SPRAY STRUCTURE OF A PRESSURE-SWIRL ATOMIZER FOR COMBUSTION APPLICATIONS

In the present work, global as well as spatially resolved parameters of a spray produced by a pressure-swirl atomizer are obtained. Small pressure-swirl atomizer for aircraft combustion chambers was run on a newly designed test bench with Jet A-1 kerosene type aviation fuel. The atomizer was tested in four regimes based on typical operation conditions of the engine. Spray characteristics were studied using two optical measurement systems, Particle Image velocimetry (PIV) and Phase-Doppler Particle Analyzer (P/DPA). The results obtained with P/DPA include information about Sauter Mean Diameter of droplets and spray velocity profiles in one plane perpendicular to the spray axis. Velocity magnitudes of droplets in an axial section of the spray were obtained using PIV. The experimental outputs also show a good confirmation of velocity profiles obtained with both instruments in the test plane. These data together will elucidate impact of the spray quality on the whole combustion process, its efficiency and exhaust gas emissions. Pressure-swirl atomizers as relatively old type of atomizing devices are nowadays often being replaced in many applications by twin-fluid atomizers. But they are still very common parts of present combustion systems mainly for low power demands. Their popularity is based on simple design and operation without additional expensive devices that could lead to unwanted increase of weight in mobile applications and also to reduction of reliability, which are important factors not only in aircraft industry. Research works focused on improvement of atomization characteristics of the pressure-swirl atomizers are persistent despite long-lasting history of their development and utilization in many industrial sectors. Today research effort stems from changes in the legislative, reflects more frequent usage of less refined fuels and answers requirements for more efficient combustion devices. In general, a swirl-flow of the liquid in a pressure-swirl atomizer is induced by feeding the liquid into a swirl chamber through one or several tangential ports, that give it high angular velocity, thereby creating an air-cored vortex. In this manner, the air-core blocks a part of the nozzle outlet orifice. Under both axial and radial forces emerges the fuel through this orifice in the form of a hollow conical sheet. As the sheet expands, its thickness decreases and it soon becomes unstable and disintegrates into ligaments and then drops in the form of a well-defined hollow-cone spray. Disintegration of the sheet depends mainly on the liquid discharge velocity and thus on the liquid injection pressure. Description of the spray development with increasing injection pressure is presented, for


INTRODUCTION
Abstract: In the present work, global as well as spatially resolved parameters of a spray produced by a pressure-swirl atomizer are obtained.Small pressure-swirl atomizer for aircraft combustion chambers was run on a newly designed test bench with Jet A-1 kerosene type aviation fuel.The atomizer was tested in four regimes based on typical operation conditions of the engine.Spray characteristics were studied using two optical measurement systems, Particle Image velocimetry (PIV) and Phase-Doppler Particle Analyzer (P/DPA).The results obtained with P/DPA include information about Sauter Mean Diameter of droplets and spray velocity profiles in one plane perpendicular to the spray axis.Velocity magnitudes of droplets in an axial section of the spray were obtained using PIV.The experimental outputs also show a good confirmation of velocity profiles obtained with both instruments in the test plane.These data together will elucidate impact of the spray quality on the whole combustion process, its efficiency and exhaust gas emissions.
Pressure-swirl atomizers as relatively old type of atomizing devices are nowadays often being replaced in many applications by twin-fluid atomizers.But they are still very common parts of present combustion systems mainly for low power demands.Their popularity is based on simple design and operation without additional expensive devices that could lead to unwanted increase of weight in mobile applications and also to reduction of reliability, which are important factors not only in aircraft industry.Research works focused on improvement of atomization characteristics of the pressure-swirl atomizers are persistent despite long-lasting history of their development and utilization in many industrial sectors.Today research effort stems from changes in the legislative, reflects more frequent usage of less refined fuels and answers requirements for more efficient combustion devices.In general, a swirl-flow of the liquid in a pressure-swirl atomizer is induced by feeding the liquid into a swirl chamber through one or several tangential ports, that give it high angular velocity, thereby creating an air-cored vortex.In this manner, the air-core blocks a part of the nozzle outlet orifice.Under both axial and radial forces emerges the fuel through this orifice in the form of a hollow conical sheet.As the sheet expands, its thickness decreases and it soon becomes unstable and disintegrates into ligaments and then drops in the form of a well-defined hollow-cone spray.Disintegration of the sheet depends mainly on the liquid discharge velocity and thus on the liquid injection pressure.Description of the spray development with increasing injection pressure is presented, for example in [1].For low-viscosity fuels, the lowest injection pressure for achieving atomization using pressure-swirl atomizer is about 0.1 MPa.Evaluation of the atomization performance is based on the knowledge of spray parameters such as spray velocity, spray area, droplet size, distance and uniformity.The spray velocity depends on the driving pressure, volume flow rate and the nozzle geometry.The axial and radial velocity components also affect the spray cone angle and the spray range.The most fundamental index for atomization performance evaluation is the droplet size.Smaller droplet size positively influences the effect of heat and mass transfer and accelerates the chemical reactions.Lefebvre [1] describes various methods employed in spray characteristics measurement based on mechanical and electrical principles, which were used before the deployment of digital processing.In the recent decades, considerable advances have been made in the development of laser diagnostic and imaging techniques for measuring particle size and velocity in sprays such as Phase-Doppler Particle Analyzer (P/DPA), Planar Laser Induced Fluorescence (PLIF) and Particle Image Velocimetry (PIV).Every technique has its own advantages and drawbacks, depending on the application, therefore verification of the results obtained with one method by another one is desirable, which is also aim of this work.Several teams deal with pressure-swirl atomizer research by optical methods.Recent experimental work is focused on optimization of the spray characteristics.The aim of Chu et al. [2] was to support a theoretical model of a pressure-swirl atomizer with experiments performed by optical methods.Musemic and Walzel provided an estimation of drop size in the region of sheet formation [3].Muliadi and Sojka [4] compared patternation information of a pressure-swirl atomizer derived P/DPA measurements with values measured using PLIF.A pressure-swirl atomizer with new design of its internal mixing chamber is intended to replace the original atomizer within an update process of a combustion chamber of an aircraft engine.Micro-and macroscopic spray characteristics are to be gained as an important step during the engine innovation process.Combustion chamber adaptation and consequent numerical simulations will be based on these data.Spray structure of the new atomizer for several typical operational pressures is described within the text.Arbitrary results of two optical diagnostic systems, PIV and P/DPA, are presented and analyzed.

EXPERIMENTAL APPARATUS
The experimental equipment includes pressure-swirl atomizer, cold test bench with fluid supply system and mist extraction, Phase/Doppler Particle Analyzer and Particle Image Velocimetry.Description of our Dantec 1D P/DPA used for drop size and velocity measurements can be found in [5].

Atomizer description and operation
Single fluid pressure-swirl atomizer (Fig. 1) atomizes kerosene into a still ambient air.The newly designed atomizer is placed into a segment of kerosene feeding line, it is continuously operated and studied in the vertical downward position of the main axis (Fig. 2).The atomized fuel type is Jet A-1 aviation turbine fuel (kerosene type) with dynamic viscosity 2 mPa.s, density 810 kg/m 3 and surface tension 26 mN/m at room temperature [6].Liquid inlet temperatures, gauge pressure and volumetric flow rate were measured.The temperature was kept at 22 ± 2 °C during the experiments.

Description of the test bench and the fuel supply system
Fuel supply circuit is mounted into a mobile frame with the footprint of 600x600 mm, assembled from industrial aluminium profiles (Fig. 4).This solution offers easy transportability and enables quick alteration by fitting the elements to the frame according to our needs.The circuit consists of a stainless-steel fuel tank from which fuel

Operational regimes
The nozzle was tested in four regimes with following gauge pressure values: 150 kPa, 340 kPa, 690 kPa, 1 MPa.These values are based on typical operational conditions of the engine, from start and idle to maximum power regime.Required gauge pressure values were reached by regulating the pump speed.

DATA ACQUISITION AND PROCESSING
Particle Image Velocimetry (PIV) equipment and setup Particle Image Velocimetry (PIV) serves for planar droplet velocity measurement.A laser light sheet of approximately 1 mm thickness was produced by a dual-head pulsed Nd:YAG laser (NewWave Research Gemini, 50 mJ per pulse, max.repetition rate 15 Hz) conditioned through a cylindrical lens.The light sheet illuminated the spray in the spray axis.The image capture system consisted of TSI PIVCAM 13-8 CCD camera (1.3 Mpx) fitted with Nikon 14 mm extension tube PK-12 and Nikon 60mm f/2.8DAF Micro-Nikkor lens.

PIV processing setup
For the experiment detailed herein, 250 paired images yielding two-dimensional velocity fields (u, v) along the plane crossing the spray axis (x, z) were captured for each experimental run using straddle mode with 3.66 Hz laser pulse frequency.Required time GHOD\ ƩW EHWZHHQ ODVHU SXOVHV IRU VDWLVIDFWRU\ LPDJH SURFHVVLQJ RI WKH ZKROH ILHOG RI YLHZ ZDV VHW WR ǋV Pre-processing involved background subtraction for each image sequence.Each frame was normalized using the minimum and maximum intensity before the correlation analysis.For the image processing, ensemble PIV algorithm was selected.The PIV image pairs were interrogated using a recursive cross-correlation.For highest accuracy, deformation grid was selected.This method processes the image in multiple passes and performs image deformation, where the two frames are shifted in opposite directions and the total amount of shift should equal to the local velocity.The starting and final spot dimensions were selected to 64x64 px and 32x32 px respectively.As correlation engine, FFT Correlator was selected, followed by Gaussian peak engine for the peak location in the correlation map.The resulting instantaneous velocity fields were then validated using mean and median filters followed by Rohaly-Hart analysis [7] in one pass to replace invalid vectors, whereas the PIV processor adds the correlation maps from neighbor spots on top of the correlation map from current spot to get a good correlation peak in the summed correlation map.

Spray morphology, influence of the liquid injection pressure
In the figure 5, axial sections of the spray illuminated with a laser sheet are presented.At the lowest gauge pressure, 150 kPa, fuel emerges from the nozzle orifice in the form of a conical liquid sheet, but is contracted by surface tension forces into a closed bubble.Presented pictures (Fig. 5a, b, c, d) are axial sections of the spray According to the established nomenclature [1,8], this stage is described as the onion stage (Fig. 5a).During the film propagation, Kelvin-Helmholtz type of instability together with turbulent deformations leads to the primary break-up of the liquid sheet.Spray is very narrow, which is given by the nozzle design and also by the collapse of the liquid sheet in a close distance from the outlet orifice.Mass flow is then concentrated in the proximity of the nozzle axis.No inhomogeneities in liquid concentration are visible.In vertical position of the nozzle axis at this pressure, low velocity magnitude and influence of gravity may lead to asymmetrical spray formation.
With increasing gauge pressure, liquid sheet bubble opens into a hollow tulip shape terminating in a ragged edge, where the liquid disintegrates into drops.Spray cone angle increases as well.(Fig. 5b) Increasing kinetic energy and higher pressure differences between the emerging fluid and ambient air (Fig. 5c, d) lead to straightening of the liquid sheet and diminishing its thickness.The liquid sheet disintegrates into ligaments and drops in very short distance from the nozzle orifice in the form of well-defined hollowcone spray.At this stage, spray cone angle is defined by the inner geometry of the nozzle.Formed drops have higher initial velocity magnitude and thus the range.This behaviour complies with other pressure-swirl atomizers [1,8].In the figures 5b, c, d is evident, that the spray core is formed mainly by smaller drops, and in the radial distance drop size increases.In the spray core, more or less distinctive droplet clusters are present, which is the result of the interaction between spray and ambient air.
Description of P/DPA results P/DPA measurement was performed in the (x,y) plane perpendicular to the nozzle axis in 25 mm distance from the outlet orifice.Sauter mean diameter of droplets D 32 (Figure 6) changes significantly with increasing gauge pressure and radial distance.For low pressure, particles with the greatest diameter are concentrated close to the spray axis (area of the greatest mass flux).With increasing gauge pressure, mean drop diameter drops significantly and the difference between minimal and maximal D 32 increases.For higher pressures, Sauter mean diameter values in the axis and on the periphery of the spray vary more than threefold.At 150 kPa gauge pressure, droplet size asymmetry is evident, which may be caused by the aforementioned collapse of the liquid sheet envelope at lower gauge pressures.We assume that this phenomenon is spatially unstable.For low gauge pressures, droplets reach the highest velocity magnitude in the spray axis.With increasing gauge pressure and thus changing spray characteristics, local velocity maxima are formed in the areas of highest mass flux values, i.e. areas of the main flow of droplets generated by the breakup of the liquid sheet.For high gauge pressures, these maxima become dominant (Fig. 8).

Description of PIV results
Processing of the resultant vector files generated with Insight 3G software package, calculation of variables and graphical output were done with the Tecplot 360 2010 software.Velocity magnitudes below 0.1 m/s were cut off from the graphs.Processed PIV images provide visualization of the droplet velocity magnitude in the whole field of view and complement the P/DPA measurement.For the lowest gauge pressure, fluid mass is concentrated around the spray axis (Fig. 7a) and the highest velocity magnitude is reached in the distance interval z=15..25 mm on the spray axis.This is the area exactly under the break-up spot of the liquid sheet bubble.With increasing gauge pressure, local velocity maxima in the interconical region at the spray periphery are dominant.Droplet dynamics leads to deformation of the velocity profile with increasing axial distance.Influence of the ambient air leads to significant deceleration of drops with increasing distance from the nozzle orifice.As the figure 7 shows, velocity profile and thus liquid mass in the spray cone is unevenly distributed.For gauge pressures 340 kPa and higher, we can describe the spray shape as a hollow cone.Although the spray core is formed by large number of small drops, due to their small volume is the mass flux in this area only a fraction of the whole mass flux perpendicular to the spray axis.With increasing gauge pressure, cone angle increases first significantly (Fig. 7a, b) and then only moderately (Fig. 7c, d), what is evident from the spray photographs as well.Due to higher number of points in case of PIV measurement, smooth average velocity profiles are obtained.However, velocity profiles obtained with P/DPA are more symmetrical at higher pressures.

CONCLUSIONS
In the present work, characteristics of spray generated with a newly designed pressureswirl atomizer for a jet-engine combustion chamber are obtained using two optical diagnostic methods -Particle Image Velocimetry and Phase/Doppler Particle Analyzer.Significant changes of spray characteristics for lower gauge pressures (150 kPa, 340 kPa) and less significant for higher gauge pressures (690 kPa, 1 MPa) were observed.Liquid mass is concentrated around the spray axis for lower gauge pressures.For higher gauge pressures, local velocity and mass flow maxima in the interconical region at the spray periphery are dominant, forming a hollow-cone spray.Certain agreement of droplet velocity profiles obtained from both instruments for 25 mm distance from the outlet orifice is evident from the presented graphs.Detailed PIV measurements will be performed in the next phase of our research.Spray images will be captured in various areas of the spray in closer distance from the nozzle with different processing settings according to the droplet velocity in each area.
Stereoscopic PIV measurement will be performed as well.PC-based data gathering system of pressure, flow rate and temperature values and fuel pre-heating/cooling device will be added to the presented test bench in a short time.

Figure 2 :Figure 3 :
Figure 2: Photograph of the atomizer on cold test bench.