The Structure of the Very Near Field for a Diesel Spray: Results from a Ballistic Imaging Study

1. The Structure of the Very Near Field for a Diesel Spray: Results from a Ballistic Imaging Study M. Linne, M. Paciaroni Lund Institute of Technology T. Hall, E. B. Walters, T. Parker Engineering Division Colorado School of Mines Presented at: ILASS Americas 18th Annual Conference on Liquid Atomization and Spray Systems Irvine, CA, May 2004

2. Acknowledgements • Support for this work is provided by a grant from the Army Research Office via ARO Project Number DAAD19-02-1-0221. The equipment used and partial student support were funded by an NSF Major Research Instrumentation Grant number CTS-9711889. The authors wish to thank Lambda Research for free use of the OSLO software through the University Gratis program. • Facility development and ongoing research supported by National Science Foundation (#CTS-9502481 and (#CTS-0072967), Dr. Farley Fisher, Contract Monitor • Graduate student support, GAANN award, Department of Education • Custom drilling of injector nozzle, Raycon Corporation, Ann Arbor, Michigan • New injection system, Sturman Industries, Woodland Park, Colorado

3. Classic measurement methods do not address optically thick sprays • Diesel sprays are transient and optically thick • Diffraction based instruments • Ensemble measurement based on line-of-sight diffraction, collects scattering in the forward direction • Refractive index insensitive, maximum optical depth approximately 0.7 • Ineffective for the dense region of diesel sprays • Single particle measurements • PDPA, LDA, monitor single droplet/particle in probe volume • Diameter range 0.5 to 2.0 mm (extremes, dynamic range typically smaller), can provide velocity and number density (time averaged) • Ineffective for the dense region of diesel sprays

4. Traditional Techniques and Infrared scattering produce quantitative measures of diameter, volume fraction, etc. Ballistic imaging produces image of the breakup region for the spray Ballistic imaging offers measurement capability at the spray outlet

5. Infrared scattering for sequential spray events allow construction of “movies” of spray properties Typical Diesel Spray • CSM has developed measurement methods that allow monitoring of droplet sizes and volume fractions in a combusting diesel spray • Results are for combusting spray, initial T 873 K and Initial P 12.5 atm. Experimental Droplet Diameter movie, 0.1 msec per frame

6. Multiple scattering corrections are required closer to the nozzle outlet • Mirrored volume fraction profiles of cold and evaporating sprays Cold spray Possible Multiple Scattering zone Evaporating spray

7. Cross polarization measurements can be used to quantify the error due to multiple scattering • Error bars show the effect of multiple scattering • For combusting sprays measurements on centerline and 10 mm from the tip are possible • Goal is to bring the infrared scattering and ballistic measurements together to produce a more complete description of the dense region 25 mm below orifice

8. Optical Measurements Are Made in a Unique Diesel Simulator • Capable of operation up to 1000 K and 50 atm • Simulator is a cold-wall pressure vessel with a heated air core • Operated at 873 K and 12.5 atm • Injection System: Pressure amplifier type (built by Sturman), single hole nozzle, 160 micron diameter, L/D approximately 6 • Today’s data at room ambient conditions • Full description: Labs, J.E., J. Filley, E. Jepsen, and T.E. Parker, "A Constant Volume Diesel Spray Combustion Facility and the Corresponding Experimental Diagnostics," Review of Scientific Instruments, 2005.

9. Ballistic imaging is a line-integrated measurement capable of producing optical density images in highly scattering fields • Highly scattering fields “scramble” an image • Ballistic photons are unscattered • “Snake” photons are minimally scattered • Diffuse photons are the result of many scattering events • Image information is contained in ballistic and snake photons only • Diffuse photons can be rejected with polarization, spatial filters, and temporal filters

10. Yoo and Alfano, Optics Letters, 1990 Animation plots from previous slide

11. Crossed Polarizers: No Transmission Laser activated ½ waveplate, rotates polarization so system transmits Ballistic imaging requires aggressive elimination of diffuse photons • Spatial filtering and polarization selection remove many of the diffuse photons • Time gating is used to eliminate the majority of the remaining diffuse photons • The time gate relies on the Optical Kerr Effect • Optical Kerr Effect uses an ultra-fast laser pulse to create temporary birefringence in the medium (CS2) • Birefringence is used to produce a temporal ½ waveplate Laser input

12. The field under investigation sets the time scale requirement for the optical “shutter” • Effective time gating is required to separate ballistic from diffusive photons • Estimate the required gate time as between 0.5 and 2 times the field dimension divided by the speed of light • Gate time of less than 2 ps would be optimal

13. Ballistic imaging requires an ultra-fast laser and careful synchronization of probe and switching beams • Switching beam used to turn optical kerr cell on • Imaging beam and switching beam must arrive at the Kerr cell at the same time • Overall timing controlled by the laser; injector timing is slaved to the laser master clock

14. Investigation of the gate width shows thatsome of the diffuse photons will leak into the image • Measurement of OKE gatewidth confirms 2 ps halfwidth • Measurement of spray system response confirms that diffuse photons are not widely spaced in time • Tail of the OKE gate will admit some of the diffuse photons

15. 1 ps 2 ps Animation plots from previous slide

16. Images of the “steady” spray time show significant features along the jet edge • Image resolution approximately 50 microns • Image is line integrated • Dark regions indicate large liquid mass levels • Liquid “blobs” or high volume fraction regions • Periodic structure indicative of vigorous mixing at spray edge High Liquid Volume fraction

17. Precise measurement and control of injector start allows examination of a time sequence of sprays • Developing spray shows a vortical rollup and does not mix as vigorously at the edges as its “steady” state counterpart 2 ms 10 ms 94 ms 980 ms

18. Three distinct time periods are present in thespray: development, steady state, and shutoff • Steady spray mixes and entrains vigorously at the edges • Spray shutoff shows poor atomization 1.98 ms 2.01 ms 0.02 ms 0.001 ms before shutoff

19. Image processing can enhance data presentation: what is “reasonable?” • False color scale can enhance edge effects or other features • Quantitative results should use a mapping of the optical depth • Optical Depth is Nsl • Values at each pixel used to optical depth for that pixel

20. Optical Depth representation is limited by the dynamic range of the measurement • Optical depth scales nonlinearly with intensity • Noise and/or background limit dynamic range in image • Optical depths greater than 4 should be treated with caution • Edges in images subject to diffraction blurring

21. Optical depth image indicates internal structure for the fuel jet • Emphasizes small optical depths

22. The start-up injection time does not have significantinternal jet structure and edge mixing is not as vigorous

23. Initiation, steady state, and shut-off periods behave differently Initiation Surrounding gas flow must be started Entrainment appears to be less effective interior jet structure not “dramatic” Steady state Vigorous mixing and entrainment at the spray edges Spray edges and interior point toward a shear layer view of the jet Shutoff Poorly atomized flow Image analysis All images are line-integrated Voids are indicative of significant three dimensional structures Jet edge features are two-dimensional projections of a three dimensional object Optical depth is the correct scaling for the image Edge features can be emphasized through a choice of color scale Initial results indicate distinctbehaviors for the different spray time periods

24. Conclusions • Single shot ballistic imaging can penetrate the droplet fog that surrounds the diesel jet • Spray initiation, steady state, and shutoff behave in very different ways • Steady state shows very significant air entrainment at the spray edge • Results at steady state are consistent with an atomized spray with central high momentum, high liquid volume fraction region that behaves as a mixing layer (mixing between central region and surrounding, low liquid volume fraction ambient gas)

25. Future work • Image analysis • Noise limits and optical depth, 2-d Fourier transform • Ballistic imaging using “cheaper” ultra- fast laser system • Critical for dissemination of the technique • Uses Continuum Leopard (multicolor, 15 ps, 30+mJ) • Results for diesel-like conditions • Reacting and non-reacting

26. Laser scattering measurements Nd:Yag (1.06 mm) at 90° Tunable CO2 (9.27 mm) at 11° Scattering detectors calibrated for absolute measurements Beam power monitored to compensate for power fluctuations Lasers are focused to an experimentally verified diameter of 150 mm Measurements utilize focused and co-aligned infrared lasers

27. Infrared lasers replace the more classic visible light sources • To decrease optical thickness effects, the optical “probe” wavelengths have been shifted into the infrared • Lower Attenuation Levels • System must avoid hydrocarbon absorption features near 10.6 mm • Extinction and angular scattering techniques are used to determine droplet sizes and volume fraction as a function of position and time in the near field spray region

28. The governing scattering equation produces as a function of measured signal A ratio of two signals produces the droplet size Number density cancels out for common probe volume Number density is calculated using the diameter from corrected scattering signals (which gives the differential scattering cross section) and the 9.27 mm signal This signal is used because thickness correction is smaller and well known Result is the average volume fraction over the probe volume Scattering signals are proportional to the product of number density and optical cross section

29. Scattering Measurements Provide Spatial Resolution and an Increased Sizing Range • Ratio of Scattering Measurements at Different Wavelengths and Angles Can Be Used to Produce Spatially Resolved Droplet Size Measurements • Modeling Indicates Reported Diameters are Sauter Mean • Insensitive to Distribution Width • Diameter values greater than 16 mm cannot be uniquely sized due to multiple roots

30. Calculated Error Due to Multiple Scattering in the 1.06 mm Data • Error due to multiple scattering results in reported droplets that are small • Relative change in volume fraction can be either positive or negative • Competing mechanisms

31. Measured Ratio is a Function of Position Within the Spray • High pressure data shows ratio values of approximately 0.5 for both axial locations near centerline • Ratio quickly drops off at an axial distance of 10 mm • Ratio remains relatively high throughout the spray at 25 mm, falling with distance from centerline • Room ambient data exhibits highest ratio near centerline, with both axial positions falling quickly to below 0.2