Determination Of Distribution Of Flame Temperature And Soot Volume Fraction Based On Multispectral Analysis

With the growing global energy crisis and increasing emphasis on environment issues, the combustion efficiency improvement and combustion emission reduction of fossil fuel become an important research content. While flame temperature is one of the important parameters to determine physical status and reflect the process of combustion, its accurate measurement is of great significance for developing combustion theory and understanding combustion process. The flame temperature is coupled with radiation parameters such as emissivity, absorption coefficient, scattering coefficient and soot volume fraction, so the reconstruction of temperature and radiation parameters become an effective method, but also increase the difficulty. The multispectral measurement of flame can provide plenty of radiation intensity with flame temperature and radiation information. The researchers not only can calculate temperature and radiation parameters directly from spectral intensity information, but also reconstruct the distributions of temperature and radiation parameters based on the calculation method for radiation inversion problem. In this paper, the simulation and experimental researches about the two problems were carried on. The detailed description is as below.A simple method to determine the distribution of temperature and emissivity with wavelength and judge the gray property based on the media boundary multi-spectral radiation intensity has been demonstrated. For the gray, one-dimensional plane-parallel medium and two-dimensional axisymmetric medium, the effects of different distribution of temperature, absorption and scattering coefficients in the medium to distribution of boundary temperature and emissivity with wavelength were analyzed using the proposal method and DRESOR method. The results showed that:the boundary radiation may still display non-gray body property and the temperature may change with wavelength even for gray medium, so the validity of temperature measurement methods based on the gray body assumption should be further validated, especially for multi-wavelength temperature measurement method based on spectral intensity in a large wavelength range.A calibrated spectrometer system was used for the data acquisition of the radiation intensity profile of a flame over a range of wavelengths (200～1100 nm). The wavelength range that meets the gray body assumption, in which the emissivity can be assumed as a constant, can be determined from the emissivity profile obtained by the proposal method. Then the temperature and emissivity of the flame and their relative mean square deviations were calculated within that range. Experiments were conducted on solidified gasoline, red phosphorus flames, ethylene diffusion flame and pulverized coal-fired flames in a commercial 420 t/h boiler furnace. The results show that the radiation of the gasoline flame can be assumed as a gray body in the range between 550 and 900 nm, for the coal-fired flame the range is between 500 and 1000 nm; while the radiation of the red phosphorus flame and ethylene diffusion flame cannot be assumed as a gray body within the measurement wavelength range. The temperature and emissivity calculation results of coal-fired flames are found to be in reasonable agreement with results using other methods from the literature.A simulation investigation for simultaneous reconstruction of distributions of temperature and soot volume fraction from multi-wavelength emission in an axisymmetric sooting flame using the stochastic particle swarm optimizer (SPSO) algorithm is presented. The self-absorption of the flame is considered. The selection of parameters of the SPSO algorithm and detection wavelengths is analyzed. The effects of measurement errors and optical thickness of the flame on the accuracy of the reconstruction are investigated. A comparison between SPSO algorithm and the other literatures mentioned methods, such as Abel inversion algorithm and Tikhonov regularization method, has been made. It proved that the SPSO algorithm is robust and can obtain accurate distributions of temperature and soot volume fraction from line-of-sight intensities in only several wavelengths, especially in the flame with large optical thickness, while other methods, such as Abel inversion and Tikhonov regularization method, neglecting self-attenuation of the flame will take more errors than the SPSO algorithm.Experimental study of simultaneous reconstruction of distribution of temperature and soot volume fraction at different height of the flame from emission intensity spectra using SPSO algorithm is presented. Measurements are carried out on an axisymmetric, laboratory grate, ethylene/air diffusion flame using the visible spectroscopy. The influence of fiber aperture to the detection intensity distribution is also analyzed. The results show that:the usage of smaller optical fiber aperture can improve the spatial resolution of detection intensity distribution, and the aperture of 0.5 mm is selected. For the visible emission intensity the temperature gradually increases from the center to the edge of the flame along the radial direction, the soot volume fraction shows peak distribution along the radial direction and the peak gradually close to the flame center with the height increases. Comparison of reconstruction results with the literature shows that under the premise of the certain difference of the flame height with that of literatures, the overall distributions of flame temperature and soot volume fraction are consistent with those of the literatures and the maximum of soot volume fraction is very close.