Search

The onset of transition to cellularity and self-similar propagation of centrally ignited, expanding spherical flames in a reactive environment of H2/O2/N2 and H2/O2/He mixtures at initial pressures up to 15 bar were experimentally investigated using a newly developed, constant-pressure, dual-chamber vessel and were theoretically interpreted based on linear stability theory. The experiments were well-controlled to identify the separate and coupled effects of Darrieus–Landau instability and diffusional–thermal instability. Results show that the critical radius, rcr, for the onset of cellular instability varies non-monotonously with initial pressure for fuel-lean and stoichiometric H2/O2/N2 flames. This non-monotonous pressure dependence of rcr is well captured by linear stability theory for stoichiometric flames. The experimental critical Peclet number, Pecr = rcr/δf, increases non-linearly with the Markstein number, Ma, which measures the intensity of diffusional–thermal instability. However, a linear dependence of Pecr on Ma is predicted by linear stability theory. Specifically, the theory shows well quantitative agreement with the experimental results for mixtures with near-unity Leeff; however, it under-predicts the Pecr for mixtures with off-unity Leeff. In addition, there exists three distinct propagation stages for flames subjected to cellular instability, namely, smooth expansion, transition propagation, and self-similar propagation. The acceleration exponent, α, in the self-similar propagation stage was extracted based on the power-law of drf/dt = αA1/αrf(1 − 1/α), where rf is the instantaneous mean flame radius, and A is a constant. The values of α are located between 1.22 and 1.40, which are smaller than the suggested value (1.5) for self-turbulization. [ABSTRACT FROM AUTHOR]

A self-developed gas explosion experimental system was used to detonate methane/air mixtures of different concentrations by different ignition methods, and the propagation processes of spherical flames and irregular flames formed by the explosions were recorded using a high-speed camera. Based on image processing and the fractal interpolation method, the propagation velocity of the explosion-stretched flame was calculated. The flame image was processed in MATLAB to optimize the traditional calculation method. Based on fractal theory, a fractal interpolation function was constructed to calculate the propagation velocity at each point on the front of the spherical stretched flame, and then, the laminar combustion velocity of the methane/air mixed balloon flame was obtained. Compared with the numerical simulation and empirical formula, the error in the laminar flame velocity was less than 0.2%, which verifies the correctness of the image processing and fractal interpolation method to calculate the flame propagation velocity. Using this method, the propagation velocity of the irregular fold stretching flame of the methane/air mixture was calculated, and the propagation velocity at each point along the flame front was obtained. The reason why the flame speed of the gas explosion with a high-temperature source was higher than the flame speed of the gas explosion with an electric spark was preliminarily explained. [ABSTRACT FROM AUTHOR]

The propagation of transient, turbulent premixed flames in a vented explosion chamber in the presence of a series of obstacles is numerically investigated by a dynamic formulation for the Flame Surface Density (FSD) with the Large Eddy Simulations (LES) technique. The chemistry is modelled by a one-step overall reaction, which simulates the reaction of a stoichiometric propane-air mixture. The FSD modelling in the reaction rate model is numerically employed with two different sub-grid scale (SGS) models. The first one is based on an empirical correlation of the SGS velocity fluctuations and the second one is based on similarity ideas involved in solving the wrinkled flame front considered as a fractal surface. The numerical predictions are analyzed and compared against an algebraic, simple FSD model together with experimental data. The calculations show that the dynamic FSD models provide superior results as compared with the algebraic FSD model. The comparisons demonstrate the importance of the contributions from the unresolved FSD and provide good agreement with experimental data for the flame structure, overpressure, and burning velocities. [ABSTRACT FROM AUTHOR]

The propagation velocity of turbulent premixed flames with real thermal expansion in a time-dependent external flow is studied for an infinitely thin flame front and for flames of small but finite thickness. It is shown that the influence of temporal pulsations is usually small and can be neglected for reasonable values of the pulsation frequency. The role of pulsations is even smaller for realistic thermal expansion in comparison with the artificial case of zero thermal expansion studied previously. The results obtained indicate that the Taylor hypothesis of “ stationary” turbulence is a good approximation for turbulent flames. The role of pulsations is substantial only if the integral turbulent length scale is close to the cut-off wavelength of the Darrieus-Landau instability. In this particular case, temporal pulsations can be important for explaining recent experiments on turbulent burning. [ABSTRACT FROM AUTHOR]

The paper describes a large eddy simulation model of gaseous deflagration in a closed vessel and simulation results for stoichiometric hydrogen-air premixed combustion initiated at the center of a closed 6.37-m³ spherical vessel. The model is based on the large eddy simulation approach to turbulence modeling and the gradient method to model the mass burning rate in premixed combustion. The method for simulated flame-front thickness reduction is suggested, and its performance is investigated. The solution-adaptive mesh refinement is used to decrease the CPU time required for simulation. The simulated deflagration-pressure dynamics is in agreement with published experimental data, and the flame-front velocity is in agreement with simulation results obtained according to the lumped parameter model for the same experiment. [ABSTRACT FROM AUTHOR]