The work presented in this thesis is an investigation of the dynamics of unconfined hydrogen-air flames in the presence of buoyant effects and the determination of an ignition criterion for flame propagation between adjacent pockets of reactive gas separated by air. The experimental work was conducted using the soap bubble technique and visualized with high speed schlieren or large scale shadowgraph systems. A study was first conducted to determine the most suitable soap solution additive among glycerol, guar and polyethylene oxide for conducting the experiments, isolating guar as the best candidate. The soap solution was then used to study the dynamics of flames in single or multiple soap bubbles filled with reactive mixtures of different compositions. The soap bubble method was also further improved by designing a soap dispenser that can maintain a bubble indefinitely and a method to burst the soap solution prior to an experiment using timed heated wires. In the experiments with single bubbles, it was found that for sufficiently lean hydrogen-air mixtures, buoyancy effects become important at small scales. The critical radius of hemispherical flames that will rise due to buoyancy was measured and estimated using a model comparing the characteristic burning speed and the rise speed of the flame kernel. Excellent agreement was found between the model predictions and the measured critical flame radii. The experiments with multiple bubbles provided the scaling rules for flame transition between neighboring pockets of hemispherical or spherical shape separated by an inert gas. The test results demonstrated that the separation distance between the bubbles is mainly determined by the expansion ratio when the buoyancy effects are negligible, corresponding to near stoichiometric mixtures. For leaner mixtures with stronger buoyant effects, the critical separation distance was no longer governed by the expansion ratio alone, as buoyancy forces render the flame propagation across the inert gas more difficult. Visualization of the ignition dynamics confirmed that buoyancy forces tend to accelerate the first kernel up before ignition of the second kernel can be achieved.

The thesis comprises effects of electric fields on a fundamental study of spheri­cal premixed flame propagation.Outwardly-propagating spherical laminar premixed flames have been investigated in a constant volume combustion vessel by applying au uni-directional electric potential.Direct photography and schlieren techniques have been adopted and captured images were analyzed through image processing. Unstretched laminar burning velocities under the influence of electric fields and their associated Markstein length scales have been determined from outwardly prop­agating spherical flame at a constant pressure. Methane and propane fuels have been tested to assess the effect of electric fields on the differential diffusion of the two fuels.The effects of varying equivalence ratios and applied voltages have been in­vestigated, while the frequency of AC was fixed at 1 KHz. Directional propagating characteristics were analyzed to identify the electric filed effect. The flame morphology varied appreciably under the influence of electric fields which in turn affected the burning rate of mixtures.The flame front was found to propagate much faster toward to the electrode at which the electric fields were supplied while the flame speeds in the other direction were minimally influenced. When the voltage was above 7 KV the combustion is markedly enhanced in the downward direction since intense turbulence is generated and as a result the mixing process or rather the heat and mass transfer within the flame front will be enhanced.The com­bustion pressure for the cases with electric fields increased rapidly during the initial stage of combustion and was relatively higher since the flame front was lengthened in the downward direction.

No description available.

No description available.

No description available.

In this work, turbulent premixed combustion was analyzed through CH (methylidyne) planar laser induced fluorescence (PLIF). Flame topography measurements in terms of flame surface density and curvature were calculated based on the flame front detected by the CH PLIF signal. The goal of this work was to investigate turbulent flames with extremely high turbulence intensity using a recently developed HiPilot burner (a Bunsen-type burner). The studies were first conducted on a series of piloted jet flames to validate the methodology, and then conducted on the highly turbulent flames generated by the HiPilot burner. All flames were controlled by combusting methane and air under a fuel to air equivalence ratio of Φ=1.05, and the Reynolds number varied from 7,385 to 28,360. Flame surface density fields and profiles for the HiPilot burner are presented. These flame surface density measurements showed an overall decrease with height above the burner. In addition, curvature statistics for the HiPilot flames were calculated and probability density functions of the curvature samples were determined. The probability density functions of curvature for the flames showed Gaussian-shaped distributions centered near zero curvature. To conclude, flame topography measurements were verified on jet flames and were demonstrated on the new HiPilot flames. / Master of Science / Optical diagnostics are powerful techniques that enable the study of turbulent flames without physical interruption. The optical diagnostic technique in this thesis implemented planar laser induced fluorescence. In planar laser induced fluorescence, a laser is used to excite a specific molecular species present within a two-dimensional plane in the flame. The excited species releases the extra energy by emission of light which is the signal captured on a camera. One useful purpose of using optical diagnostics, such as planar laser induced fluorescence, is the ability to image the flame structure of turbulent flames. The flame structure is significant for two reasons. First, the flame structure details how the chemistry of the flame interacts with the turbulent flow field. Second, the flame structure is directly related to the burning rate of the reactants. The primary contribution of this thesis investigated the two-dimensional flame structure of a newly designed burner named the HiPilot burner. However, in order to strengthen the fidelity of the methods for determining certain flame structure quantities a precursive analysis on the classical jet flame was completed. The results acquired show structural measurements of the HiPilot flames which contribute to the repository of data for the combustion community

Thin filament pyrometry is a proven technique used to measure flame temperature by capturing the spectral radiance produced by the immersion of silicon carbide filaments in a hot gas environment. In this study a commercially available CMOS color camera was used, and the spectral response of each color channel was integrated with respect to the assumed graybody radiation spectrum to form a look up table between color ratio and temperature. Interpolated filament temperatures are then corrected for radiation losses via an energy balance to determine the flame temperature. Verification of the technique was performed on the Holthuis and Associates Flat Flame Burner, formerly known as the Mckenna Burner, and the results are directly compared to literature values measured on a similar burner. The results are also supported by radiation corrected measurements taken using a type B thermocouple on the same burner setup. An error propagation analysis was performed to determine which factors contribute the most to the final measurement uncertainty and confidence intervals are calculated for the results. Uncertainty values for a single point measurement were determined to be between ±15 and ±50 K depending on the color ratio and the total uncertainty associated with day-to-day changes in the measurement setup was found to be ±55 K. / Master of Science / Determination of flame temperature is an important aspect of combustion research and is often critical to the evaluation of combustion systems as well as the integration of those systems into more complex devices. In this thesis the technique of thin filament pyrometry was implemented and verified through the use of a well characterized calibration flame. This technique involves placing thin filaments usually made from silicon carbide into the flame and capturing the spectrum of light they emit with a detector. Since the amount of light emitted as well as which wavelengths the light is concentrated in is a strong function of temperature, this methodology may be used to calculate the temperature of the flame. Thin filament pyrometry has the advantage compared to other techniques in that it is extremely cheap to implement and requires no advanced scientific equipment. The SiC filaments have been shown to have a very high resistance to the flame environment and do not face many of the same challenges that can cause problems for other techniques. A statistical analysis of the method implemented in this work was also performed and the expected uncertainty was similar to many of the alternative techniques which necessitate a more complex or expensive setup.

No description available.

No description available.

Thesis (Sc.D.)--Massachusetts Institute of Technology, Dept. of Chemical Engineering, 1981. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND SCIENCE. / Bibliography: leaves 330-335. / by Koon Gee Neoh. / Sc.D.