PLIF Flow Visualization of Methane Gas Jet from Spark Plug Fuel Injector in a Direct Injection Spark Ignition Engine

Transcription

1 PLIF Flow Visualization of Methane Gas Jet from Spark Plug Fuel Injector in a Direct Injection Spark Ignition Engine TAIB ISKANDAR MOHAMAD and HOW HEOY GEOK Department of Mechanical and Materials Engineering, Faculty of Engineering, Universiti Kebangsaan Malaysia, (National University of Malaysia) Bangi, Selangor, MALAYSIA Abstract: - A Spark Plug Fuel Injection (SPFI), which is a combination of a fuel injector and a spark plug was developed with the aim to convert any gasoline port injection spark ignition engine to gaseous fuel direct injection [1]. A direct fuel injector is combined with a spark plug using specially fabricated bracket connected to a fuel pipe and a fuel path is drawn along periphery of a spark plug body to deliver the injected fuel to the combustion chamber. The cross-sectional area of SPFI injection nozzle is significantly larger than the normal direct fuel injector nozzles. Therefore, investigating the effect of such parameter on the injection process and subsequently the air-fuel mixing behaviour inside the combustion chamber is important. A laser diagnostic flow visualization using planar laserinduced fluorescent (PLIF) technique was used. Due to safety and close proximity of properties, nitrogen was used as methane substitute. Nitrogen at 50, 60 and 80 bar pressure were seeded with acetone as flow tracer and injected into a bomb containing nitrogen. Bomb pressure was varied to simulate the pressure inside combustion cylinder during the compression stroke where actual injections in engine experiments will take place. The plume shape and depth of penetration of jet tip were measured. Results show that the gas jet follows the behaviour suggested by vortex ball model [2]. The cone angle and the maximum jet width of the fully developed gas jets from the SPFI injection are 23 o and 25 mm respectively regardless of the injection pressures. The penetration lengths of the fully developed jets are between 90 and 100 mm at 8 to 12 milliseconds after the start of injections, depending on the bomb and injection pressures. Jet penetration is directly proportional to the injection pressure but inversely proportional to the cylinder or bomb pressure. The penetration lengths indicate that sufficient distance would be travelled by the gas jet for air-fuel mixing in the engine during experiment. Key-Words: - Air-fuel mixing, Direct fuel injection, Laser-Induced Fluorescent, Flow visualization, Gaseous fuel 1 Introduction The flow patterns of air and fuel in an internal combustion engine play an important role in determining many operating characteristics such as air-fuel mixing, combustion, fuel economy, performance and emissions. The absence of detailed knowledge of the flow will lead to engine development based mainly on empiricism and intuition rather than well established scientific principles [3]. Most flow visualization of fuel injection studies deal with gas jet or liquid spray coming out from injection nozzle or injector orifice. It this study, a component known as Spark Plug Fuel Injector (SPFI) was developed and the gas jet from injection nozzle, which located at some distance from injector orifice, was investigated. The cross-sectional area of SPFI injection nozzle is significantly larger than those of conventional fuel injector. Furthermore, the injection penetrates into a high pressure turbulent air in the engine cylinder. Thus it is an interesting subject of study with regard to the effect of these parameters on the structure and behaviour of gas jet. The objective of this study is to investigate the shape and depth of penetration of methane jet from a Spark Plug Fuel Injection (SPFI) in a direct injection engine. Laser diagnostic approach was considered to visualize the gas jet developed in the bomb. Because compressed nitrogen is injected into nitrogen, the imaging method must be able to differentiate the injection and surrounding gas when injection takes place. Therefore, the method with a tracer element included is necessary. The planar-lif (PLIF) technique was chosen over all other methods in this experiment due to this reason and also due to its simplicity and local availability of equipment. In the PLIF method, the tracer must be able to produce a visible effect when excited by the laser sheet and its saturation properties must be equivalent to the fluid. The applications of LIF method in methane-fuelled engine were mainly found in flame and emission studies. ISSN: ISBN:

2 A quantitative study on temperature, OH, and CH 2 O effects on the stabilization of lifted turbulent methane flames issuing into a high temperature vitiated co-flow was carried out using LIF method [4]. Quantitative measurements of NO concentration along the centerline of atmospheric pressure, methane air, and counter-flow partially premixed flames using the laser-induced fluorescence technique have also been carried out [5]. A study on micro-jet in a turbulent confined cross flow using PLIF was carried out to investigate the concentration of methane in the mixing region of the flow [6]. A number of tracer materials have been used in PLIF imaging including acetone, pentanone-3, NO 2 [7], and rhodamine [8]. Acetone was chosen to be the LIF gas tracer in this experiment as it has been successfully utilised in gaseous flow measurement as reported by a number of sources [9, 10, 11]. All these PLIF-related studies were focused on flame and emission analysis, liquid fuel injection and injection process of a gas into another gas. Limited studies were done on both injection and environment of the same gas. This paper will discuss the success of imaging and qualitative analysis of single-species gasto-gas high pressure injection using LIF technique for the analyzing the SPFI. The results will be used for analyzing and design optimization of this novel product. Further studies on SPFI regarding the flame and emission will be carried out in the future. 2 Spark Plug Fuel Injector Fig. 1shows the image of Spark Plug Fuel Injector (SPFI) and all the components explained. The technical drawings and connection to a gasoline direct injector (GDI) encapsulated in a bracket is shown in Fig. 2. SPFI consists of a spark plug with a 1 mm by 2 mm square cross-section fuel path cut out along the periphery of its threaded section and a steel tube soldered to the end of the cut section. A GDI is connected to it using a specially developed bracket to the end of the fuel path. The distance from the GDI injection nozzle to the SPFI nozzle is 11 cm. The GDI injector is connected to a 230 bar methane bottle through a pressure regulator where methane pressure is reduced to the desired pressure. A specially developed injection control was used to regulate fuel injection by referencing crank angle signals from a camshaft encoder. The length of injection pulse determines fuel mass delivered, therefore air/fuel ratio to the cylinder. Fig. 1 Spark plug fuel injector Fig. 2 (a) SPFI connected to Injector Housing and (b) Technical drawing of SPFI 3 Experimental methods and procedures 3.1 Injection Chamber and LIF Imaging The main components involved in the setup of this experiment can be categorised into four sub-groups; a laser-optical lenses system, an imaging system, a fuel supply system and a fuel injection chamber. Details of the complete experimental components are shown schematically in Fig.3. The fuel injection driver circuit was synchronized with the camera and laser systems. A mosfet, acted as a gate/bridge between a power supply unit and the SPFI, is excited by the output of a pulse generator which functioned as the main driver to the experiment. At the same time, the pulse generator output was connected to another pulse/delay generator where its signal was sent to the laser source. The laser source simultaneously sent a signal to the camera so laser pulse was synchronised with the opening of the shutter. An oscilloscope was connected to both pulse/delay generator to measure the delay between initiation of fuel injection and image capturing. The plume of fuel emerging from the injector nozzle was imaged. The plume was illuminated with a 100 mm high, 0.5mm thick light sheet from an XeCl excimer laser (308nm) with mJ/pulse. The plume was imaged with a LaVision SprayMaster 3 camera with image intensifier and 105mm Nikkor UV lens. Two sets of images were taken. Laser Induced Fluorescence (LIF) images were taken of fluorescence ISSN: ISBN:

3 emitted by the acetone mixed with nitrogen. The LIF is emitted by both droplets and vapour and the signal is proportional to the local mass concentration of fuel independent of physical state. Images were taken at several different times after the start of the injector current pulse. At each time, 20 images were taken from successive cycles. The camera settings were f/8, intensifier gate 200ns, intensifier gain 60, camera gain 95. The images were corrected for background intensity and light sheet intensity variations with LaVision DaVis 6.2 software. The nitrogen acetone doping mechanism is shown in Fig. The nitrogen is bubbled through the acetone and trapped in the upper part of the threshold bottle. As bubbling progresses, the trapped gas increases its pressure and as it reaches the desired injection pressures, the supply valve is closed. An outlet gas pipe is attached to the bottle positioned above the surface of acetone, thus only acetone-saturated gas is delivered to the SPFI. The pressure gauge on the gas pipe near SPFI was used to determine the injection pressures. The experiments were carried out by referencing the cylinder pressure of a motorised engine. For 80 bar injection pressure, the optimal injection time is at 215 o BTDC (compression). For 50 bar and 60 bar injections, the optimal injections were at 170 o BTDC (compression). Fuel injection durations were 6 ms, 10 ms and 12 ms for 80 bar, 60 bar and 50 bar injection pressures respectively. At 1100 rpm, these injections cover 40 o CA, 66 o CA and 79 o CA respectively, as shown in the figure. Taking into account the fuel delivery delay due to the lengthy fuel path in the SPFI, it was decided to perform injections at three bomb pressures: 1 bar, 3 bar and 10 bar. The 1 and 3 bar pressure represent the range of actual injection cylinder pressure, while the 10 bar pressure was chosen to investigate the effect of gas jet if injection is delayed at later stage of compression stroke. Fig. 4 Nitrogen acetone doping mechanisms for SPFI spray imaging 3.2 Image Calibration It is important to quantify the injection penetration and the width of gas plume from the images. The bomb viewing window allows 110 mm circular diameter visual access to the injection gas flow. Calibration of gas jet dimensions was achieved by taking of a scaled image attached to the fuel injection vertical plane as shown in Fig.5. Fig. 6 shows a calibration image captured by the CCD camera and compared with a PLIF image with no gas injection. The bright spots on the PLIF image are the scattering from the injector holder, which could not be avoided. Fig also shows a grainy texture to the PLIF image. Much effort was applied to avoiding this background noise and improvements were made. However, the need to pick out the very slight fluorescent caused by the light doping (stronger doping proved impossible) also means picking out this background noise. The design of the doping system was limited by a need to operate it at high pressures (up to 80 bar) and limited supply of acetone (hence the depth of bubbling was limited). Fig. 3 Schematic of the PLIF imaging of SPFI injection flow visualization Fig. 5 Calibration of fuel injection measurement ISSN: ISBN:

4 Fig. 6 Calibration image (left) and PLIF image with no gas injection (right) 4 Results and discussion The images showed in Fig. 7 are the consecutive images of gas jet at various injection and bomb pressures. The bright white areas penetrating from the top of each image indicates the injected gas presence. The intensities of brightness can be used to describe qualitatively the gas concentration. It is important to note that because the acetone-doped nitrogen stayed almost statically inside the injection fuel line due to slow rate of injection in this experiment (once in every 1 second), it was possible that some of the acetone settled or condensed to the pipe walls. This can also be a cause of the background noise problem. Fig. 7 Consecutive images of various injection pressures and bomb pressures The results show that in all injection pressures, the first appearance of gas jet can be seen at 2.5 ms but fully developed (are shown by ) at different times. Higher injection pressure produce earlier fully-developed gas plume. Besides, injection pressures have strong influence on the effective fuel delivery time which can be implied from the time of detachment of gas plume from the injection nozzle where higher injection produces less delivery time. However, increasing bomb pressure will result shorter jet penetration and slower effective delivery time. Fig. 8 Fully developed gas jet from SPFI at 60 bar injection and 3 bar bomb pressure It can be seen that the fully developed gas jets from the SPFI injection were relatively narrow, with about 23 o cone angle and 25 mm of maximum width in a fully developed gas jet as shown in Error! Reference source not found.. The penetration length of the fully developed gas jet is between 90 and 100 mm at 8 to 10 milliseconds after the start of injections. The injection durations for 50, 60 and 80 bar fuel pressures on the Ricardo E6 engine were 12 ms, 10 ms and 6 ms respectively.. The image on t Fig.8 shows the fully developed gas jets from a 60 bar injection pressure into 3 bar bomb pressure. A significant amount of background scattering on the image can be seen. The scattering was believed to be the result of minor illumination of the background gas in the bomb by the laser sheet. It was made worse by the need to pick out slight fluorescent from a lightly doped injection gas. This however, did not prevent the visualization of gas jet development and qualitative determination of penetration length and plume width. The penetrations of gas jet based on the PLIF imaging experiments are shown in Fig. 9 and Fig. 10Error! Reference source not found.. Jet penetration is proportional to square root of injection time. Bomb pressure has noticeable effects on jet penetration. As the pressure increases, jet penetrations were reduced. However, for the same bomb pressure, the variation in penetration lengths with changing fuel pressure is relatively small. This is because fluid flows exiting the SPFI injection nozzle were at sonic conditions for all injection pressures. Sonic conditions cause the flow to ISSN: ISBN:

5 be choked, which means the gas velocity remains constant except when flow temperature increases. However, the mass flow rate at the sonic conditions not only depends on flow temperature but also depends on the injection densities which are proportional to pressures. Therefore mass flow rate determines the length of injection duration and in certain degree affected the length of jet penetration. compression stroke. However, the width of the jet and the direction of injection away from the point of ignition could be detrimental to the combustion and engine performance. Therefore, an appropriate calibration of injection timings must be undertaken to compensate this Jet tip penetration, mm bar injection 60 bar injection 80 bar injection Fig.11 A fully developed gas jet inside Ricardo E6 combustion chamber with piston at BDC Time after start of injection, ms Fig. 9 Effects of injection pressure on the SPFI jet tip penetration Jet tip penetration, mm Time after start of injection, ms 80 bar inj / 3 bar bomb 80 bar inj / 10 bar bomb Fig. 10 Effect of bomb pressure on the SPFI jet tip penetration SPFI utilizes a fuel injector (Audi-FSI direct injector) which is optimised for direct injection and stratified charge operations. However, the optimization of nozzle design and orientation to give best effect on stratified charge direct injection was not taken as an advantage when used as a part of SPFI. As a result, the effective fuel injection behaviour is determined mainly by the injection nozzle as well as the fuel path and this work has given a qualitative understanding of this behaviour which is useful for design optimization process of SPFI. Fig.11 shows the representation of a projected fully developed gas jet at 60 bar injection pressure and 1 bar bomb pressure inside the Ricardo E6 combustion chamber with piston at BDC where the engine experiments were carried out. It indicates the sufficient jet penetration length especially during the later part of 4 Conclusion Visualization of gas jet development from a spark plug fuel injector has been successfully implemented using planar laser-induced fluorescent method with acetone as the flow tracer. Imaging gas jet in a gas environment which has been difficult to achieve previously is now a reality. The gas jet agrees with the vortex ball model [2, 12]. The jet is generally narrow with 23 o cone angle and 25 mm of maximum width in a fully developed gas jet. The penetration length of the fully developed gas jet is between 90 and 100 mm at 8 to 12 milliseconds after the start of injections depending on the injection and bomb pressures. Penetration length is proportional to square root of time and directly proportional to the injection pressure but inversely proportional to the bomb pressure. The images and measurements of tip penetration as well as tip velocity will be compared with the vortex ball model through simulation. The outcome of the work will be used the design optimization of SPFI and the calibration of injection timing during engine experiments. References: [1] Mohamad, T. I., Development of a Spark Plug Fuel Injector for Direct Injection of methane in Spark Ignition Engine, PhD Thesis, 2006, School of Engineering, Cranfield University, Cranfield. [2] Turner, J. S., The Starting Plume in Neutral Surroundings, Journal of Fluid Mechanics Vol. 13, 1962, pp [3] Cole, J. B. and M. D. Swords, Laser Doppler Anemometry Measurements in an Engine, Applied Optics, Vol. 18, No. 10, [4] Gordon, R. L., A. R. Masri, et al., Simultaneous Rayleigh temperature, OH- and CH 2 O-LIF imaging ISSN: ISBN:

6 of methane jets in a vitiated co-flow, Combustion and Flame, Vol. 155, No. 1-2, 2008, pp [5] Ravikrishna, R. V. and N. M. Laurendeau, Laserinduced fluorescence measurements and modeling of nitric oxide in counterflow partially premixed flames, Combustion and Flame, Vol. 122, No. 4, 2000, pp [6] Kelman, J. B., D. A. Greenhalgh, et al., Micro-jets in confined turbulent cross flow, Experimental Thermal and Fluid Science, Vol. 30, No. 4, 2006, pp [7] Hayashida, M., T. Yamato, et al., Investigation of Performance and Fuel Distribution of a Direct Injection Gas engine using LIF Measurement. SAE [8] Yamakawa, M., S. Isshiki, et al., Measurement of Ambient Air Motion of D.I. Gasoline Spray by LIF- PIV, in The Fifth International Symposium on Diagnostics and Modeling of Combustion in Internal Combustion Engines (COMODIA 2001), 2001, Nagoya, Japan. [9] Choi, H. J., Y. S. Ko, et al, Visualization of Concentration Field in a Vortex Ring Using Acetone PLIF, Journal of Visualization, Vol. 5, No. 2, 2002, pp [10] Lozano, A., B. Yip, et al., Acetone: a tracer for concentration measurements in gaseous flows by planar laser-induced fluorescence, Experiments in Fluids, Vol. 13, No. 6, 1992, pp [11] Olsen, D. B. and B. D. Willson, The impact of cylinder pressure on fuel jet penetration and mixing." ICE, Vol. 39, 2002, pp [12] Boyan, X. and M. Furuyama, Jet Characteristics of CNG Injector with MPI system, JSAE Review, Vol. 19, 1998, pp ISSN: ISBN: