SEARCH · Search NASA

Results for “Fuel distribution”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9

All sources NASA NTRS DOE OSTI

NASA NTRS1982

Analytical modeling of operating characteristics of premixing-prevaporizing fuel-air mixing passages. Volume 1: Analysis and results

A model for predicting the distribution of liquid fuel droplets and fuel vapor in premixing-prevaporizing fuel-air mixing passages of the direct injection type is reported. This model consists of three computer programs; a calculation of the two dimensional or axisymmetric air flow field neglecting the effects of fuel; a calculation of the three dimensional fuel droplet trajectories and evaporation rates in a known, moving air flow; a calculation of fuel vapor diffusing into a moving three dimensional air flow with source terms dependent on the droplet evaporation rates. The fuel droplets are treated as individual particle classes each satisfying Newton's law, a heat transfer, and a mass transfer equation. This fuel droplet model treats multicomponent fuels and incorporates the physics required for the treatment of elastic droplet collisions, droplet shattering, droplet coalescence and droplet wall interactions. The vapor diffusion calculation treats three dimensional, gas phase, turbulent diffusion processes. The analysis includes a model for the autoignition of the fuel air mixture based upon the rate of formation of an important intermediate chemical species during the preignition period.

Anderson, O. L.↗

NASA NTRS1978

Nitric oxide formation in gas turbine engines: A theoretical and experimental study

A modified Zeldovich kinetic scheme was used to predict nitric oxide formation in the burned gases. Nonuniformities in fuel-air ratio in the primary zone were accounted for by a distribution of fuel-air ratios. This was followed by one or more dilution zones in which a Monte Carlo calculation was employed to follow the mixing and dilution processes. Predictions of NOX emissions were compared with various available experimental data, and satisfactory agreement was achieved. In particular, the model is applied to the NASA swirl-can modular combustor. The operating characteristics of this combustor which can be inferred from the modeling predictions are described. Parametric studies are presented which examine the influence of the modeling parameters on the NOX emission level. A series of flow visualization experiments demonstrates the fuel droplet breakup and turbulent recirculation processes. A tracer experiment quantitatively follows the jets from the swirler as they move downstream and entrain surrounding gases. Techniques were developed for calculating both fuel-air ratio and degree of nonuniformity from measurements of CO2, CO, O2, and hydrocarbons. A burning experiment made use of these techniques to map out the flow field in terms of local equivalence ratio and mixture nonuniformity.

Mikus, T.↗

NASA NTRS1983

Spontaneous Ignition Characteristics of Hydrocarbon Fuel-air Mixtures

Although the subject of spontaneous ignition of liquid fuels has received considerable attention in the past, the role of fuel evaporation in the overall spontaneous ignition process is still unclear. A main purpose of this research is to carry out measurements of ignition delay times, using fuels of current and anticipated future aeronautical interest, at test conditions that are representative of those encountered in modern gas turbine engines. Attention is focused on the fuel injection process, in particlar the measurement and control of man fuel drop size and fuel-air spatial distribution. The experiments are designed to provide accurate information on the role of fuel evaporation processes in determining the overall ignition delay time. The second objective is to examine in detail the theoretical aspects of spontaneous ignition in order to improve upon current knowledge and understanding of the basic processes involved, so that the results of the investigation can find general and widespead application.

Lefebvre, A. H.↗

NASA NTRS2014

NASA Numerical and Experimental Evaluation of UTRC Low Emissions Injector

Computational and experimental analyses of a PICS-Pilot-In-Can-Swirler technology injector, developed by United Technologies Research Center (UTRC) are presented. NASA has defined technology targets for near term (called "N+1", circa 2015), midterm ("N+2", circa 2020) and far term ("N+3", circa 2030) that specify realistic emissions and fuel efficiency goals for commercial aircraft. This injector has potential for application in an engine to meet the Pratt & Whitney N+3 supersonic cycle goals, or the subsonic N+2 engine cycle goals. Experimental methods were employed to investigate supersonic cruise points as well as select points of the subsonic cycle engine; cruise, approach, and idle with a slightly elevated inlet pressure. Experiments at NASA employed gas analysis and a suite of laser-based measurement techniques to characterize the combustor flow downstream from the PICS dump plane. Optical diagnostics employed for this work included Planar Laser-Induced Fluorescence of fuel for injector spray pattern and Spontaneous Raman Spectroscopy for relative species concentration of fuel and CO2. The work reported here used unheated (liquid) Jet-A fuel for all fuel circuits and cycle conditions. The initial tests performed by UTRC used vaporized Jet-A to simulate the expected supersonic cruise condition, which anticipated using fuel as a heat sink. Using the National Combustion Code a PICS-based combustor was modeled with liquid fuel at the supersonic cruise condition. All CFD models used a cubic non-linear k-epsilon turbulence wall functions model, and a semi-detailed Jet-A kinetic mechanism based on a surrogate fuel mixture. Two initial spray droplet size distribution and spray cone conditions were used: (1) an initial condition (Lefebvre) with an assumed Rosin-Rammler distribution, and 7 degree Solid Spray Cone; and (2) the Boundary Layer Stripping (BLS) primary atomization model giving the spray size distribution and directional properties. Contour and line plots are shown in comparison with experimental data (where this data is available) for flow velocities, fuel, and temperature distribution. The CFD results are consistent with experimental observations for fuel distribution and vaporization. Analysis of gas sample results, using a previously-developed NASA NOx correlation, indicates that for sea-level takeoff, the PICS configuration is predicted to deliver an EINOx value of about three for the targeted supersonic aircraft. Emissions results at supersonic cruise conditions show potential for meeting the NASA goals with liquid fuel.

Hicks, Yolanda R.↗

NASA NTRS2014

NASA Numerical and Experimental Evaluation of UTRC Low Emissions Injector

Computational and experimental analyses of a PICS-Pilot-In-Can-Swirler technology injector, developed by United Technologies Research Center (UTRC) are presented. NASA has defined technology targets for near term (called "N+1", circa 2015), midterm ("N+2", circa 2020) and far term ("N+3", circa 2030) that specify realistic emissions and fuel efficiency goals for commercial aircraft. This injector has potential for application in an engine to meet the Pratt & Whitney N+3 supersonic cycle goals, or the subsonic N+2 engine cycle goals. Experimental methods were employed to investigate supersonic cruise points as well as select points of the subsonic cycle engine; cruise, approach, and idle with a slightly elevated inlet pressure. Experiments at NASA employed gas analysis and a suite of laser-based measurement techniques to characterize the combustor flow downstream from the PICS dump plane. Optical diagnostics employed for this work included Planar Laser-Induced Fluorescence of fuel for injector spray pattern and Spontaneous Raman Spectroscopy for relative species concentration of fuel and CO2. The work reported here used unheated (liquid) Jet-A fuel for all fuel circuits and cycle conditions. The initial tests performed by UTRC used vaporized Jet-A to simulate the expected supersonic cruise condition, which anticipated using fuel as a heat sink. Using the National Combustion Code a PICS-based combustor was modeled with liquid fuel at the supersonic cruise condition. All CFD models used a cubic non-linear k-epsilon turbulence wall functions model, and a semi-detailed Jet-A kinetic mechanism based on a surrogate fuel mixture. Two initial spray droplet size distribution and spray cone conditions were used: 1) an initial condition (Lefebvre) with an assumed Rosin-Rammler distribution, and 7 degree Solid Spray Cone; and 2) the Boundary Layer Stripping (BLS) primary atomization model giving the spray size distribution and directional properties. Contour and line plots are shown in comparison with experimental data (where this data is available) for flow velocities, fuel, and temperature distribution. The CFD results are consistent with experimental observations for fuel distribution and vaporization. Analysis of gas sample results, using a previously-developed NASA NOx correlation, indicates that for sea-level takeoff, the PICS configuration is predicted to deliver an EINOx value of about 3 for the targeted supersonic aircraft. Emissions results at supersonic cruise conditions show potential for meeting the NASA goals with liquid fuel.

Hicks, Yolanda R.↗

NASA NTRSDate not supplied

Analytical and Experimental Demonstration of an Alternate Mixing Performance Metric for High-Speed Fuel Mixing Studies

To experimentally assess the fuel/air mixing performance of high-speed fuel injectors, one-dimensional metrics that quantify the degree of mixing completeness downstream of the fuel injection location are required. The most accurate assessment of mixing performance is achieved with the mixing efficiency parameter. In order to experimentally determine the mixing efficiency parameter, the spatial distributions of both mass flux and fuel mass fraction must be measured. In-stream gas sampling techniques are commonly used to measure the fuel mass fraction distribution; however, the mass flux distribution is not easily determined because it requires the measurement of three independent aerothermodynamic variables in addition to the gas composition. Therefore, to experimentally determine the mixing efficiency parameter, the spatial distributions of four independent properties must be measured, with each property generally requiring its own unique probe. Because of this difficulty, it is commonly assumed that alternate metrics, which rely solely on the fuel distribution, are good indicators of mixing performance. However, since these alternate metrics do not provide a mass flux-weighted measure of mixing completeness, they can lead to incorrect conclusions being drawn about the mixing performance of the studied fuel injector configuration. Recognizing this shortcoming, this work proposes two new alternate mixing performance metrics that are easier to obtain than the mixing efficiency parameter. The analytical development of the new metrics, as well as their application to relevant CFD and experimental data of high-speed fuel injector configurations, is presented in this work. For two different experiments, the new metrics are shown to provide an excellent representation of the true mass flux-weighted mixing performance, unlike the traditionally-used alternatives. The results presented herein suggest that the new metrics can serve as accurate surrogates for mixing efficiency in future high-speed fuel/air mixing studies.

Cody R Ground↗

NASA NTRS2010

Optical Characterization of a Multipoint Lean Direct Injector for Gas Turbine Combustors: Velocity and Fuel Drop Size Measurements

Performance of a multipoint, lean direct injection (MP-LDI) strategy for low emission aero-propulsion systems has been tested in a Jet-A fueled, lean flame tube combustion rig. Operating conditions for the series of tests included inlet air temperatures between 672 and 828 K, pressures between 1034 and 1379 kPa and total equivalence ratios between 0.41 and 0.45, resulting in equilibrium flame temperatures approaching 1800 K. Ranges of operation were selected to represent the spectrum of subsonic and supersonic flight conditions projected for the next-generation of commercial aircraft. This document reports laser-based measurements of in situ fuel velocities and fuel drop sizes for the NASA 9-point LDI hardware arranged in a 3 3 square grid configuration. Data obtained represent a region of the flame tube combustor with optical access that extends 38.1-mm downstream of the fuel injection site. All data were obtained within reacting flows, without particle seeding. Two diagnostic methods were employed to evaluate the resulting flow path. Three-component velocity fields have been captured using phase Doppler interferometry (PDI), and two-component velocity distributions using planar particle image velocimetry (PIV). Data from these techniques have also offered insight into fuel drop size and distribution, fuel injector spray angle and pattern, turbulence intensity, degree of vaporization and extent of reaction. This research serves to characterize operation of the baseline NASA 9- point LDI strategy for potential use in future gas-turbine combustor applications. An additional motive is the compilation of a comprehensive database to facilitate understanding of combustor fuel injector aerodynamics and fuel vaporization processes, which in turn may be used to validate computational fluid dynamics codes, such as the National Combustor Code (NCC), among others.

Heath, Christopher M.↗

NASA NTRS1973

Experimental detailed power distribution in a fast spectrum thermionic reactor fuel element at the core/BeO reflector interface region

A zero-power critical assembly was designed, constructed, and operated for the prupose of conducting a series of benchmark experiments dealing with the physics characteristics of a UN-fueled, Li-7-cooled, Mo-reflected, drum-controlled compact fast reactor for use with a space-power conversion system. The critical assembly was modified to simulate a fast spectrum advanced thermionics reactor by: (1) using BeO as a reflector in place of some of the existing molybdenum, (2) substituting Nb-1Zr tubing for some of the existing Ta tubing, and (3) inserting four full-scale mockups of thermionic type fuel elements near the core and BeO reflector boundary. These mockups were surrounded with a buffer zone having the equivalent thermionic core composition. In addition to measuring the critical mass of this thermionic configuration, a detailed power distribution in one of the thermionic element stages in the mixed spectrum region was measured. A power peak to average ratio of two was observed for this fuel stage at the midplane of the core and adjacent to the reflector. Also, the power on the outer surface adjacent to the BeO was slightly more than a factor of two larger than the power on the inside surface of a 5.08 cm (2.0 in.) high annular fuel segment with a 2.52 cm (0.993 in. ) o.d. and a 1.86 cm (0.731 in.) i.d.

Klann, P. G.↗

NASA NTRS2004

Experimental Investigation of a Multiplex Fuel Injector Module With Discrete Jet Swirlers for Low Emission Combustors

A low-NOx emissions combustor concept has been demonstrated in flame-tube tests. A lean-direct injection (LDI) concept was used where the fuel is injected directly into the flame zone and the overall equivalence ratio of the mixture is lean. The LDI concept described in this report is a multiplex fuel injector module containing multipoint fuel injection tips and multi-burning zones. The injector module comprises 25 equally spaced injection tips within a 76 by 76 mm area that fits into the flame-tube duct. The air swirlers were made from a concave plate on the axis of the fuel injector using drilled holes at an angle to the axis of the fuel injector. The NOx levels were quite low and are greater than 70 percent lower than the 1996 ICAO standard. At an inlet temperature of 810 K, inlet pressure of 2760 kPa, pressure drop of 4 percent and a flame temperature of 1900 K with JP8 fuel, the NOx emission index was 9. The 25-point injector module exhibited the most uniform radial distribution of fuel-air mixture and NOx emissions in the flame tube when compared to other multipoint injection devices. A correlation is developed relating the NOx emissions to inlet temperature, inlet pressure, equivalence ratio and pressure drop.

Tacina, Robert↗

NASA NTRS1993

The effect of incomplete fuel-air mixing on the lean limit and emissions characteristics of a Lean Prevaporized Premixed (LPP) combustor

Results are presented from an experimental study of the effect of incomplete fuel-air mixing on the lean limit and emissions characteristics of a lean, prevaporized, premixed (LPP), coaxial mixing tube combustor. Two-dimensional exciplex fluorescence was used to characterize the degree of fuel vaporization and mixing at the combustor inlet under non-combusting conditions. These tests were conducted at a pressure of 4 atm., a temperature of 400 C, a mixer tube velocity of 100 m/sec and an equivalence ratio of .8, using a mixture of tetradecane, 1 methyl naphthalene and TMPD as a fuel simulant. Fuel-air mixtures with two distinct spatial distributions were studied. The exciplex measurements showed that there was a significant amount of unvaporized fuel at the combustor entrance in both cases. One case, however, exhibited a very non-uniform distribution of fuel liquid and vapor at the combustor entrance, i.e., with most of the fuel in the upper half of the combustor tube, while in the other case, both the fuel liquid and vapor were much more uniformly distributed across the width of the combustor entrance. The lean limit and emissions measurements were all made at a pressure of 4 atm. and a mixer tube velocity of 100 m/sec, using Jet A fuel and both fuel-air mixture distributions. Contrary to what was expected, the better mixed case was found to have a substantially leaner operating limit. The two mixture distributions also unexpectedly resulted in comparable NO(x) emissions, for a given equivalence ratio and inlet temperature, however, lower NO(x) emissions were possible in the better mixed case due to its leaner operating limit.

Santavicca, D. A.↗

NASA NTRS2016

CFD Analysis of Mixing Characteristics of Several Fuel Injectors at Hypervelocity Flow Conditions

CFD analysis is presented of the mixing characteristics and performance of three fuel injectors at hypervelocity flow conditions. The calculations were carried out using the VULCAN-CFD solver and Reynolds-Averaged Simulations (RAS). The high Mach number flow conditions match those proposed for the planned experiments conducted as a part of the Enhanced Injection and Mixing Project (EIMP) at the NASA Langley Research Center. The EIMP aims to investigate scramjet fuel injection and mixing physics, improve the understanding of underlying physical processes, and develop enhancement strategies and functional relationships relevant to flight Mach numbers greater than eight. Because of the high Mach number flow considered, the injectors consist of a fuel placement device, a strut; and a fluidic vortical mixer, a ramp. These devices accomplish the necessary task of distributing and mixing fuel into the supersonic cross-flow albeit via different strategies. Both of these devices were previously studied at lower flight Mach numbers where they exhibited promising performance in terms of mixing efficiency and total pressure recovery. For comparison, a flush-wall injector is also included. This type of injector generally represents the simplest method of introducing fuel into a scramjet combustor, however, at high flight Mach number conditions, the dynamic pressure needed to induce sufficient fuel penetration may be difficult to achieve along with other requirements such as achieving desired levels of fuel-to-air mixing at the required equivalence ratio. The three injectors represent the baseline configurations planned for the experiments. The current work discusses the mixing flow field behavior and differences among the three fuel injectors, mixing performance as described by the mixing efficiency and the total pressure recovery, and performance considerations based on the thrust potential.

Drozda, Tomasz G.↗

NASA NTRS2000

Structure and Soot Properties of Nonbuoyant Ethylene/Air Laminar Jet Diffusion Flames

The structure and soot properties of round, soot-emitting, nonbuoyant, laminar jet diffusion flames are described, based on long-duration (175-230/s) experiments at microgravity carried out on orbit In the Space Shuttle Columbia. Experiments] conditions included ethylene-fueled flames burning in still air at nominal pressures of 50 and 100 kPa and an ambient temperature of 300 K with luminous Annie lengths of 49-64 mm. Measurements included luminous flame shapes using color video imaging, soot concentration (volume fraction) distributions using deconvoluted laser extinction imaging, soot temperature distributions using deconvoluted multiline emission imaging, gas temperature distributions at fuel-lean (plume) conditions using thermocouple probes, not structure distributions using thermophoretic sampling and analysis by transmission electron microscopy, and flame radiation using a radiometer. The present flames were larger, and emitted soot men readily, than comparable observed during ground-based microgravity experiments due to closer approach to steady conditions resulting from the longer test times and the reduced gravitational disturbances of the space-based experiments.

Urban, D. L.↗

NASA NTRS2001

Structure and Soot Properties of Nonbuoyant Ethylene/Air Laminar Jet Diffusion Flames

The structure and soot properties of round, soot-emitting, nonbuoyant, laminar jet diffusion flames are described, based on long-duration (175-230-s) experiments at microgravity carried out on orbit in the Space Shuttle Columbia. Experimental conditions included ethylene-fueled flames burning in still air at nominal pressures of 50 and 100 kPa and an ambient temperature of 300 K with luminous flame lengths of 49-64 mm Measurements included luminous flame shapes using color video imaging soot concentration (volume fraction) distributions using deconvoluted laser extinction imaging, soot temperature distributions using deconvoluted multiline emission imaging, gas temperature distributions at fuel-lean (plume) conditions using thermocouple probes, soot structure distributions using thermophoretic sampling and analysis by transmission electron microscopy, and flame radiation using a radiometer.The present flames were larger, and emitted soot more readily, than comparable flames observed during ground-based microgravity experiments due to closer approach to steady conditions resulting from the longer test times and the reduced gravitational disturbances of the space-based experiments.

Urban, D. L.↗

NASA NTRS1993

Fuel-rich catalytic combustion of a high density fuel

Fuel-rich catalytic combustion (ER is greater than 4) of the high density fuel exo-tetrahydrocyclopentadiene (JP-10) was studied over the equivalence ratio range 5.0 to 7.6, which yielded combustion temperatures of 1220 to 1120 K. The process produced soot-free gaseous products similar to those obtained with iso-octane and jet-A in previous studies. The measured combustion temperature agreed well with that calculated assuming soot was not a combustion product. The process raised the effective hydrogen/carbon (H/C) ratio from 1.6 to over 2.0, thus significantly improving the combustion properties of the fuel. At an equivalence ratio near 5.0, about 80 percent of the initial fuel carbon was in light gaseous products and about 20 percent in larger condensable molecules. Fuel-rich catalytic combustion has now been studied for three fuels with H/C ratios of 2.25 (iso-octane), 1.92 (jet-A), and 1.6 (JP-10). A comparison of the product distribution of these fuels shows that, in general, the measured concentrations of the combustion products were monotonic functions of the H/C ratio with the exception of hydrogen and ethylene. In these cases, data for JP-10 fell between iso-octane and jet-A rather than beyond jet-A. It is suggested that the ring cross-linking structure of JP-10 may be responsible for this behavior. All the fuels studied showed that the largest amounts of small hydrocarbon molecules and the smallest amounts of large condensable molecules occurred at the lower equivalence ratios. This corresponds to the highest combustion temperatures used in these studies. Although higher temperatures may improve this mix, the temperature is limited. First, the life of the present catalyst would be greatly shortened when operated at temperatures of 1300 K or greater. Second, fuel-rich catalytic combustion does not produce soot because the combustion temperatures used in the experiments were well below the threshold temperature (1350 K) for the formation of soot. Increasing the temperature above this value would remove the soot-free nature of the process. Since all the fuels studied show a similar breakdown of the primary fuel into smaller molecular combustion products, this technique can be applied to all hydrocarbon fuels.

Brabbs, Theodore A.↗

NASA NTRS1976

Research on plasma core reactors

Experiments and theoretical studies are being conducted for NASA on critical assemblies with one-meter diameter by one-meter long low-density cores surrounded by a thick beryllium reflector. These assemblies make extensive use of existing nuclear propulsion reactor components, facilities, and instrumentation. Due to excessive porosity in the reflector, the initial critical mass was 19 kg U(93.2). Addition of a 17 cm thick by 89 cm diameter beryllium flux trap in the cavity reduced the critical mass to 7 kg when all the uranium was in the zone just outside the flux trap. A mockup aluminum UF6 container was placed inside the flux trap and fueled with uranium-graphite elements. Fission distributions and reactivity worths of fuel and structural materials were measured. Finally, an 85,000 cu cm aluminum canister in the central region was fueled with UF6 gas and fission density distributions determined. These results are to be used to guide the design of a prototype plasma core reactor which will test energy removal by optical radiation.

Jarvis, G. A.↗

NASA NTRS2006

A Theoretical Solid Oxide Fuel Cell Model for System Controls and Stability Design

As the aviation industry moves towards higher efficiency electrical power generation, all electric aircraft, or zero emissions and more quiet aircraft, fuel cells are sought as the technology that can deliver on these high expectations. The Hybrid Solid Oxide Fuel Cell system combines the fuel cell with a microturbine to obtain up to 70 percent cycle efficiency, and then distributes the electrical power to the loads via a power distribution system. The challenge is to understand the dynamics of this complex multi-discipline system, and design distributed controls that take the system through its operating conditions in a stable and safe manner while maintaining the system performance. This particular system is a power generation and distribution system and the fuel cell and microturbine model fidelity should be compatible with the dynamics of the power distribution system in order to allow proper stability and distributed controls design. A novel modeling approach is proposed for the fuel cell that will allow the fuel cell and the power system to be integrated and designed for stability, distributed controls, and other interface specifications. This investigation shows that for the fuel cell, the voltage characteristic should be modeled, but in addition, conservation equation dynamics, ion diffusion, charge transfer kinetics, and the electron flow inherent impedance should also be included.

Kopasakis, George↗

NASA NTRS1937

Fuel Spray and Flame Formation in a Compression-Ignition Engine Employing Air Flow

The effects of air flow on fuel spray and flame formation in a high-speed compression-ignition engine have been investigated by means of the NACA combustion apparatus. The process was studied by examining high-speed motion pictures taken at the rate of 2,200 frames a second. The combustion chamber was of the flat-disk type used in previous experiments with this apparatus. The air flow was produced by a rectangular displacer mounted on top of the engine piston. Three fuel-injection nozzles were tested: a 0.020-inch single-orifice nozzle, a 6-orifice nozzle, and a slit nozzle. The air velocity within the combustion chamber was estimated to reach a value of 425 feet a second. The results show that in no case was the form of the fuel spray completely destroyed by the air jet although in some cases the direction of the spray was changed and the spray envelope was carried away by the moving air. The distribution of the fuel in the combustion chamber of a compression-ignition engine can be regulated to some extent by the design of the combustion chamber, by the design of the fuel-injection nozzle, and by the use of air flow.

Rothrock, A M↗

NASA NTRS1999

Formation of Sprays From Conical Liquid Sheets

Our objective is to predict droplet size distributions created by fuel injector nozzles in Jet turbines. These results will be used to determine the initial conditions for numerical simulations of the combustion process in gas turbine combustors. To predict the droplet size distribution, we are currently constructing a numerical model to understand the instability and breakup of thin conical liquid sheets. This geometry serves as a simplified model of the liquid jet emerging from a real nozzle. The physics of this process is difficult to study experimentally as the time and length scales are very short. From existing photographic data, it does seem clear that three-dimensional effects such as the formation of streamwise ligaments and the pulling back of the sheet at its edges under the action of surface tension are important.

Peck, Bill↗