ILASS Americas, 25 th Annual Conference on Liquid Atomization and Spray Systems, Pittsburgh, PA, May 213 Performance of an -Assist Nozzle Using Heated Timothy E. Lane *, Brian R. Collett, and Richard J. Byers Battelle Memorial Institute 55 King Avenue Columbus, OH 4321 USA Abstract assist nozzles are capable of producing a range of sprays with varying mass median diameters by controlling the liquid and air flow rates. For spray configurations that require a set liquid flow rate, the mass median diameter (MMD) of spray is again easily controlled by the air flow rate. However, if the spray design requires a set liquid AND set air flow rate, varying the MMD of the spray becomes difficult. Varying the temperature of the air allows for some control over the MMD of a spray with fixed liquid/air flow rates. During this study, the median droplet size of the spray was measured using an Artium Phase Doppler Interferometer (PDI). Using a Spraying Systems Company PF285 Fluid Nozzle with PA14111 Cap, the mass and air flow rates were fixed for each test. The air flow was measured using a Sierra Smart Trak 2 M1M1 flow meter, which has an accuracy of ±2.%. The droplet size was controlled by varying the temperature of the air using an in-line Farnam Cool Touch CT5 air heater. Increasing the temperature of the air from room temperature to 45K caused the median diameter of the spray to decrease approximately 15%. Increasing the air temperature had less effect on sprays with higher liquid flow rates (larger droplets) than lower liquid flow rates (smaller droplets) for similar air flow rates. It was determined in this testing that increasing the air temperature entering an air-assist nozzle reduces the size of the droplets without impacting the flux of the liquid downstream. * Corresponding author: lanet@battelle.org
ILASS Americas, 25 th Annual Conference on Liquid Atomization and Spray Systems, Pittsburgh, PA, May 213 Introduction assist nozzles are capable of producing a range of sprays with varying MMDs by controlling the liquid and air flow rates. For spray configurations that require a set liquid flow rate, the MMD of the spray is again easily controlled by adjusting the air flow rate. However, if the spray design requires a set liquid AND set air flow rate, varying the MMD of the spray becomes difficult. Again, any system that requires a set liquid flowrate can reduce the droplet size of the spray by increasing the air flowrate. However, in many cases, there is a limit to the amount of air that can be supplied to the nozzle. For example, set liquid flowrates and droplet sizes are needed to generate the appropriate icing clouds, which have air flowrate limitations [1]. To further reduce the droplet size once the maximum air flowrate is reached, increasing the air temperature is a possible solution. During this study, droplet size and velocity was measured using an Artium PDI, designed and constructed by Artium Technologies, Inc. of Sunnyvale, CA. The techniques used by the PDI to measure droplet size and velocity have previously been described in literature [2,3]. The probe volume corrected (PVC) fluxes were also recorded in order to accurately assess the overall spray plume size distribution. The Artium PDI PVC flux measurements have already been shown to have good agreement with traditional mechanical patternation local volume flux measurements [4]. Using a Spraying Systems Company PF285 Fluid Nozzle with PA14111 Cap mounted to a 1/4J nozzle body (SS-1), the mass and air flow rates were fixed for each test. The air flow was measured using a Sierra Smart Trak 2 M1M1 flow meter, which has an accuracy of ±2.%. The droplet size was controlled by varying the temperature of the air using an in-line Farnam Cool Touch CT5 air heater. Test Procedure A two-dimensional Artium Technologies PDI 2MD instrument was used to acquire drop size and velocity measurements across the spray plume. The PDI system was operated in a 1-D orientation for these measurements, resulting in a purely stream-wise velocity component. The transmitter and receiver were mounted on a rail assembly with rotary plates at a 4 degree forward scattering collection angle. The 5 mm lenses were used for both the transmitter and receiver, which allowed for measurement of droplets in the range of 1.5 to 16 µm [5]. The SS-1 nozzle was placed 12 inches from the PDI measurement location. The nozzle was sprayed horizontally while affixed to a traversing system. The nozzle was traversed both vertically and horizontally (always 12 inches from the PDI measurement volume) to fully analyze the spray plume. The center of the spray plume was positioned at,. The nozzle was moved in both the positive and negative x- and y-directions until the edge of the spray plume was reached. The spray plume was also measured at several off-axis locations. For each test configuration, the spray plume was measured at ~25-3 measurement locations. Figure 1 shows a schematic of an example test grid. 2.5 2 1.5 1.5 -.5-1 -1.5-2 -2.5-2.5-2 -1.5-1 -.5.5 1 1.5 2 2.5 Figure 1. Example test grid for spray plume measurements with coordinates in inches. At each measurement location, at least 1, droplets were measured as they passed through the PDI laser intersection. The PVC distribution is used to provide the D V.1, D V.5, and D V.9 diameters as well as the volume flux. The D V.1 diameter is the value where 1% of the total volume sprayed is made up of drops with diameters less than or equal to the D V.1. The D V.5, or volume median diameter (VMD), is the diameter where 5% of the total volume of liquid sprayed is made up of droplets with diameters smaller than the D V.5. Finally, the D V.9 is the value where 9% of the total volume of liquid sprayed is made up of droplets with diameters smaller than the D V.9. The volume flux (cm 3 /cm 2 /s) is a measurement of the liquid volume (cm 3 ) that passes through the probe volume (cm 2 ) per unit time (s). The Sierra Smart Trak flow meter was used to set the air flowrate for each nozzle configuration with and without heated air. The heated air was controlled using an in-line Farnam Cool Touch CT5 air heater. When heated, the air temperature was set to 45K. City water, which was allowed to equilibrate to ambient temperature, was sprayed into an environment with a temperature of approximately 2 C and a relative humidity in the range of 3-5%. Two liquid water flow rates were tested (.11 and.3 lpm) at three different air flowrates. The nozzle operating conditions used for each test are shown in Table 1.
Ch1 Volume Flux (cm/s) ILASS Americas, 25 th Annual Conference on Liquid Atomization and Spray Systems, Pittsburgh, PA, May 213 Data Analysis To determine the overall flux of the entire spray plume, a surface was fit to the volume fluxes measured at each x,y location. The surface was integrated over the measurement range to provide the overall flux, which was then compared to the known liquid flow rate. The D V.1, D V.5, and D V.9 diameters measured at each x,y location were then multiplied by the volume flux measured at the same location. A surface was also fit to the flux*diameter values and integrated over the same range. The resulting values were then divided by the calculated volume flux for the entire spray plume to provide the D V.1, D V.5, and D V.9 diameters for the entire spray. TableCurve 3D Version 4. was used to fit each surface with the best fit equation that provided the highest r 2 value (Equation 1) and F-statistic value (Equation 2). The Sum of Squares due to Error (SSE, or Sum of Residuals Squared) (Equation 3) is the actual leastsquares measure of fit. The z data value (e.g., in Figure 2 z is the measured flux) is and the estimated z value is. In the Sum of Squares about the Mean (SSM) calculation (Equation 4), is the mean of the z data values. In these equations, n is the total number of data points and m is the number of coefficients in the equation. After selecting the surface that maximized both the r 2 and F-statistic values, the equation was then integrated over the range of the measurement data using TableCurve 3D. For consistency, the same equation was used to fit the flux and flux*diameter surfaces for a given test. As an example, Figure 2 shows the best fit equation to the fluxes measured for Test 14. The best fit equation (Equation 239 in TableCurve 3D) is shown in Equation 5, with the four coefficients (a, b, c, and d) listed in Figure 2. (1) (2) (3) (4) (5).25.2.15.1.5 Test 14: Flux, x, y Rank 1 Eqn 239 z=logisticx(a,b,c)*logisticy(1,d,c) r^2=.9826988 DF Adj r^2=.97955313 FitStdErr=.17189 Fstat=435.46252 y 1.5 1.5 -.5 a=.24282579 b=-.14421883 c=.33489921 d=-.2939276-1 -1.5-2 -.5-1 -1.5-2 1.5 1.5 Figure 2. Best fit surface equation for the fluxes measured for Test 14. Results and Discussion With the air heated to 45 K, the nozzle became hot to the touch. To determine if the increased nozzle temperature could be heating the liquid, a simple heat transfer equation was used to calculate the temperature increase in the water. In the worst case scenario, assuming the nozzle walls are 45 K, the water temperature was only expected to increase 8 K and 3 K for the low and high water flowrates, respectively. However, only a temperature of 377 K was measured for the nozzle body. Using this temperature, the water temperature was only expected to increase 4 and 1.5 K for the low and high water flowrates, respectively. The only time the liquid was exposed to the heated air was in the internal mixing chamber of the nozzle. The time each droplet spent in the chamber was minimal, on the order of 1 ms, so the droplets were not expected to increase in temperature within this timeframe. For the worst case scenario where the air and liquid are assumed to come to an equilibrium temperature, the liquid temperature would only increase 2-4 K for the low liquid flowrates and 1-2 K for the high liquid flowrates. However, equilibrium was never achieved based on the flowrates of air and liquid. Again, the temperature was not expected to significantly decrease the density, viscosity, and surface tension of the water. The most significant difference between the nonheated and heated test conditions was the exit velocity of the air. Using the measured mass flowrate (measured flow rate in standard liters per minute multiplied by the density of air at 1 atmosphere and 294 K), the volumetric flowrate of air was calculated using the measured pressures (listed in Table 1) and temperature (45 K). The orifice diameter of the SS-1 air cap was 2.79 mm. The calculated air velocities exiting the nozzle are also listed in Table 1. Clearly, heating the air increases the exit velocity of the air. x.15.1.5.25.2
Droplet Size (µm) Droplet Size (µm) Test # ILASS Americas, 25 th Annual Conference on Liquid Atomization and Spray Systems, Pittsburgh, PA, May 213 Treatment Nozzle Operating Conditions Liquid lpm Liquid Pressure, psig Pressure, psig slpm Exit Velocity, m/s D V.1, μm PDI Measurements D V.5, μm D V.9, μm Flux, gph 1 Heated.11 12 25 113 174 11.6 17.2 25.7.6 2 Heated.11 9 15 83 171 12.8 21.1 33.5.6 3 Heated.11 5.5 1 65 161 15.9 28.8 46.3.8 4 Not Heated.11 14 2.5 113 128 12.1 19.3 29.8.7 5 Not Heated.11 8.5 13 83 12 15.9 27.4 43.2.9 6 Not Heated.11 6 9 65 19 18.9 35.9 62.5.12 7 Not Heated.11 17 25 13 131 11. 17.4 26.6.7 8 Not Heated.11 1 15 92 124 14.7 24.4 38.2.8 9 Not Heated.11 6.5 1 72 116 17.3 32.5 54.2.11 1 Heated.3 4 4 152 17 12.1 2.6 34..17 11 Heated.3 33 3 124 17 15.5 26.2 44.3.17 12 Heated.3 25 2 96 17 18.4 35.9 65.6.27 13 Not Heated.3 37 34.5 152 123 14.8 24.8 41.5.22 14 Not Heated.3 3 25.5 124 123 16. 28.9 51.8.19 15 Not Heated.3 23 17 96 121 2.2 41.1 75.2.28 16 Not Heated.3 4 4 174 127 13.1 21.6 35.7.27 17 Not Heated.3 33 3 139 124 14.7 25.5 43.5.17 18 Not Heated.3 26 2 16 121 18.6 36. 66.6.2 Table 1. Nozzle operating conditions for each test and measured droplet size and flux. 8 7 a) 1.7 gph Liquid Flowrate 8 7 b) 4.8 gph Liquid Flowrate 6 6 5 5 4 4 3 3 2 2 1 1 6 7 8 9 1 11 12 Flowrate (slpm) 9 1 11 12 13 14 15 16 Flowrate (slpm) Figure 3. D V.1 (squares), D V.5 (triangles) and D V.9 (circles) diameters in microns plotted versus the air flowrate (slpm). The red lines correspond to the heated air (45 K) tests and the blue lines correspond to the non-heated air tests. Initially, the heated air was thought to evaporate some of the liquid while in the mixing chamber. The measured flux values appeared to reinforce this notion. However, as the droplets got smaller, the spray plume became denser making the PDI measurements more difficult. The measured flux measurements are most likely inaccurate for the denser sprays and the actual flux is expected to match the liquid flow rates. For the dense sprays, the probably of measuring two or more droplets at the same time increased and the validation decreased. By not measuring every droplet passing through the PDI measurement volume, the flux was expected to decrease. This decrease in flux can be seen for both the heated and non-heated tests. Also, due to the limited exposure time and heat transfer rates, evaporation did not appear to be the rea-
Volume Median Diameter (µm) D V.9 (µm) ILASS Americas, 25 th Annual Conference on Liquid Atomization and Spray Systems, Pittsburgh, PA, May 213 son for the reduced droplet sizes. As shown in Figure 3, the D V.9 diameters decreased more than the D V.1 diameters (the percent difference is also shown in Table 2). If evaporation was playing a role in reducing the droplet size, it should affect the smaller droplets more than the larger droplets. The larger droplets clearly have a larger percent change in diameter as compared to the smaller droplets (Table 2). Again, this is counterintuitive for a reduction in the droplet sizes due to evaporation. The smaller droplets should undergo a higher percent reduction in size than the larger droplets over the same period of time if evaporation where the dominant mode of mass transfer. Liquid gph 1.7 4.8 slpm Percent Decrease in Droplet Size by Heating the (%) D V.1 D V.5 D V.9 113 4 11 14 83 19 23 22 65 16 2 26 152 18 17 18 124 3 9 14 96 9 13 13 Table 2. Percent decrease of D V.1, D V.5, and D V.9 diameters as air temperature increased to 45 K. Since the larger droplets were experiencing a larger percentage of the size reduction, the dynamics of the nozzle were investigated to determine if the increased temperature was causing increased shear stress associated with the higher air velocities. As mentioned earlier, the velocity of the air leaving the nozzle orifice was significantly higher for the heated air (Table 1). While the velocities for each condition look similar, the amount of air is quite different due to the change in air density based on the pressure and temperature. Higher velocities lead to increased energy dissipation rates. The maximum stable droplet size, d max, is proportional to the energy dissipation rate, ε (Lasheras et. al., 1998), which is provided in Equations 6. (6) where ρ and U are the density and velocity of the liquid, l, and gas, g, respectively. D g is the orifice diameter of the nozzle. The D V.9 diameters are compared to the total initial flux of kinetic energy (contained mainly in the gas) per unit total mass (gas plus liquid) in Figure 5 [6]. 8 7 6 5 4 3 2 1.5.1.15.2 [U g3 /(1+ρ l U l /ρ g U g )] -2/5 (s 1.2 m -.8 ) Figure 5. D V.9 diameters in microns plotted versus the flux of total kinetic energy of the air jet per unit total mass of the system (m -.8 s 1.2 ). The red lines correspond to the heated air (45 K) tests and the blue lines correspond to the non-heated air tests. The circles represent the 1.7 gph liquid flowrate data and the triangles represent the 4.8 gph liquid flowrate data. 45 4 35 3 25 2 15 1 5 6 8 1 12 14 16 18 Flowrate (slpm) Figure 6. D V.5 diameters in microns plotted versus the air flowrate (slpm). The red lines correspond to the heated air (45 K) tests and the blue lines correspond to the non-heated air tests. The circles represent the 1.7 gph liquid flowrate data and the triangles represent the 4.8 gph liquid flowrate data. Similar to Figure 5, Figure 6 shows that the nonheated air tests follow the same trends for both the high and low liquid flowrates. Increasing the temperature shifts the curve down to lower droplet sizes. Also, there was a larger reduction in the size of the larger droplets than the smaller droplets. Due to the short exposure time to the heat, the droplets are not expected heat-up or evaporate. The reduction in the MMDs for the heated air tests was caused by any increase in the exit velocity of the air which reduced the size of the droplets.
ILASS Americas, 25 th Annual Conference on Liquid Atomization and Spray Systems, Pittsburgh, PA, May 213 When an application requires a fixed air and liquid flowrate, it is possible to reduce the size of the droplets by increasing the temperature of the air supplied. By heating the supply air, the exit velocity of the air increases to further reduce the size of the droplets in the spray plume. Acknowledgements The authors would like to thank Darrell Wright and Luis Piñeiro of U.S. Army Program Executive Office for Simulation, Training, and Instrumentation (PEO STRI) for encouraging and supporting this work. References 1. Byers, R. J.; Lane, T. E.; Collett, B. R. 25 th Annual Conference on Liquid Atomization and Spray Systems, Spray Interaction of Two Atomizing Nozzles in High Velocity, Pittsburgh, PA, May 213. 2. Bachalo, W. D. (198). A Method for Measuring the Size and Velocity of Spheres By Dual Beam Light Scatter Interferometry, Applied Optics 19 (3): 363-37. 3. Bachalo, W. D. and Houser, M. J. (1984). Phase Doppler Spray Analyzer for Simultaneous Measurements of Drop Size and Velocity Distributions, Optical Engineering 23 (5): 583-59. 4. Bade, K. M. and Schick, R. J. (211). Phase Doppler Interferometry Volume Flux Sensitivity to Parametric Settings and Droplet Trajectory, Atomization and Sprays 21 (7): 537-551. 5. Artium Technologies, Inc. (212). PDI-2 MD User Manual, Sunnyvale, CA 9486. 6. Lasheras, J. C.; Villermaux, E.; Hopfinger, E. J. (1998). Break-up and Atomization of a Round Water Jet by a High-Speed Annular Jet, Journal of Fluid Mechanics 357: 351-379.