2. Schlieren technique is widely used to visualize in-homogeneities in refractive index and large discrete changes in density that occur across the shock wave [ 1-3, 5]. A 150 W mercury-xenon arc lamp with a small (1.5 mm) arc distance system was used as a white light source. Light from the arc was collected with a 50- mm focal length f/1.2 Nikkor camera lens, offering superior broad wavelength performance compared to a simple achromatic condensing lens. The light was passed through a 1-mm aperture to form a high-quality point source that was collimated by a 115-mm diameter f/8 mirror. After passing through the combustion vessel, the collimated beam was refocused using the other f/8 mirror. The converging light reached a focal point 100 mm before entering a high-speed CMOS camera (Phantom v7.1) equipped with a 50-mm focal length Nikkor lens. A knife edge was placed at the focal point to form a schlieren system which is sensitive to the Table 1. Experimental conditions Injector conditions Type Bosch common-rail Nozzle Single-hole, KS1.5/0.86 Nozzle diameter 90 µm Injection pressure 150 MPa Ambient conditions Oxygen 0 or 15 % Temperature 455 ~ 950 K Density 0.81 ~ 22.8 kg/m 3 Fuel conditions Type No. 2 Diesel Density at 288 K 843 kg/m 3 Temperature 373 K first-derivative of refractive-index gradients. Images were collected at 75,471 frames per second, 2-µs exposure time, and with 256 by 64 pixel resolution. If needed, framing rate was set to higher 95,238 frames per second, sacrificing resolution down to 256 by 32 pixels. To improve the clarity of the shock waves in the schlieren imaging, different background correction methods were implemented depending on the uniformity of schlieren effects in the ambient gases. Figure 3 shows an example of shock wave images before and after the image post-processing. In Method A, background subtraction was performed using an average of images prior to the injection event. In Method B, the background correction was made using the previous image frame. As shown, method A worked

4 Method A Method B Figure 3. Images post-processing methods. From the top, the raw image {Img(k)}, background image {Img(0) or Img(k-1)}, and background-subtracted image are shown. well for raw images with a relatively uniform background while method B was needed to distinguish shock waves from non-uniform backgrounds created by refractive index gradients at high ambient temperature and density conditions. Results and Discussions The first part of this section shows a temporal development of shock waves at low ambient temperature and density to give a basis for the later discussion. Next, shock waves at high ambient temperature are discussed. We then show results with various ambient densities at both low ambient temperature and high ambient temperature. In the final section, shock waves generation at engine conditions are discussed based upon these results. Temporal Development of Shock Waves Figure 4 shows the diesel spray penetration and generation of shock waves at a fairly low ambient temperature and density, conditions expected for earlyinjection timing far before TDC in an engine. Applying image correction Method A, the schlieren images Figure 4. Schlieren imaging time sequence after start of injection (ASI). The spray penetration (black region) and shock waves generation are shown. Ambient conditions: N 2 gas, 455 K, 2.55 kg/m 3 exhibit a dark spray region over a gray background. The images shown are a high-speed sequence of a single injection. The first indications of shock wave generation by the spray occur at 40 µs after start-of-injection (ASI), as a compression appears as a bright spot at the leading edge of the spray. By 80 µs, a strong shock wave develops. The shock wave remains attached to the leading edge of the spray. At 119 µs ASI, a bow shock wave is found that is detached from the tip of the spray. The shock wave continues to separate from the spray tip with increasing time ASI, indicating that the spray tip has decelerated below the speed of sound. Behind the leading bow shock wave, there are also new shock waves appear to be generated upstream of the spray tip in faster-moving fluid closer to the injector. Conceivably, many expansion or rarefaction waves follow the shock waves but those are not easily differentiated in the image. Nakahira et al. [ 2] proposed that these upstream shock waves originate from the injector nozzle. On the other hand, Pianthong et al. [ 5] reported that this upstream shock wave is observed alongside the spray body. The latter observation appears to be more consistent with our results. We show more evidence of these side-spray-generated shock waves at high ambient density conditions later in Fig. 10. Figure 5 shows the axial penetration distance of spray tip determined from the images in Fig. 4. Also

5 shown is the shock wave penetration after it detaches from the spray tip. The shock wave detaches at 105 µs ASI and then travels at m/s, close to the sonic speed for ambient gases at 455 K (432 m/s). An interesting finding from Fig. 5 is that the penetration data points before 40 µs ASI show a lower slope than those that immediately follow. This hesitation is caused by injector-opening effects, where the needle valve lift causes a ramping rate of injection velocity [ 11]. Figure 6 shows no evidence of shock Figure 5. Axial penetration distance of spray tip and shock wave determined from Fig. 4. Figure 6. Spray penetration time sequence after start of injection during the start-up period (injection-rate ramp-up) wave generation during this start-up time period. The first shock wave at the spray tip is measured only after the end of this start-up time. Indeed, the calculated spray tip velocity is subsonic during this start-up time, as shown in Fig. 7. The spray velocity then accelerates when the needle valve of the fuel injector is fully lifted. Note that the shock wave appears as a bow shock wave during this time period (Fig. 4) because the spray tip exceeds the speed of sound. Afterwards, the spray tip velocity declines as the spray tip is decelerated by the ambient gas. Figure 8 shows images captured well after the time of shock wave detachment from the spray tip. The leading shock wave has now reflected against the back wall of the vessel and is traveling opposite to the spray direction. At 305 µs ASI, the leading shock wave crosses the head of the spray. Afterwards, the reflected shock waves are superimposed on newly generated shock waves at the edge of the spray. As the shock wave exists around the spray or even acts as a reflected energy wave, there is potential for the shock wave to affect the liquid breakup or mixing processes. For example, a shock wave creates additional drag force on the leading edge of objects [ 7-8] or sprays [ 5]. However, the reflected shock waves noted in this study did not have a conclusive effect on the penetrating spray. Although shock waves are expected to affect the spray at some level of detail, the effect of the shock wave on spray penetration was less significant than typical spray penetration fluctuations before interaction with the reflected shock wave. Effect of Ambient Temperature Because the speed of sound increases with increasing ambient temperature, the possibility of shock wave generation decreases if other spray and ambient parameters remain constant. Figure 9 shows schlieren Figure 7. Calculated spray tip velocity and Mach number from Fig. 5. Also shown is an ideal gas speed of sound computed using Eq. 1. Figure 8. Schlieren imaging time sequence after start of injection (ASI) well after the shock wave detachment time. Superposition of reflected and newly generated shock waves are seen.

6 images for various ambient temperatures at a fixed, and relatively low, ambient density of 3.56 kg/m 3. The fuel injection pressure was held constant at 150 MPa. For experiments at higher ambient temperature and density, unlike the low-temperature results given in Fig. 4, the ambient gas consisted of 75% N 2, 15 % O 2, 6% CO 2, and 4% H 2 O. Similar ambient gas constituents are generated when using exhaust-gas recirculation in advanced combustion regimes of a diesel engine [ 9]. Schlieren images processed by Method B are shown at early (42 µs ASI) and later (137 µs ASI) spray development times in Fig. 9. Figure 9 shows that the type of compression wave generated by the spray tip changes with increasing temperature. The images at 137 µs ASI show detached shock waves at 600 K and 700 K. The shock waves are characterized by sharp refractive index gradients at the leading edge of the shock wave and spray-bodygenerated, upstream shock waves similar to those in Fig. 4. These features are especially clear when viewing the schlieren image time sequence, rather than individual images. By contrast, at temperatures of 800 K or higher, only very weak waveforms are generated, as indicated by arrows in Fig. 9. These waves are not easily discernable in individual schlieren images, but the image time sequence shows that they do exist and that they propagate at the speed of sound. For these higher temperature conditions, spray-generated waves are not detectable, unlike the 600 K and 700 K conditions. Furthermore, the appearance of the waveform in the schlieren images is similar to that generated during autoignition heat release in a reacting spray in our facility. The pressure wave caused by heat release in a reacting diesel spray would not be characterized as a shock wave. From the observations above, we therefore conclude that the weak waveform at temperatures of 800 K or higher is an acoustic pressure wave, which is generated even if the spray is not supersonic. Further support for this conclusion will be given for hightemperature, high-density conditions discussed later. The initial development of the shock or acoustic waves is shown in the schlieren images at 42 µs ASI. A refractive index (density) gradient characteristic of compression at the spray tip is visible for all ambient temperatures shown. The waves quickly detach from spray as the spray decelerates and then travel at the speed of sound across the chamber. The waves therefore travel faster at higher ambient temperature, as seen by the location of the leading wave at 137 µs ASI. This difference was more significant than the initial spray penetrations fluctuation shown at 42 µs ASI. Effect of Ambient Density Ambient density affects shock wave generation in diesel sprays because ambient density affects the spray deceleration. As shown in Figs. 4-5, the spray quickly interacts with the ambient gas after the start of injection, causing spray deceleration to velocities lower than the speed of sound. As a result, the shock wave formation transitions from attached, oblique shock waves to detached, bow shock waves. This section addresses how ambient density affects shock wave generation at both low-temperature, and high-temperature conditions representative of an engine. Figure 9. Schlieren imaging for various ambient temperatures at fixed density of 3.56 kg/m 3. Ambient is 15 % O 2 gas. Times shown at the image lower-left are the time after start of injection. The locations of the leading shock or acoustic wave are annotated by arrows.

7 Figure 10. Schlieren imaging for three different ambient densities at low ambient temperature of 455 K. Ambient is N 2 gas. Times shown at the image lower-left are the time after start of injection. Figure 11. Calculated spray tip velocity and Mach number from Fig. 10. The speed of sound is computed using Eq. 1. Figure 10 shows a time sequence of the diesel spray and shock waves for three different ambient densities at a fixed ambient temperature of 455 K. At this relatively low ambient temperature, shock waves are generated at all conditions, but the type of shock wave, and the spray penetration rate, changes with ambient density. At low ambient density of 0.81 kg/m 3, an attached oblique shock wave is seen at the spray tip for times less than 119 µs ASI. At 2.55 kg/m 3, the shock wave is attached at 80 µs ASI, but it detaches from the spray tip to form a bow shock wave by 119 µs ASI. At 11.7 kg/m 3, the shock wave is already detached by 80 µs ASI. These results show that an increase in ambient density tends to slow the spray penetration rate, thereby affecting the type of shock wave generated. The computed spray tip velocities, shown in Fig. 11, confirm that the spray tip velocity decreases more quickly after the start of injection with increasing ambient density. The spray tip velocities are also affected by ambient density during the subsonic start-up period as well as the maximum velocity during injection. The start-up velocities, as well as the maximum velocity, tend to decrease with increasing ambient density. One thing notable from Fig. 11 is that the maximum spray tip velocity is about 10% lower than the speed of sound at the highest density of 11.7 kg/m 3. Despite calculated spray velocities that fall below the speed of sound, shock waves are clearly generated (Fig. 10). We believe this apparent contradiction is the result of lack of resolution in the high-speed imaging technique to measure local velocity differences between the spray and ambient gases. For example, the schlieren imaging technique provides the spray tip penetration relative to the start of injection, but the imaging method does not measure instantaneous velocities of the spray. The local velocity difference between the spray tip and the ambient may indeed be supersonic during stages of injection, but fluctuation in the tip of the penetrating spray would cause lower mean penetration rates. The time resolution of the high-speed imaging may also not be adequate to resolve stages of injection that are temporarily supersonic. Figure 11 shows that the increase in spray tip velocity during injection-rate ramp-up is met by rapid deceleration at 11.7 kg/m 3 after only 40 µs ASI, resulting in few measurement points at

8 the location of maximum velocity. The actual spray velocity may therefore be temporally unresolved for this condition. In addition to the local supersonic spray velocity, another possibility that can account for the shock wave generation alongside the subsonic spray body is a transonic effect. It is well known that a normal shock wave is easily generated some portion of the airflow over the wing although the airplane velocity is subsonic [ 12]. However, the shock waves seen at the spray tip (Fig. 10) can not be explained with this transonic flow pattern. The effect of ambient density on spray deceleration continues with increasing ambient density, but the impact on shock wave generation is more profound when the temperature is also increased, as would be expected in an engine. Figure 12 shows a schlieren imaging time sequence of spray, shock waves, and acoustic waves over a wide range of ambient densities at a fixed, elevated temperature of 700 K. This higher temperature condition has a higher speed of sound (524 m/s) than the conditions from Fig. 10. Due to the high ambient temperature and density, the schlieren background was non-uniform and background subtraction Method B was used for post-processing. The locations of the leading shock or acoustic wave are annotated by arrows in Fig. 12 at 126 µs ASI. Consistent with the discussion for Fig. 9, shock wave generation is clearly shown for 3.56, 5.22, and 7.27 kg/m 3 densities. That is, a relatively strong shock wave generated from the head of the spray is followed by further generation of shock waves upstream in the spray. However, only weaker acoustic pressure waves are visible at densities of 10.5 kg/m 3 or higher, similar to the high-temperature conditions of Fig. 9 at 3.56 kg/m 3. Yet unlike the ambient temperature variation of Fig. 9, the position of the shock or acoustic waves follows a non-monotonic trend with increasing ambient density. As shown by the arrows in Fig. 12, the wave position decreases, and then increases with increasing ambient density. Since ambient temperature is constant, the speed of the shock or acoustic wave does not vary with ambient density. However, the axial position and timing of the wave generation does depend on ambient density. Figure 12 shows that the spray penetrates farther downstream at the time of shock wave generation for low ambient density conditions. For example, when a shock wave is first noticeable at 42 µs ASI, the spray penetration distance is approximately 15 mm for 3.56 kg/m 3 density, compared to 10 mm for 7.27 kg/m 3. This approximate 5-mm difference persists as the shock wave penetrates downstream at 95 µs and 126 µs ASI. To explain the increase in the acoustic wave axial position with increasing ambient density, we need to consider the analyzed spray tip velocities corresponding to the images in Fig. 12, as shown in Fig. 13. Like Fig. 11, the time ASI of maximum spray velocity, as well as the maximum velocity itself, tends to decrease with increasing density. It is reasonable to assume that the time of the acoustic wave generation corresponds to the time of maximum spray velocity (also close to the Figure 12. Schlieren imaging for various ambient densities at high ambient temperature of 700 K. Ambient is 15 % O 2 gas. Times shown at the image lower-left are the time after start of injection. The locations of the leading shock or acoustic wave are annotated by arrows.

9 Figure 13. Calculated spray tip velocity and Mach number from Fig. 12. The speed of sound is computed using Eq. 1. Peak Spray Head Velocity [m/s] measured modeled 20-µs ramp 1.31*exp fit 800 K sound speed 700 K sound speed Ambient Density [kg/m3] Figure 14. Measured peak spray head velocity (average of the top three velocities for each ambient density in Fig. 13) for various ambient densities. Also shown are estimated peak spray head velocity using the jet entrainment model [ 10] and predicted local spray velocity with weighting factor of Dashed lines are ideal gas speed of sounds computed using Eq. 1 for 700 K and 800 K. time of maximum deceleration of the spray). Once formed, the acoustic waves would travel at sonic speeds away from the slower-moving spray. Therefore, acoustic waves are expected to form more quickly at high ambient density, and propagate faster than the spray tip. The net result is that the axial position of the acoustic wave increases with increasing ambient density. For further analysis of the spray tip velocities for the conditions of Fig. 12, we averaged the top three velocities for each ambient density condition (to avoid experimental noise) and plotted these values against ambient density in Fig. 14. Similar to Fig. 11, the results from Fig. 14 (and Fig. 13) show that the jet tip velocities are lower than the speed of sound for the entire ambient density range. As discussed above, the observation of shock wave generation when the calculated spray tip velocity is lower than the speed of sound can be explained by differences between the local spray velocity difference and that derived from the average spray penetration. Figure 14 also shows model predictions of the maximum spray velocity from a simple transient jet entrainment model [ 10] using the experimental injection-rate ramp-up at the start of injection of 20 µs. The model was exercised in this study to explore how various injection-rate ramp-ups can affect the maximum spray velocity, especially for more realistic multi-hole injectors, as discussed in the next section. Further details of this model will be forthcoming in a future publication. The modeling results show generally good agreement with the measured spray tip velocity, indicating that a deceleration of the spray with increasing ambient density is expected and wellfounded. Still, the maximum jet head velocity was predicted subsonic. To consider differences between the actual local spray velocity and the measured jet tip penetration velocity, we propose a weighting factor be applied to the experimental jet tip velocities of Fig. 14 to represent the maximum local velocity of the spray. Recall that lack of adequate time resolution, exacerbated by using an average of three points near the maximum velocity in Fig. 14, as well as natural fluctuation in the jet tip, would tend to underpredict the actual local maximum velocity. The measurement of shock waves at given conditions, and their corresponding sonic speeds, are used here to determine the appropriate weighting factor. For example, it was clear that the shock waves were generated at the ambient density of 7.27 kg/m 3 and an ambient temperature of 700 K. A weighting factor of 1.31 to the maximum jet tip penetration velocity is required to reach supersonic jet velocities at this condition. The resulting maximum local spray velocity is shown in Fig. 14. In addition, Fig. 9 showed that shock waves were not generated at an ambient temperature of 800 K or higher using an ambient density of 3.56 kg/m 3. Therefore, we set the sonic speed at 800 K (556 m/s) as the threshold velocity at low density. This velocity is only 10% lower than the Bernoulli injection velocity. Shock Wave Generation in a Diesel Engine In Fig. 1, the potential for shock wave generation at diesel engine conditions was discussed based on the speed of sound at engine temperatures and the Bernoulli injection velocity. In this section, we summarize the

10 results of this study by presenting a diagram for shock wave generation in a diesel engine in a similar format. Figure 15 shows the maximum spray velocity during injection as a function of ambient temperature and density, compared to the speed of sound at a given crank angle of compression in diesel engine. For reference, the Bernoulli injection velocity at the experimental injection pressure of this study (150 MPa) is shown. The measured spray tip (head) maximum velocity, and the expected maximum spray local velocity from the 1.31 weighting in Fig. 14 are plotted again in Fig. 15. The region of shock wave generation is indicated in the figure where the spray velocity exceeds the speed of sound. An important finding of the study was that different types of shock waves are generated depending on the spray velocity. We often observed shock waves even though the maximum spray head velocity was lower than the speed of sound. Therefore, the maximum local spray velocity must be considered. At early crank angles, where both the maximum spray head velocity and local spray velocity are higher than the speed of sound, an oblique shock wave is observed at the tip of the spray. On the other hand, a bow shock wave that detaches from the spray tip is generated at crank angles where only the maximum local velocity of the spray is faster than the speed of sound. The limit for the local spray velocity was found at a sonic speed of 800-K charge gas. This velocity is only 10% lower than the Bernoulli velocity and the decreased jet velocity can easily be explained by pressure losses through the nozzle, as well as jet expansion (spreading) outside of the nozzle. At near top-dead-center conditions, only weak acoustic waves are observed at jet velocities less than the speed of sound. In a real engine, however, the ramp-up profile of the fuel injection-rate is smoother (or slower) than the injector used in this study. This is because the conventional injector has a nozzle with multiple holes, typically 5 to 8 holes compared to the single hole of the current injector. To project how such a change in rate of injection would affect shock wave generation, the modeled maximum local spray velocity with a longer (100 µs) ramp-up time is shown. The slower injectionrate profile is shown to significantly decrease the predicted maximum spray velocity. The local velocities fall well below the speed of sound limit. Accordingly, conventional diesel engines would have a limited chance of shock wave generation, particularly if injection occurs at timings near TDC. Conclusion High-speed schlieren imaging was performed to investigate the potential for shock waves generation at diesel engine conditions. Diesel sprays were injected into high-temperature, high-pressure environments in an optically accessible chamber. Results show that increasing ambient temperature and density both inhibit shock wave generation. High ambient temperature increases the speed of sound and therefore hinders potential shock wave generation. High ambient density decelerates the spray during rate-ofinjection ramp-up such that the spray may never reach sonic velocities, particularly if the ambient temperature Velocity [m/s] Charge Density [kg/m 3 ] Bernoulli Injection Velocity at: 150 MPa Nozzle Losses Oblique Shock Waves Bow Shock Waves Max. Local Jet Velocity Charge Gas Speed of Sound Acoustic Waves Max. Local Jet Velocity Max. Jet HEAD Velocity 20-µs Ramp µs Ramp Crank Angle of Injection [ ] Figure 15. Shock waves generation diagram in a diesel engine. Engine conditions: 348-K BDC temperature, 16 compression ratio, and 150-MPa injection pressure Charge Temperature [K]

11 is also high. For these reasons, spray-generated shock waves are not expected at injection timings typical of a diesel engine. Acknowledgement Support for this research was provided by the U.S. Department of Energy, Office of Vehicle Technologies. The research was performed at the Combustion Research Facility, Sandia National Laboratories, Livermore, California. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy s National Nuclear Security Administration under contract DE-AC04-94AL The authors thank Mark Musculus for providing transient jet entrainment model. References 1. Settles, G.S., American Scientist 94:22-31 (2006). 2. Nakahira, T., Komori, M., Nishida, M., and Tsujimura, K., SAE Tech. Pap. Ser (1992). 3. MacPhee, A.G., Tate, M.W., Powell, C.F., Yue, Y., Renzi, M.J., Ercan, A., Narayanan, S., Fontes, E., Walther, J., Schaller, J., Gruner, S.M., Wang, J., Science 295: (2002). 4. Im, K-S, Lai, M-C, and Wang, J., SAE Tech. Pap. Ser (2004). 5. Pianthong K., Zakrzewski, S., Milton, B.E., and Behnia, M., Seventeenth Annual Conference on Liquid Atomization & Spray Systems, Zurich, Switzerland, September 2001, pp Engine Combustion Network Experimental Data archive, 7. Rotman, D., Phys. Fluids A 3(7): (1991). 8. Hermanson, J.C., and Ceregen, B.M., Proc. Combust. Inst. 27: (1998). 9. Kook, S., Bae, C., Miles, P.C., Choi, D., and Pickett, L.M., SAE Tech. Pap. Ser (2005). 10. Kook, S., Pickett, L.M., Musculus, M.P.B., and Gehmlich, R.K., The 7 th International Symposium on diagnostics and Modeling of Combustion in Internal Combustion Engines (COMODIA 2008), Hokkaido, Japan, July 2008 (in presse). 11. Powell, C.F., Yue, Y., Poola, R., Wang, J., and Lai, M.-C., SAE Tech. Pap. Ser (2001). 12. U.S. Department of Transportation Federal Aviation Administration, Airplane Flying Handbook, U.S. Government Printing Office, 2004, p