Abstract In the present work, an experimental investigation was performed by varying the fuel/air ratio of a liquid-fuel gas turbine derived burner in the non-premixed mode, until an ultra-lean combustion condition was reached. In this condition, flame instabilities occur with negative impacts on combustion efficiency. Two high speed visualization systems in the visible range and in the infrared spectral region were used. Moreover, they were supported by an OH⁎ chemiluminescence measurement and by gas exhaust measurements. Different techniques were used starting from the luminosity signal of each pixel: the Wavelet Decomposition to calculate the wavelet energy, the frequency analysis of pixel intensities of the flame images to estimate the dominant frequency, finally the statistical analysis to calculate the pixel intensity variance. Both the statistical and frequency analyses were applied to the OH⁎ chemiluminescence data. One of the most important results of the present work regarded the capability of imaging techniques to individuate the instability insurgence and to be used as a predictive tool. Furthermore 2D maps of some parameters, extracted by the wavelet-based analysis of flame images, permitted to investigate local unsteadiness in the flame area.
Image processing for the characterization of flame stability in a non-premixed liquid fuel burner near lean blowout
DE GIORGI, Maria Grazia;SCIOLTI, ALDEBARA;CAMPILONGO, STEFANO;FICARELLA, Antonio
2016-01-01
Abstract
Abstract In the present work, an experimental investigation was performed by varying the fuel/air ratio of a liquid-fuel gas turbine derived burner in the non-premixed mode, until an ultra-lean combustion condition was reached. In this condition, flame instabilities occur with negative impacts on combustion efficiency. Two high speed visualization systems in the visible range and in the infrared spectral region were used. Moreover, they were supported by an OH⁎ chemiluminescence measurement and by gas exhaust measurements. Different techniques were used starting from the luminosity signal of each pixel: the Wavelet Decomposition to calculate the wavelet energy, the frequency analysis of pixel intensities of the flame images to estimate the dominant frequency, finally the statistical analysis to calculate the pixel intensity variance. Both the statistical and frequency analyses were applied to the OH⁎ chemiluminescence data. One of the most important results of the present work regarded the capability of imaging techniques to individuate the instability insurgence and to be used as a predictive tool. Furthermore 2D maps of some parameters, extracted by the wavelet-based analysis of flame images, permitted to investigate local unsteadiness in the flame area.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
