Content uploaded by Anubhav Sinha
Author content
All content in this area was uploaded by Anubhav Sinha on Oct 08, 2016
Content may be subject to copyright.
ICLASS 2015, 13th Triennial International Conference on Liquid Atomization and Spray Systems, Tainan, Taiwan, August 23-27, 2015
1
Spray in preheated crossflow – Effect of air density
A. Sinha*1, S. Prakash R.2, and R.V. Ravikrishna1
1Department of Mechanical Engineering, Indian Institute of Science, Bangalore, India
2Department of Aerospace Engineering, Indian Institute of Science, Bangalore, India
*Corresponding author email – er.anubhav@gmail.com
Abstract
The present study focuses on an airblast spray of ethanol injected into a crossflow of preheated air. Air tem-
perature is varied from 295 to 418 K. Correlation for the spray trajectory is derived for the wind-ward spray
edge. All previously reported studies have emphasized on the role of momentum ratio on the spray trajectory,
whereas the importance of air density on spray atomization has not been addressed, possibly due to lack of ex-
perimental data. In this study, for a constant momentum ratio at different crossflow temperatures, the spray tra-
jectory has been measured. From this data, a general correlation is derived which accounts for the density of the
crossflow air by adding a correction factor to the correlation. This correlation is found to be in reasonable
agreement with experimental results.
In the spray in crossflow configuration, used in present study, droplet diameters and velocities are measured
downstream of the atomizer at various locations. It is observed that droplet size distribution is shifting towards
smaller diameters. This decrease in droplet size is attributed to droplet evaporation. Droplets horizontal velocity
is also found to increase as the droplets travel downstream. This behavior is attributed to the drag force exerted
by the crossflow air which accelerates the droplets till they attain crossflow air velocity.
Keywords: Airblast spray, spray trajectory, preheated crossflow, droplet evaporation.
Introduction
Spray in crossflow is a phenomenon of recent interest for its applicability on gas turbine combustors. This
configuration is found to be useful in previous studies on agricultural sprays [1-3], spray painting [4, 5] and vari-
ous other engineering processes [6, 7]. The applicability of this configuration to gas turbine combustors, particu-
larly for low NOx combustion was realized first by Leong et al. [8, 9] and since then there have been a few stud-
ies on this aspect. Leong et al. [8, 9] carried out experimental studies on airblast spray subjected to a crossflow
using laser-based diagnostics and derived a correlation for the spray trajectories. Their main focus was on the
effect of ambient pressure on the spray trajectory; and they introduced an extra term to account for pressure vari-
ation. Although this study was first ever study on spray trajectory in a crossflow, it was limited to a single liquid
(Jet-A) and over a small range of operating conditions. Sinha et al. [10] conducted experiments on water and
ethanol sprays in the presence of a crossflow and derived a general trajectory equation which incorporated liquid
physical properties and was found to predict trajectories for both water and ethanol sprays. The generalized cor-
relation was also found to be in reasonable agreement with the Jet-A results of Leong et al. [8, 9]. Although these
two studies focused on deriving correlations from experimental data for spray trajectory, both these studies used
crossflow air at ambient conditions. Among the very few studies on preheated crossflow, Lynch et al. [11] car-
ried out an experimental study on a pressure-swirl atomizer injecting JP-8 into a crossflow of preheated air. They
obtained spray structure images using laser sheet illumination. Coherent structure velocimetry and PDPA tech-
niques were used to measure droplet diameters and velocities. Brandt et al. [12] conducted an experimental study
on kerosene spray injected into a preheated crossflow using a prefilming airblast atomizer. The crossflow was
maintained at elevated pressure of 9 bar, temperature of 750 K and a velocity of 120 m/s for the baseline case.
Air pressure was varied from 3 bar to 14.5 bar, and temperature was varied from 650 K to 850 K. The PDA
technique was used to measure droplet size and velocity. There are a few experimental studies on jet in crossflow
(JICF) under high temperature and high pressure conditions [13-16]. Comparison with a JICF trajectory correla-
tion is made with the spray correlation derived in the present work in results section. It has been observed that
the general form of equation, having dependence only on momentum ratio is not adequate for preheated cross-
flow for JICF studies. The reason is that the lower density of the preheated air leads to a decrease in Weber
number, which affects the secondary atomization and in turn the spray trajectory. Hence, it becomes important to
conduct experiments for preheated crossflow and assess the importance of air temperature. To the best of au-
thors’ knowledge, there is no reported literature available providing any correlation for predicting spray trajecto-
ry in the presence of a preheated crossflow. The objective of the present work is to extend the analysis done in
the previous work [10] to preheated crossflow, and derive a correlation which accounts for temperature change
ICLASS 2015, 13th Triennial International Conference on Liquid Atomization and Spray Systems, Tainan, Taiwan, August 23-27, 2015
2
in its formulation. The next few sections describe the experimental setup and imaging techniques, results on the
spray trajectory and droplet evaporation, conclusions and suggestions for future work.
Experimental Setup and Imaging Techniques
A schematic of the experimental facility used in this study is shown in Fig. 1. It consists of three air tanks
(with a total capacity of 3000 liters) filled with compressed air, which is passed through a 15-kW electrical heat-
er before entering the experimental setup. The volume of air entering the heater is controlled by a pressure regu-
lator, which is responsible for maintaining a constant air flow rate during the experiment. Air velocity is meas-
ured before and after experiments and is observed to remain unchanged.
The setup is made of a divergent section, a settling chamber and a rectangular test section. The test sec-
tion consists of metallic plates for top and bottom walls and quartz plates for the side walls. An airblast injector
(Make: Spraying systems, Model: SU11) is flush fitted to the bottom wall. Nitrogen is used as the airblast gas
and is fed from pressurized vessel as shown in Fig. 1. Ethanol is also kept in a pressurized vessel and is supplied
to the injector through a mass flow controller.
Spray structure is obtained using shadowgraphy technique, wherein a laser beam incident on a diffuser plate
forms a bright background illumination and resulting images are captured using a CCD-camera. A large number
of instantaneous images are captured and then processed suitably to obtain trajectory points. For droplet size and
velocity measurements Particle/ Droplet Imaging Analysis (PDIA) and Particle Tracking Velocimetry (PTV)
techniques are used. More details about the experimental setup and imaging techniques used can be found else-
where [10].
Figure 1. Schematic of experimental facility
Results
The results are presented under two sections: (i) Spray trajectory and (ii) Droplet size and velocity meas-
urements.
(i) Spray trajectory
The objective of the present set of experiments is to access the effect of temperature on the spray trajectory
equation. In previous studies on spray and jet in crossflow under ambient conditions, it is observed that the spray
trajectory depends upon momentum ratio [8-10]. To isolate the effect of momentum ratio, and temperature, it
was decided to maintain the same crossflow momentum for all the cases studied in this study for different tem-
peratures. Although the air density of crossflow air changed with change in temperature, the momentum was
Compressed
Air
Test Section
Liquid
cylinder
Nitrogen
cylinder
Converging
section
Injector
Settling
Chamber
Diffuser
Y
X
Heater
Pressure
regulator
ICLASS 2015, 13th Triennial International Conference on Liquid Atomization and Spray Systems, Tainan, Taiwan, August 23-27, 2015
3
maintained constant by varying the crossflow air velocity proportionately. The momentum change in the study is
effected by varying the spray momentum while maintaining a fixed crossflow momentum. The details of exper-
imental conditions maintained for various cases are given in Table 1.
liquid (g /s)
0.4
GLR
0.1- 0.2
Temperature (K)
293 - 418
q2
3.2 – 10.9
Table 1. Conditions used in the experimental study
A large number of instantaneous images (300-500) are used to study spray structure and trajectory. Some typical
instantaneous images are shown in Fig. 2. These instantaneous images are binarized using the Otsu’s algorithm
[17]. These binarized images are further added and processed to obtain spray trajectory points. A processed im-
age is shown in Fig. 3. More details about the image processing can be found in [10].
(a) T=393 K, q2 = 3.2, GLR=0.1
(b) T=393 K, q2 = 10.9, GLR=0.2
Figure 2. Instantaneous images of ethanol spray in crossflow for typical conditions
Figure 3. Processed image of ethanol spray in crossflow obtained using 300 instantaneous images shown in Fig.
2. This image corresponds to the conditions for the case shown in Fig. 2(a), for T=393 K, q2 = 3.2, GLR=0.1.
The trajectory equation obtained from previous cold flow studies [10] is of the form:
(1)
where q2 represents the two phase momentum ratio, X and Y are the horizontal and vertical locations (cf. Fig. 1),
d is the diameter of the injector orifice, σliq is the surface tension of the test liquid and σwater is the surface tension
of water. It shows that the spray trajectory in a crossflow is primarily dependent on the two phase momentum
ratio (q2), which is defined by Leong et al. [9]. The momentum of airblast air is also taken into account in defin-
ing q2 which makes it different from momentum ratios used in jet in crossflow studies. Moreover, the spray tra-
jectory is also dependent on the surface tension on the liquid and air interface. Spray of liquid with higher sur-
face tension will penetrate further than a liquid with lower surface tension, as per the correlation given above
(Eq. 1); provided all other parameters are maintained constant. This can be attributed to the fact that the liquid
will lower surface tension will have a higher Weber number and hence will produce smaller droplet diameters
after secondary breakup. Smaller droplets tend to get entrained faster and follow the crossflow velocity direction.
20 mm
ICLASS 2015, 13th Triennial International Conference on Liquid Atomization and Spray Systems, Tainan, Taiwan, August 23-27, 2015
4
Hence, the spray of liquid with lower surface tension will have a smaller penetration. For further details on the
trajectory behavior in cold flow, the reader can refer to [10].
On the basis of the above equation (Eq. 1) and a multivariable regression analysis [18] using the temperature
variation used, and trajectory points obtained from the present set of experiments, a new trajectory correlation is
obtained:
(2)
(a) T=293 K, q2 = 3.2, GLR=0.1
(b) T=293 K, q2 = 10.9, GLR=0.2
(c) T=373 K, q2 = 3.2, GLR=0.1
(d) T=373 K, q2 = 10.9, GLR=0.2
(e) T= 418 K, q2 = 3.2, GLR=0.1
(f) T=418 K, q2 = 10.9, GLR=0.2
Figure 4. Comparison of proposed trajectory correlation with experimental data for ethanol at different tem-
peratures
ICLASS 2015, 13th Triennial International Conference on Liquid Atomization and Spray Systems, Tainan, Taiwan, August 23-27, 2015
5
It is to be noted that the effect of temperature is accounted for by using an additional term. In this term, T is
the crossflow air temperature; Tref is the reference temperature of 300 K. Since increase in temperature lowers
air density, Weber number and consequently degree of secondary atomization, increase in temperature is ex-
pected to increase spray penetration. The predicted trajectory using the above derived correlation (Eq. 2) and the
experimentally obtained trajectory points are compared in Fig. 4. It is observed that the correlation is able to give
a reasonable match for a wide range of crossflow air temperatures and momentum ratios. Figure 5 shows the
comparison of experimentally obtained trajectory for decane and Jet-A1 with crossflow air at 145K. This com-
parison further supports the applicability of the derived correlation (Eq. 2) and extends it usage to multi-
component liquid sprays.
(a) Decane, T= 418 K, q2 = 3.3, GLR=0.1
(a) Jet-A1, T= 418 K, q2 = 3.3, GLR=0.1
Figure 5. Comparison of proposed trajectory correlation with experimental data for Decane and Jet-A1
sprays at 418K temperature
In another related experimental study [19], this equation (Eq. 2) is also found to give reasonably good pre-
diction for other liquids over a wide range of liquid properties and operating conditions, thus validating the ap-
plicability of this equation for spray in crossflow configuration. It would be interesting to compare the present
correlation derived for spray in crossflow to that correlations used for jet in crossflow studies. Elsamian et al. [13]
have formulated the following equation:
(3)
Assuming that the Reynolds number and momentum ratios are same for both jet and spray experiments, it ap-
pears from the power of x, that the spray trajectory will be steeper as compared to jet trajectory. Moreover, the
dependence of trajectory is momentum ratio is more significant in the case of jet than the spray. Further investi-
gation is required to compare both the configurations and formulate a general correlation common for both.
(ii) Droplet size and velocity measurements
Droplet sizes and velocities are measured using PDIA and PTV techniques, as described in the previous sec-
tion. Measurements are taken at two stations (P1 and P2). The locations of these stations in the test section are
shown in Fig. 6. The conditions used for droplet size and velocity measurements are summarized in Table 2.The
measurement stations are located in the mid plane of the test section. It is to be noted that the droplets are en-
trained into the crossflow and have lost most of their initial vertical velocity by the time they reach locations P1
and P2, which has been confirmed by the velocity measurements (not shown here). Hence, it can be assumed
that the same set of droplets is observed at P1 and P2. The droplet sizes are thus expected to reduce from P1 to
P2 due to evaporation. As observed, this is indeed the case, the Sauter mean diameter (SMD) of droplets are 32.2
µm at P1 and 21.5 µm at P2. The droplet diameter distribution at P1 and P2 is shown in a form of normalized
probability distribution function (PDF) in Fig. 7 and the droplet horizontal velocity at those points is shown in
Fig. 8.
ICLASS 2015, 13th Triennial International Conference on Liquid Atomization and Spray Systems, Tainan, Taiwan, August 23-27, 2015
6
liquid (g /s)
0.4
GLR
0.1
Temperature (K)
373
Crossflow air velocity (m/s)
27
q2
3.18
Table 2. Experimental conditions used for the droplet diameter and velocity measurements.
Figure 6. Locations of stations for droplet size and velocity measurements.
From P1 to P2, it can be observed that the PDF is shifted towards smaller diameter values showing the effect
of droplet evaporation. Another reason for decrease in droplet diameters could be due to secondary breakup. The
measured droplet sizes and velocities are used to calculate the droplet aerodynamic Weber number. It is found
that the droplet Weber numbers, for the present set of conditions are smaller than 0.1 (cf. Fig. 9), and hence are
much smaller than the critical value of Weber number (12), required for secondary breakup initiation. Thus, it
can be inferred that the decrease in droplet diameters is due to evaporation and not due to secondary breakup.
The low values of Weber number can be attributed to two factors. First, the air density has decreased as it is pre-
heated. Secondly, as the droplets travel downstream, their horizontal velocity increases and hence the relative
velocity with the crossflow air decreases which also reduces Weber number. This increase in velocity is more
prominent for the smallest droplets which attain a velocity very close to the crossflow velocity. For the larger
droplets, the velocity is increasing from P1 to P2 (cf. Fig. 6) due to drag force exerted by the crossflow air, how-
ever, it has not yet attained the crossflow velocity by location P2. The phenomenon of smaller droplets attaining
the crossflow velocity earlier than the larger droplets is observed also in previous experimental studies [10]. The
reason for this behavior is attributed to the larger momentum to drag ratio of larger droplets (cf. Fig. 17 of [10]).
Conclusions
The present study presents experiments carried out on airblast spray of ethanol injected into a crossflow of
preheated air. The first part of the study is focused on spray structure, particularly the spray trajectory in the
presence of crossflow. It is found that the decrease is air density caused by the increase of temperature affects the
spray trajectory. To account for the effect of temperature, experiments are conducted at different temperatures
but constant crossflow air momentum. The crossflow air momentum is maintained same by changing the air ve-
locity proportionately with the change of temperature. The variation in q2 is brought by the variation in spray
momentum. A generalized correlation which accounts for crossflow air temperature and liquid physical proper-
ties is derived. The effect of temperature is suitably accounted for by incorporating a term for temperature in-
crease in the equation. The predictions of the correlation are found to give reasonably good match with experi-
mental results. The second part of this study is focused on droplet size and velocity measurements, which are
conducted at two downstream locations. The effect of droplet evaporation is evident from the droplet SMD and
normalized PDF measured at these two locations. The phenomenon of smaller droplets attaining the crossflow
velocity earlier than the larger droplets is clearly quantified, and the detailed droplet size and velocity data serve
as a useful validation dataset for simulations of droplet evaporation under convective conditions.
Injector
Air
Flow
250 mm
25 mm
P1
P2
50 mm
40 mm
95 mm
ICLASS 2015, 13th Triennial International Conference on Liquid Atomization and Spray Systems, Tainan, Taiwan, August 23-27, 2015
7
Figure 7. Droplet normalized PDFs at various distances from the injector.
Figure 8. Droplet horizontal velocity (along the
crossflow direction)
Figure 9. Droplet Weber number at x=40 mm and
X=95 mm from the injector.
Acknowledgements
The authors acknowledge the financial support provided by the GATET programme of the Aeronautics
Research & Development Board (ARDB), Government of India, for this work.
ICLASS 2015, 13th Triennial International Conference on Liquid Atomization and Spray Systems, Tainan, Taiwan, August 23-27, 2015
8
References
[1] Ghosh, S. and Hunt, J.C.R., Spray Jets in a Cross Flow, Journal of Fluid Mechanics, Vol. 365, no.5, pp.
109-136, 1998.
[2] Philips, J.C., and Miller, P.C.H., Field and Wind Tunnel Measurements of the Airborne Spray Volume
Downwind of Single Flat-Fan Nozzles, Journal of Agriculture Engineering Research, vol. 72, pp. 161-170,
1999.
[3] Philips, J.C., Miller, P.C.H., and Thomas, N.H., Air Flow and Droplet Motions Produced by the Interaction
of Flat-Fan Sprays in Cross Flows, Atomization and Sprays, vol. 10, pp. 83-103, 2000.
[4] Li, L.K.B., Green, S.I., Davy, M.H., and Eadie, D.T., Viscoelastic Air-blast Sprays in a Cross-Flow. Part 1:
Penetration and Dispersion, Atomization and Sprays, vol. 20, pp. 697-720, 2010.
[5] Li, L.K.B., Green, S.I., Davy, M.H., and Eadie, D.T., Viscoelastic Air-blast Sprays in a Cross-Flow. Part 2:
Droplet Velocities, Atomization and Sprays, vol. 20, pp. 721-735, 2010.
[6] Bai, B., Zhang, H., Liu, L., and Sun, H., Experimental Study of Turbulent Mixing of Spray Droplets in a
Crossflow, Experimental Thermal and Fluid Science, vol. 33, pp. 1012-1020, 2009.
[7] Zhang, H., Bai, B., Liu, L., Sun, H., and Yan, J., Droplet Dispersion Characteristics of the Hollow Cone
Sprays in Crossflow, Experimental Thermal and Fluid Science, vol. 45, pp. 25-33, 2013.
[8] Leong, M.Y., McDonell, V.C. and Samulesen, S., Mixing of an Airblast Atomized Fuel Spray Injected into
a Crossflow of Air, NASA/CR-2000-210467, 2000.
[9] Leong, M.Y., McDonell, V.C. and Samulesen, S., Effect of Ambient Pressure on an Airblast Spray Injected
into a Crossflow, Journal of Propulsion and Power, vol. 17, no. 5, 2001
[10] Sinha, A., Surya Prakash, R., Madan Mohan, A., and Ravikrishna, R.V., Airblast Spray in Crossflow -
Structure, Trajectory and Droplet Sizing, International Journal of Multiphase Flows , vol. 72, pp. 97-111,
2015.
[11] Lynch, A., Batchelor, R.G., Kiel, B., Miller, J., Gord. J., and Reeder, M., Spray Characteristics of a Pres-
sure-Swirl Fuel Injector Subjected to a Crossflow and a Coflow, Atomization and Sprays, vol. 21, pp. 625-
643, 2011.
[12] Brandt, M., Rchner, M., and Schmitz, G., An Experimental and Numerical Study of Kerosine Spray Evap-
oration in a Premix Duct for Gas Turbine Combustors at High Pressure, Combustion Science and Technol-
ogy, vol. 138, issue 1, pp. 313-348, 1998.
[13] Elsamian, M., Amighi, A., and Ashgriz, A., Atomization of Liquid Jet in High-Pressure and High-
Temperature Subsonic Crossflow, AIAA Journal, Vol. 52, No. 7, 2014, pp. 1374-1385
[14] Becker, J., and Hassa, C., “Breakup and Atomization of Kerosene Jet in Crossflow at Elevated Presure,
Atomization and Sprays, Vol. 12, Nos. 1–3, 2002, pp. 49–67.
[15] Bellofiore, A., Cavaliere, A., and Ragucci, R., “Air Density Effect on the Atomization of Liquid Jets in
Crossflow,” Combustion Science and Technology, Vol. 179, Nos. 1–2, 2007, pp. 319–342.
[16] Surya Prakash, R., Sinha, A., Raghunandan, B.N., Tomar, G., and Ravikrishna, R.V., Breakup of Volatile
Liquid Jet in Hot Cross Flow, Procedia IUTAM (in press)
[17] Otsu, N., A Threshold Selection Method from Gray-level Histograms, IEEE Trans. Sys., Man., Cyber, vol.
9 (1), pp. 62–66, 1979.
[18] Ross, S.M., Introduction to Probability and Statistics for Engineers and Scientists, Academic Press, Fourth
edition, 2009.
[19] Sinha, A., Surya Prakash, R., Madan Mohan, A., and Ravikrishna, R.V., Studies on Evaporation of Fuel
Droplets under Forced Convection Using Spray in Crossflow Methodology, Fuel (under consideration)