An Experimental and Numerical Investigation on Characteristics of Methanol and Ethanol Sprays from a Multi-hole DISI Injector Yajia E 1, Min Xu 1, Wei Zeng 1, Yuyin Zhang 1, David J. Cleary 2 1 Inst. of Automotive Engineering, Shanghai Jiao Tong University, Shanghai 200240, mxu@sjtu.edu.cn 2 Powertrain Systems Research, GM China Research & Development, No 99 Fucheng Rd., Shanghai 200120 Abstract The objective of this study is to investigate the characteristics of methanol and ethanol sprays from a multi-hole injector for DISI engines. An optical diagnostic technique based on instantaneous LaVision laser sheet imaging system was used to observe the sprays under various injection pressures and chamber ambient pressures in a constant volume chamber. The effects of fuel type on liquid atomization and spray characteristics in terms of penetration and cone angle were examined and analyzed. On the other hand, a spray model based on AVL FIRE software was used to predict the spray characteristics of methanol, ethanol and gasoline fuels. The model was validated by use of these experimental results. As a result, the difference in spray characteristics of these three fuels and the effect of the in-cylinder flow on fuel distributions were clarified. Keywords: Spray, Methanol, Ethanol, Gasoline, CFD, Laser sheet imaging 1. Introduction Energy and environment have been more and more concerned world-widely. Therefore, further improving the efficiency and decreasing the pollution of internal combustion engines have become increasingly vital. Alcohol as one of the alternative fuels can be produced from many sources, such as coal, nature gas, biomass, etc. On the other hand, alcohols are oxygenated fuels with properties such as a higher octane number, broader flammability limits, higher flame speeds, higher latent heats when compared to gasoline. Therefore, a higher compression ratio, shorter burn duration, and cleaner combustion are easy to realize at a spark ignition engine operating on alcohol fuels [1]. Direct injection spark ignition engine (DISI) has merits of high thermal efficiency, low emissions, and high torque output. Combining alcohols with DI could draw benefits from both and offer a better solution [2]. Up to date, experimental studies on engine performances operating on methanol or ethanol fuels were reported many times [3, 4], however, the complete investigation on the methanol and ethanol spray characteristics from typical DI injector nozzles are rarely seen. In this work, we investigated the characteristics of alcohols and gasoline sprays injected from an 8-hole DI injector by using the planar laser imaging technique MIE-scattering. By using AVL FIRE software, we also predicted the spray SMD for three fuels, and investigated airflow effect on spray atomization. 2. Experiment Setup Figure 1 shows the schematic diagram of the experimental setup, which includes a constant volume pressure chamber, the fuel supply system and camera planar laser imaging system. An eight-hole DI injector was installed in the constant volume pressure chamber. The injector nozzle orifices are the step-holes with the L/D of 2, the metering orifice diameter of 0.15mm, and the outlet diameter of 0.3mm. Three hydraulic piston accumulators were used to generate various injection pressures of methanol, ethanol and gasoline fuels respectively, and supply to the injector. The injection pressure and chamber ambient pressure were controlled by two different nitrogen pressurization systems. The spray was illuminated by the laser sheet, which is generated by an Nd: YAG laser system, and photographed by the LaVision instantaneous imaging system in a direction of 90 degree to the laser sheet. A Programmable Timing Unit (PTU) was used to
synchronize the laser, the camera and the injector. Figure 1. Experimental apparatus Table 1. Experiment Conditions 97# gasoline, ethanol, Test fuels methanol Injection pressure (MPa) 5, 10 Ambient pressure (MPa) 0.1, 0.25, 0.45, 0.8, 1.0 Ambient temperature 20 C Fuel temperature 20 C 3. Experiment Conditions and Fuel Properties The experimental conditions to investigate the fuel spray are listed in Table 1. Nearly a full range of DISI engine injection and ambient pressures at the room temperature was considered. The injection pressures were selected as 5MPa and 10MPa, and the chamber ambient pressures were varied as 0.1MPa, 0.45MPa, 0.85MPa and 1.0MPa. Pressure regulators on the top of the nitrogen gas cylinders regulated those two pressures. The tested fuels were gasoline (octane number 97), pure ethanol fuel and pure methanol fuel. The physical properties of these fuels are listed in Table 2. Table 2. Fuel properties Test fuels 97# Methanol Ethanol gasoline Density (g/ml, 20 C) 0.7406 0.796 0.79 Viscosity (MPa*s, 20 C) 0.42 0.75 1.2 Surface tension (mn/m, 20 C) 18.93 22.18 22.05 Latent heat (kj/kg) 310 1109 904 Equivalent fuel/air ratio 1:14.8 1:6.5 1:9.0 Ignition limit (vol %) 6.7-36 4.3-19 1.4-7.6
Figure 2. Sprays of gasoline, methanol and ethanol (P inj =5MP, P a =0.1MP, t=0.28msasoi) 4. Experiment Results and Discussion To examine the spray characteristics of different fuels from a multi-hole injector, spray images were taken under different injection and ambient pressures in the constant pressure chamber without airflow. Fig. 2 shows the sprays of methanol, ethanol and gasoline at an injection pressure of 5MPa and ambient pressure of 0.1MPa, photographed at 0.28ms after the start of injection. The penetration and spray angle were measured from the images. Fig. 3 shows the spray penetration development at injection pressure of 10MPa, and ambient pressure of 0.1MPa, 0.45MPa, and 1.0MPa. It is obviously seen that the penetration decrease as the increase of ambient pressure. The penetration of gasoline spray is the longest, while ethanol spray penetration is the shortest. Fig. 4 shows the spray penetrations at injection pressure of 5MPa, and ambient pressure of 0.1MPa, 0.45MPa, and 1.0MPa. The trend of penetration of gasoline, ethanol and methanol spray is similar to the case of P inj =10 MPa. Based on the report of Hiroyasu and Arai [5], before the jet starts breaking up, the liquid density plays important role on axial spray penetration. Viscosity affects the penetration all the time. At the same fuel injection and ambient conditions, a smaller density and viscosity could cause a longer penetration. Therefore, the gasoline spray penetration is the longest and the ethanol spray liquid is the shortest. On the other hand, the fluid with higher viscosity (such as ethanol) has higher frictional losses inside the nozzle, resulting in lower initial injection velocity, which leads to a shorter penetration. Figure 5 shows the spray angles of ethanol, methanol and gasoline with different injection pressures and ambient pressures. The higher injection pressure results in the bigger spray angle, because higher injection pressure leads a bigger relative velocity and faster droplet breakup, both of these help to improve the air-fuel mixing, and to distribute the droplet widely. Higher ambient pressure also increases the spray angle at a constant injection pressure due to increased resistance. The effect of different fuel on spray angle is not significant. Figure 3. Spray penetration (P inj =10MPa) Figure 4. Spray Penetration (P inj =5MPa)
Figure 5. Spray Angle Table 3. Computation Condition Ambient temperature [K] 293 Ambient pressure [MPa] 0.1 Fuel Ethanol/methanol /gasoline Fuel temperature [K] 293 Injection quantity [mg] 15 Injection pressure [MPa] 5 Injection duration [ms] 1.5 Time step [ms] 0.05 Calculation period [ms] 1 Turbulence model k ε Evaporation model Breakup model Dukowicz Huh_Gosman injector in FIRE software manuals. The initial droplet size was set to be the nozzle diameter, and the start velocity was calculated from the experiment results. For boundary conditions, all surface of the cylindrical domain were set as walls, and all velocities in all directions were set to 0 m/s. A constant temperature of 293 K was set for the entire wall. Fig. 6 shows the penetration comparison between simulation and experiment results. The difference in penetration is proved to be quite small, and the trend that gasoline has the largest penetration while ethanol has the shortest is also predicted. 4. Computational Model Setup Numerical simulation was performed using the CFD code AVL FIRE. A cylindrical calculation domain with a diameter of 50 mm and a height of 100 mm was meshed using Fame Hybrid, and the number of cells is 192,000. The computation conditions were listed in Table 3. The nozzle was set at the top center of the cylinder. The Huh-Gosman breakup model was used in this simulation, which assumes the flow turbulence inside the nozzle as the decisive parameter for the breakup mechanism. Based on the work of Huh and Gosman as well as Huh, Lee and Koo [6], this model was recommended as a good option for multi-hole DI Figure 6. Comparison of spray penetration between simulation and experiment results
6. Numerical Results and Discussion This part discussed the SMD of different fuels. The simulation accuracy could not be assessed directly due to unavailability of SMD experimental data. However the close agreement between predicted and measured spray penetrations for the three test fuels gives some indication that the simulation results can be trusted to a certain extent. Thus, the predicted SMD results for different fuels were compared and analyzed. The influences of fuels and airflow on the spray structure were also discussed. SMD of Different Fuels Figure 7 shows the SMD as a function of time for ethanol, methanol and gasoline sprays. The simulation injection pressure was 5MPa, and the ambient pressure was 0.1MPa. The simulation results show that ethanol has the biggest SMD, while gasoline has the smallest SMD. Since the breakup and atomization of droplets are dictated by the Weber number of liquid droplets and ambient air, a larger Weber number leads to a faster breakup process. From the definition of Weber number ρu a σ W e 2 =, the larger velocity and smaller surface tension results in a larger Weber number. From the above discussion, the gasoline has the largest initial injection velocity, and the surface tension coefficient of gasoline is the smallest. All these factors lead to the largest Weber number of gasoline and hence the smallest SMD. Figure 7. Numerical results of SMD of different spray fuels injected into quiescent air (P inj =5MPa P a =0.1MPa) Airflow Effects on SMD and Spray Structure To investigate the airflow effect on the spray atomization, an initial swirl air motion was added in the cylinder. The axis of the swirl was set as the axis of the cylinder, and the swirl angular velocity was 12000r/min, which corresponds to a swirl ratio of 4 at an engine speed of 3000 r/min. Figure 8 shows the ethanol spray structure distortion due to the influence of the swirl at 1ms ASOI. The upper picture illustrates the spray structure in the quiescent ambient air, while the lower pictures are the structure of the spray injected into the swirling airflow. A larger area of droplet distribution in radial direction and shorter axial penetration can be observed in the spray injected into the swirling airflow compared to the quiescent ambient. Airflow speed=0r/min, t=1.0m ASOI Airflow speed=12000r/min, t=1.0ms ASOI Figure 8. Spray Structure in strong Airflow (P inf =5MPa, P a =0.1MP) Figure 9 shows the SMD of different fuel sprays injected into the swirling airflow. Gasoline has the largest SMD. Methanol has almost the same SMD as gasoline, but ethanol has the smallest SMD. By comparing with Fig. 6, we can find that SMD of the three fuel sprays in the swirling airflow is smaller than SMD of spray without airflow, and the difference is about 8µm for ethanol and gasoline. The airflow tends to increase the relative velocity between the fuel sprays and surrounding air, hence results in a larger Weber number and a smaller SMD. The SMD of methanol spray changed less comparing to the other two fuels,
probably due to relative poorer initial vaporization caused by a large latent heat. spray distributes more widely, and the axial penetration becomes shorter. Acknowledge Authors would like to appreciate Mr. Hao Chen, Mr. Ming Zhang, and Mr. Gaoming Zhang for their great help on experiments. The research was carried out at National Engineering Laboratory for Automotive Electronic Control Technology and sponsored by General Motors Company. References Figure 9. Numerical results of SMD of different fuel sprays injected into an ambient gas with swirl flow (P inj =5MPa, P a =0.1MPa) 5. Conclusion The characteristics of methanol and ethanol sprays from multi-hole injector for DISI engines were examined both experimentally and numerically. The conclusions are summarized as following: 1. Experimental results show that the spray structure of alcohols and gasoline are similar. Gasoline spray has the longest penetration, while ethanol has the shortest penetration in all experiment conditions, and the effect of fuel type on spray angle was not obvious. 2. The spray penetration increases by about 5 mm at 1 ms ASOI when injection pressure increase from 5 MPa to 10 MPa for all three fuels. 3. The spray penetration depends strongly on the ambient pressure, and the penetration increases by 30 mm at 1.0 ms ASOI when decreasing the ambient pressure from 1.0 MPa to 0.1 MPa. The dependence of spray penetration on ambient pressure is similar for gasoline, methanol, and ethanol fuels. 4. The spray angle increases with increase of injection pressure and ambient pressure 5. The numerical results show that SMD of gasoline spray is the smallest, and ethanol is the largest due to the larger Weber number. 6. Compared to the quiescent ambient air, SMD becomes much smaller in the swirling airflow. The [1] Heather L, MacLean, Lester B. Lave. Evaluating automobile fuel/propulsion system technologies. Progress in Energy and Combustion Science 2003.29:1-69. [2] M Xu and L E Markle. CFD-Aided Development of Spray for an Outwardly Opening Direct Injection Gasoline Injector. SAE Paper 1998. 980493. [3] Shi Shaoxi, Fu Maolin, Yang Jingyun, Zhou Zuoheng. Study on Burning Pure Methanol in D.I.Diesel Engines by Hot-Surface Ignition. Transactions of CSICE. 940029. [4] Qi Dong-hui, Liu Sheng-hua, Li Hui, Lu Sheng-chun. Performances of electronic fuel injection engine fueled with methanol-gasoline blended fuel. Journal of Traffic and Transportation Engineering. 167121637 (2006) 0220043204. [5] H. Hiroyasu, M. Arai, Structures of Fuel Sprays in Diesel Engines, SAE Paper No. 850126, 1985. [6] Huh, K. Y. and Gosman, A. D, A Phenomenological Model of Diesel Spray Atomization, Proceedings of the International Conference on Multiphase Flows, 1991.