ICLASS-2006 Aug.27-Sept.1, 2006, Kyoto, Japan

1. INTRODUCTION

Effervescent atomizers became a commonplace in industrial applications. The technology seems promising, although several operation-related issues are known; in particular spray unsteadiness observed under certain operating conditions. Its negative effects are apparent mainly in combustion process.

The effervescent spray unsteadiness has been documented by several investigators [1-7]. Their results indicate that the spray unsteadiness is connected with two-phase flow inside the atomizer mixing chamber. Spray tends to be unsteady in regimes where big time changes of the void fraction in the two-phase mixture flowing through the exit orifice are detected [7].

Capability of spray unsteadiness quantification is necessary to detect unstable nozzle operation as well as to enable distinguishing unsteady atomizer version in the design phase or to give insight in spray unsteadiness. So far, several methods have been developed to quantify spray unsteadiness. Edwards & Marx’s method [8] uses a comparison of the measured interparticle arrival time distribution with a theoretical distribution function modelled as an inhomogeneous Poisson process. A planar droplet sizing technique, based on combined measurement of Laser Induced Fluorescence intensity and scattered light intensity of spray droplets was used by Domann & Hardalupas [9] to quantify the spray unsteadiness. The mean and rms of liquid volume, surface and Sauter Mean diameter (SMD) were used to evaluate the unsteadiness. An alternative method is based on measurement of two-phase flow pressure fluctuations in the vicinity of the exit orifice. A simplified relation between the pressure fluctuations and liquid flow rate fluctuations at the exit orifice is used [7].

This paper describes another application of a planar imaging technique. We applied a combined PIV-PLIF

technique to a simple effervescent atomizer. Light heating oil was atomized, with air as the atomizing medium. Set of LIF image pairs was used to calculate instantaneous 3D velocity fields and liquid concentration in the plane perpendicular to the spray axis. Measurements were taken over a wide range of liquid flow rates and air gauge pressures. Results show spatial distribution of fluctuation of liquid mass flux and liquid concentration and assessment of the influence of operation conditions upon the spray unsteadiness. The transparent atomizer used here allowed to study the mixing process and internal two-phase flow. Two-phase flow patterns predicted by published two-phase flow map and actual patterns visualized using digital camera are documented here. They are used to explain the relation between behaviour of the mixture inside the atomizer and resulting spray fluctuations. The results acquired here are compared with results of some earlier investigations. 2. EXPERIMENTAL FACILITY

The experimental equipment includes an effervescent atomizer, test bench with fluid supply system and the combined PIV-PLIF system.

Atomizer description. A simplified transparent Plexiglas version of an industrial effervescent atomizer was used for the experiments. Fig. 1 is a view of the nozzle installed in the test bench. This optically accessible atomizer was mainly designed to allow an internal two-phase flow visualization. The liquid (oil) enters the central orifice on top, while the air enters by side orifices and is injected into the liquid, from each side, through a set of 15 holes with the diameter of 1 mm. The internal diameter of the mixing chamber is 8 mm and the length, downstream of the last row of air holes, is 80 mm. The two fluids form a mixture, flow downstream and exit the atomizer through an orifice to

Paper ID ICLASS06-147 EFFERVESCENT ATOMIZER – TEMPORAL AND SPATIAL VARIATION OF SPRAY

STRUCTURE

Jan Jedelsky 1, Miroslav Jicha 2

1Brno University of Technology, Technicka 2, 61669 Brno, Czech Republic, E-mail: jedelsky@fme.vutbr.cz 2Brno University of Technology, Technicka 2, 61669 Brno, Czech Republic, E-mail: jicha@fme.vutbr.cz

ABSTRACT An effervescent atomizer generated spray is studied using combined PIV-PLIF technique. Liquid phase concentration, 3D velocity and thus mass flux distribution in a plane perpendicular to the spray axis is measured. Spatial distribution of temporal fluctuations of the liquid concentration and mass flux are evaluated. Both fluctuations show annular spatial structure with low value near spray axis and maximum close to the spray edge. A transparent Plexiglass atomizer version is used in this study. Internal atomizer two-phase flow patterns are predicted by published maps and actual two-phase flow structure is visualized by a digital camera. The atomizer is operated in the range of air gauge pressure 0.1 – 0.5 MPa and Gas-to-Liquid-Ratio 2 – 50%. This range of operation conditions leads to various two-phase flow patterns from bubbly to dispersed. Results show that in case of high GLR the internal mixture is relatively homogeneous and decreasing GLR under certain value can lead to a poor mixing. Influence of operation conditions as well as influence of internal two-phase flow pattern on the fluctuations of the liquid concentration and mass flux is documented. The operation of the atomizer under low GLR leads to an inhomogeneous mixture inside atomizer mixing chamber and results in unsteady spray. Keywords: Effervescent Atomization, Liquid Concentration, Liquid Mass Flux, Spray Structure, Spray Unsteadiness, Combined PIV-PLIF System

the ambient atmosphere forming a spray. The exit orifice made of brass plate has a diameter of 2.5 mm and a length of 0.7 mm. There is a conical junction with the apical angle of 90° between the orifice and the mixing tube.

Fig. 1: Plexiglas atomizer model.

The atomizer described and used here is based on an

early design with a long mixing chamber. It is necessary to remark that its spray is highly unsteady mainly in case of low GLR values as it will be seen later. This atomizer version is unfit for regular use, but it is suitable for new measurement method application and to observe the influence of operation conditions on the spray steadiness.

Atomizer operation. The atomizer was continuously operated and studied in the vertical downward position of the main axis. The test bench and the fluid supply system description can be found thereinafter. Physical properties of the atomized liquid – light heating oil and the atomizing medium – air are documented in Table 1. The air and oil supplies are controlled separately. Twin-fluid atomizer operation conditions can be basically described by two independent parameters. The air gauge pressure and GLR were chosen in our case. Thus other values, liquid pressure, liquid and air flow rates, are dependent upon them. Since one of the goals of this study was to investigate how the spray unsteadiness depends on the atomizer operation conditions, experiments were performed for several air gauge pressures and GLR values. Both fluids temperatures, gauge pressure and volumetric flow rates were measured. Table 1: Physical Properties of Fluids at Room Temperature.

Fluid µ σ n ρ kg/(m.s) kg/s2 - kg/m3

Light Heating Oil 0.0185 0.0297 1.488 874

Air 1.23 (normal)

Water (ref. medium) 0.0010 0.0727 1.333 1000

Combined stereoscopic PIV-PLIF system. The system lay-out is seen in Fig. 2. Double pulsed Nd:YAG Continuum SLII-10 laser (1-3) equipped with 4th harmonic generator (2) provides two pulses of 265nm radiation with duration of app. 4ns. The laser light passing through mirrors (8) is converted by a light sheet generator (6) to a light sheet of approximately 100mm width and 2.5mm thickness, which defines the measurement plane. The light sheet illuminates the spray in a cross section perpendicular to the spray axis in the distance 152 mm from the nozzle (4) exit. The image capture system consists of two TSI PIVCAM 13-8 cameras with Scheimpflug optics and 28mm lenses with appropriate optical filters (5L, 5R). The cameras are arranged in a stereoscopic setup tilted about 45° from a perpendicular view. The cameras and laser are controlled by PC (7) using synchroniser (9). Both cameras receive two images, each from one of the laser pulses. The double images are processed with a stereo PIV algorithm to provide a 3-component velocity vector field. The same images are also used for LIF intensity field, which represents the liquid density information. A series of 256 images is recorded for each experiment. Data are processed using TSI Insight and MATLAB7 software.

Fig. 2: Combined stereoscopic PIV-PLIF system.

Test bench. A schematic layout of the experimental

facility is shown in Fig. 3. It consists of a gear pump (14) that supplies light heating oil from a main fuel tank (16) through filters (15), control valves (9, 10) and flowmeters (6) into the atomizer (7). A chiller (11) ensures constant oil temperature and hence also its viscosity. The compressed air is delivered, either from the central plant, or from a two stage compressor (1), depending on the required pressure through an air chamber (2), filters (4), control valve, check valve (5) and flowmeters (6) into the atomizer. Spray is collected in a vessel (12) and returned to the main supply tank by the pump (13). The collector is connected to an oil mist separator that keeps the spray zone free of aerosol but does not disturb the spray. The gear pump (14) delivers the oil with a pressure up to 3 MPa. The fuel piping is equipped with a hydraulic shock absorber placed in front of the atomizer. Pressure (3) and temperature (17) readings are

Air entries

Oil entry

Mixing chamber

Spray

taken at the atomizer inlets for both the fuel and air. The pressure measurements are complemented by the pressure difference measurement (8). Uncertainty in fluid pressure measurements is 2 kPa. Uncertainty of fluid pressure difference measurement is 1 kPa. Uncertainty of the fluid temperature measurements is 0.7°C. Uncertainty of the volumetric flow rate measurements is 5% of measured value. Uncertainty values of the different instruments are taken from their technical specification sheets.

Fig. 3: Schematic layout of experimental facility.

3. DATA EVALUATION

Instantaneous space-resolved distributions of liquid mass flux ( )yxmi ,& and concentration ( )yxci , (Fig. 9a) are employed to provide statistical information about the magnitudes of fluctuations expressing the spray unsteadiness in the cross section (x,y) perpendicular to the spray axis. Recorded set of n=256 images is used to calculate the mean image representing the local mean liquid mass flux ( )yxm ,& and concentration ( )yxc , :

( ) ( )∑=

=n

ii yxm

nyxm

1

,1, && , (1)

( ) ( )∑=

=n

ii yxc

nyxc

1

,1, , (2)

see Figs. 9c, 9e. Spatial distribution of root-mean-square of liquid mass flux and concentration fluctuations normalized by the local mean value is as follows:

( ) ( ) ( ) ( )[ ]∑=

−=′n

ii yxmyxm

nyxmyxm

1

2,,1

,1, &&

&& , (3)

( ) ( ) ( ) ( )[ ]∑=

−=′n

ii yxcyxc

nyxcyxc

1

2,,1,

1, . (4)

These quantities serve as a scale of spray unsteadiness level. Instantaneous value of liquid flow-rate through the measurement plane can be computed from the liquid mass flux distribution:

( )[ ]∑ ⋅=

yxii syxmM

,

,&& , (5)

where s is a pixel area. RMS of flow-rate fluctuations normalized by the mean flow-rate M& ′ is useful for global characterisation of the spray unsteadiness. Mean value of normalized mass flux fluctuations:

( )∑ ′=′

yx

yxmm,

,&& (6)

shows an overall level of mass flux fluctuations in the measurement plane. 4. ATOMIZER INTERNAL TWO-PHASE FLOW

As mentioned above, the mixing process and two-phase flow behaviour inside the atomizer mixing chamber has a direct impact on the spray steadiness. Description of the effervescent atomizer internal two-phase flow is usually made by means of published maps for fully developed two-phase flow [10-15]. Another possibility is direct two-phase flow visualization by high-speed cameras. It enables a qualitative description of the two-phase flow patterns and mixing process [4, 11, 12, 15]. A velocity field inside the mixing chamber can be determined by PIV based methods. A PDA technique was employed to describe the bubble diameter distribution and velocity in [16]. An optic, acoustic or electric property of the mixing fluids is used to characterize bubbles distribution, size and void fraction [11]. In this section we document the predicted two-phase flow patterns using a modified Baker’s map and also employ a digital camera made instantaneous images of internal atomizer two-phase flow to document real atomizer behaviour. 4.1 Two-phase Flow Map

The two-phase flow regimes for fully developed flow reported in the literature are often correlated empirically using parameters based on various combinations of fluid flow rates, fluid physical properties and pipe dimensions [17-21]. For known (measured) gas and liquid volumetric flow rates gV& and lV& and the corresponding densities ρg and ρl, the gas (liquid) mass flux gm& ( lm& ) per channel cross-sectional area A is:

AV

m ggg

&&

⋅=ρ

, AVm ll

l

&&

⋅=ρ

. (7, 8)

The frequently used Gas-to-Liquid mass ratio, GLR, reads:

ll

gg

l

g

l

g

VV

mm

MM

GLR&

&

&

&

&

&

⋅

⋅===

ρρ

. (9)

From many published two-phase flow maps [17-21] we

have used modified Baker’s map for vertical downward two-phase flow with transformed coordinates [18] (see Fig. 4). The coordinates in the map are:

ψλψλ ⋅⋅⋅=⋅⋅⋅= GLRkmm

kX xl

gx &

& (10)

and λ

gy

mkY

&= , (11)

where kx and ky are constants needed for conversion from Imperial System to SI system. The other parameters read:

2/1

⋅=

w

l

a

g

ρρ

ρρ

λ (12)

and

3/12

⋅=

l

w

w

l

l

w

ρρ

µµ

σσ

ψ . (13)

It is clear that for constant fluid physical properties (given in Tab. 1) the value of the X coordinate depends on the GLR and on the gas pressure (change of ρg). The Y coordinate depends on the gas mass flux and the gas pressure. The influence of the gas pressure and the gas mass flux (or GLR) can be expressed by a direction of position change in the two-phase flow map. If we assume:

gl m

m&

&1

≅ , (14)

than we obtain the influence of the gas pressure on the angle β of position change measured from the X axis (skipping constants):

[ ]g

g

ggl

gg

g

gp p

mpmm

pmpXpY

tgg

3

2/11

2/3

5.05.0 &

&&

&−=

⋅⋅

⋅−=

∂∂

∂∂=

−−

−

β , (15)

and the gas mass flux influence is

[ ]g

g

gg

g

g

gm p

mpm

pmXmY

tgg 22

3

2/13

2/1 &

&&

&& −=

⋅−=

∂∂

∂∂=

−

−

β . (16)

The gas pressure increase, while keeping gas mass flux constant means an increment of X coordinate and reasonable decrement of Y coordinate, see Fig. 4. The gas pressure increase, while keeping GLR constant means an increment of both X and Y coordinates. The gas mass flux increase (or GLR increase) while pressure is kept constant leads to a decrement of X coordinate and an increment of Y coordinate. 4.2 Specific Two-phase Flow Patterns

The Baker’s map with transformed coordinates for vertical parallel fully developed flow presented in Fig. 4 shows that for very low GLR values (GLR < 0.3%, while pg = 0.2 MPa) the two-phase flow pattern should be bubbly or

nearly plug flow, for low GLR values (0.3% < GLR < 15%) it should be bubble – froth. With GLR increase it should change to annular (15% < GLR < 35%) and further to dispersed flow (GLR > 35%). Decrease of pressure (keeping GLR = 10%) leads to shift from bubble – froth to slug – plug flow, but for pressure range 0.1 - 0.5 MPa in our experiment the position change in the chart is not as significant as in case of GLR change in measured range 0.7-80%.

Fig. 4: Modified Baker’s map for vertical downward two-phase flow with transformed coordinates according to [18] with marked influence of operation conditions on the two-phase flow regimes.

Fig. 5: Predicted two-phase flow patterns for atomizer operation with pg=0.4 MPa and GLR of 0.5%, 3%, 14% and 82%.

It is important to highlight that published two-phase flow maps and also the Baker’s map with transformed coordinates used here are derived for fully developed flow. The two-phase mixture generated inside the effervescent atomizer leaves the nozzle shortly after the intermixture and real two-phase flow patterns can differ from the patterns predicted by published two-phase flow maps [4].

Results of two-phase flow visualization using the digital camera show different two-phase flow patterns acquired under atomizer operation with pg=0.4 MPa and variable GLR of 0.5%, 3%, 14% and 82%. The predicted two-phase flow patterns are marked up in Fig. 5. Actual images of the

two-phase flow can be seen in Fig. 6.

GLR=0.5%: Air penetrates into the liquid through a set of small holes from both sides with low velocity, bubbles ”slide“ along the wall, and air from the downstream holes increases the volume of the bubbles. The two-phase mixture is non-homogenous in the longitudinal direction (Fig. 6a) and it could be a reason for the spray unsteadiness. The observed flow pattern is bubbly-froth, in agreement with Baker’s map.

GLR=3.0%: The air penetrates into the liquid more quickly than in the previous case. The two-phase flow pattern can be described as bubbly-frothy, the mixture downstream the air holes is also composed of large volumes of separated phases (Fig. 6b). The Baker’s map also indicates bubbly-frothy regime.

GLR=14%: The two-phase mixture is more homogeneous than in the previous two cases (Fig. 6c). The observed flow pattern is frothy-annular; the Baker’s map shows an annular pattern.

GLR=82%: The air enters into the mixing tube in small bubbles at almost perpendicular direction towards the tube axis. The liquid travels partly as an annular film on the walls of the tube and partly is distributed in the air which flows in the centre of the tube (Fig. 6d). The two-phase flow is not developing or changing downstream of the air holes. In accordance with the Baker’s map the flow pattern should be dispersed.

For detailed description of the vertical and horizontal two-phase flow inside this effervescent atomizer using high-speed camera see [4].

Fig. 6: Two-phase flow visualization by digital camera. Operation pressure pg=0.4 MPa, a) GLR = 0.5%, b) GLR = 3.0%, c) GLR = 14%, d) GLR = 82%. 5. SPRAY STRUCTURE

Typical 3D velocity field in perspective view, averaged per 256 shots is documented on Fig. 7. Dominant component is the axial component (the nozzle axis is identical to the Z-axis, which is perpendicular to the measurement plane) Averaged image of the Z-axis velocity component is seen on Fig. 8. The droplets have maximum velocity in the spray axis. For air gauge pressure 0.2 MPa and GLR=5% it is 35m/s.

Actual concentration of liquid in the measurement plane ( )yxci , is documented in Fig. 9. The spray consists of

small droplets in case of GLR 5% and higher (Fig. 9a). In case of very low GLR, about 2% also ligaments of the nonatomized liquid mainly near the spray axis may be observed (Fig. 9b). This phenomenon is connected with inhomogeneous two-phase mixture (plug flow) prior to discharge. Time-averaged (mean) liquid mass concentration image ( )yxc , (Fig. 9c) shows its maximum in spray axis. 90% of liquid mass is covered inside circle less than 60mm in diameter. Image of RMS fluctuations of the liquid mass concentration ( )yxc ,′ normalized by local mean liquid mass concentration displayed in Fig. 9d has generally annular shape; the fluctuations are low near the spray axis and increase with the distance from the axis. After reaching a maximum it decreases again. This pattern was also observed in PDPA measurement of spray unsteadiness in [7]. The mean mass flux and mass flux fluctuation distribution in the measurement plane ( )yxm ,& , ( )yxm ,′& (see Figs. 9e, 9f) are analogous to corresponding patterns of the mass concentration distribution.

Influence of atomizer operation conditions on mean values and fluctuations of liquid mass flux and concentration has been studied. The results are correlated with atomizer operation conditions described by air gauge pressure and GLR and also with atomizer internal two-phase flow patterns.

Fig. 10 (left) documents radial distribution of the mean liquid mass flux for air gauge pressure 0.2 MPa. Only half

Air entry

Mixture discharge

Flow direction

a) b) c) d)

Liquidentry

Fig. 7: Typical spatial distribution of mean liquid velocity, 3D vector field in perspective view, p=0.2MPa, GLR=5%.

Fig. 8: Spatial distribution of mean liquid velocity, z-component, p=0.2MPa, GLR=5%.

profiles have been plotted. The data are normalized by their respective maximum value to enable comparison. About 90% of the liquid mass is contained in the area of 44mm in diameter in case of GLR = 2%. Spray cone angle decreases with increasing GLR; more than 90% of liquid mass is included inside circle with 29mm in diameter for GLR = 50%. Results for different air gauge pressure for GLR 5%

are seen on Fig. 10 (right). The pressure change does not show significant influence on the liquid distribution.

Radial profiles of RMS fluctuations of the liquid mass flux normalized by the local mean liquid mass flux are displayed in Fig. 11. The fluctuations have a low value near the spray axis and increase with the distance from the axis. The peak value lies near the spray edge. This specific

Fig. 10: Radial profiles of the normalized mean liquid mass flux ( )yxm ,& for different GLR at pa=0.2 MPa (left) and for different air gauge pressure at GLR = 5% (right).

Fig. 9: a) Typical spatial distribution of instant liquid mass concentration ( )yxci , , p=0.2MPa, GLR=5%, b) Spatial distribution of instant liquid mass concentration with ligaments of nonatomized liquid near spray axis ( )yxci , , p=0.2MPa, GLR=2%, c) Spatial distribution of mean liquid mass concentration ( )yxc , , p=0.2MPa, GLR=5%, d) Temporal fluctuations of liquid mass concentration, local RMS norm. by the local mean ( )yxc ,′ . p=0.2MPa, GLR=5%, e) Spatial distribution of mean liquid mass concentration ( )yxm ,& , p=0.2MPa, GLR=5%, f) Temporal fluctuations of liquid mass concentration, local RMS norm. by the local mean ( )yxm ,′& . p=0.2MPa, GLR=5%.

a) b) c)

d) e) f)

pattern of the unsteadiness was also observed in our PDPA measurement of the spray unsteadiness [7]. Similar results are also published in paper of Luong and Sojka [5]. The spray unsteadiness shows strong influence on GLR (Fig. 11, left). The fluctuations are low in case of high GLR and tend to increase with GLR decrease. This can be observed on the entire radial profile. In case of GLR = 2 % also growth of fluctuations near spray axis is seen. It could be caused by inferior atomization documented by areas of nonatomized liquid as mentioned above. Influence of the air gauge pressure on the fluctuations of liquid mass flux is not as significant with the exception of the results at 0.1 MPa (Fig. 11 right). Increase of the pressure mainly leads to a shift of the fluctuation maximum to higher radial distance from the spray axis. Liquid mass concentration distribution and influence of GLR on this distribution is very similar to the distribution of liquid mass flux and is not plotted here.

Overall values of fluctuations calculated using Eq. (5, 6) are correlated with predicted internal two-phase flow patterns. Increase of GLR leads to a decrease of the value of the normalized liquid mass flux fluctuations ( )yxm ,′&

and corresponding overall value m′& (Fig. 12). RMS of

flow-rate fluctuations normalized by mean flow-rate M& ′ (Fig. 13) also shows decreasing tendency with GLR

increase. This is in good agreement with results acquired in [7]. Increase of air gauge pressure while keeping GLR constant leads to a decrease of value m′& . Value M& ′ does not show a distinct tendency which is also comparable with results in [7].

6. CONCLUSIONS

Spray structure of simple plexiglass effervescent atomizer was studied with respect to two-phase flow patterns in the atomizer mixing chamber. Mean liquid mass concentration and liquid mass flux distribution in plane perpendicular to spray axis show bell shaped profiles with maximum in the spray axis. Normalized fluctuations of the liquid mass concentration and normalized fluctuations of the mass flux distribution in the plane perpendicular to the atomizer axis have annular distribution with a maximum near the spray edge.

Atomizer internal two-phase flow patterns and mixture homogeneity depend significantly on the atomizer operation conditions. Simple effervescent atomizer under study shows an inhomogeneous mixture in case of low GLR. This leads to high level of fluctuations of the liquid mass concentration and fluctuations of the mass flux - an

Fig. 12: Normalized RMS of overall mass flux fluctuations in the measurement plane m′& vs. the two-phase flow pattern.

Fig. 13: RMS of flow-rate fluctuations normalized by mean flow-rate M& ′ vs. the two-phase flow pattern.

Fig. 11: Radial profiles of temporal fluctuations of the liquid mass flux, local RMS normalized by the local mean ( )yxm ,′& for different GLR at pa=0.2MPa (left), for different air gauge pressure at GLR 5% (right).

unsteady spray. The internal two-phase mixture tends to be more homogeneous with increasing GLR and hence also spray fluctuations decrease. Note that our ongoing research shows a strong effect of the atomizer internal design on the spray steadiness and capability to acquire steady spray also in case of low GLR. 7. ACKNOWLEDGEMENT

We greatly acknowledge financial support from projects GA 101/03/P020 and GA 101/06/0750 funded by the Czech grant agency. 8. NOMENCLATURE

A cross-sectional area [m2] c mass concentration [kg/m3] GLR gas-to-liquid ratio by mass [-] kx, ky constants [-] M& mass flow rate [kg/s] m& mass flux [kg/m2 s] n refraction index [-] p pressure [Pa] RMS root-mean-square value s pixel area [m2] SMD Sauter Mean Diameter [µm] V& volumetric flow rate [m3/s] X, Y Cartesian coordinates

Greek letters β angle of position change [deg] λ argument of two-phase flow [-] µ dynamic viscosity [kg/s m] ρ density [kg/m3] Ψ argument of two-phase flow [-] σ surface tension [kg/s2]

subscripts l atomized liquid g atomizing gas a air w water

superscripts

( )’ fluctuating value ( ¯ )’ root-mean-squared fluctuating value ( ¯ ) time-averaged mean value

9. REFERENCES 1. Roesler, T. C. and Lefebvre, A. H., Photographic studies

on aerated-liquid atomization, combustion fundamentals and applications, Proceedings of the Meeting of the Central States section of the Combustion Institute, Indianopolis, Indiana, Paper 3, 1988.

2. Whitlow, J. D. and Lefebvre, A. H., Effervescent Atomizer Operation and Spray Characteristics, Atomization and Sprays, Vol.3, No.2, pp137-155, 1993.

3. Chin, J. S. and Lefebvre, A. H., A Design Procedure for Effervescent Atomizers, Journal of Engineering for Gas Turbines and Power, Vol. 117, pp. 266-271, 1995.

4. Jedelsky, J. and Jicha, M., Characteristics of a

Two-phase Flow Inside the Mixing Chamber of an Effervescent Atomizer, Proceedings of 3rd International Symposium on Two-Phase Flow Modelling and Experimentation, Pisa Italy, 2004.

5. Luong, J. T. K. and Sojka, P. E., Unsteadiness In Effervescent Sprays, Atomization and Sprays, Vol. 9, No. 1, pp. 87-109, 1999.

6. Bush, S.G. and Sojka, P.E., Entrainment by Effervescent Sprays at Low Mass Flowrates. Proceedings of the 6th International Congress on Liquid Atomization and Spray Systems, 609-615, Rouen, France, July, 1994.

7. Jedelsky, J. and Jicha, M., Unsteadiness of Effervescent Atomizer generated sprays. 20th Annual Conference on Liquid Atomization and Spray System (ILASS Europe), pp. 253-258, 5-7 September 2005, Orléans – France.

8. Edwards, C. F. and Marx, K. D., Multi-Point Statistical Structure of the Ideal Spray, Part II: Evaluating Steadiness Using the Interparticle Time Distribution, Atomization and Sprays, Vol. 5, pp. 475-506, 1995.

9. Domann, R. and Hardalupas, Y., Planar Droplet Sizing for Quantification of Spray Unsteadiness, Proceedings of 18th International Conference on Liquid Atomization and Spray Systems - ILASS-Europe, 2002.

10. C.A. Catlin and J. Swithenbank, Physical Processes Influencing Effervescent Atomizer Performance in the Slug and Annular Flow Regimes, 2001, Atomization and Sprays, Vol. 5, pp. 575-595, 2001.

11. M. Lorcher, F. Schmidt and F. Mewes, Flow Field and Phase Distribution Inside Effervescent Atomizers, 9th International Conference on Liquid Atomization and Spray Systems - ICLASS 2003, pp. 12-9. Italy, 2003.

12. J.S. Chin and A.H. Lefebvre, Flow Regimes in Effervescent Atomization, 5TH ILASS AMERICAS, pp. 210-214, 1992.

13. J.S. Chin and A.H. Lefebvre, Flow patterns in internal-mixing, twin-fluid Atomizers, Atomization and Sprays, Vol. 3, pp. 463-475, 1993.

14. J.S. Chin and A.H. Lefebvre, A Design Procedure for Effervescent Atomizers, Journal of Engineering for Gas Turbines and Power, Vol. 117, pp. 266-271, 1995.

15. P.J. Santangelo and P.E. Sojka, A Holographic Investigation of The Near Nozzle Structure of an Effervescent Atomizer Produced Spray, Atomization and Sprays, Vol. 5, pp. 137-155, 1995.

16. G. Brenn, H. Braeske and F. Durst, Investigation of the Unsteady Two-Phase Flow with Small Bubbles in a Model Bubble Column using Phase-Doppler Anemometry, Chemical Engineering Science, Vol. 57, pp. 5143-5159, 2002.

17. P.B. Whaley, Two-phase Flow and Heat Transfer, chap. 3, Oxford university press, 1996.

18. Process Associates of America, http://www.processassociates.com/process/fluid/2faz_xy.htm, 2003.

19. P.L. Golan and A.H. Stenning, Two-phase Flow Vertical Flow maps, Proc. Instn. Mech. Engnrs, Vol. 184 no. 3C, pp. 110-16, 1969-70.

20. T. Oshinowa and M.E. Charles, Vertical Two-Phase Flow, 1: Flow-Pattern Correlations, Can J Chem Engng, Vol. 52, pp. 25-35, 1974.

21. Y. Taitel and A.E. Dukler, A Model for Predicting Flow Regime Transitions in Horizontal and near Horizontal Gas-Liquid Flow, AIChE J., Vol. 22, pp. 47-55, 1976.