Due to limited reserves of crude oil, development on alternative engine fuel
has attracted more concern in the engine community. Alternative fuels are usually
cleaner fuels compared to conventional liquid fuels such as gasoline and diesel
fuel in the combustion process of engines. Natural gas is one of the most important
alternative fuels and the use of natural gas as engine fuel has been studied
for many years and realized in both spark-ignition engine and the compression-ignition
engine (Cho and Bang-Quan, 2007).
Internal combustion engine using hydrogen is considered as a suitable pathway
to hydrogen economy, because fuel cell technologies are considered to be more
mature and cost effective (Wolf and Nordheimer, 2000).
Hydrogen has been proposed as an alternative fuel due to its unique properties.
Its wide flammability range allows higher efficiency. Natural gas and hydrogen
gain importance as an alternative fuel. The use of natural gas blended with
hydrogen is a viable alternative to pure fossil fuels because of the expected
reduction of the total pollutant emissions and the increase in thermal efficiency
(Wang et al., 2007; Huang
et al., 2007; Blarigan and Keller, 2002).
Since they have several advantages, blending fuels of CNG and H2
may come into use in SI engine. Before that, fundamental research is necessary
to clarify the fuel spray characteristics of those two alternative fuels.
The spray behavior plays an important role in engine air-fuel mixing and combustion process, which in turn influences the engine performance. To control the fuel injection, optimize combustion and reduce emissions of the engines, it is necessary and important to understand the characteristics of fuel sprays.
Fuel spray shape, spray tip penetration and spray angle are terms used to characterize
the over all spray structure. Spray angle, a parameter which is the most commonly
used to describe spray distribution, is important because it affects the axial
and radial distribution of the fuel (Kim and Moon, 2001).
Varde, (Varde, 1985; Dan et al.,
1994) have studied the characteristics and performance of spray cone angles
of traditional diesel injectors. It was discovered that the cone angle increased
as the ambient pressure climbed from 0.1 to 3.3 MPa.
The spray image can be obtained using Schlieren photography technique while the processing of the images can be carried out using Digital Image Processing and the Particle Image Velocimetry (PIV) can be used to measure the spray velocity.
Schlieren systems are usually considered to be a qualitative instrument. Greig
et al. (1997) has studied qualitative spray characterization in direct
injection spark ignition engine. The paper concludes with examples of recorded
images of fuels sprays inside the engine.
Several researchers used Schlieren photography technique to measure the spray
characteristics of alternative fuels. The spray characteristics of ethanol-gasoline
blends, as well as pure gasoline were studied under various ambient conditions
by means of high-speed Schlieren photography technique (Gao
et al., 2007). The results showed that when adopting fuel blends
with variable ethanol-gasoline fractions in the swirl-type injector sprays at
low pressure, the main spray tip penetration decreased and spray angle increased
with the increased of ethanol fraction, after 3.5 m sec-1 start of
injection, which meant a better vaporization.
The characteristics of methanol sprays were studied under different opening
pressure, ambient density and nozzle diameter using Schlieren photography techniques
(Yanfeng et al., 2007). The results showed that
with an increased ambient density, the penetration length and tip velocity decreased
and cone angle widened quickly.
Some researchers (Zhao et al., 1999; Lee
et al., 2001) also made use of Schlieren photographs to investigate
the spray characteristics of a high-pressure swirl injector for a gasoline direct
injection engine under various ambient and injection conditions. The results
showed that the spray tip penetration increased with the increase in injection
pressure, while little influence from the injection pressure on the spray cone
angle was observed except for an injection pressure of 2.0 MPa.
Recently, many digital imaging and image processing tasks are being conducted
to measure the spray characteristics. The application of a direct photographic
imaging and image processing system can also be employed to find the quantitative
characterization of sprays. Image processing software has been developed by
a few researchers. They are used to study the effect of varying pressure, tip
penetration, spray angle and fuel area density (Shao and
Yan, 2006; Shao and Yan, 2008; Jeong
et al., 2007; Petit et al., 2007).
In order to quantify the performance of the spray characteristics, Particle
Image Velocimetry (PIV) technique can be implemented. It is used to obtain instantaneous
velocity measurement and related properties in fluid. Lee
and Lee (2007) analyzed the spray characteristics according to the injection
duration under ambient pressure conditions and the injection timing of an optical
engine. The spray velocity can be calculated through the PIV method and the
vorticity, in turn, can be calculated from the spray velocity component.
In general, the previous study, investigated only on detailed behavior of fuel sprays of alternative fuels such as gasoline, ethanol, methanol and diesel. However, very few works were done in area of spray evolutions and mixing behaviors in a direct CNG injection.
This study focuses on the studies of the spray characteristics of compressed
natural gas and hydrogen fuels which were visualized under ambient conditions
by means of high-speed schlieren photography technique. The main objectives
of this work were to investigate the effects of Wide cone angles (WAI) and narrow
cone angle (NAI) injectors and injection pressure on spray behaviors such as
spray shape, tip penetration and spray angle. A comparative analysis of CNG
and H2 sprays is also illustrated in detail.
MATERIALS AND METHODS
Experimental set-up: In this study, two types of fuel samples were prepared for the experiments, i.e. compressed natural gas and hydrogen. The compositions of natural gas are listed in Table 1. The compressed natural gas used in the experiment was the commercial product from Malaysia. The fuel properties of natural gas and hydrogen are listed in Table 2.
According to Table 2, the lean burn capability of hydrogen
is much better than that of natural gas and the laminar burning velocity of
hydrogen is 7 times faster than the natural gas. Hence, the addition of hydrogen
to natural gas is expected to increase the flame propagation speed and stabilize
the combustion process, especially under lean mixture combustion. The quenching
distance of hydrogen is one-third that of natural gas and this is beneficial
to reducing the unburned HC (Liu et al., 2008).
Figure 1 illustrates the schematic diagram of the schlieren
imaging experimental set-up which consists of high-speed video camera, source
of light (lamp), a mirror, high pressure direct injector, regulator pressures
for CNG and H2 and knife edge (razor blade). The fuel rail which
pressures ranging from 1.2 to 1.8 MPa was injected openly at the ambient temperature
and pressure during the experiment.
Experimental procedure: The schematic diagram for the injector is in
Fig. 3. For the case of pure CNG spray, both inlet gas 1 and
2 are supplied with CNG.
|| Composition of natural gas
|| Fuel properties of natural gas and hydrogen
|| Schematic diagram of experimental setup
|| Fuel injector used for testing
This also applies for the case of pure hydrogen. The injector used in the experiment
was a special one because it did not allow the two gases to mix thoroughly.
In order to avoid back flow, the pressure of gas 1 and 2 maintained to be the
The fuel injection of the CNG-DI system as in (Fig. 2) is
fitted either with a wide angle or narrow angle injectors at injection pressures
ranging from 1.2 to 1.8 MPa. The experiment was conducted in sequence with pure
CNG being the first and pure hydrogen the second. Two pressure regulators were
used to control each inlet pressures and this was done manually.
Both wide angle and narrow angle injectors were used one at a time to inject
the fuels in pseudo motoring of 1000 rpm in an open spray.
|| Definitions of spray penetration and spray angle
The capture of images was performed by a high speed video camera (Photron,
FASTCAM-APX) operated at a speed of 4,000 frames sec -1 with effective
pixel size of 640x128. A Nikon 60 mm f/2.8D Micro-Nikkor lens was used to accompany
the camera. In order to get clear images, the knife-edge was adjusted to the
focal point of the light.
Image processing: In order to find the edges of an image Canny edge
detection algorithm was used (Canny, 1986).
The Canny edge detection algorithm, which works by looking for local maxima
of the gradient of Image intensity I, was used to find the spray boundary. The
intensity gradient was calculated using the derivative of a Gaussian filter.
|| Schlieren image for pure CNG at (1.2 MPa)
The method uses two thresholds, to detect strong and weak edges and includes
the weak edges in the output only if they are connected to strong edges.
Spray expansion area can be parameterized by using spray cone angle and tip penetration measurements. In order to convert pixel measurements to SI metric measurements a calibration image of a graduated scale was taken.
The value of one pixel was found out to be equivalent to 0.238 mm.
Figure 4 provides how spray tip penetration and spray cone
angles are defined. The spray angles were obtained by drawing horizontal lines
at 20 mm downstream from nozzle tip and measuring the angle between the edges
of the spray (Dan et al., 1994). And the spray
tip penetration can be measured along the axial distance between injector nozzle
end and the furthest spray point along the spray axis of symmetry.
RESULTS AND DISCUSSION
Effect of wide and narrow cone angle injectors: Figure 4 shows the visualization results of the Schlieren spray images for CNG under ambient conditions when the fuel was injected with injection pressure in the range of 1.2 MPa. For comparison purposes, Wide Angle (WAI) and narrow cone angle (NAI) injectors were used. It was found that the intensity of the injected gas for narrow cone angle injector was higher than the wide cone angle injector before the time reached 5.0 m sec-1. This phenomenon occurred because the fuel for the wide angle case had already been mixed with surrounding air. At the time of 6.25 m sec-1, the images of a wide cone angle injector shows that the gas has already disappeared while the narrow angle injector shows some residual gaseous. In addition, the tip penetration for NAI was longer than WAI.
Effect of injection pressure: Figure 5 and 6 show the Schlieren photographs of spray images for CNG and H2 respectively. Wide angle injector was used under ambient conditions and injection pressure of 1.2 and 1.8 MPa with the range of time from 0.75 to 6.25 m sec-1. As shown in the spray images, the injection pressure has a large influence on the spray structure, the sprays has basically cone structures under all conditions. However, for the case of low injection pressure (1.2 MPa), a weak cone shape and lower tip penetration were observed. With an increase of the injection pressure to (1.8 MPa), a strong vortex in the spray tip plume was found to develop. Generally, both the penetration and cone angle have shown an increase with the increase of injection pressures. The increase of injection pressure leads to the increase in initial speed of the spray and thus increases in the flow rate of injection.
Figure 6 illustrates hydrogen has a larger cone angle compared to CNG. The wide spray angle of H2 could improve mixing rate of mixture due to the larger distribution area for the injected gas.
Figure 7 is spray cone angle measurement against the time after the start of injection. The spray cone angle was measured at a distance of 20 mm from end of injector nozzle. It shows that, hydrogen has a larger cone angle for the entire injection time whereas CNG have lower cone angles throughout the injection time due to the fact that hydrogen has a lower density than CNG and it can easily mix with air.
The spray tip penetrations of two types of fuels were computed and plotted with respect to their injection time as it is depicted in Fig. 8. The injection pressure of fuels was 1.2 MPa and it is injected into ambient environment.
It is obvious that tip penetration increases as time after start of injection
increases witnessed in Fig. 8. Also it can be seen that the
tip penetration of two types of fuels were found to be similar up to time 0.5
m sec-1 except for time after 0.5 m sec-1 where the penetration
of H2 is higher than CNG.
|| Schlieren image for pure CNG at (1.2-1.8 MPa)
|| Schlieren image for pure H2 at (1.2-1.8 MPa)
||Spray cone angles of CNG and H2 against the time
after start of injection at (1.2 MPa)
||Spray tip penetration of CNG and H2 against the
time after start of injection at (1.2 MPa)
The Schlieren images show the intensity of the injected gas for Narrow Angle
Injector (NAI) was higher than the Wide Angle Injector (WAI) before the time
reached 5.0 m sec-1.
At the time of 6.25 m sec-1, Wide Angle Injector (WAI) showed that the gas had already disappeared while the Narrow Angle Injector (NAI) showed some residual gases, Wide Angle Injector (WAI) has larger spray cone angle than the narrow cone angle and this is a behavior believed to be good for mixing.
The spray images have also shown, the injection pressure has a large influence on the spray structure. Generally, both the penetration and cone angle has increased with the increase of injection pressure.
As for the spray of pure H2, the images showed that the spray spreads faster than the pure CNG and the spray also has a larger cone angle and higher tip penetration due to the lower density of hydrogen relative to CNG.
The Authors of this study would like to gratefully acknowledge Center for Automotive
Research (CAR) and Department of Mechanical Engineering at Universiti Teknologi
Petronas for the provision of lab facilities and financial support for this