Donghun Kim and Yudaya Sivathanu, **
An Information Delivery System for Visually Impaired People in a Dynamic Environment**,
IEEE Internation Conference on Systems, Man, and Cybernetics, October 2011, Achorage, Alaska.

In this paper, we propose an information delivery system for visually impaired people to examine any object in their environment. This system provides a magnified view of any region that is pointed at by the user to monitor it. The human pointing gesture has the ability to extract an element from the environment. The two important problems in codifying this gesture are to estimate the direction to which the finger is pointing to, and then to segment the region of interest according to this direction. To solve these two problems, we present a hand and a finger detection method using color and disparity data from a stereo camera system. Then, we estimate the direction in 3-D using geometric and textural features. We utilize the estimated direction to discriminate potential candidates in the 3-D space. Finally, we segment the region of interest by backprojecting the candidate areas while filtering signal noises of the disparity data. Experimental results that validate the proposed system in a dynamic environment is also presented.

Yudaya Sivathanu, Jongmook Lim, and Deepa Divakaran, **
Temporal Variation of Aerosol Characteristics in Nebulizers**,
Proceedings of the Respiratory Drug Delivery X, 2006, Boca Raton, FL.

Nebulizers are widely used by patients who either cannot correctly use a metered dose inhaler, or simply prefer a nebulizer [1,2]. Various types of nebulizers and their associated performance have been reviewed with a view of establishing standards [3]. One important parameter that affects the drug output rate from a nebulizer is the design of the nozzle. Subsequently the effect of different nozzle designs [4] or multiple copies of the same nozzle design [5] on drug output have been studied, but not in ways that are efficient enough to make them suitable for use as a quality assurance tool. This article describes a simple method to evaluate transient variations in the performance of individual nebulizers (including consistency of delivery over the entire operation period) using statistical extinction tomography.

Jongmook Lim, Yudaya Sivathanu, and Paul Sojka, **
Optical Patternation of a Multi-hole Fuel Spray Nozzle**, Proceedings of the ICLASS.

Patternation for a multi-hole spray nozzle using an extinction tomography based optical patternator is reported. Path integrated transmittances were obtained from a six-hole injector nozzle using water under cold flow conditions. The path-integrated transmittances were deconvoluted using the maximum likelihood estimation method with a new grid scheme. The current optical patternator uses only six view angle measurements because of limitation in the physical size of the instrument. However, a theoretical study with the new grid scheme showed that the angular resolution could be increased significantly by compromising radial resolution. Furthermore, an increase in angular resolution improves deconvolution error caused by the limited spatial resolution. The new technique was applied to measurements from a six hole injector nozzle and successfully resolved the droplet surface area spatial distribution.

Y. Zheng, J. P. Gore, and Y. R. Sivathanu, **
Infrared Spectral Imaging of Jet Flames**,
Proceedings of the 34th National Heat Transfer Conference, Pittsburgh, PA.

An infrared (IR) camera in conjunction with a spectrograph and a digital frame grabber were used to obtain rapid simultaneous measurements of spectral radiation intensities along chord-like paths in laminar jet flames. The results from two different laminar co-flow jet flames with an identical nominal heat release rate of 2.1 kW are reported. A non-premixed C2H4/air flame and a partially premixed C2H4/air flame with Φ = 1.6 were considered to investigate the effects of partial premixing. The camera captured multi-spectral infrared line images within a wavelength range from 2.2 μm to 4.9 μm for each flame. Separate blackbody images were used to calibrate the sensitivity, wavelength and radial location associated with each pixel of the 316x240 array. After calibration, spectral radiation intensities for the chord-like paths with high spatial resolution were obtained. The spectra reveal the combined effects of temperature, CO2, H2O, fuel, and soot distributions.

J. Ji, J. P. Gore, Y. R. Sivathanu, and J. Lim, **
Fast Infrared Array Spectrometer for Radiation Measurements of Lean Premixed Flames**,
Proceedings of the 34th National Heat Transfer Conference, Vol. 2, pp. 73-78.

A Fast Infrared Array Spectrometer (FIAS) is used to simultaneously measure spectral radiation intensities over the 1.8 - 4.9 μm wavelength ranges at 390 Hz. A three-point second-degree Lagrange interpolation polynomial is employed to calibrate the non-linear response of the FIAS to spectral radiation intensity. This calibration method gives excellent measurements of blackbody radiation except at low wavelength around 2 μm due to the room humidity and at 4.3 μm due to the absorption by room carbon dioxide. The calibrated FIAS is used to measure spectral radiation intensities from premixed laminar flames and premixed turbulent flame for which previous baseline data are available. The agreement between the FIAS measurements and the previous data is very good except for the carbon dioxide band due to absorption effects in the test chamber.

Fu, C., Sojka, P. E., and Sivathanu, Y. R., **
Water Mist Impingement onto a Heated Surface**,
Proceedings of the 5th Joint ASME/JSME Thermal Engineering Conference, ASME Paper No: AJTE99-6367.

An experimental study on the interaction of a water mist with a heated surface is described. The long term objective is to produce experimental data that can be used to validate submodels for four key physical phenomena involved in the interaction of sprays with burning surfaces: (1) the effect of buoyancy (caused by the hot combustion products) on the trajectory of a single droplet, (2) the effect of evaporation on the trajectory of a single droplet, (3) the cessation of reaction and reduction in flame spread caused by the droplets on flaming surface combustion, and (4) the reduction in surface temperature caused by the effect of drop impingement, spreading, and evaporation, on surface combustion. The short term objective was to complete experiments that mapped the size history of water droplets as they approached a hot surface using a Malvern Spray Analyzer to study the effect of both buoyancy and evaporation on droplet trajectory. Surface temperature profiles were recorded using fine-wire thermocouples. The influence of water mass flow rate and drop size distribution on the hot surface temperature profile is presented and discussed.

W. M. Glasheen, C. A. DeMilo, and Y. R. Sivathanu, **
Turbine Inlet Gas Temperature Sensor**,
Proceedings of the ASME Turbo Expo '99, ASME, NY.

A practical turbine inlet gas temperature sensor (TIGTS) has been designed, fabricated, and evaluated both numerically and experimentally. The sensor remotely measures the hot spectral radiation intensities at several discrete wavlengths and obtains the Planck-function weighted temperature of the optical path using a look-up-table. A high temperature optical head collects the radiation intensities emitted by hot gases along a narrow path. The radiaton is transmitted using optical fibers to a quadrant detector which measures intensities at four discrete wavelengths with the spectral bands of water vapor and carbon dioxide. After analog signal processing, under microprocessor control, the signals are digitized. They are input to a look-up-table which was obtained with stochastically generated gas concentrations and temperatures in conjuction with the narrow band radiation calculation model, RADCAL

**
Characterization of Particulate from Fires Burning Silicone Fluids**,
Proceedings of the 7th AIAA/ASME Joint Thermophysics and Heat Transfer Conference, HTD-Vol. 357, pp. 191-196.

Y. R. Sivathanu, J. P. Gore, and N. M. Laurendeau, **
Effects of Radiation on NO Kinetics in Turbulent Hydrogen/Air Diffusion Flames**,
Proceedings of the 32nd National Heat Transfer Conference, HTD-vol. 341, pp. 1-6, ASME, New York.

We describe a coupled radiation and NO kinetics calculation of turbulent hydrogen/air diffusion flame properties. Transport equations for mass, momentum, mixture fraction, enthalpy (sensible + chemical) including gas band radiation, and NO mass fraction are solved. NO kinetics is described by a one step thermal production mechanism. The local temperature is obtained by solving the enthalpy equation taking radiation loss from H2O into consideration. Radiation/turbulence and chemical kinetics/turbulence interactions are treated using a clipped Gaussian probability density function (PDF) for mixture fraction and a delta PDF for enthalpy. The source terms in the enthalpy and mass fraction of NO equations are treated using assumed PDF integration over the mixture fraction space. The results of the simulation are compared with existing measurements of the Emission Indices of NO (EINO) in turbulent H2/air diffusion flames. The major conclusion of the paper is that coupled turbulence/radiation interactions should be taken into account while computing EINO.

Y. J. Zhu, Y. R. Sivathanu, and J. P. Gore, **
Diagnosis of Gas Species Concentrations and Temperatures in Laminar Flames**,
Heat Transfer in Combustion and Fire, HTD-vol. 310, pp. 120-127, ASME, New York.

The structure of premixed laminar methane/air flames at atmospheric pressure was studied in this paper. In the post flame region of these flat flames, the emission spectroscopy was used to obtain the temperature and gas species concentrations information which are very important in the study of combustion. Based on the thermal emission by gas molecules in lean flames, the emission data obtained by using he multi-wavelength infrared detection method were operated in an iterative algorithm to get the gas species mole fractions and temperatures. At the same positions in the flame, temperature measurements were obtained from thermocouple and H2O and CO2 concentrations were obtained from gas chromatography. The emission spectroscopy provided accurate temperature and mole fraction data compared to the theoretical value with consideration of reasonable heat loss.

M. Franaszek, C. Hagwood, E. Simiu, and Y. Sivathanu, **
Mean Zero Upcrossing Time of a Melnikov Process as a Lower Bound for Mean Escape Time from a Potential Well: Gaussian and Non-Gaussian Excitations**,
Proceedings of The International Conference on Structural Safety and Reliability," ICOSSAR, 97, Kyoto, Japan.

The mean zero upcrossing time of the Melnikov process is a lower bound for the system's mean time of escape from a well. We present results which show that, for small perturbations by Gaussian or dichotomous noise, this lower bound is weak. The results consist of estimates of the mean zero uprcrosiing time of the Melkinov process on the one hand, and of the mean time of escape from a well on the other. We interpret the results obtained for color noise excitation in light of Melnikov theory.

Y. R. Sivathanu and J. P. Gore, **
Radiative Heat Transfer Inside a Cylindrical Enclosure with Nonparticipating Media Using a Deterministic Statistical Method**,
Proceedings of the ASME Heat Transfer Division, HTD-vol. 332, pp. 145-152, ASME, New York.

The radiative heat transfer inside a cylindrical enclosure is modeled using a discrete probability function (DPF) method. The discrete probability function method involves solution of he equation of radiative heat transfer using Lagrangian simulations of representative photon trajectories on a discrete spatial grid. The DPF method is applied to radiation exchange in a cylindrical tube which has a hot source on end and a detector at the other end. The cylindrical wall absorbs and reflects (both diffusely and specularly) the radiation incident on it. The calculations are used to simulate the effect of collimating tubes used in intrusive multi-wavelength emission spectroscopy. Results highlight the effect of surface properties on the apparent source temperature determined by the detector. The calculation procedure has application to the measurements of spectral absorption and reflection coefficient of the cylindrical surface using an inverse method.

Y. R. Sivathanu and J. P. Gore, **
A Discrete Probability Function Method for a Turbulence Mixing Layer**,
Fire, Combustion, and Hazardous Waste Processing, HTD-Vol. 296, pp. 99-106, ASME, New York.

Scalar mixing in turbulent layer (with no shear) is simulated using discrete probability function (DPF) and Monte-Carlo solutions to the Probability Density Function (PDF) transport equations. The DPF method consists of solving the Lagrangian equations for notional parcels in the turbulent flow field for discrete realization of the boundary and initial conditions, and velocity field. The probabilities of the solutions are calculated from the probability density functions (PDFs) of the boundary and initial conditions, and the velocity field. The velocity statistics throughout the flow field and the scalar statistics at the initial station are input from available experimental data. The simulation provides the evolution of the probability density function of the scalar at downstream stations. The simulation requires a specification of the Taylor micro-scale which is provided using the experimental data at present. An exponential parabolic form of the spatial correlation coefficient is assumed. The present simulation correctly captures the change of the RMS fluctuations with downstream distance and the growth of the scalar layer. Two different sets of experimental data are used to evaluate the simulations in the present work.

L. G. Blevins, Y. R. Sivathanu, J. P. Gore and M. A. Shahien, **
Radiometric Measurements of Wall Temperature in the 800K to 1150K Range for a Quartz Radiant Heating Tube**,
Radiative Heat Transfer: Current Research, HTD-Vol. 276, pp. 115-120, ASME, New York.

Combustion and heat transfer in a vertical straight-through radiant heating tube are being studied experimentally. Since the tube is made of quartz, it is difficult to measure wall temperatures using thermocouples. The objective of the work described in this paper is to use a non-intrusive radiometric technique to measure the wall temperature of a quartz radiant heating tube. The quartz temperature is deduced by measuring the emission intensity and transmittance of a selected path through the quartz tube and comparing it graphically with solutions of the radiative transfer equation for various temperatures. Carbon dioxide and H2O gas-band interference is avoided by selecting proper wavelengths for comparison. Tube wall temperature profiles for radiant tubes operating at nominal heat inputs of 10, 15, and 20 kW with the overall equivalence ratio of 0.8 are presented. In-situ high temperature transmittance measurements improve the accuracy of the technique by up to 10%.

Y. R. Sivathanu and J. P. Gore, **
A Discrete Probability Function Method for Radiation in Enclosures and Comparison with the Monte Carlo Method**,
Radiative Heat Transfer: Current Research, HTD-Vol. 276, pp. 213-218, ASME, New York.

A statistical method, recently developed in this laboratory, based on the application of physical laws to discrete values of boundary conditions and source functions and appropriate statistical laws to associated probabilities is applied to radiation transfer in enclosure. The new approach is termed the Discrete Probability Function (DPF) method and has been applied to simulation of radiative transfer in flames with significant advantages over the Monte Carlo technique. The DPF method simulates a limited number of representative photon trajectories in the enclosure. The number is selected based on the number of bins required for the resolution of the Probability Density Function (PDF). From these simulations, results of engineering interest are obtained from the statistics of the interaction of the photons with the walls of the enclosure. However, instead of relying on obtaining converged statistics from a large number of photon trajectories, the DPF method carries additional information concerning the probabilities associated with the limited photons. This results in substantial saving in computer time since only a limited number of photon trajectories need to be simulated. In this paper, the DPF method is applied to radiation exchange in cylindrical tubes and the results are compared with those obtained using the conventional Monte Carlo method. Results are presented for both direct view factors and indirect radiation from reflecting surface. The Monte Carlo results have very little error in calculating the direct view factors. However, when the indirect radiation from reflecting surface is calculated, the DPF method is more accurate than the Monte Carlo method which shows effects of statistical noise.

Y. R. Sivathanu and J. P. Gore, **
Coupled Radiation and Soot Kinetics Calculations in Laminar Acetylene/Air Diffusion Flames**,
First ISHMT/ASME Heat and Mass Transfer Conference," in press.

L. G. Blevins, Y. R. Sivathanu, M. A. Shahien, and J. P. Gore, **
A Study of the Global Properties of Radiant Tube Flames**,
First ISHMT/ASME Heat and Mass Transfer Conference," in press.

P. Dutta, J. P. Gore, Y. R. Sivathanu, and P. E. Sojka, **
Global Properties of High Liquid Loading Turbulent Crude Oil + Methane/Air Jet Diffusion Flames**,
Heat and Mass Transfer in Fires and Combustion Systems, HTD-Vol. 250, pp. 37-42, ASME, New York.

Y. R. Sivathanu and J. P. Gore, **
DPF Deconvolution of Turbulent Flame Data**,
Fluid Dynamics and Thermodynamics-1993, pp. 1627-1634.

J. H. Jang, Y. R. Sivathanu, and J. P. Gore, **
Lagrangian Simulation of the Radiative Cooling Process in an Acetylene/Air Jet Flame**,
Heat and Mass Transfer in Fires and Combustion Systems, HTD Vol. 223, pp. 81-86, ASME, New York.

Y. R. Sivathanu and J. P. Gore, **
A Discrete Probability Function Method for the Equation of Radiative Transfer**,
Heat and Mass Transfer in Fires and Combustion Systems, HTD Vol. 223, pp. 73-80, ASME, New York.

A discrete probability density (DPF) method for the equation of radiative transfer is derived. The DPF is defined as the integral of the probability density (PDF) over a discrete interval. The derivation allows the evaluation of the DPF of intensity leaving desired radiation paths including turbulence-radiation interactions without the use of computer intensive stochastic methods. The DPF method has a distinct advantage over conventional PDF methods since the creation of a partial differential equation from the equation of radiative transfer is avoided. Further, convergence of all moments of intensity is guaranteed at the basic level of simulation unlike the stochastic method where the number of realizations for convergence of higher order moments increases rapidly. The DPF method is described for a representative path with approximately integral length scale sized spatial discretization. The results show good agreement with measurements in a propylene/air flame except for the effects of intermittency resulting from highly correlated realization. The method can be extended to the treatment of spatial correlations as described in Appendix A. However, information regarding spatial correlations in turbulent flames is needed prior to the execution of this extension.

Y. R. Sivathanu, J. P. Gore, J. Jenssen, and D. W. Senser, **
A Study of Insitu Emission Coefficients of Soot Particles in Laminar Flat Flames**,
Heat Transfer in Fires and Combustion Systems, HTD Vol. 199, pp. 1-7., ASME, New York. Also presented at the Annual Conference for Fire Research 1992, Washington D. C., p.111.

Simultaneous measurements of absorption at 632 nm and emission at 800 nm and 900 nm along diametric paths are used to infer soot volume fractions based on both absorption and emission and temperatures based on emission in laminar, premixed flat flames. Agreement between soot volume fractions based on emission and absorption is used as an indicator of optical path homogeneity. Emission measurements at 2300 nm and 4000 nm for homogeneous paths are used to infer the optical properties of soot at these wavelengths. The results for methane, propane and ethylene mixed with nitrogen and oxygen show insensitivity to the C/H ratio and weak dependence on the temperature for the optical properties of soot. Insitu check of homogeneity is crucial in obtaining reliable measurements of optical properties.

M. E. Kounalakis, Y. R. Sivathanu, and G. M. Faeth, **
Infrared Radiation Statistics of Non luminous Turbulent Diffusion Flames**,
Proceedings of the ASME/JSME Thermal Engineering Joint Conference, HTD Vol. 5, pp. 3-12, ASME, New York.

Mixture fraction and radiation statistics were studied for radiation paths through turbulent carbon monoxide/hydrogen diffusion flames burning in still air. Measurements include Mie scattering for mixture fraction statistics and a fast-response infrared monochromator for radiation statistics. Measured mixture fraction statistics also were used to predict radiation statistics based on stochastic time series methods, the laminar flamelet approximation and a narrow-band radiation model. Measured intensities of radiation fluctuations were in the range of 10 to 40 percent, which cause mean radiation levels to be 1.2 - 4.2 times larger than estimates based on means scalar properties in the flames. In contrast, stochastic predictions of mean and fluctuating radiation properties were generally in excellent agreement with measurements. An exception was the temporal integral scales of radiation fluctuations, where differential diffusion errors of the Mie scattering measurements were identified as the source of the discrepancies.

U. O. Koyulu, Y. R. Sivathanu, and G. M. Faeth, **
Carbon Monoxide and Soot Emissions from Buoyant Turbulent Diffusion Flames,**,
Fire Safety Science, Proceedings of the Third International Symposium, (G. Cox and B. Langford eds.) Elsevier Science Publishers Ltd., London. pp. 625-634.

Carbon monoxide concentrations and mixture fractions were measured in the fuel-lean (over fire) region of turbulent acetylene, propylene, ethylene, propane and methane diffusion flames burning in still air. Three burners (having exit diameters of 5,50 and 234 mm) were used to study conditions ranging from buoyant jet flames to pool-like fires. Carbon monoxide generation factors (mass of CO emitted per unit mass of fuel carbon burned) were uniform throughout the over fire region for a given flame condition. Additionally, CO generation factors of sooting fuels approached asymptotic values for flame residence times roughly an order of magnitude longer than the normal smoke point residence time, similar to earlier measurements of soot generation factors for similar conditions. Finally, processes of carbon monoxide and soot emission appear to be closely related due to the good correlation between their emission factors in the asymptotic regime: 0.34 (standard deviation of 0.09) kg CO/kg soot. However, non sooting methane/air flames still emitted low levels of CO so that there is a component of CO emissions that is not associated with soot.

Y. R. Sivathanu and J. P. Gore, **
A Discrete Probability Function Method for the Equation of Radiative Transfer**,
J. Quant. Spec. & Rad. Trans., vol. 49, pp. 269-280.

Radiative heat transfer inside a cylindrical tube is modeled using a statistical method called the discrete probability function (DPF) method. The DPF method involves solution of the equation of radiative heat transfer using Lagrangian simulations of representative photon trajectories on a discrete spatial grid. The DPF method is different from the Markov Chain method in terms of associating a probability with each state of the photon rather than a transition from one state to another. The advantages and disadvantages of the DPF method in comparison to the Markov Chain method are demonstrated in this paper using two practical applications of the cylindrical tube radiative heat transfer problem. The cylindrical tube has a hot source at one end and a detector at the other end. The cylindrical wall absorbs and reflects (both diffusely and specularly) the radiation incident on it. The present calculations have applications in: (1) intrusive pyrometry with collimating light guides, and (2) measurement of the spectral absorption and reflection coefficients of coatings using two, coated cylindrical tubes as specimen. The results show that: (1) the effect of light guide surface properties on errors in pyrometry must be carefully assessed, and (2) the method can be used for a convenient evaluation of radiative properties of coatings.

Y. R. Sivathanu, J. P. Gore, and J. Dolinar, **
Transient Scalar Properties of Strongly Radiating Diffusion Flames**,
Heat and Mass Transfer in Fires and Combustion Systems, HTD Vol. 148, pp. 45-56, ASME, New York.

An experimental and a theoretical study of transient radiation properties of strongly radiation turbulent diffusion flames is described. Simultaneous transient measurements of the temperatures and volume fractions of soot obtained using a three wavelength emission/absorption probe are reported. Soot volume fractions inferred from the absorption measurements are significantly higher than those inferred from emission data suggesting the presence of large quantities of relatively cold soot within the flames. The results show the effects of progressive radiative cooling with distance from the injector exit. Reasonably good predictions of monochromatic intensities were obtained using a bivariate stochastic analysis.

J.J.A. Green, Y. Sivathanu ,Optical Force for Aerosol Particle Diagnostics, ICLASS 2018, **
14th Triennial International Conference on Liquid Atomization and Spray Systems, Chicago, IL, USA, July 22-26, 2018**,
ILASS – Europe 2016, 27th Annual Conference on Liquid Atomization and Spray Systems, Sep. 2016, Brighton, UK.

J. Lim, Y. Sivathanu, M. Wolverton, and J. Green, Conversion of Scattering Phase Function to SMD in Scattering/Extinction Tomography, ICLASS 2018, **
14th Triennial International Conference on Liquid Atomization and Spray Systems, Chicago, IL, USA, July 22-26, 2018**

A single frequency laser was arranged to pump a free space, passive, symmetric, confocal, unstable FabryPerot etalon cavity resonator. Non-Laminar and Laminar humidifier particles were configured to flow through the cavity at an oblique angle. A high speed camera was configured to view the flow at an orthogonal angle, and to detect scattered light from the particles. When mechanically expanding or contracting the cavity, image processing analysis indicated that optical force was applied en masse to non-laminar humidifier water particles whose initial particle velocity was 0.5 m/s and size 14 um. Whereas a mechanically static (motionless) and 1-pass (no cavity) did not show optical force. In the Laminar case, with other factors involved, the static cavity also seemed to apply a small amount of force, yet still expanding and contracting cavities showed more force. With a calibration standard of optical force to photo-detector voltage, a particle of unknown size can potentially be measured. And, this technique can potentially be scaled to measure particle sizes below 1 micron, which is currently inaccessible by diffraction and phase-doppler based instruments, nor by ion mobility spectrometry

The scattering phase function of droplets at a specific angle is a unique function of the drop size. This paper describes the conversion of the scattering phase function to planar droplet size (Sauter Mean Diameter, SMD) of particles in the spray. The conversion process proposed utilizes Mie theory, where, the scattering angle must be known a priori. To convert the phase function to drop sizes, a calibration method is required. A mono-disperse droplet generator with a known size for the particles was used to calibrate the optical efficiency of the system. An experiment was conducted to measure a pressure type spray with droplets of unknown size at two different pressures. Path integrated extinction of a laser beam, and the scattering intensities 30 degrees from the incident beam, were measured simultaneously, to estimate the planar scattering phase function. The planar scattering phase functions were converted to planar drop sizes using the calibration. PDA measurements were then conducted to validate the extinction and scattering tomography method. The SMD results were in good quantitative agreement with the PDA measurement for both pressure conditions. The uncertainty of conversion from the phase function to SMD was also analyzed theoretically from Mie theory with the assumption of lognormal distribution for the drop sizes. The results show that the SMD is not only a function of scattering phase function, but also a weak function of second order moment of the drop size distribution.

Lim, J., Sivathanu, Y.,Green, J.,2017, **
Tomographic Reconstruction of scattering Phase Function of Droplet in an Urea Injector**,
ILASS--Americas 2017, 29th Annual Conference on Liquid Atomization and Spray Systems, May. 2017, Atlanta, GA.

Mie theory relates scattering phase function of droplets to the size distribution. Line of sight scattering intensity distributions of the droplets were measured at 30 degree off the incidence angle using a modified SetScan optical patternator. A reconstruction algorithm for local scattering phase function was developed and theoretically validated. The reconstructed scattering phase function provides the first moment of the PDF of drop size distribution, and can be used to estimate the spatial distribution of Sauter Mean Diameter (SMD) of the droplets. The SMD distribution in a pressure atomizer was estimated from the local scattering phase function was validated against 2-D grid scan of PDPA measurements At higher pressure condition, the local scattering phase functions are more uniform than the distribution at lower pressure condition. The SMD distribution from the scattering phase function agrees well with the PDA measurement when the shape of PDF of drop size distribution is maintained.

Sivathanu, Y., Lim, J., Mi, Y., Muliadi, A., Nitelscu, O., Shieh, T., 2015, **
Comparison of Velocity Measurements in a GDI Injector obtained using Statistical Pattern Imaging Velocimetry and Phase Doppler Interferometry**,
ILASS – Europe 2016, 27th Annual Conference on Liquid Atomization and Spray Systems, Sep. 2016, Brighton, UK.

Sivathanu, Y, Lim, J., Green, J., Wolverton, M.,2016 **
Combined Visible and X-Ray Extinction Measurements in a Water Spray from a GDI Injector**,
ILASS--Americas 2016, 28th Annual Conference on Liquid Atomization and Spray Systems, May. 2016, Dearborn, MI.

Statistical Pattern Imaging Velocimetry is a new technique to estimate the planar velocity field from the high speed videos. The technique utilizes an ensemble of either backlit or side lit videos to obtain full planar velocities in sprays and flames. Unlike conventional particle imaging velocimetry, statistical pattern imaging velocimetry does not require well resolved images of particles within turbulent flows. Instead, the technique relies of patterns formed by coherent structures in the flow. The implementation of SPIV technique quite is relatively straight forward since high speed videos can be readily obtained either in a laboratory or production floor setting. The biggest challenge for the SPIV techniques is that the procedure is computationally expensive even with an ordinary mega-pixel camera. To improve the computation speed, a successive partitioning scheme was employed. In addition, to improve spatial resolution to sub pixel dimensions, a weighted central averaging scheme was used. With these two enhancements, the SPIV method was used to obtain planar radial and axial velocities in a spray emanating from a GDI injector. The velocity measurements were validated with those obtained using a Phase Doppler Interferometry.

Characterization of GDI injectors in the near injector region using conventional optical techniques is difficult due to the high levels of obscuration present in these sprays. This paper describes the characterization of a GDI injector using combined X-Ray and laser extinction tomography. One of the advantages of using X-Rays is that the near injector region of dense sprays can be characterized reliably. A single axis X-Ray system was used to obtain laser extinction measurements in a water spray from a GDI injector. The injector was operated at several pressures representative of those used in the automotive industry. The injector was rotated to provide data a multiple view angles. The data was deconvoluted using a Maximum Likelihood Estimation method to provide the local mass concentrations in the spray. For validation purposes, additional measurements were obtained at several locations downstream of the injector where the path integrated extinction was less than 95%. At these locations, the GDI injector was also evaluated using a SETscan optical patternator. The SETscan optical patternator provided the planar local surface area densities in the spray. The Sauter Mean Diameter (SMD) of the spray was estimated from these combined X-Ray and laser extinction measurements. A diffraction based instrument was also used to obtain the SMD in the spray at the same location. The results obtained from the diffraction system and the combined visible and X-Ray extinction measurements agreed with each other within the uncertainty bands of the methods. Therefore, a combination of soft X-Ray and visible extinction tomography promises to provide a unique method for characterizing sprays from GDI injectors.

Y. Sivathanu, J. Lim, and V. Kulkarni **
Transient Characteristics of GDI injectors using Statistical Extinction Tomography**,
ICLASS 2015, 13th Triennial International Conference on Liquid Atomization and Spray Systems, Tainan, Taiwan, August 23-27, 2015

Multi-hole fuel injectors are used in a variety of automotive applications. For such injectors, it is desirable to obtain a variety of spray characteristics such as total spray angle, fuel drop surface areas, individual plume locations, and plume angles so as to optimize their efficiency. Simultaneous light-of-sight extinction tomography measurements at two axial locations were used to obtain transient characteristics of two GDI injectors. The measurements were obtained in a pressure chamber with opposing windows. Therefore, the nozzle was rotated twelve times and data was obtained synchronously with the injection pulse so as to enable deconvolution of the line-of-sight measurements. The fuel temperature and ambient pressure was found to be the two most important characteristics in determining the plume locations and angles. The plume centroid locations, obtained using extinction tomography, were compared with those obtained with a mechanical patternator. The centroid locations of the plumes obtained from the two instruments matched within the uncertainty band of these instruments.

Eric Coker, Alvaro Cruz-Cabrera, Frank van Swol, Walter Gill, David Surmick, Leland Sharp, Lim Jongmook, Edward Bystrom and Aren Haug, **
Measurement of Solid Rocket Propellant Exhaust Gas Temperatures Using Molecular Spectroscopic Methods**,
42nd Annual Conference of NATAS, Sante Fe, New Mexico, September 2104.

Future deep-space missions will likely include a nuclear reactor (radioisotope thermal generator, RTG) as part of the vehicle's payload. The risk of a breach to an RTG during a launch abort is believed to be minimal in most cases; however, there is a concern that the use of solid propellants as part of the boost stage could increase the chance of an RTG breach during an abort. It is expected that the temperatures from the propellant exhaust will be 2000 to 3000 K, and the burn will persist until all the material is consumed. A series of controlled field tests burning solid rocket propellants were performed at Sandia National Laboratories with the purpose of understanding the interaction of the propellant exhaust with other materials. Because the exhaust temperatures could exceed witness materials melting points, spectral measurements of different gas emitting species were performed with the purpose of determining their temperatures. One of these species is hydrochloric acid (HCl) gas, whose spectrum was measured at the mid-wave infrared (MWIR) using a Fourier Transform Infrared (FTIR) spectrometer. Temperatures were obtained by fitting the P-Branch (3.5 to 4 micrometer) of the HCl spectrum to a Boltzmann distribution. Calibration and verification tests were performed in a laboratory setting where HCl was heated to known high temperatures. The emitted spectra from hot HCl were used as a reference in the temperature calculation of the same gas during an actual propellant burn. We expect to get more accurate temperatures by using a diatomic line strength model.

Lim, J., Sivathanu, Y., and Kim, J., **
Improving the spatial resolution and Computational Performance of Statistical Pattern Imaging Velocimetry**,
ILASS Americas, 26th Annual Conference of Liquid Atomization and Spray Systems, Portland, OR May 2014.

Statistical Pattern Imaging Velocimetry (SPIV) is a technique to estimate the planarvelocity field from the successive images. SPIV utilizes an ensemble of videos to obtain fullplanar velocity. Unlike particle imaging velocimetry (PIV) technique that uses micro-seconddelayed double shots, SPIV does not require well resolved particle images. Therefore, SPIV iswell suited for the estimating velocity in sprays and turbulent flames, both of which have welldefined patterns embedded in the flow videos. The implementation of SPIV technique quite issimple because successive images can be used obtained with an ordinary background lightingand a high speed camera.

One challenge to utilizing SPIV for routine spray analysis is that the procedure iscomputationally expensive even with an ordinary mega-pixel camera. The typical computationtime for a M x M pixel array camera scales as M2 x N where M is the number of pixels and N isnumber of frames in the video. A successive partitioning technique with nearest neighboringsearch (SPM-NNS) was successfully implemented. The SPM-NNS technique essentiallyreduced the computational time such that it scales as M X N, while maintaining same spatialresolution of the measurement.

One potential shortcoming of the SPIV is the limited spatial resolution when the velocityestimation is based on pixel-to-pixel correlation. A correlation from a pixel to a correlation-coefficient-weighted center was tested and found that sub-pixel resolution could be achieved.The planar velocities for a urea doser injector with SPIV were compared with Phase DopplerParticle Anemometer (PDPA) and showed good agreement.

Lim, J., Sivathanu, Y., and Wolverton, M., **
Evaluation of Soft X-Ray Tomography for the Near Injector Characterization of Dense Sprays**,
ILASS Americas, 25th Annual Conference of Liquid Atomization and Spray Systems, Pittsburgh, PA.

The recent development of high energy synchrotron based X-Ray tomography for high pressure and high flow rate sprays is adding to the in-depth understanding for the nature of the spray core region. X-Ray tomography is found to be valuable tool in optimization of automotive and aero-engine nozzles. However, the synchrotron based CT system is not portable, and is not widely inaccessible due to the nature of the X-Ray source. A computed tomography system with a commercial soft X-Ray source and detectors was developed and tested for the quantitative mass density measurement. The X-Ray tomography system consists of simultaneous multi-view measurements with high spatial (0.1 mm) and high temporal (1.1 KHz) resolution. The system can be applied for the non-repeating transient sprays as well as cyclic transient sprays such as those present in automotive applications. The planar mass fluxes for two high flow nozzles were estimated using the X-Ray tomography system in conjunction with a statistical image correlation velocimeter. The angular distribution of the planar mass flux obtained was validated using a mechanical patternator. In addition, the radial distribution was validated using an optical patternator. The integrated planar mass flux was found to be in excellent agreement with the inlet flow of the two nozzles. The Soft X-Ray tomography system has the potential to be used for the quality audit of large scale - high flow industrial nozzles as well as short duration automotive injectors where the high levels of obscuration renders conventional optical techniques ineffective.

S. E. Parrish, R. J. Zink, Y. Sivathanu, and J. Lim, **
Spray Patternation of a Multi-Hole Injector Utilizing Line-of-Sight Extinction Tomography**,
ILASS Americas, 22st Annual Conference on Liquid Atomization and Spray Systems, Cincinnati, OH, May 2010.

Spray characteristics of a multi-hole injector intended for use in direct-injection gasoline engines have been eva-luated for a variety of operating conditions including those that promote flash boiling. Planar line-of-sight extinc-tion tomography was employed to evaluate spray patterns to determine the influence of ambient pressure and tem-perature and fuel pressure and temperature on spray structure. Under severe flash boiling conditions, sprays were found to collapse and change shape significantly.

Yudaya Sivathanu and Jongmook Lim, **
Structure Measurements in Plumes burning Aluminized Propellants using Fan Beam Emission Tomography**,
Joint Army-Navy-NASA-Air Force (JANNAF) 43rd Combustion Subcommittee, Paper No. 793, 2009.

Fan Beam Emission Tomography is emerging as a valuable tool for determining the planar structure of fires and plumes. This paper provides an overview of the structure information in plumes obtained by burning small and large propellant chunks. This information is required to validate numerical codes as well as to design and develop safer solid rocket motors. The basic hardware required for fan beam emission tomography consists of two high-speed imaging spectrometers. The spectrometers are typically used to measure the path integrated spectral radiation intensities from multiple view angles. The spectrometers used in this study measured the spectral radiation intensity at 128 wavelengths and at 256 view angles. Each spectra was measured at 1320 Hz, providing for a total planar measurement at 11 Hz. These measurements are deconvoluted to provide the local gas and particulate concentrations as well as the temperature. The algorithm uses a Maximum Likelihood Estimation method in conjunction with a linearized form of the equation of radiation transfer. The basic assumptions involved in the deconvolution software are discussed followed by a sensitivity analysis for some key physical parameters used. The effects of multiple scattering of radiation on the estimated temperatures were also studied using numerical simulations. The numerical simulations used the estimated plume properties in conjunction with a Discrete Ordinate Method to calculate the radiation that would be measured by a spectrometer, taking scattering into consideration. The results show that the estimated particulate and gas temperatures do not change significantly due to the effects of multiple scattering within the plume.

Suo-Anttila, J., Gill, W., Sivathanu, Y., and Lim, J., **
Characterization of Aluminized Solid Propellant Surface Combustion**,
Joint Army-Navy-NASA-Air Force (JANNAF) 55th Propulsion Meeting, Paper No. 974, 2008.

The properties of burning aluminized solid propellant surfaces were characterized using a mid-infrared scanning spectrometer. The spectral dependence of the radiation energy emitted from the plume and the surface of the burning solid propellant was examined using a Spectraline ES-200 model spectrometer. The ES-200 measured spectral radiation intensity at 390 Hz in the 1.3 to 4.8 micron wavelength range with a resolution of 22 nm. The propellant was a 12.7 cm diameter by 3.5 cm tall sample of inhibited propellant (AP/HTPB/Al composite), which was ignited inside a facility with controlled airflow. The spectrometer was placed above the propellant, at an angle, and it scanned a 20 cm line centered about the surface of the burning aluminized solid propellant. The spectrometer was protected from the environment using radiation shields and a small nitrogen purge to prevent particulate from collecting on the window. Spectra obtained from the diagnostic include emission from both the plume and the propellant surface and analysis of the data to provide information on the surface properties requires knowledge of the independent plume emission. Previous propellant fire spectra have shown that the plume radiation has two very prominent peaks corresponding to water vapor at 2.7 microns and carbon dioxide at 4.5 microns. Analysis of the plume spectra relative to the surface spectra from calorimeters in past experiments resulted in the selection of three bands, 1.25 to 1.65 micron, 2.17 to 2.28, and 3.60 to 3.90, for the data reduction. The temperature and emissivity of the surface determined from the measured radiation intensities was approximately 1700 K and 0.2, respectively. The experimental data provide valuable information on the properties of the burning propellant surface for phenomenon discovery and the development and validation computational models.

Sivathanu, Y., Lim, J., Reinhart, L. E., Bowman, R. C., Gill, W., and Figueroa, V., **
Structure of Aluminized Propellant Plumes Obtained using Multi-Angular Infrared Emission Spectroscopy**,
Joint Army-Navy-NASA-Air Force (JANNAF) 55th Propulsion Meeting, paper No. 941, 2008.

Understanding the structure of solid propellant plumes is important for the design and safe operation of solid rocket motors. This paper demonstrates the use of multi-angular emission spectroscopy for estimating spatially resolved gas and particulate temperatures in plumes obtained by burning large propellant chunks. The multi-angular spectral radiation intensities from the plume were measured using two high speed imaging spectrometers. The imaging spectrometers measured the spectral radiation intensities from 1 to 5 microns at a frequency of 1320 Hz. The imaging spectrometers had linear scanners integrated in their optics which enables measurement at 128 different view angles from each spectrometer. Therefore, the radiation emitted by the plume was measured at 256 view angles at a frequency of 11 Hz. The path integrated measurements were deconvoluted using a maximum likelihood estimation method in conjunction with a linearized equation of radiative transfer. The gas and particulate properties required for the method were obtained from the literature. Several different sizes of propellant chunks, ranging from six to twenty inches in diameter, were used in the experiments. The results obtained at two axial locations during the burn tests are reported. The results show that in most cases, it is possible to estimate the gas and particulate temperatures as well as their concentrations using emission spectroscopy. In addition, emission spectroscopy can be used to estimate the surface temperatures of solid objects that are immersed in the plume.

D. A. Feikema, W. Kim, and Y. Sivathanu, **
A Small Angle Scattering Sensor System for the Characterization of Combustion Generated Particulate**,
45th Aerospace Sciences Meeting & Exhibit, Paper No. 2007-0, AIAA, Washington, DC.

One of the critical issues for the US space program is fire safety of the space station and future launch vehicles. A detailed understanding of the scattering signatures of particulate is essential for the development of a false alarm free fie detection system. This paper describes advanced optical instrumentation developed and applied for fire detection. The system is being designed to determine four important physical properties of disperse fractal aggregates and particulates including size distribution, number density, refractive indices, and fractal dimension. Combustion generated particulate are the primary detection target; however, in order to discriminate from other particulate, non-combustion generated particles should also be characterized. The angular scattering signature is measured and analyzed using two photon optical laser scattering. The Rayleigh-Debye-Gans (R-D-G) scattering theory for disperse fractal aggregates is utilized. The system consists of a pulsed laser module, detection module and data acquisition system and software to analyze the signals. The theory and applications are described.

Paul Hicks, Yudaya Sivathanu, and Jongmook Lim, **
Optical and Mechanical Patternation of an High Flow Rate Industrial Gas Turbine Nozzle**,
ILASS Americas, 21st Annual Conference on Liquid Atomization and Spray Systems, Orlando, Florida, May 2008.

One of the key sources of uncertainty in patternation is the misalignment of the nozzle to the center of the patternation device. A systematic study of this uncertainty was completed using a mechanical and an optical patternator. Two types of misalignment were explored during the study. The first is the position misalignment of the nozzle to the center of the patternator. The second is an angular misalignment of the nozzle to the plane of patternation. The latter results in a much lower uncertainty than the former for mechanical patternators. For optical patternators, corrections can be applied for the first type of misalignment. Therefore, angular misalignment results in a greater uncertainty than the position misalignment for optical patternators. A simulation of the spray mass flux for the first type of misalignment also supports these conclusions.

Yudaya Sivathanu and Jongmook Lim, **
Visible Light Extinction Tomography for Dense Sprays**,
ILASS Americas, 20th Annual Conference on Liquid Atomization and Spray Systems, Chicago, May 2007.

This study examines the feasibility of visible light extinction tomography for understanding the structure of dense sprays. The SETScan patternator was used to obtain local surface area densities in aero-engine nozzles, flowing fuel at very high flow rates. The contour map of the local surface area of the drops within the spray showed a hollow cone behavior. The results indicate that the SETscan patternator can reliably obtain the spray pattern and surface area densities in high flow rate aero-engine nozzles.

Yudaya Sivathanu, Jongmook Lim, Ariel Muliadi , Paul E. Sojka, Yong Chen, Nitin Sharma, and Prabodh Varanasi, **
Measurement of Spatially Resolved Mean Velocities in a Transient Spray using Statistical Image Correlation Velocimetry**,
ILASS Americas, 20th Annual Conference on Liquid Atomization and Spray Systems, Chicago, May 2007.

Statistical Image Correlation Velocimetry (SICV) was used to obtain mean velocity information in the spray issuing from a consumer fragrance injector. Statistical Image Correlation Velocimetry consists of obtaining several ensembles of the entire transient spray event at 1000 Hz. The spatial correlation coefficient between all the pixels in the image plane is then calculated from the ensemble of images. Based on the magnitude of the spatial correlation coefficient at different locations in the image plane, the mean velocity distribution is obtained. These velocities can be estimated from spatial correlation of streaklines in the images, without needing to resolve individual droplets. Furthermore, it is also not necessary to have well defined streaklines to estimate the velocity because the correlation is based on the statistics of the successive images, rather than on a single set of successive shots. The velocities obtained using SICV were compared with those obtained using a conventional Phase Doppler Anemometer (PDA). The mean velocities estimated from the fragrance dispenser compare reasonably well with those obtained using conventional PDA.

Ariel R. Muliadi, Paul E. Sojka, Yudaya R. Sivathanu, and Jongmook Lim, **
Comparison of Particle Dynamics Analyzer (PDA) and Setscan Optical Patternator Results**,
Proceedings of 2007 ASME International Mechanical Engineering Congress and Exposition, November 11 - 15, 2007, Seattle, Washington.

The primary goal of this study was to determine when patternation information derived from Particle Dynamics Analyzer (Dantec Dynamics dual-PDA) measurements of volume flux, velocity and mean drop size agreed with corresponding values measured using an optical patternator (En�Urga, Inc SetScan OP-600). To achieve this, data from each instrument was transformed into spatially resolved absorptances (equivalent to drop surface area per unit volume) and compared. The secondary goal of this study was to explain the cause of any discrepancies in comparison of the two absorptance sets when they occurred. Key conclusions drawn from this study are: absorptance agreement to within 20% can be achieved in many cases; however, the difference between the PDA-calculated and optical patternator-measured absorptances becomes larger as the drop arrival rate increases, as the drop size distribution becomes wider, and when a significant drop size-velocity correlation is present. These discrepancies are attributed to an underestimation of the volume flux (which becomes more important with increasing droplet arrival rate), an over-reporting of the mean drop diameter (which is the result of the restrictive data acquisition scheme applied when ensuring mass closure), by the limited PDA dynamic range (which can preclude simultaneously accounting for both the largest and smallest drops in the spray), and by the optical patternator�s number-density based measurement scheme (which is known to be sensitive to drop size-velocity correlations).

Yudaya Sivathanu, Jongmook Lim, and Paul Hicks, **
Optical and Mechanical Patternation of an High Flow Rate Industrial Gas Turbine Nozzle**,
Proceedings of 2007 ASME International Mechanical Engineering Congress and Exposition, November 11 - 15, 2007, Seattle, Washington.

Many industrial gas turbine nozzles have a very high flow rate of fuel. Performing optical and mechanical charac-terization of these high flow rate nozzles is of interest for quality assurance and nozzle improvement programs. In particular, the objective of the study is to determine the impacts on performance of a nozzle tip re-design initiated to improve fuel injector performance. Towards this end, optical and mechanical patternation of the original and a pro-posed redesign are discussed. The SETscan optical patternator that is used in this study is based on statistical de-convolution of path integrated extinction measurements obtained at six view angles and 512 parallel paths at each view angle. The local drop surface area per unit volume is obtained from the deconvolution. The drop surface areas are directly proportional to the local mass, momentum, energy, and species transfers. Although obscuration caused by the fuel drops was greater than 90% at the center of the spray, the theoretically calculated path integrated extinc-tion based on the local surface areas reported by the patternator agrees within 1% of the measured path integrated extinction. Therefore, the nozzles provide an efficacy test of the SETscan optical patternator for highly obscuring sprays. Also, mechanical patternation is performed using a twenty four-sector patternator. The angular distribution of mass fluxes obtained from both the mechanical patternator compares reasonably well with the angular distribution of surface area densities obtained from the optical patternator. The differences between the original and the re-designed nozzle are immediately apparent with the optical patternator. Based on the study, the feasibility of evaluat-ing design and development iterations of high flow rate nozzles using optical patternation is demonstrated.

Sivathanu, Y., Lim, J., Reinhart, L., and Bowman, R., **
Emission Tomography in Propellant Strands**,
presented at the 2006 JPL/JANNAF Ambient Atmosphere Solid Propellant Combustion Workshop, Florida.

Fan beam tomography was used to determine the temperature, gas concentrations, and particulate volume fractions in fires from solid propellant strands. Two Spectraline Model ES-100 spectrometers were used to obtain path integrated radiation intensities from the fires. The measurements were deconvoluted to show the gas and particulate temperatures as well as their concentrations.

Yudaya Sivathanu, Jongmook Lim, Henrik Linden and Preben Noerskov, **
Pressure Effects on Drop Diameters and Surface Areas in Oil Nozzles**,
ILASS Americas, 19th Annual Conference on Liquid Atomization and Spray Systems, Toronto, Canada, May 2006.

An experimental study of the characteristics of the sprays from four oil nozzles, operated at different pressures, is reported. The four oil nozzles have the same mass flow rate, but different spray angles. Three instruments were used to characterize the sprays from the oil nozzles. The drop diameters were measured at two axial locations using Fraunhofer diffraction. Planar drop surface area densities were measured at the same two axial locations using an optical patternator. The optical patternator provided the planar drop surface area densities from path integrated extinction measurements obtained at six view angles and 256 parallel projects per view angle. The planar drop surface area densities were used to estimate spray angles as well as the angular and radial uniformity. In addition, the optical patternator provided data at 1000 Hz, which was used to obtain the turbulence intensities for the nozzles. A mechanical patternator was used to estimate the mass based angular and radial uniformity of the sprays. The Sauter Mean drop diameters generally decrease with pressure. The total planar drop surface area densities for the nozzles are highly correlated with pressure. The total planar drop surface area densities increases linearly with pressure, the linearity of the correlation being greater than 99%. A normalization scheme was used to compare the results obtained from the mechanical and optical patternators. The mechanical and optical patternator provided consistent spray uniformity results at most locations within the spray. However, the mechanical patternator could not provide an accurate estimate of the spray mass flux at the center of the oil nozzles due to flow stagnation at the center of the patternator. This implies that radial patternation of oil nozzles should not be performed using mechanical patternators. The turbulence intensities of the oil nozzles studied were very low, and less than 5%. The turbulence intensities of the sprays decrease with increase in pressure.

D. A. Feikema, J. Lim, and Y. Sivathanu, **
Flame Structure and Scalar Properties in Microgravity Laminar Flames**,
Proceedings of the 2006 Technical Meeting of the Central States Section of the Combustion Institute, Cleveland, OH pp., 2006.

Recent results from microgravity combustion experiments conducted in the Zero Gravity Facility (ZGF) 5.18 second drop tower are reported. Emission mid-infrared spectroscopy measurements have been completed to quantitatively determine the flame temperature, water and carbon dioxide vapor, radiative emissive power, and soot concentrations in a microgravity laminar ethylene/air flame. The ethylene/air laminar flame conditions are similar to previously reported experiments including the Flight Project, Laminar Soot Processes (LSP). Soot concentrations and gas temperatures are in reasonable agreement with similar results available in the literature. However, soot concentrations and flame structure dramatically change in long duration laminar diffusion flames as demonstrated in this paper.

J. Lim, Y. Sivathanu, and D. Feikema, **
Fan Beam Emission Tomography**,
44th Aerospace Sciences Meeting & Exhibit, Paper No. 2006-0436, AIAA, Washington, DC.

A new method of estimating temperature and gas species concentrations from mid infrared spectral radiation intensity measurements in laminar, axisymmetric flames is reported. The spectral radiation intensities emitted from a co-flow burner are measured at multiple wavelengths and at multiple view angles using a high speed mid infrared spectrometer in conjunction with a scanner. The spectrometer with the scanner obtains the spectral radiation intensities from 1.3 to 4.8 μm at 128 discrete view angles. Radiation from the H2O and CO2 gas bands are used to estimate gas temperature and mole fractions of CO2 and H2O gas. The deconvolution algorithm is based on a Maximum Likelihood Estimation method in conjunction with a linearized equation of radiative transfer. The algorithm is first verified with synthetic data. The deconvolution algorithm obtains spatially resolved temperatures and gas concentrations from the synthetically generated spectral radiation intensities. The algorithm is then applied to measurements obtained from the co-flow burner in a laminar methane/air diffusion flame. The data obtained from the co-flow burner using the present methods is in reasonable agreement with values available in the literature.

Jongmook Lim, Yudaya Sivathanu, and Paul E. Sojka, **
Comparison of Drop Surface Area Measurements obtained using Extinction and Diffraction**,
ILASS Americas, 18th Annual Conference on Liquid Atomization and Spray Systems, Irvine, CA, May 2005.

Comparison of spray characteristics using different techniques is important to establish the validity of the techniques, as well as to provide for accurate estimates of uncertainties. This paper compares line-of-sight drop surface areas obtained from a spray using two different optical methods, namely statistical extinction tomography and Fraunhofer diffraction. Statistical extinction tomography obtains path integrated extinction measurements at multiple view angles and at multiple slices at each view angle. The path integrated extinction measurements are deconvoluted using a Maximum Likelihood Estimation method to provide the local extinction coefficients. For liquids with drop sizes greater than 1 micron, and refractive indices close to unity, the local extinction coefficients are identical to the drop surface area per unit volume. Fraunhofer diffraction measures the light scattered at multiple angles from the laser line through the spray. An inverse algorithm is then used to provide the drop sizes and extinction values from the Mie-Scattering theory, assuming negligible absorption of the incident laser light. A water spray from a commercial nozzle was used to compare the measurements using the two different techniques. Measurements using both techniques were obtained at two axial locations (30 and 60 mm from the injector exit). To minimize errors in the diffraction measurements due to multiple scattering effects, the flow rate was chosen to be very low. Diffraction measurements were obtained at several parallel paths through the spray, by moving the injector radially at a fixed axial location. The statistical extinction tomography measurements were obtained over the entire plane simultaneously. The extinction measured by the Fraunhofer diffraction system tends to be higher than those measured by the extinction tomography system at all radial locations. Surface area per unit volume measurements showed reasonable agreement at the upstream location (X=30 mm), i.e. to within about 25%. However, the measurements showed considerable differences at the downstream location (X=60 mm).

Jongmook Lim, Yudaya Sivathanu, and Douglas Feikema, **
Fan Beam Emission Tomography**,
Proceedings of the Western States Section of the Combustion Institute, Paper No. 04S-21, Davis, CA, 2004.

A new method of estimating temperature, soot volume fraction, and gas species concentrations from mid infrared spectral radiation intensity measurements in laminar axisymmetric flames is reported. The spectral radiation intensities emitted from a co-flow burner are measured at multiple wavelengths and at multiple view angles using a high speed mid infrared spectrometer in conjunction with a scanner. The spectrometer with the scanner obtains the spectral radiation intensities from 1.3 to 4.8 microns at 128 discrete view angles. Radiation from the soot, H2O and CO2 gas bands are used to estimate soot temperature, soot volume fraction, gas temperature and mole fractions of CO2 and H2O gas. The deconvolution algorithm is based on a Maximum Likelihood Estimation method in conjunction with a linearized equation of radiative transfer. The algorithm is first verified with synthetic data. The deconvolution algorithm obtains spatially resolved temperatures, soot volume fractions, and gas concentrations from the synthetically generated spectral radiation intensities. The algorithm is then applied to measurements obtained from the co-flow burner. The data obtained from the co-flow burner using the present methods agrees reasonably well with values available in the literature.

Yudaya Sivathanu, Jongmook Lim, and Ye Mi, **
Optical Patternation of a Diesel Engine Injector under Cold Conditions**,
Proceedings of the 16th Annual Conference on Liquid Atomization and Spray Systems, Monterey, CA , 2003.

Patternation of diesel injectors with an extinction tomography based optical patternator is reported. Path integrated transmittances were obtained from a diesel injector under cold conditions. The path-integrated transmittances were deconvoluted using the maximum likelihood estimation method. The total drop surface area per unit volume across a horizontal plane was obtained at a frequency of 1 KHz. For modeling the combustion process of a diesel spray, the three important parameters that are required are the drop surface area per unit volume and the characteristic length and time scales. The temporal resolution of 1 KHz used for the experiments was not sufficient to obtain the characteristic length and time scales for the diesel injector. However, spatially resolved drop surface areas per unit volume were obtained using the optical patternator. The drop surface areas per unit volume were used to obtain patternation numbers for the injector. The results show that deconvolution of path-integrated transmittances provide a very reliable method of obtaining total surface area per unit volume across any plane in the spray. The patternator also provides an easy method to determine the quality of diesel injectors for quality control applications.

Yudaya Sivathanu, Jongmook Lim, and Douglas Feikema, **
Fan Beam Emission Tomography for Laminar Fires,**,
Seventh International Workshop on Microgravity Combustion and Chemically Reacting Systems, NASA/CP-2003-212376, pp. 237-240.

Obtaining information on the instantaneous structure of turbulent and transient flames is important in a wide variety of applications such as fire safety, pollution reduction, flame spread studies, and model validation. Different experimental techniques such as Tunable Laser Absorption Spectroscopy, Fourier Transform Infrared Spectroscopy, and Flame Emission Spectroscopy to mention a few. Diagnostics with high power lasers are difficult to implement in microgravity environment. Most flames emit significant radiation signatures that are used in various applications such as fire detection, light-off detection, flame diagnostics, etc. Radiation signatures can be utilized to maximum advantage for determining structural information in turbulent flows. Emission spectroscopy is most advantageous in the infrared regions of the spectra, principally because these emission lines arise from transitions in the fundamental bands of stable species such as CO2 and H2O. Based on the above, the objective of this work was to develop a fan beam emission tomography system based on mid infrared spectrometers to obtain the local scalar properties such as temperature and mole fractions of major gas species from path integrated multi-wavelength infrared radiation measurements.

M. Bundy, G. W. Mulholland, S. Manzello, J. Yang, J. H. Scott, and Y. Sivathanu, **
Microgravity Superagglomerates Produced by Silane and Acetylene**,
41st Aerosapce Sciences Meeting and Exhibit, January 2003, Reno, NV, AIAA Paper: 2003-986.

A study of acetylene and silane diffusion flames was performed in normal gravity (1-g) and microgravity (μ-g) to determine the evolution of soot and silica agglomerates. The objective of this study was to understand the process of gas phase agglomeration leading to superagglomerates and a gel-like structure for 1-g silane and μ-g acetylene flames. Silane flames were observed in normal gravity. Long fibrous particles were produced that consisted of partially fused spherule structures as well as cage-like structures at high silane flow. Microgravity acetylene soot plume development was observed at fuel flow rates of 0.6 cm3/s, 1.0 cm3/s, and 2.0 cm3/s. Primary particle size distributions were measured using TEM analysis on soot particle agglomerates sampled using a thermophoretic probe in the post-flame region of the plume. Soot superagglomerates were observed in microgravity as large as 1 mm using direct imaging of the soot plume. For the first time, visible particles are imaged both in the luminous flame zone and the plume. A numerical flame model was used to predict the profiles for particle volume fraction, temperature, species, and velocity. The computations showed good qualitative agreement with the visible flame and plume structure from the 1-g and μ-g acetylene experiments. The flame model results were used as input conditions for a simulation to predict the agglomerate growth and gelation. The agglomeration simulations predict that the formation of large soot agglomerates (>100 μm) are possible at μ-g conditions but not at 1-g based on free molecular agglomeration including gelation.

Suleyman Gokoglu, Matthew Bundy, George W. Mulholland, Samuel Manzello, Jiann Yang, John Henry Scott, Yudaya Sivathanu, **
Microgravity Superagglomerates Produced by Silane and Acetylene**,
Seventh International Workshop on Microgravity Combustion and Chemically Reacting Systems, NASA/CP-2003-212376, pp. 57-60.

The size of the agglomerates produced in the upper portion of a flame is important for a variety of applications. Soot particle size and density effect the amount of radiative heat transfer from a fire to its surroundings. Particle size determines the lifetime of smoke in a building or in the atmosphere, and exposure hazard for smoke inhaled and deposited in the lungs. The visibility through a smoke layer and dectectability of the smoke are also greatly affected by agglomerate size. Currently there is limited understanding of soot growth with an overall dimension of 10 m and larger. In the case of polystyrene, smoke agglomerates in excess of 1 mm have been observed raining out from large fires. Unlike hydrocarbon fuels, silane has the advantage that silica particles are the major combustion product resulting in a particle volume fraction a factor of ten greater than that for a carbonaceous smoke. There are two very desirable properties of silica aero-gels that are important for both space and earth based applications. The first important property is its inertness to most oxidizing and reducing atmospheres. Therefore, silica aero-gels make excellent fire ablatives and can be used in very demanding applications. The second important property is that silica aero-gels are expected to have very high porosity (greater than 0.999), making them lightweight and ideal for aerospace applications. The added benefit of the high porosity is that they can be used as extremely efficient filters for many earth based applications as well. Evidence of the formation of superagglomerates in a laminar acetylene/air diffusion flame was found by Sorensen et al. [1]. An interconnecting web of super-agglomerates was observed to span the width of the soot plume in the region just above the flame tip and described as a gel state. It was observed that this gel state immediately breaks up into agglomerates as larges as 100 m due to buoyancy induced turbulence. Large soot agglomerates were observed in microgravity butane jet diffusion flames by Ito et al.[2]. Several other works to date have studied the effect of flame structure on soot volume fraction and agglomeration size in a microgravity environment.[3-4]. In microgravity the absence of buoyant convective flows increases the residence time in the flame and causes a broadening of the high temperature region in the flame. Both of these factors play a significant role in gas phase radiation and soot formation.

Yudaya Sivathanu and Jongmook Lim, **
Estimating Gas Temperature and Mole Fractions from Spectral Radiation Measurements in a Homogeneous Hot Gas Layer**,
2003 Propulsion Measurement Sensor Development Workshop, Huntsville, AL.

A mid infrared spectrometer was used to obtain path integrated emission intensities from a flat flame diffusion burner using premixed methane/air flames. The emission intensities were deconvoluted using a Linear Equation of Radiative Transfer to estimate the gas species concentrations and temperatures. Measurements were also obtained using a thermocouple and a gas chromatograph. The values obtained from emission spectroscopy was found to be in good agreement with those obtained from the thermocouple and the gas chromatograph. The utility of using mid infrared spectrometers to obtain flame strucutre in premixed hydrocarbon flames was thus successfully demonstration.

J. Lim, Y. Sivathanu, and D. Feikema, **
Fan Beam Emission Tomography for Estimating Scalar Properties in Laminar Flames**,
Proceedings of the 3rd Joint Meeting of the U.S. Sections of The Combustion Institute, Paper No. D06.

A new method of estimating temperatures and gas species concentrations (CO2 and H2O) in a laminar flame is reported. The path-integrated, spectral radiation intensities emitted from a laminar flame at multiple wavelengths and view angles are calculated using a narrow band radiation model. Synthetic data, in the form of radial profiles of temperature and gas concentrations, are used in these calculations. The calculations mimic measurements that would theoretically be obtained using a mid-infrared spectrometer with a scanner. The path�integrated spectral radiation intensities are deconvoluted using a maximum likelihood estimation method in conjunction with an iterative scheme. The deconvolution algorithm accounts for the self-absorption of radiation by the intervening gases, and provides the local temperature and gas species concentrations. The deconvoluted temperatures and gas concentrations are compared with the synthetic data used for calculating the spectral radiation intensities. The deconvoluted temperatures and gas species concentrations are within 0.5 % of the synthetic data. The deconvolution algorithm is expected to provide combustion researchers with an easy method of obtaining the radial profiles of major gas species concentrations and temperatures in laminar flames non-intrusively using a mid-infrared spectrometer with a scanner.

J. Lim and Y. Sivathanu, **
Tomographic Reconstruction of Spatial Correlation in a Turbulent Paint Spray**,
Proceedings of the 2002 Painting Technology Workshop, Lexington, Kentucky.

One of the important spray parameters that influences the quality of painted products is the spatial correlation of drops within the spray. For instance, if the spatial correlation between drops in a flat fan spray are highly negative, then uniformity of the coating over an object that is transported on a conveyor belt will be compromised. High speed path integrated extinction measurements were obtained from a paint spray at multiple view angles and projects. These measurements were deconvoluted to provide the spatial correlation of drop surface areas in a flat fan spray. This paper describes the mathematical method used to obtain the spatial correlations, and validation of the method using synthetically generated data. The measurements, which were obtained with the SETscan optical patternator, also showed that for some fan sprays, the spatial correlation coefficeint is negative at certain separation distances.

Y. Sivathanu and J. Lim, **
Optical Patternation of Paint Sprays Using Extinction Tomography,**,
Proceedings of the 2002 Painting Technology Workshop, Lexington, Kentucky.

Spray uniformity is very important in painting applications. In the furniture and automotive industry, painting is often accomplished by moving the object along a conveyor belt with a fixed spray painting station. If the spray is non-uniform, the resulting coating will have several imperfections. This paper discusses the use of an Statistical Extinction Tomography based optical patternator to infer the uniformtity of sprays. The optical patternator uses extinction measurements obtained over six view angles and 256 projections at each view angle to estimate the local surface area density of the drops. The patternator quickly provides the major and minor spray angle as well as the ellipcity of the spray. When the paint nozzle is worn, the spray angles will become larger resulting in wastage of paint due to overspray.

Kim, W. C., Sivathanu, Y. R. and Gore, J. P., **
Characterization of Spectral Radiation Intensities from Standard Test Fires for Fire Detection**,
12th International Conference on Automatic Fire Detection, March 26-28, 2001, National Institute of Standards and Technology, Gaithersburg, MD.

Spectral radiation intensities at 160 mid infrared wavelengths leaving six standard test fires specified in the guidelines of the European Committee for Standardization (ECN Fires) were measured. A recently developed Fast Infrared Array Spectrometer (FIAS) was utilized to acquire spectral radiation intensities in the 1.8 � 4.9 μm wavelength range from the transient as well as the steady burning fires. The mean and root mean square, the Probability Density Function (PDF), the Cumulative Probability Density Function (CPDF), and the Power Spectral Density (PSD) of the spectral radiation intensities were studied to characterize the fires. The spectral radiation intensity in the 4.3 μm micron carbon dioxide band is larger than that at other wavelengths for the open fires. The continuum radiation from soot particles dominates gas band radiation for the smoldering fires. The statistical characteristics of the fires depend on the fuel type, time after ignition and measurement wavelength and can be used in fire detection algorithms.

J. Lim, Y. Sivathanu, Y. Xin, and J. Gore, **
Measurements of Spectral Radiation from an Ethylene Pool Fire using Fast Array Infrared Spectrometer**,
The Third Asia-Pacific Conference on Combustion, June 24-27, 3C303.

Spectral radiation intensity from a 7.1 cm ethylene pool fire was measured using Fast Array Infrared Spectrometer over the 1.3 to 4.8 μm spectral range at 390 Hz. The mean and RMS of the spectral radiation intensities were computed based on 5000 samples at different radial and axial positions. In addition, the PDF of spectral radiation intensities for the 2.7 micron water vapor band was also obtained. The results shows that the spectral radiation intensity have relatively small fluctuations near the base of the flame. At axial locations downstream from the base, flame intermittency causes very strong fluctuations in the measured spectral radiation intensities. The radiation intensities are closely related to the local flame temperatures, gas and particulate concentrations, and the turbulence intensities. Therefore, a systematic study of these intensities will provide a better understanding to pool fire structure as well as provide validation data for numerical codes.

Y. Sivathanu, J. Lim, V. Narayanan, and R. Joseph, **
Fan Beam Emission Tomography for Transient Fires**,
Sixth International Microgravity Combustion Workshop NASA/CP-2001-210826, NASA Glenn Research Center, pp. 369-372.

Radiation signatures can be utilized to maximum advantage for determining structural information in turbulent flows. Emission spectroscopy does not require a light source, and alignment and data collection are relatively straightforward. Emission spectroscopy is most advantageous in the infrared regions of the spectra, principally because these emission lines arise from transitions in the fundamental bands of stable species such as CO2 and H2O. Emission spectroscopy also currently offers the most accurate method of temperature determination. However, emission spectroscopy is a line-of-sight method and deconvolution of measurements has to be addressed with all emission spectroscopy methods. The objective of the present work is to develop Fan Beam Emission tomography methods for the transient fires that are currently being supported by the NASA microgravity program.

G. W. Mulholland, J. C. Yang, J. H. Scott, and Y. Sivathanu, **
KISS: Kinetics and Structure of Superagglomerates Produced by Silane and Acetylene**,
Sixth International Microgravity Combustion Workshop NASA/CP-2001-210826, NASA Glenn Research Center, pp. 289-292.

The objective of this study is to understand the process of gas phase agglomeration leading to superagglomerates and a gel-like structure for microgravity (0-g) silane and acetylene flames. Ultimately one would apply this understanding to predicting flame conditions that could lead to the gas phase production of an aero-gel. The approach is to burn acetylene and silane and to analyze the evolution of the soot and silica agglomerates. Acetylene is chosen because it has one of the highest soot volume fractions and there is evidence of superagglomerates being formed in laminar acetylene flames. Silane has the advantage that silica particles are the major combustion product resulting in a particle volume fraction a factor often greater than that for a carbonaceous smoke.

Y. Zheng, J. Gore, and Y. Sivathanu, **
Infrared Spectral Imaging of Jet Flames**,
Paper No. 53, Technical Meeting of The Central States Section of the Combustion Institute.

An infrared (IR) camera in conjunction with a spectrograph and a digital frame grabber were used to obtain rapid simultaneous measurements of spectral radiation intensities along chord-like paths in laminar jet flames. The results from two differnet laminar co-flow jet flames with an identical nominal heat release rate of 2.1 kW are reported. A non-premixed C2H4/air flame and a partially premixed C2H4/air flame with Φ = 1.6 were considered to investigate the effects of partial premixing. The camera captured multi-spectral infrared line images within a wavelength range from 2.2 to 4.9 μm for each flame. Separate blackbody images were used to calibrate the sensitivity, wavelength and radial location associated with each pixel of the 316 x 240 array. After calibration, spectral radiation intensities for chord-like paths with high spatial resolution were obtained. The spectra revealed the combined effects of temperature, CO2, H2O, fuel, and soot distributions.

G. W. Mulholland, A. Hamins, and Y. Sivathanu, **
Kinetics and Structure of Superagglomerates Produced by Silane and Acetylene**,
Fifth International Microgravity Combustion Workshop NASA/CP-1999-208917, NASA Glenn Research Center, pp. 491-494.

The evolution of smoke in a laminar diffusion flame involves several steps. The first step is particle inception/nucleation in the high-temperature fuel-rich region of the flame followed by surface growth and coagulation/coalescence of the small particles. As the primary spheres grow in size and lose hydrogen, the colliding particles no longer coalesce but retain their identity as a cluster of primary spheres, termed an agglomerate. Finally, in the upper portion of the flame, the particles enter an oxidizing environment which may lead to partial or complete burnout of the agglomerates. Currently there is no quantitative model for describing the growth of smoke agglomerates up to superagglomerates with an overall dimension of 10 microns and greater. Such particles are produced during the burning of acetylene and fuels containing benzene rings such as toluene and polystyrene. In the case of polystyrene, smoke agglomerates in excess of 1 mm have been observed "raining" out from large fires. Evidence of the formation of superagglomerates in a laminar acetylene/air diffusion flame has been recently reported. Acetylene was chosen as the fuel since the particulate loading in acetylene/air diffusion flames is very high. Photographs were obtained by Sorensen using a microsecond xenon lamp of the "stream" of soot just above the flame. For low flow rates of acetylene, only submicrometer soot clusters are produced and they give rise to the homogeneous appearance of the soot stream. When the flow rate is increased to 1.7 cu cm/s, soot clusters up to 10 microns are formed and they are responsible for the graininess and at a flow rate of 3.4 cu cm/s, a web of interconnected clusters as large as the width of the flame is seen. This interconnecting web of superagglomerates is described as a gel state by Sorensen et al (1998). This is the first observation of a gel for a gas phase system. It was observed that this gel state immediately breaks up into agglomerates due to buoyancy induced turbulence and gravitational sedimentation.

J. Ji, Y. R. Sivathanu, and J. P. Gore, **
Thin Filament Spectral Pyrometry for Flame Temperature Measurements,**,
First Joint Meeting of The United State Sections of The Combustion Institute, Washington, DC.

The drive for higher efficiency and lower pollutant emissions has necessitated lean premixed combustion on commercial gas turbine engines. Turbulent flames are complex because of its spatial and temporal inhomogenieties of scalars in the flow field. Modeling such complex phenomena needs experimental diagnostics with high spatial and temporal resolution and high accuracy as well as precision for temperature measurements. The objective of this study was to develop, calibrate, and validate a multiwavelength thin filament pyrometric technique to obtain spatially and temporally resolved temperatures in a premixed turbulent flame.

C. Fu, P. E. Sojka, and Y. R. Sivathanu, **
On the Interaction Between Evaporating Sprays and Heated Surfaces,**,
Proceedings of the 12th Annual Conference on Liquid Atomization and Spray Systems, Indianapolis, IN pp. 271-276, 1999.

An experimental study mapping the size and velocity history of water droplets in a polydispersed spray as they approach a heated surface was conducted. A Phase Doppler Particle Analyzer (PDPA) was use to record the size and velocity histories. Different mass flow rate were applied to study the effect of both buoyancy and evaporation on droplet trajectory. The initial surface termperature was about 600 OC. Surface temperature profiles were recorded to study the influence of water mass flow rate and drop size distribution on the hot surface temperature profile. The influence of water mass flow rate and drop size distribution on the hot surface temperature profile, and on the variation of size and velocity distribution, are presented and discussed. The long term objective is to produce experimental data that can be used to validate submodels for four four key physical phenomena involved in the interaction of sprays with burning surfaces: (1) the effect of buoyancy (caused by hot combustion products) on the trajectory of a single droplet; (2) the effect of evaporation on the trajectory of a single droplet; (3) the cessation of reaction and reduction in flame spread caused by the droplets on flaming surface combustion; and (4) the reduction surface temperature caused by the effect of of drop imingement, spreading and evaporation on surface combustion.

C. Hagwood, G. W. Mulholland, and Y. Sivathanu, **
The DMA Transfer Function with Brownian Motion: A Trajectory/Monte Carlo Approach,**,
Proceedings of the American Association of Aerosol Research, AAAR�99-18th Annual Conference, Paper 6C1.

The transfer function for the Differential Mobility Analyzer (DMA) is derived based on particle trajectories for both non-diffusing particles and diffusing particles. The effect of particle diffusion is assessed by using a Monte-Carlo method for particles of sizes 1,3,10,30, and 100 nm. This approach includes both the effect of wall losses and axial diffusion. The range of validity of the Stolzenburg analysis is assessed by comparing his transfer function, the peak of his transfer function, and its dimensionless width with similar calculations based on the Monte-Carlo. For particle sizes smaller than 10 nm, the Monte-Carlo method indicates large wall losses, which result in a reduction in the peak of the transfer function by as much as a factor of 10 to 30, sensitivity to the flow-field, and skewness of the transfer function. It is shown that Stolzenburg's approximate formula for the standard deviation of the width of the transfer function agrees with Monte-Carlo simulations for particle sizes of 3 nm and larger.

Y. R. Sivathanu, J. P. Gore, and Y. J. Zhu, **
Monte Carlo Simulations of Radiative Transfer in a House Including Specular and Diffuse Reflections for the Evaluation of Two Wavelength Optical Fire Detectors**,
Annual Conference on Fire Research, Gaithersburg, MD, pp. 5-6.

Fire detectors based on sensing thermal radiation from a strating fire have the potential to provide early warning and improve fire safety and reduce fire loss. However, these detectors must rely on sensing and discriminating radiation signals of a starting fire from those of common sources such as the sun, room lights and others. The radiaiton signal incident on a detector may have undergone multiple relfections at the walls with complex properties and configuration factors. Effective radiation temperature based on measurement of intensities at two wavelengths is one of the fire characteristics used in many flame sensors.

S. D. Sovani, P. E. Sojka, and Y. R. Sivathanu, **
Prediction of Drop Size Distributions from First Principles: Joint-pdf Effects**,
Eleventh Annual Conference on Liquid Atomization and Spray Systems, Ontario, Canada, pp. 376-380.

The first principles analytical method for predicting drop size distributions in sprays developed in a previous study, and used to investigate the separate influences of fluctuations in liquid physical properties and relative velocity on the width of the drop size distribution, was extended to study how joint-pdf fluctuations impact the width of the drop size distribution. Two types of fluctuations were considered: those due to relative velocity and those due to physical properties. These fluctuations were coupled through joint-pdfs of velocity and viscosity, velocity and surface tension, and velocity and density. Results demonstration how joint fluctuations lead to a distribution of drop sizes.

Y. Zhu, Y. Sivathanu, and J. Gore, **
Fire Detection System Design Using Statistical Modeling of Enclosure Radiative Heat Transfer**,
Proceedings of the 1998 Technical Meeting of the Central States Section of the Combustion Institute, The Combustion Institute, Pittsburgh, PA.

A statistical technique based on the Discrete Probability Function (DPF) concept was used to model radiative heat transfer in a simple house with a small fire. Ray tracing algorithms were constructed to treat enclosures of different shapes. The DPF technique accurately calculated shape factors and intensity distribution in rectangular and cylindrical enclosures without participating media. The DPF method was found to be more efficient than the Monte Carlo technique with equal number of photons. Monte Carlo technique is, however, easier to code particularly for enclosures with complex shapes, such as real buildings. Studies for selecting the number and optimum locations for radiation based fire detectors are in progress.

R. O. S. Prasad, R. N. Paul, Y. R. Sivathanu, and J. P. Go, **
Flamelet Models for Non-Isenthalphic Turbulent Premixed Jet Flames**,
34th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit, Paper No. AIAA 98-3495, Cleveland, OH.

A computational study of a non-isenthalpic premixed turbulent jet flame is described. The flame burns a homogeneously premixed stoichiometric methane-air mixture injected into a co-flow of air. The enthalpy (Chemical + sensible) varies because of mixing between the jet fluid and the co-flow. The performance of the Bray-Moss (BM) model and three flame surface density (FSD) models is evaluated by comparing the predictions of mean velocity and temperature profiles with recent experimental data. The reaction progress variable approach, which is established for isenthalpic flames, is extended to the present non-isenthalpic flames by including mean and mean square mixture fraction equations. The joint probability density function (PDF) of the reaction progress variable and the mixture fraction is modeled in terms of two statistically independent PDFs. The time average reaction rate term is modeled using the BM and FSD models. All models yielded reasonable predictions of mean velocity. The MB and MB models provided the best agreement with the mean temperature data but the other FSD models wth slight running of the constants could provide similar agreement as well. The results show that a simple extension of the FSD models is promising for the treatment of non-isenthalpic flames. It appears that the differences in the conceptual framework of the FSD models disappear in their implementation using basically the same turbulence properties of kinetic energy and dissipation rates.

Y. J. Zhu, A. Lloyd, Y. Sivathanu and J. Gore, **
Experimental and Numerical Evaluation of a Near Infrared Fire Detector**,
Proceedings of the Second International Conference on Fire Research and Engineering, NIST, Gaithersburg, MD.

Near Infrared fire detectors work on the principle of detecting fires based on a statistical analysis of the apparent source temperatures of fires. The apparent source temperatures are estimated from the radiation intensity incident on the fire detector at two near infrared wavelengths. However, in some instances, the fires are not in the direct view of the detector, and most of the radiation which are incident on the detector reaches it after multiple reflections from the walls of the building. An experimental and numerical study of the effects of these reflections on the temperatures inferred by a near infrared fire detector are presented. The experimental evaluation was conducted using three open and two smoldering fires. The results show that the near infrared fire detector is capable of discriminating open fires from reflected radiation. However, for smoldering fires, the intensities obtained from reflected radiation are too low to be successfully discriminated from background noise. Numerical evaluation of the performance of the near infrared fire detector in cylindrical and rectangular enclosures were conducted utilizing a photon tracing algorithm in conjunction with the discrete probability function method. The numerical evaluation confirms that the detector can successfully detect fires from reflected radiation if its sensitivity is sufficiently high.

S. D. Sovani, P. E. Sojka, and Y. R. Sivathanu, **
Predictions of Dropsize Distributions from First Principles: �Effect of Relative Velocities**,
Tenth Annual Conference on Liquid Atomization and Spray, Ontario, Canada.

A first principles analytical method has been developed for predicting drop size distribution in sprays. This technique will be especially useful in the process of injector design due to its ability to predict the effect of two important parameters on drop size distribution: mean relative velocity and relative velocity fluctuations. Since this approach is theoretical and does not depend on empirical correlations it will considerably decrease developmental experimentation. This technique is based on solving the equation of growth of fluid mechanic disturbances in cylindrical liquid ligaments which relates the magnitude of the relative velocity to the size of the drops formed. Solution proceeds by applying the Discrete Probability Function (DPF) method to the fluid mechanic instability model, which in turn describes the breakup of liquid ligaments into drops. Velocity fluctuations are introduced into the instability model as a discrete probability function and the disturbance equation of growth solved to obtain a drop size distribution. Results indicate that the width of the drop size pdf is strongly dependent on the level of velocity fluctuations. Furthermore, the width of the drop size increases with the level of velocity fluctuations in a slightly non-linear manner. Finally, a distinct upper limit is found to exist for the drop size distribution, which cannot be exceeded by any amount of turbulent fluctuation.

Y. R. Sivathanu and L. K. Tseng, **
Fire Detection Using Near-IR Radiation and Source Temperature Discrimination**,
Annual Conference on Fire Research, NIST, Gaithersburg, MD, pp. 117-118.

New fire detection concepts and algorithms are justified only if they improve upon existing ones with lower false alarm rates and greater sensitivity to starting fires. In addition, the detectors and signal processing instruments should be easy to operate and maintain, have high flexibility and be relatively inexpensive. Currently residential fire detectors include optical smoke sensors, ionization smoke sensors and temperature sensors. Conventional smoke sensors utilize light scattering or smoke ionization measurements, while temperature sensors utilize thermocouple measurements. The disadvantages with conventional single sensor detectors are that there is a significant time delay between the start of the fire, and the transport of either combustion products or smoke to positions close enough to enable detection and single sensor detectors involve a high rate of false alarms due to changes in the operating environment. Combinations of smoke sensors and odor sensors which involve multiple fire signatures are less prone to false alarms, but involve greater initial and maintenance costs. The objective of the present work was to investigate whether these two characteristics of natural fires could be exploited in a near-infrared fire detector operating on the principle of source temperature discrimination.

L. A. Gritzo, Y. R. Sivathanu, W. Gill, and T. Y. Chu, **
Absorption/Emission Spectroscopy Measurements of Soot Volume Fraction and Soot Emission in Large Fires**,
Annual Conference on Fire Research, Giathersburg, MD.

The hazard posed by a large, fully turbulent (i.e. > 1.3 m diameter) hydrocarbon fuel fire is defined by the history of the net heat flux from the fire to the system of interest. Although the radiative transport depends on the knowledge of soot absorption and emission within the fire, relevant data have not been obtained for large fires. Data are therefore needed to validate soot generation and combustion submodels presently employed in fire field model. Due to the turbulent nature of these environments, highly time-resolved measurements are required. To satisfy these requirements, a new optical fiber-based soot probe has recently been developed at Sandia in collaboration with En�Urga Inc. and urdue University to perform highly-transient measurements of the soot extinction coefficient, soot emission temperature, and soot volume fraction. A description of the technique employed in large fires, as well as recently acquired data, are presented.

A. S. Kelkar, Ch. Ramakrishna, Y. R. Sivathanu, and J. P. Gore, **
Temperature and Velocity Statistics of Lean Premixed Jet Flames for NOX calculations**,
34th Aerospace Sciences Meeting & Exhibit, Paper No. 96-0818, AIAA, Washington, DC.

An experimental and computational study of premixed turbulent combustion in a jet configuration is described. Measurements of the instantaneous velocity and temperature for two different flow rates and two different equivalence ratios were completed. The measurements were obtained using Laser Doppler Velocimetry and Thin Filament Pyrometry. The velocity and temperature statistics were used as input to a probability density function based model along with a chemical kinetics model to predict the NOX emission index from the flames. The predictions are evaluated using measurements of EINOX obtained from a chemiluminescent analyzer.

Y. R. Sivathanu and L.-K. Tseng, **
Characterization of Radiation Properties of Fires Using Multi-Wavelength Measurements**,
Proceedings of The Central States Section of The Combustion Institute, pp. 305-310.

A new type of fire detector which operates on a statistical analysis of the near infrared radiation incident on it is presented. The spectral radiation intensities incident on the fire detector are continuously measured at two near-infrared wavelengths (900 and 1000 nm), and a time series of apparent source temperatures is obtained from these measurements. The power spectral density and the probability density function of the apparent source temperatures are sufficient to determine the presence of a fire in the vicinity of the detector. The characteristics of the near-infrared radiation emanating from a heptane pool, a polyurethane foam and a wooden crib fire were studied to determine an efficient detection algorithm. The detector can also indicate the presence of a fire in an adjoining room from the radiation which is incident on it due to reflections from common building materials.

Y. R. Sivathanu, M. S. Klassen, A. S. Kelkar, Ch. Ramakrishna, J. P. Gore, and N. M. Laurendeau, **
Predictions and Measurements of NO Concentration in Lean Premixed Turbulent Jet Flames**,
Proceedings of The Central States Section of The Combustion Institute, pp. 80-85.

An experimental and computational study of NO emissions from premixed turbulent jet flames which operate in the distributed reaction regime is reported. Measurements of radial and axial velocities and temperature were used as inputs to a statistical model based on discrete probability functions along with a reduced chemical kinetics scheme to predict the distribution of NO number density in the flame. Local measurements of instantaneous NO number density were made in the post-flame zone of a lean methane/air flame via laser-saturated fluorescence. Predictions of the global NO emission index showed good qualitative agreement with measurements obtained using sampling and chemiluminescent analysis. The predictions also successfully captured the wide range of NO number densities found experimentally at local positions in the post-flame zone.

R. Wade, Y. R. Sivathanu and J. P. Gore, **
Soot Volume Fraction and Temperature Properties of High Liquid Loading Spray Flames**,
Proceedings of the Joint Technical Meeting of the Central and Western States (USA) Sections and the Mexican National Section of the International Combustion Institute and American Flame Research Committee, pp. 791-796.

The relationship between burning rate, visible flame length, and sooting properties of spray flames is investigated. Multiwavelength emission/absorption spectroscopy was applied to the measurement of soot volume fractions and temperatures for high liquid loading effervescent atomized flames. The statistics of the emission and absorption data were interpreted in terms of the statistics of the local properties using a novel discrete probability function based deconvolution method. The results show the coupled effects of soot volume fractions and temperature on the radiative heat loss from the spray flames. The effervescent atomized burner configuration allows a study of the radiation properties over a wider range of soot and temperature combinations than that allowed by gas jet flames. Comparison between conventional deconvolution techniques and the present method show that consideration of turbulence/radiation interactions is essential in applying tomography to time varying fields.

Y. R. Sivathanu, A. Hamins, C. Hagwood and T. Kashiwagi, **
Tomographic Reconstruction of the Local PDFs of Soot Volume Fraction and Temperature**,
Proceedings of the Joint Technical Meeting of the Central and Western States (USA) Sections and the Mexican National Section of the International Combustion Institute and American Flame Research Committee, pp. 92-97.

Deconvolution of local properties from line-of-sight measurements is important in a wide variety of applications such as x-ray tomography, nuclear magnetic resonance imaging, atmospheric sciences, optical interferometry and flow field diagnostics. The Radon Transforms form the theoretical basis for retrieving local properties from path integrated measurements under steady state conditions. These methods have found wide-spread application in tomographic spectroscopy of laminar flames. For turbulent flow fields, conventional deconvolution algorithms cause greater difficulty due to the transient nature of the phenomena being studied. Progress has been made in obtaining ultra-fast multiple angle and multiple ray measurements in a turbulent flow field over a small time interval. This technique has limited temporal resolution and suffers from a high degree of deconvolution noise due to the asymmetric nature of the instantaneous flow field. Recently, a discrete probability function (DPF) method was developed to deconvolute path integrated measurements in order to obtain the local PDFs of soot volume fractions in turbulent flames. The objective of this work is to extend the DPF method to obtain local PDFs of soot volume fraction and temperature from path integrated measurements of emission intensities. The deconvolution method is evaluated by synthetic noise-free data as well as experimental data obtained using an intrusive optical pyrometer.

Y. R. Sivathanu and J. P. Gore, **
Soot Kinetics/Radiation Interactions in Methane/Air Diffusion Flames**,
Annual Conference on Fire Research, NIST, Gaithersburg, MD. pp. 25-126.

Radiation heat transfer from flames depends on the instantaneous soot volume fractions and temperatures. The major obstacle to obtaining accurate soot volume fraction predictions in flames is the strong coupling between the finite rate kinetics of soot processes and radiation. Detailed models of soot kinetics cannot be incorporated in turbulent flame studies due to the limitations of computer resources. Therefore, many studies have concentrated on simplified global kinetics models incorporating soot nucleation, growth and oxidation mechanisms. The local temperatures needed for the soot kinetics calculations were obtained using simplified global radiation models. However, in strongly radiating flames it has been shown that the radiation and soot kinetics calculations need to be coupled to accurately predict observed soot volume fractions. The degree of this coupling in weakly radiating flames, where radiation is predominantly from gas species molecules rather than soot, has not been studied. The objective of the present work is evaluate this coupling of radiation and soot kinetics in weakly radiating flames by using the simplified soot kinetics model of Ref. 4 along with a narrow-band radiation model.

Y. R. Sivathanu and J. P. Gore, **
DPF Simulation of Reacting Flow in a Scalar Mixing Layer**,
Proceedings of the Spring Technical Meeting, Central States Section of the Combustion Institute, Paper N0. 60, pp. 366-371.

Scalar mixing in a turbulent layer (with no shear) is simulated using the discrete probability function (DPF) method. The DPF method consists of solving the Lagrangian equations for notional parcels in the turbulent flow field for discrete realization of the boundary and initial conditions, and velocity field. The probabilities of the solutions are calculated from the probability density functions (PDFs) of the boundary and initial conditions, and the velocity field. The velocity statistics throughout the flow field and the scalar statistics at the initial station are input from available experimental data. The simulation provides the evolution of the probability density function of the scalar at downstream stations. The present simulation correctly captures the change of the RMS fluctuations with downstream distance and the growth of the scalar layer.

Y. R. Sivathanu and J. P. Gore, **
Oxidation and Thermophoresis of Soot in Laminar Diffusion Flames**,
Proceedings of the Fall Technical Meeting of the Eastern States Section of The Combustion Institute, Paper No. 66, pp. 311-314, Princeton, New Jersey.

Radiation heat transfer from turbulent flames depends on the instantaneous soot volume fractions and temperatures due to the high degree of turbulence/radiation interactions. Most simplified practical flames are turbulent and consequently there is great interest in models of the soot kinetics processes that are easy to include in turbulent flow calculations. However predictions of soot volume fractions in strongly radiating flames remains a challenge in spite of increase understanding of the soot formation and oxidation process. The objective of the present work is to consider the effect of thermophoresis and two phase effects on the overall soot formation and oxidation in laminar acetylene/air diffusion flames. In addition, the relative importance of soot oxidation and formation on the overall soot emission from these flames are highlighted.

J. Dylla, Y. R. Sivathanu, and J. P. Gore, **
Multi--Variate Spatial Correlation Measurements in Turbulent Jet Flames**,
31st Aerospace Sciences Meeting, Paper No. 93-0801, AIAA, Washington, DC.

The important properties that determine the radiative heat loss from strongly radiating flames are the instantaneous temperatures and soot volume fractions. Since radiation heat loss depends on the net result of absorption and emission along a path, values of temperature and soot volume fractions at multiple points are involved. The present work describes the first measurements of two-point statistics of these properties in a strongly radiating flame. A two-point three-wavelength optical probe was used to measure the spatial and temporal correlation coefficients of soot volume fractions and temperature in an acetylene/air jet diffusion flame. The spatial and temporal correlation coefficients of soot volume fractions and temperature were approximated quite well b exponentially decaying functions. Soot volume fractions and temperatures were negatively correlated. There are differences between the present measurements of length and time scales of temperature and soot volume fraction and existing measurements of length and time scales of mixture fraction in weakly radiating flames. While the differences are small near the jet axis, intermittency of the jet near the flame sheet is found to affect the measurements of integral time and length scales substantially. The measurements justify the spatial correlation function and the radiation path discretization used in past simulations of the equation of radiative transfer.

R. Mital, Y. R. Sivathanu, and J. P. Gore, **
Effect of Scattering by Soot Aggregates on Radiative Heat Transfer**,
Proceedings of the Spring Technical Meeting, Central States Section of the Combustion Institute, Ohio, pp. 388-393. Also presented at the Annual Conference for Fire Research, NIST, Gaithersburg, MD., pp.113-114.

Radiation heat transfer studies in luminous flames have assumed negligible effects of scattering by soot particles. This assumption is based on the small diameter of soot particles compared to the wavelengths of interest at flame temperatures. The objective of the present work was to evaluate the effects of scattering process on radiative heat transfer along a representative path in a luminous diffusion flame. The results show that for the present conditions and the representative selection of the optical constants, the net effect of in- and out- scattering on the radiation intensity leaving a path is approximately 5%.

M. Klassen, Y. R. Sivathanu, and J. P. Gore, **
The Role of Reflection and Fuel Vapor Absorption in Radiation Heat Feedback to Pool Flames**,
Annual Conference on Fire Research, NIST, Gaithersburg, MD, pp. 107-108.

Radiation from liquid fueled flames is an important heat transfer mechanism in combustion systems. Radiative heat flux from fires surrounding object determines the possibility of ignition, flame spread and flashover. In liquid pool flames, the energy required for fuel vaporization is received entirely from the flame. Measurements of radiative heat feedback to the fuel surface show a large discrepancy between the measurements and the heat flux required to support the burning rate. Important factors that explain this discrepancy include reflection of radiation off the fuel surface and the absorption of radiation by fuel vapor. This paper describes measurements of incident radiative heat flux and simultaneous emission/absorption measurements above the pool surface in an effort to explain the discrepancy.

Y. R. Sivathanu and J. P. Gore, **
A Method for Multiray Radiative Source Term Calculations Including Turbulence Radiation Interactions**,
Twenty-Fourth Fall Technical Meeting of the Eastern States Section, The Combustion Institute, Ithaca, New York, pp. 94.1-94.4.

Turbulent fluctuations are known to increase the radiative heat flux from luminous non-premixed flames over the value estimated from average temperature and absorption coefficients. Stochastic simulations of the temperature and species concentration profiles have been used to treat these effects. The simulations involve construction a time series of scalar properties for each point within the flame and integrating the equation of transfer over each realization in a time series. Such calculations are expensive since several realizations of scalars are needed to obtain a statistically significant mean. The objective of the present work was to develop a simpler method for calculating radiative source term for individual points in strongly radiation flames while accounting for the effects of non-homogeneous distribution of scalars and of turbulence radiation interactions.

Y. R. Sivathanu and J. P. Gore, **
Simultaneous Multiline Emission-Absorption Measurements in Optically Thick Turbulent Flames**,
Twenty-ninth Aerospace Sciences Meeting, Paper No. 91-00833, AIAA, Washington, D. C.

Simultaneous transient emission/absorption measurements at five wavelengths in strongly radiating, optically thick, turbulent diffusion flames burning acetylene in air were completed. The data were processed to obtain CO2 mole fractions, temperatures and soot volume fractions spatially resolved to the estimated local integral length scales of mixture fluctuations. Temperatures and soot volume fractions based on emission intensities showed strong negative correlation due to radiative cooling effects. Probability density functions of soot volume fractions conditioned on CO2 mole fractions showed similarities with position. However, probable diffusion effects of negligible diffusitive of soot particles were observed. Probability density functions of soot CO2 concentrations and temperature illustrate the important role of radiative heat transfer in determining the flame structure. A multivariate stochastic analysis resulted in good predictions of radiation intensities.

Y. R. Sivathanu and G. M. Faeth, **
Soot properties in the Fuel-rich Region of Turbulent Diffusion Flames**,
Proceedings of the Spring Technical Meeting, Central States Section of The Combustion Institute, Dearborn, paper no. 36, pp. 103-109.

Practical buoyant flames are generally luminous and emit significant continuum radiaton from soot. The Rayleigh limit of soot particles is generally acceptable for estimating the spectral absorption coefficients of soot and scattering can be ignored; therefore, the mean soot properties needed for radiation analysis are the refractive indices of soot and distributions of soot volume fractions within the flame. There is some controversy concerning the refractive indices of soot in flames, but they are relatively independent of composition of the original fuel and estimates are available in the literature. On the other hand, little information is available concerning soot volume fractions in turbulent flames. Therefore, these properties were studied during the present investigation.

Y. R. Sivathanu, G. M. Faeth, and J. P. Gore, **
Soot Properties in the Overfire Region of Turbulent Diffusion Flames**,
Proceedings of the Spring Technical Meeting, Central States Section of The Combustion Institute, Indianapolis, paper no. 47, pp. 275-280.

The laminar flamelet concept of turbulent nonpremixed flames implies that scalar properties can be correlated solely as functions of mixture fraction, called state relationships, independent of local values of flame stretch and the past history of the flow. This approach has been widely used since the time of Burke and Schumann and is justified by the tendency of flames to maintain local thermodynamic equilibrium, at least near stoichiometric conditions. Bilger extended the concept by noting that although local thermodynamic equilibrium was not maintained for fuel-rich conditions in hydrocarbon/air diffusion flames, departure from equilibrium of the concentrations of major gas species tended to be universal for wide ranges of flame stretch. The objective of the present investigation was to extend the laminar flamelet concept to evaluate soot volume fraction state relationship in the overfire region of turbulent diffusion flames.

Y. R. Sivathanu, J. P. Gore, and G. M. Faeth, **
Scalar Structure and Soot Properties in the Overfire Region of Turbulent Ethylene and Acetylene/Air Diffusion Flames**,
Proceedings of the Spring Technical Meeting, Central States Section of The Combustion Institute, Chicago, paper no. 49, pp. 372-377.

Radiant energy transport influences the burning rate, the rate of spread, and the structure of unwanted fires � particularly soot containing luminous flames. In order to better understand these phenomena, the structure and radiation properties of the flaming region of turbulent ethylene and acetylene/air diffusion flames were recently studied in this laboratory. The present investigation extends this work to the overfire region.

G. M. Faeth, J. P. Gore, and Y. R. Sivathanu, **
Radiation from Soot Containing Flames**,
AGARD Conference Proceedings No. 422, Paris, pp. 17: 1-12.

The thermal radiation properties of turbulent soot-containing diffusion flames are considered, treating: scalar structure, excluding soot; soot properties; and turbulence/radiation interactions. The laminar flamelet concept, implying universal correlations of scalar properties as a function of mixture fraction (found from analysis or measurements of laminar flames) appears to be effective for estimates of scalar structure needed for analysis of flame radiation. Extension of laminar flamelet concept to soot volume fractions (the soot property needed for radiation predictions) has also been encouraging; however, universality is cruder than for gas properties due to effects of finite rate chemistry. Turbulence/radiation interactions were studied by comparing radiation predictions based on statistical simulation of scalar properties along a radiation path (stochastic analysis) with convention results using time averaged properties along the path. Stochastic predictions of spectral radiation intensities are found to be 40-100 percent higher than mean-property predictions in luminous turbulent diffusion flames, suggesting significant effects of turbulence radiation interactions.