**1. Introduction**

58 Advanced Fluid Dynamics

Vejahati, F. 2006. CFD simulation of gas-solid bubbling fluidized bed. Master's thesis,

Veluswamy, Ganesh K., Rajesh K. Upadhyay, Ranjeet P. Utikar, Geoffrey M. Evans, Moses

Vogt, C., R. Schreiber, G. Brunner, and J. Werther. 2005. "Fluid dynamics of the supercritical

Vreman, Bert, Bernard Geurts, and Hans Kuerten. 1997. "Large-eddy simulation of the

van Wachem. 2000. Derivation, Implementation, and Validation of Computer Simulation

Wang, H. G, W. Q Yang, P. Senior, R. S Raghavan, and S. R Duncan. 2008. "Investigation of

Werther, J., B. Hage, and C. Rudnick. 1996. "A comparison of laser Doppler and single-fibre

Wiesendorf, Volker. 2000. *The Capacitance Probe: A Tool for Flow Investigations in Gas-solids* 

Wu, C., Y. Cheng, M. Liu, and Y. Jin. 2008. "Measurement of axisymmetric two-phase flows

Yang, Wen-ching. 2003. *Handbook of fluidization and fluid-particle systems*. New York: Marcel

Ye, S., X. Qi, and J. Zhu. 2009. "Direct Measurements of Instantaneous Solid Flux in a CFB

Zhou, Hao, Guiyuan Mo, Jiapei Zhao, Jianzhong Li, and Kefa Cen. 2010. "Experimental

Zhu, Haiyan, Jesse Zhu, Guozheng Li, and Fengyun Li. 2008. "Detailed measurements of

Zhu, Jesse, Bo Leckner, Yi Cheng, and John Grace. 2005. Fluidized Beds. In *Multiphase Flow Handbook*, ed. Clayton Crowe, 20052445:5-1-5-93. CRC Press, September.

*Fluidization Systems*. Shaker Verlag GmbH, Germany, September 8.

*Technology* 32 (4) (April 1): 580-589. doi:10.1002/ceat.200800361.

fluidized bed." *Powder Technology* 158 (1-3): 102-114.

35 (5) (October): 381-391. doi:16/0255-2701(96)80018-3.

O. Tade, Michael E. Glenny, Shantanu Roy, and Vishnu K. Pareek. 2011. "Hydrodynamics of a Fluid Catalytic Cracking Stripper Using γ-ray Densitometry." *Industrial & Engineering Chemistry Research* 50 (10) (May 18): 5933-

turbulent mixing layer." *Journal of Fluid Mechanics* 339 (May): 357-390.

Models for Gas-Solid Fluidized Beds. Ph.D. Thesis, Delft University of Technology.

batch fluidized‐bed drying by mathematical modeling, CFD simulation and ECT measurement." *AIChE Journal* 54 (2) (February 1): 427-444. doi:10.1002/aic.11406. Wang, R.-C., and Y.-C. Han. 1999. "Momentum dissipation of jet dispersion in a gas-solid fluidized bed." *Journal of the Chinese Institute of Chemical Engineers* 30 (3): 263-271. Wang, Zhengyang, Shaozeng Sun, Hao Chen, Qigang Deng, Guangbo Zhao, and Shaohua

Wu. 2009. "Experimental investigation on flow asymmetry in solid entrance region of a square circulating fluidized bed." *Particuology* 7 (6): 483-490.

reflection probes for the measurement of the velocity of solids in a gas-solid circulating fluidized bed." *Chemical Engineering and Processing: Process Intensification*

by an improved X-ray-computed tomography technique." *Industrial and Engineering* 

Riser using a Novel Multifunctional Optical Fiber Probe." *Chemical Engineering &* 

Investigations on the Performance of a Coal Pipe Splitter for a 1000 MW Utility Boiler: Influence of the Vertical Pipe Length." *Energy & Fuels* 24 (9): 4893-4903.

flow structure inside a dense gas-solids fluidized bed." *Powder Technology* 180 (3):

University of Regina.

5941. doi:10.1021/ie1021877.

doi:10.1017/S0022112097005429.

doi:16/j.partic.2009.07.004.

*Chemistry Research* 47 (6): 2063-2074.

339-349. doi:16/j.powtec.2007.02.043.

Dekker.

doi:10.1021/ef1007209.

Injection of the liquid fuel across the incoming air flow is widely used in gas turbine engine combustors. Thus it is important to understand the mechanisms that control the breakup of the liquid jet and the resulting penetration and distribution of fuel droplets. This understanding is needed for validation of Computational Fluid dynamics (CFD) codes that will be subsequently incorporated into engine design tools. Additionally, knowledge of these mechanisms is needed for interpretation of observed engine performance characteristics at different velocity/altitude combinations of the flight envelope and development of qualitative approaches for solving problems such as combustion instabilities (Bonnel et al., 1971). This chapter provides an introduction and literature review into the subject of cross-flow fuel injection and describes the fundamental physics involved. Additionally highlighted are experimental technique and recent experimental data describing the variables involved in fuel spray penetration and fuel column disintegration.

In recent years, there has been a great drive to reduce harmful emissions of oxides of Nitrogen oxides (NOx) from aircraft engines. One of the several approaches to achieve low emissions is to avoid hot spots in combustors by creating a lean homogeneous fuel-air mixture just upstream of the combustor inlet. This concept is termed as Lean Premixed Prevaporized (LPP) combustion. Creating such a mixture requires fine atomization and careful placement of fuel to achieve a high degree of mixing. Liquid jet in cross flow, being able to achieve both of these requirements, has gained interest as a likely candidate for spray creation in LPP ducts (Becker & Hassa, 2002). Since the quality of spray formation directly influences the combustion efficiency of engines, it is important to understand the fundamental physics involved in the formation of spray.

As seen in Fig. 1, the field of a spray created by a jet in cross flow can be divided into three modes: 1) Intact liquid column, 2) Ligaments, and 3) Droplets. The liquid column develops hydrodynamic instabilities and breaks up into ligaments and droplets (Marmottant & Villermaux, 2004; Madabushi, 2003; Wu et al., 1997). This process is referred to as primary breakup. The location where the liquid column ceases to exist is known as the column breakup point (CBP) or the fracture point. The ligaments breakup further into smaller droplets and this process is called secondary breakup.

The most relevant parameter for drop breakup criterion is the Weber number, *We airUairD fuel* / <sup>2</sup> (in this formula *ρair* and *Uair* - density and velocity of the crossing air respectively, *D* - diameter of the injection orifice and *Ϭfuel* is the surface tension of the fuel).

Fuel Jet in Cross Flow – Experimental Study of Spray Characteristics 61

breakup of the droplets was minimal. The main function of the shock wave is to produce the high speed convective flow that is responsible for the disintegration of droplets. This prompted subsequent researchers to use this characteristic time (t\*) for droplets in subsonic

> \* 1/2 5.0 (/) /

*d u* (1)

*la a*

Lower values of *tb/t\**=3.44 were reported later (Wu et al., 1997) for liquid jet disintegration in the cross flow with Weber number in the range of *We*=71 – 200. The column breakup location for higher *We* flows could not be determined. They also found that the CBP was located at about eight diameters downstream of the orifice in the direction of airflow for the

Other researchers (Sallam et al., 2004) measured column breakup point at *We* range of 0.5- 260. Their studies yielded different value of *tb/t\** = 2.5. However, the uncertainties became high as *We* of the flow was increased. This can be explained by the fact that the experimental methods that have been employed so far for measuring the CBP position involve the analysis of the spray images obtained by back illumination technique. This method works reasonably well for low *We* flows in the absence of shear breakup. In the shear breakup regime, that is relevant for the gas turbine applications it becomes very difficult to analyze the spray images and find the location of CBP because of the presence of droplets in high density around the liquid column. This paper demonstrates a method to overcome this

Method used in the current study was first suggested by (Charalompous et al., 2007) who developed a novel technique to locate the CBP for a co-axial air blast atomizer. In this atomizer high density of droplets around the liquid jet column limited optical access to the jet. To overcome this problem, they illuminated the liquid jet column seeded with fluorescent Rhodamine WT dye with a laser beam from the back of the injector. The liquid jet acted as an optical fiber up to the point it breaks up. The jet is visible due to florescence of the dye until the location of the CBP and the light gets scattered beyond that location giving the precise location of the CBP. The current study aims at extending this technique to locate

Spray penetration into the cross flow have received significant attention by the experimentalists hence placement of fuel in a combustor is significant for its design. In 1990s researchers (Chen et al., 1993, Wu et al.*,* 1997) have carried out experiments at different momentum flux ratios of water jets and developed a correlation of the dependence of the upper surface trajectory of jets in a cross flow with liquid to air momentum flux ratio. Later (Stenzler et al., 2003) a Mie scattering images were used to find the effect of momentum flux ratio, Weber number and liquid viscosity on jet penetration. As in other previous studies, they found that increasing momentum flux ratio increased penetration. Increasing the Weber number decreased the average droplet size and since smaller droplets decelerate faster, the overall penetration of the spray decreased. However, many of these correlations are applicable to specific operating conditions, injector geometries and measurement

It was also found (Tamaki et al., 1998, 2001) that the occurrence of cavitation inside the nozzle significantly influences the breakup of the liquid jet into droplets. The collapse of cavity bubbles increased the turbulence of the liquid jet accelerating its breakup into

*b b*

*t t*

 

*t*

flows as well by.

cases reported.

shortcoming.

techniques.

the CBP of liquid jets in cross flow.

*We* is the ratio of disruptive aerodynamic force to capillary restoring force. The critical *We* above which a droplet disintegrates is *We*=10 (Hanson et al., 1963). When Weber number is high (*We* >200), another mode of breakup called the shear breakup becomes dominant. During shear breakup, aerodynamic forces exerted by the flow on the surface of the liquid jet or ligaments strip off droplets by shear. Though both modes of breakup contribute to atomization of the liquid jet, the domination of one mechanism over the other is dependent on *We* and on liquid jet momentum flux to air momentum flux ratio, *q.*

Fig. 1. Schematic of spray created by a liquid jet in cross flow (from Ann et. al., 2006)

Currently two parameters that characterize disintegration of the fuel jet in the cross flow are subjects of great interest among the users of the experimental data. They are (1) column breakup point (CBP) and (2) penetration of spray into the cross flow. The location of CBP is important for the development of computational models for the prediction of spray behavior. Since the aerodynamic drag for the liquid jet is significantly different from that of droplets, it is crucial to know the exact location of jet disintegration into droplets to be able to predict the extent to which the droplets penetrate into the air stream. On the other hand direct measurements of the spray penetration are significant for development of the design tools for use by the engine developers as well as for validation and adjustment of the spray computational models. Various researchers have measured CBP location and spray penetration with reasonable uncertainties. However, these parameters are still not explored extensively because of ambiguities in definition and due to experimental difficulties. A number of experimental studies of column breakup and spray penetration under conditions that simulate those in gas turbine engines were undertaken and are briefly reviewed below.

In the early work on the aerodynamic breakup of liquid droplets in supersonic flows researchers (Ranger & Nichollas, 1969) carried out experiments to find the time required for individual droplets dropped into a supersonic cross flow to breakup to form a trace of mist. They found this time (*tb*) to be proportional to the droplet diameter (*d*), inversely proportional to the relative velocity between the droplet and the airflow (*ua*), and proportional to the square root of liquid-to-air density ratio ( *l <sup>a</sup>* / ). Based on the images taken, they found that the constant of proportionality (*tb/t\**), defined by equation (1) to be 5. Another conclusion of their study was that the effect of the shock wave on the aerodynamic 60 Advanced Fluid Dynamics

*We* is the ratio of disruptive aerodynamic force to capillary restoring force. The critical *We* above which a droplet disintegrates is *We*=10 (Hanson et al., 1963). When Weber number is high (*We* >200), another mode of breakup called the shear breakup becomes dominant. During shear breakup, aerodynamic forces exerted by the flow on the surface of the liquid jet or ligaments strip off droplets by shear. Though both modes of breakup contribute to atomization of the liquid jet, the domination of one mechanism over the other is dependent

ua

Intact Liquid Column

Droplets

Ligaments

Zb Column Breakup Point

Fig. 1. Schematic of spray created by a liquid jet in cross flow (from Ann et. al., 2006)

engines were undertaken and are briefly reviewed below.

proportional to the square root of liquid-to-air density ratio (

Currently two parameters that characterize disintegration of the fuel jet in the cross flow are subjects of great interest among the users of the experimental data. They are (1) column breakup point (CBP) and (2) penetration of spray into the cross flow. The location of CBP is important for the development of computational models for the prediction of spray behavior. Since the aerodynamic drag for the liquid jet is significantly different from that of droplets, it is crucial to know the exact location of jet disintegration into droplets to be able to predict the extent to which the droplets penetrate into the air stream. On the other hand direct measurements of the spray penetration are significant for development of the design tools for use by the engine developers as well as for validation and adjustment of the spray computational models. Various researchers have measured CBP location and spray penetration with reasonable uncertainties. However, these parameters are still not explored extensively because of ambiguities in definition and due to experimental difficulties. A number of experimental studies of column breakup and spray penetration under conditions that simulate those in gas turbine

In the early work on the aerodynamic breakup of liquid droplets in supersonic flows researchers (Ranger & Nichollas, 1969) carried out experiments to find the time required for individual droplets dropped into a supersonic cross flow to breakup to form a trace of mist. They found this time (*tb*) to be proportional to the droplet diameter (*d*), inversely proportional to the relative velocity between the droplet and the airflow (*ua*), and

taken, they found that the constant of proportionality (*tb/t\**), defined by equation (1) to be 5. Another conclusion of their study was that the effect of the shock wave on the aerodynamic

*l* 

*<sup>a</sup>* / ). Based on the images

on *We* and on liquid jet momentum flux to air momentum flux ratio, *q.*

ul

breakup of the droplets was minimal. The main function of the shock wave is to produce the high speed convective flow that is responsible for the disintegration of droplets. This prompted subsequent researchers to use this characteristic time (t\*) for droplets in subsonic flows as well by.

$$\frac{t\_b}{\text{ft}^\*} = \frac{t\_b}{\left(\rho\_l / \rho\_a\right)^{1/2} d \quad / \text{u}\_a} = 5.0 \tag{1}$$

Lower values of *tb/t\**=3.44 were reported later (Wu et al., 1997) for liquid jet disintegration in the cross flow with Weber number in the range of *We*=71 – 200. The column breakup location for higher *We* flows could not be determined. They also found that the CBP was located at about eight diameters downstream of the orifice in the direction of airflow for the cases reported.

Other researchers (Sallam et al., 2004) measured column breakup point at *We* range of 0.5- 260. Their studies yielded different value of *tb/t\** = 2.5. However, the uncertainties became high as *We* of the flow was increased. This can be explained by the fact that the experimental methods that have been employed so far for measuring the CBP position involve the analysis of the spray images obtained by back illumination technique. This method works reasonably well for low *We* flows in the absence of shear breakup. In the shear breakup regime, that is relevant for the gas turbine applications it becomes very difficult to analyze the spray images and find the location of CBP because of the presence of droplets in high density around the liquid column. This paper demonstrates a method to overcome this shortcoming.

Method used in the current study was first suggested by (Charalompous et al., 2007) who developed a novel technique to locate the CBP for a co-axial air blast atomizer. In this atomizer high density of droplets around the liquid jet column limited optical access to the jet. To overcome this problem, they illuminated the liquid jet column seeded with fluorescent Rhodamine WT dye with a laser beam from the back of the injector. The liquid jet acted as an optical fiber up to the point it breaks up. The jet is visible due to florescence of the dye until the location of the CBP and the light gets scattered beyond that location giving the precise location of the CBP. The current study aims at extending this technique to locate the CBP of liquid jets in cross flow.

Spray penetration into the cross flow have received significant attention by the experimentalists hence placement of fuel in a combustor is significant for its design. In 1990s researchers (Chen et al., 1993, Wu et al.*,* 1997) have carried out experiments at different momentum flux ratios of water jets and developed a correlation of the dependence of the upper surface trajectory of jets in a cross flow with liquid to air momentum flux ratio. Later (Stenzler et al., 2003) a Mie scattering images were used to find the effect of momentum flux ratio, Weber number and liquid viscosity on jet penetration. As in other previous studies, they found that increasing momentum flux ratio increased penetration. Increasing the Weber number decreased the average droplet size and since smaller droplets decelerate faster, the overall penetration of the spray decreased. However, many of these correlations are applicable to specific operating conditions, injector geometries and measurement techniques.

It was also found (Tamaki et al., 1998, 2001) that the occurrence of cavitation inside the nozzle significantly influences the breakup of the liquid jet into droplets. The collapse of cavity bubbles increased the turbulence of the liquid jet accelerating its breakup into

Fuel Jet in Cross Flow – Experimental Study of Spray Characteristics 63

 Factors that vary flow conditions in the experiment inconspicuously for the researcher such as temperature of the crossing air flow which may change the temperature of the

 Turbulence of the core and boundary layer characteristics of the crossing air flow that may significantly influence spray penetration but rarely mentioned by researches. Imaging technique that was used for many years for capturing spray trajectories was static photography that typically captured superposition of sprays on one image due to the fact that time constant of such oscillatory phenomena as liquid jet disintegration in the cross flow is by several orders lower than expose rate of any available camera used

The objective of this study was to investigate the spray trajectories and determine locations of the column break up points (CBP) formed by the Jet-A fuel injected from the injectors of different geometries into a cross flow of air while the above mentioned influencing factors

 Both injectors used in the study that had the same diameter of the orifice and a different shape of the internal path were manufactured using the same equipment and technology. They were installed with orifices openings flush with the air channel wall

Crossing air flow was of the room temperature. Its turbulence level in the core was

 High speed imaging technique (~24,000fps) with spray illumination by the short laser flashes of 30ns duration was used to capture instantaneous images of the spray several times during its movement from maximum to minimum position. That allowed statistically relevant processing of the images and thus extracting information about the

Sprays penetration into the cross flow were investigated using Jet-A fuel for a wide range of momentum flux ratios between *q=5* and *q=100*. Velocity of the air flow was varied to attain Weber numbers in the range of *We=400* to *We=1600*. Air pressure and temperature in the test channel were P=5 atm and T~300K respectively. Column breakups were investigated also at higher air temperature of 550K (in addition to T=300K) and by using water injection in addition to jet fuel experiments in attempt to achieve wider range of non-dimensional

Figure 3 shows a schematic of the experimental setup used to study the injection of a liquid jet from a flat surface into the cross flow of air at elevated pressure. This setup had a plenum chamber, a rectangular air supply channel, a test section with injector under investigation and a pressurized chamber with four 38mm (1.5 inch) thick windows for optical access to

Plenum chamber was 203.2 mm in diameter and 457.2mm long. Two perforated screens were installed at the entrance and at the exit of the plenum to achieve necessary level of turbulence and flow uniformity in the test section. The rectangular supply channel was 62.3mm (2.45 inch) by 43.2mm (1.7 inch) in cross-section and was 304.8mm long. It was equipped with a "bell-mouth" air intake which was connected to the bottom of the plenum chamber to smoothen the air flow. On the other end of the channel four aerodynamically shaped plates were attached to the channel creating a test section with a cross-section 31.75 x

injector and thus surface tension and viscosity of the injected fuel.

in most of experiments.

will be isolated. For this purpose:

parameters.

the spray.

**2. Experimental setup** 

25.4mm (1.25 x 1.00 inch).

(i.e. with no spray well, or cavity).

~4%. Thickness of the boundary layer was ~3mm.

averaged spray trajectories and their RMS values.

droplets. Additional researchers (Ahn et al., 2006) explored the effect of cavitations and hydraulic flip of the orifice internal flow on the spray properties created by a jet in cross flow. They found that while spray trajectories followed the previously obtained correlations (Wu et al., 1997) in absence of cavitations and hydraulic flip, the presence of these phenomena resulted in significant disagreements between the observed trajectories and the ones reported (Wu et al., 1997). Consequently, they concluded that the design of the injector has a significant effect on the spray trajectories.

Practically all previous studies of fuel spray attempted to describe its penetration trajectory into the cross-flow of air in the form of equation that typically incorporate momentum flux ratio of the liquid jet to air flow, <sup>2</sup> <sup>2</sup> / *fuelU fuel airUair q* , Weber number and certain function

that describe shape of the outer edge of the spray. Usually, these equations incorporate a number of empiric coefficients that were obtained by processing experimental data. In spite of availability of dozens of correlations their practical use remains problematic because they all provide different results. Figure 2 shows result of application of different correlations to one spray with *q=20* and *We=1000*.

Fig. 2. Comparison of the spray penetration trajectories (x and z – coordinates in the direction of fuel injection and crossing air flow respectively, d - is diameter of the injection orifice)

It can be observed that the spray penetration trajectories differ from each other to an extent of 100%. Among factors that causes such a big difference the following ones seems to be the most important:

 Design of the injector and its position in the cross flow (i.e. *l/d*, shape and quality of the internal fuel path, presence or absence of the spray well or cavity between the injection orifice and the channel e.t.c).

62 Advanced Fluid Dynamics

droplets. Additional researchers (Ahn et al., 2006) explored the effect of cavitations and hydraulic flip of the orifice internal flow on the spray properties created by a jet in cross flow. They found that while spray trajectories followed the previously obtained correlations (Wu et al., 1997) in absence of cavitations and hydraulic flip, the presence of these phenomena resulted in significant disagreements between the observed trajectories and the ones reported (Wu et al., 1997). Consequently, they concluded that the design of the injector

Practically all previous studies of fuel spray attempted to describe its penetration trajectory into the cross-flow of air in the form of equation that typically incorporate momentum flux

that describe shape of the outer edge of the spray. Usually, these equations incorporate a number of empiric coefficients that were obtained by processing experimental data. In spite of availability of dozens of correlations their practical use remains problematic because they all provide different results. Figure 2 shows result of application of different correlations to

, Weber number and certain function

Stenzler ‐ Mie Lin ‐ Shadowgraph Lin ‐ PDPA Becker ‐ Shadowgraph Wu, Kirk. ‐ PDPA Chen ‐ Mie Tambe ‐ Shadowgraph Wu ‐ Shadowgraph Su ‐ Mie

Bellofiore ‐ Shadowgraph Amighi ‐ Shadowgraph Ragucci ‐ Shadowgraph

0 10 20 30 40 50 60

**z/d**

Fig. 2. Comparison of the spray penetration trajectories (x and z – coordinates in the direction of fuel injection and crossing air flow respectively, d - is diameter of the injection

It can be observed that the spray penetration trajectories differ from each other to an extent of 100%. Among factors that causes such a big difference the following ones seems to be the

 Design of the injector and its position in the cross flow (i.e. *l/d*, shape and quality of the internal fuel path, presence or absence of the spray well or cavity between the injection

has a significant effect on the spray trajectories.

one spray with *q=20* and *We=1000*.

0

orifice and the channel e.t.c).

10

20

**x/d**

orifice)

most important:

30

40

ratio of the liquid jet to air flow, <sup>2</sup> <sup>2</sup> / *fuelU fuel airUair q*


The objective of this study was to investigate the spray trajectories and determine locations of the column break up points (CBP) formed by the Jet-A fuel injected from the injectors of different geometries into a cross flow of air while the above mentioned influencing factors will be isolated. For this purpose:


Sprays penetration into the cross flow were investigated using Jet-A fuel for a wide range of momentum flux ratios between *q=5* and *q=100*. Velocity of the air flow was varied to attain Weber numbers in the range of *We=400* to *We=1600*. Air pressure and temperature in the test channel were P=5 atm and T~300K respectively. Column breakups were investigated also at higher air temperature of 550K (in addition to T=300K) and by using water injection in addition to jet fuel experiments in attempt to achieve wider range of non-dimensional parameters.
