Microscopic Visualization of Liquid Column Break-up Process in Gasoline PFI Injector

Transcription

1 Microscopic Visualization of Liquid Column Break-up Process in Gasoline PFI Injector Nobuyuki KAWAHARA *1, Eiji TOMITA *1, Mamoru SUMIDA *2 1: Department of Mechanical Engineering, Okayama University, Okayama, JAPAN, 2: Mitsubishi Electric Corp., Himeji, JAPAN Abstract The purpose of this study is to observe the fuel break-up process very close to an injection hole of a practical PFI injector and to examine the effect of internal structure on characteristics of spray behavior. In order to investigate the atomization process at the nozzle exit, visualization of spray structure was performed using an ultra high-speed camera (max. camera speed 1Mfps) with a long-distance microscope. The visualized experiments were carried out using PFI injector in a closed chamber at the atmospheric pressure. Three main conclusions were obtained from this study. It has been shown that the liquid fuel column was injected. During the injection period, the spray indicates the quasi-steady state mode. Surface waves of liquid column can be recognized. Liquid column very close to the nozzle exit were broken up to liquid ligament structure. These surface waves cause the break-up of liquid column to liquid ligaments. Break-up process of liquid ligaments to droplets can be investigated using high-speed magnified images. Moreover, interactions between droplets, such as penetration or coalescence, can be visualized. 1. Introduction It is necessary for internal combustion engines to improve thermal efficiency and to reduce exhaust emissions because of saving energy resource, problems of exhausting nitrogen oxide, unburned hydrocarbons, carbon monoxide, particulate matters, and carbon dioxide. The fuel Injection systems for usual spark-ignition engines inject the fuel into the intake port. The advantages of gasoline Port Fuel Injection (PFI) system are increased power and torque, more uniform fuel injection, more rapid engine response, more precise control of the air-fuel ratio during cold-start and engine warm-up. Current subject of PFI injector is decrease of unburned hydrocarbon during cold-start. To solve this subject, atomization of PFI injector should be improved (Heywood 1988, Zimmerman et al. 1999, Liu et al. 2006). In order to understand the spray characteristics formed by multi-hole injector, investigations of gasoline injection sprays using several measurement techniques like laser sheet method with laserinduced (exciplex) fluorescence (LIF), phase Doppler anemometor (PDA) (Ahao et al. 1995), particle image velocimetry (PIV) have been carried out for better control of spray and combustion characteristics. TAB (Taylor Analogy Break-up) model (O Rourke and Amsden 1987) and DDM (Discrete Droplet model) are widely used for numerical simulation (Dukowicz 1980, Glodowski et al. 1996). These results provided the detailed information about spray tip penetration, spray cone angle, distribution of the liquid/ vapor phase, or droplet velocity and diameter due to secondary break-up process. However, one of the key processes affecting spray behavior is the primary spray break-up, because it defines the starting conditions for the spray distribution, the evaporation process, and mixture formation. Experimental investigations of atomization process were restricted due to very high-speed and very small region phenomena. Although scale-up models have been used to study the primary spray structure, it is impossible to match Reynolds, Weber, and cavitation numbers and time scales simultaneously in practical high-pressure multi-hole injector. Microscopic and very fast investigation of primary spray structure of practical multi-hole injector is strongly needed

2 Fundamental mechanisms of liquid jet breakup have been discussed under extensive experimental and theoretical research for more than a century. Chigier and Reitz provided reviews of liquid jet breakup regimes (Chigier and Reitz 1996). They described the various regimes of breakup of liquid jets injected into both stagnant and coflowing gases. Hiroyasu, H., et al., (1996) and Arai, M., et al. (1988) showed the breakup behavior of liquid jet. They indicated the relationship between injection velocity and breakup length of liquid jet. Breakup regimes are classified into 5 regions, such as smooth jet region, transition flow region, wavy jet region, incomplete spray, and complete spray. They also showed the effects of internal cavitation inside nozzle on spray atomization. The effects of nozzle cavitation on primary breakup have been investigated using diesel injector (Stiesch 2003). Currently, the micro cavitaion was confirmed inside PFI injector with low injection pressure condition (Ishimoto et al. 2007). Secondary droplet breakup is basically driven by the aerodynamic forces from surrounding gas phase. Droplet breakup regimes were categolized using dimensionless droplet Weber number, such as vibrational breakup, bag breakup, bag/ streamer breakup, stripping breakup, and catastrophic breakup (Wierzba 1993). Visualizations have been traditionally carried our using photographic film recordings (Zhao and Ladommatos 2001, Frohn and Roth 2000). Recently, solid state cameras, particularly CCD camera, have been developed. The images can be stored directly in solid state memory in high-speed video camera. Jet breakup and droplets breakup have been visualized using photographic method. Droplet stroboscope techniques, which are good for freezing the movement, have been usually used for droplet deformation process and droplet breakup. However, it is difficult to obtain the time-series of deformation or liquid column breakup due to the limitation of high resolution. Microscopic and very fast investigation is needed to investigate the liquid column breakup and secondary droplet breakup for practical PFI injection. In this research, experimental investigation of atomization process of practical PFI injector were demonstrated using a high-speed video camera (maximum speed: 1Mfps) (Kawahara et al, 2003, Kawahara et al. 2004). Initial state and development of the primary liquid column breakup were discussed in only one nozzle hole of PFI injector. During the injection period, developments of the surface waves at the spray boundary were discussed. Breakup of liquid ligament into droplets was investigated at the downstream of injection. Moreover, coalescence of droplets or interaction of droplets were visualized in practical PFI injector. 2. Experimental Apparatus 2.1 PFI Injector and Experimental Set-up The gasoline engine PFI injector used in this study is prototype that was fabricated specifically for research use. Usually this kind of injector has multi-hole, such as 10 holes. In this research, in order to focus on the primary spray structure of one spray plume, therefore the special injector of 9 holes covered by the plate is produced, as shown in Fig. 1. Safety gasoline (Dry-solvent, density: 770 kg/m 3, surface tension: 2.45x10-2 N/m) is used as a fuel instead of gasoline. The injector has an inwardly opening needle. The diameter of nozzle exit is φ 0.2 mm. Figure 2 shows schematically the experimental set-up for investigation of the primary spray structure of high-pressure multi-hole injector. The closed vessel used was a steel cylinder, 180 mm in diameter and about 350 mm long. The PFI injector is mounted on the upper end plate of the closed vessel. The PFI injector is controlled with an electric injector driver, requiring a high voltage power source and external trigger input for injector pulse width control. Nitrogen gas was used to pressurize the fuel supply system and provides adjustable fuel delivery pressure at 0.3MPa

3 2.2 Measurement Technique A high-speed video camera (maximum speed: 1Mfps (1 million frames/second), maximum exposure time: μs, Shimadzu Hyper Vision HEX-108) was used with high resolving longdistance microscope, as shown in Fig. 3 (Kawahara et al, 2003, Kawahara et al. 2004). This highspeed camera has a resolution of 316 x 260 pixels. Backlighting method was used for detection of primary spray plume very close to the nozzle exit. Barlow lens was used for changing the spatial resolution. Typical captured images are shown in Fig. 4 (left image: whole spray image, right image: magnified image). The whole image obtained using the long-distance microscope indicates the spray plume as black color. In this case, the captured image was approximately 5.0 x 6.0 mm at 1Mfps as camera speed in order to obtain the whole image of one spray plume. On the other hand, the liquid column of primary spray very close to the nozzle exit was colored by black using the backlighting method. The obtained image was scaled at 1.7 x 2.0 mm using long-distance microscope with barlow lens at 1Mfps as camera speed (exposure time: 0.25μs). In order to obtain the primary spray structure, very fast camera speed should be used. However, very fast camera speed has very short exposure time as 0.25μs, therefore strong bright light source should be used. Using this backlighting method, detail structure of the primary liquid column can be obtained. It is difficult to investigate the liquid column structure, especially inside column using backlighting method, therefore LIEF (Laser Induced Exciplex Fluorescence)(Melton 2983) method were adopted to compare with backlighting images. Fuel is iso-octane with fluorobenzene and TEA (Triethylamine). We focused only on liquid phase with band-pass filter (center wavelength: 350 nm). Fig.1 Detail of nozzle shape - 3 -

4 Fig.2 Experimental set-up Constant volume chamber Long distance microscope Barlow lens Nitrogen gas cylinder Metal halide lamp High-speed video camera Fig. 3 High-speed video camera with long-distance microscope - 4 -

5 Whole image Magnified image 1mm Fig.4 Whole image and magnified image 3. Experimental Results and Discussion 3.1 Liquid Column Breakup The transient spray from one nozzle hole of PFI injector was shown in Fig. 5. Delay time denotes the elapsed time from the injection start signal. During the initial stage of spray, the spray is shaped like a liquid column as the leading edge region. During injection period, liquid column is formed very close to the nozzle exit. These liquid columns have a fluctuation of movement. The surface of the liquid column is not smooth and neat due to the ligament structures. The development of one of the spray plume can be seen. Three stages in the development of spray can be identified: (1) the initial stage (Delay=0~0.8 ms); (2) the quasi-steady state (Delay=0.8~1.6 ms); (3) the postinjection stage. Figure 6 indicates the backlighting image and LIEF (Laser induced exciplex fluorescence) image in order to understand the liquid column shape. Both whole image of PFI injection and magnified images at 1.0 mm, 4.5 mm, 7.5 mm, 11.0 mm from nozzle exit. LIEF images were obtained using laser sheet of Nd-YAG laser to understand liquid column shape. Liquid fuel forms the liquid column very close to the nozzle exit, and then makes liquid ligament and droplets due to the break-up of liquid ligament. Liquid column has Kelvin-Helmholts instability caused by the large slip velocity between the liquid jet and ambient air. It is generally accepted that an aerodynamic instability causes the liquid column break-up into ligament. Instability of liquid column causes the surface waves of liquid column very close to nozzle exit. Liquid column can be seen as the solid column from both backlighting image and LIEF image. This liquid column is broken up to the liquid ligament around 4.5mm from nozzle exit. These liquid ligaments are broken up to the liquid droplets at 7.5 mm from the nozzle exit. Using microscopic measurement method, the breakup phenomena of liquid ligament into droplets were observed. Here, the surface waves of liquid column very close to the nozzle exit can be detected using obtained high speed image at 1 Mfps (exposure time: 0.25μs) as camera speed, as shown in Fig. 7. Growth of surface wave on liquid column can be seen using high-speed images. These surface waves cause the break-up of liquid column to liquid ligaments. Time-series of these wavelength of surface wave on liquid column and diameter of liquid column are shown in Fig. 8. Development of surface wave on liquid column can be detected. Development of surface wave causes the breakup of liquid column into liquid ligament. Process of liquid column breakup into liquid ligaments can be investigated using high-speed imaging

6 Delay t, ms mm Whole image : Speed:64kfps Exposure:4μs mm Back lighting image Magnified image : Speed:500kfps Exposure:250ns Fig.5 Time-series of spray images 1mm from nozzle LIEF image Liquid column 2mm 4.5mm from nozzle Liquid ligament 7.5mm from nozzle Liquid ligament Liquid droplet 11mm from nozzle Liquid droplet 1mm Fig.6 Comparison of spray shape - 6 -

7 Delay, μs mm 0.11mm 0.13mm 0.15mm 0.18mm 0.19mm 0.10mm 0.10mm 0.08mm 0.05mm Fig.7 Surface wave of liquid column 0.3 Wavelength Liquid diameter Length, mm Delay, μs Fig. 8 Time-series of wavelength and amplitude of surface wave on liquid column 3.2 Liquid Ligament Breakup Figure 9 indicates examples of break-up of liquid ligaments to droplets in the secondary breakup using the magnified images at 7.5 mm/ 11.0 mm from nozzle exit. Liquid ligament were broken up into three droplets due to the surface tension on liquid ligament and instability of liquid ligaments as Rayleigh-Taylor breakup (Patterson and Reitz 1998, Liu et al. 1993). Secondary droplet breakup is basically driven by the aerodynamic forces from surrounding gas phase. Droplet breakup regimes were categolized using dimensionless droplet Weber number, such as vibrational breakup, bag breakup, bag/ streamer breakup, stripping breakup, and catastrophic breakup (Wierzba 1993). Here, droplet Weber number is defined as follows (Stiesch 2003); 2 llvrel ρ g We d = (1) σ - 7 -

8 where ρ g is density of gas phase, l l is representative length of liquid phase, is relative vrel velocity between liquid phase and gas phase, σ is surface tension of liquid phase. Droplet Weber number of liquid ligament is 2.48 during the breakup of liquid ligament in Fig. 9 (b) as shown in Fig.10. Thus, breakup of liquid ligament into liquid droplets was distinguished as vibrational breakup in the secondary breakup region in ref. (Wierzba 1993). In next paper, detail phenomena of breakup of liquid ligaments will be discussed. This is one example of breakup of liquid ligament, however the break-up process of liquid column and liquid ligament can be discussed using a high-speed magnified images. 3.3 Droplets Interaction Finally, examples of droplet interaction in practical PFI injector were shown in Fig. 11. Penetration of droplets was shown in (a), coalescence of droplets in (b) in Fig. 11. In the case of penetration, upper side smaller droplets have larger velocity than lower side larger droplets, thus smaller droplets penetrates larger droplet. When two droplets have almost same velocity in (b), these droplets show the coalescence of droplets. Usually modeling of injection spray was considered with turbulence and cavitation based primary breakup, breakup model of liquid column into liquid ligament, secondary droplets breakup, and droplets collision and coalescence. Using high-speed magnified images, all phenomena inside spray atomization can be discussed Delay t, ms Liquid ligament 7.5mm from nozzle exit (a) 7.5 mm from nozzle exit Delay t, ms Droplet Liquid ligament Droplet 11.0mm from nozzle exit (b) 11.0 mm from nozzle exit Fig.9 Examples of break-up of liquid ligament into liquid droplets - 8 -

9 We d =2.48 Fig. 10 Weber number of liquid ligament Delay t, ms mm from nozzle exit (a) Penetration of droplets Delay t, ms mm from nozzle exit (b) Coalescence of droplets Fig. 11 Interaction of droplets - 9 -

10 4. Conclusions The primary spray structure very close to nozzle of practical PFI injector used in port-injected spark-ignition engine has been investigated. Visualizations of primary spray structure were demonstrated using a high-speed video camera (max. 1Mfps) with a long-distance microscope. Three main conclusions were obtained from this study. (1) It has been shown that the liquid fuel column was injected. During the injection period, the spray indicates the quasi-steady state mode. (2) Surface waves of liquid column can be recognized. Liquid column very close to the nozzle exit were broken up to liquid ligament structure. These surface waves cause the break-up of liquid column to liquid ligaments. (3) Break-up process of liquid ligaments to droplets can be investigated using high-speed magnified images. Moreover, interactions between droplets, such as penetration or coalescence, can be photographed. Acknowledgments The authors appreciate Mr. Tomohiro NAKAYAMA and Mr. Hiroyuki HIGASHINO for their helps in the experimental work. REFERENCES Arai, M., Shimizu, M., Hiroyasu, H., (1988) Break-up Length and Spray Formation Mechanism of a High Speed Liquid Jet, Proc. of the 4th Int. Conf. on Liquid Atom. and Spray Systems, Chigier, N., and Reitz, R.D., (1996) Regimes of Jet Breakup and Breakup Mechanisms, Recent Advances in Spray Combustion: Spray Atomization and Drops Burning Phenomena Volume I, AIAA, Inc., Dukowicz, J.K., (1980) A Particle-Fluid Numerical Model for Liquid Sprays, Journal of Computational Physics, 35: Frank Zimmermann, John Bright, Wei-Min Ren, Bill Imoehl, (1999) An Internally Heated Tip Injector to Reduce HC Emissions During Cold-Start, SAE Paper, Frohn, A., and Roth, N., (2000) Dynamics of Droplets, Springer. Fu-Quan Ahao, Joon-Ho Yoo, Ming-Chia Lai, (1995) The Spray Characteristics of Dual-Stream Port Fuel Injectors for Applications to 4-Valve Gasoline Engines, SAE Paper, Heywood, J. B., (1988) Internal Combustion Engine Fundamentals, McGraw-Hill Book, Inc. Hiroyasu, H., Arai, M., Shimizu, M., (1996) Effect of Internal Flow Conditions Inside Injector Nozzles on Jet Breakup Processes, Recent Advances in Spray Combustion: Spray Atomization and Drops Burning Phenomena Volume I, AIAA, Inc., Ishimoto, J., Sato, F., and Sato, G., (2007) ILASS JAPAN, 16th Symposium on Atomization (JAPANESE). Kawahara, N., Tomita, E., Nakayama, T., and Sumida, M., (2003) Microscopic Observation of Primary Spray Structure of High-Pressure Swirl Injector for Gasoline Direct Injection Engine, The 9th Int. Conf. on Liquid Atomization and Spray Systems (ICLASS 2003). Kawahara, N., Tomita, E., Kasahara, D., Nakayama, T., Sumida, M., (2004) Fuel Breakup near Nozzle Exit of High-Pressure Swirl Injector for Gasoline Direct Injection Engine, SAE Paper No

11 Liu, A.B., Mather, D., and Reitz, R.D., (1993) SAE Mark L. Glodowski, Donna J. Michalek, Lawrence W. Evers, (1996) The Use of Results from Computational Fluid Dynamic Fuel Injector Modeling to Predict Spray Characteristic, SAE Paper, Melton, L.A., (1983) Spectrally separated fluorescence emissions for diesel fuel droplets and vapor, Applied Optics, 22, 14: O Rourke, P.J., and Amsden, A.A., (1987) SAE Paper No Patterson, M.A., and Reitz, R.D., (1998) SAE Stiesch, G., (2003) Modeling Engine Spray and Combustion Processes, Springer, p Wierzba, A., (1993) Deformation and Breakup of Liquid Drops in a Gas Stream at Nearly Critical Weber Numbers, Exp. in Fluid, 9: Xin Liu, Kyoung-Su Im, Yujie Wang, Jin Wang, David L.S. Hung, James R. Winkelman, Mark W. Tate, (2006) Quantitative Characterization of Near-Field Fuel Sprays by Multi-Orifice Direct Injection Using Ultra fast X-Tomography Technique, SAE Paper, Zhao, H., and Ladommatos, N., (2001) Engine Combustion Instrumentation and Diagnostics, SAE, Inc